The Current Development Status of Agricultural Machinery Chassis in Hilly and Mountainous Regions

Abstract

The scenario adaptability of agricultural machinery chassis in hilly and mountainous regions has become a key area of innovation in modern agricultural equipment development in China. Due to the fragmented nature of farmland, steep terrain (often exceeding 15°), complex topography, and limited suitability for mechanization, traditional agricultural machinery experiences significantly reduced operational efficiency—typically by 30% to 50%—along with poor mobility. These limitations impose serious constraints on grain yield stability and the advancement of agricultural modernization. Therefore, enhancing the scenario-adaptive performance of chassis systems (e.g., slope adaptability ≥ 25°, lateral tilt stability > 30°) is a major research priority for China’s agricultural equipment industry. This paper presents a systematic review of the global development status of agricultural machinery chassis tailored for hilly and mountainous environments. It focuses on three core subsystems—power systems, traveling systems, and leveling systems—and analyzes their technical characteristics, working principles, and scenario-specific adaptability. In alignment with China’s “Dual Carbon” strategy and the unique operational requirements of hilly–mountainous areas (such as high gradients, uneven terrain, and small field sizes), this study proposes three key technological directions for the development of intelligent agricultural machinery chassis: (1) Multi-mode traveling mechanism design: Aimed at improving terrain traversability (ground clearance ≥400 mm, obstacle-crossing height ≥ 250 mm) and traction stability (slip ratio < 15%) across diverse landscapes. (2) Coordinated control algorithm optimization: Designed to ensure stable torque output (fluctuation rate < ±10%) and maintain gradient operation efficiency (e.g., less than 15% efficiency loss on 25° slopes) through power–drive synergy while also optimizing energy management strategies. (3) Intelligent perception system integration: Facilitating high-precision adaptive leveling (accuracy ± 0.5°, response time < 3 s) and enabling terrain-adaptive mechanism optimization to enhance platform stability and operational safety. By establishing these performance benchmarks and focusing on critical technical priorities—including terrain-adaptive mechanism upgrades, energy-drive coordination, and precision leveling—this study provides a clear roadmap for the development of modular and intelligent chassis systems specifically designed for China’s hilly and mountainous regions, thereby addressing current bottlenecks in agricultural mechanization.

1. Introduction

Agriculture, as a foundational and strategic sector of the national economy, plays an essential role in ensuring food security, improving rural livelihoods, and promoting rural revitalization. With the full-scale implementation of the rural revitalization strategy, advancing agricultural mechanization has become a key approach to building modern agricultural industrial systems and overcoming development constraints in hilly and mountainous regions [1,2,3]. According to data from China’s Third National Land Survey, hilly and mountainous areas cover approximately 3.1908 million square kilometers, accounting for about 69.4% of the country’s total land area. The farmlands in these regions are characterized by a distinct “three-more and three-less” pattern: fragmented plots, significant slope variations, and complex topography (the “three mores”); limited contiguous arable land, low mechanization suitability, and difficulties in implementing standardized operations (the “three lesses”) [4,5,6,7,8]. These unique geographical features have rendered conventional agricultural machinery poorly adapted to local conditions, leading to persistent challenges such as reduced operational efficiency, limited terrain traversability, and insufficient adaptability. The growing mismatch between current agricultural equipment and terrain requirements has become a major constraint on achieving stable and sustainable grain production.

To address these challenges, terrain-adaptive intelligent chassis systems—incorporating innovative design features such as elevated ground clearance, four-wheel drive, and dynamic center-of-gravity adjustment—have emerged as a breakthrough solution. The core technological value of these systems lies in the synergistic integration of multi-mode locomotion mechanisms with intelligent cooperative control systems, further enhanced by contour-matching leveling and multi-sensor fusion positioning. This integrated approach enables the development of highly adaptable agricultural machinery chassis capable of overcoming complex terrain constraints, thereby ensuring stable operation on steep slopes (6–15°), muddy terrain, and rugged environments. Specifically, these systems effectively address operational stability issues in hilly farmlands (6–15° slope gradients) and sloped croplands (15–25°) while significantly improving precision during operations in irregularly shaped fields [9,10,11,12,13]. These technological advancements have notably enhanced safety performance, energy efficiency, and modular adaptability across diverse operational scenarios, positioning this technology as a key enabler for advancing agricultural modernization and ensuring food security in hilly and mountainous regions.

In recent years, extensive research has been conducted by domestic and international scholars on the intelligentization, versatility, and stability improvement of next-generation agricultural machinery. This is exemplified by key technological advancements in contour-adaptive chassis design, multi-sensor fusion positioning, and modular functional component integration [14,15,16]. These innovations have shown considerable potential in enhancing operational efficiency under complex terrain conditions, improving the reliability and stability of agricultural chassis, and advancing mechanization in hilly and mountainous regions. Nevertheless, three major limitations remain in current research: (1) Insufficient terrain coverage: Systematic studies on steep slopes (15–25°) and unstructured terrains—such as terraced fields and rugged muddy areas—are still limited; (2) Control strategy constraints: Existing algorithms inadequately address dynamic collaborative optimization across multiple subsystems (power, drive, and leveling), lacking closed-loop capabilities that integrate real-time terrain perception, decision-making, and execution; (3) Lack of alignment with Dual Carbon goals: Principles of energy efficiency optimization and lightweight design have not been sufficiently integrated into the development frameworks of intelligent chassis systems. It is essential to develop a dynamic modeling framework grounded in unified state-space equations, where the slope angle θ and slip ratio s serve as fundamental state variables. This framework must incorporate a cooperative optimization mechanism that integrates torque distribution vectors with terrain parameters (e.g., soil shear modulus), thereby improving the generalization performance of theoretical models under complex operational conditions.

In hilly and mountainous regions marked by complex terrains and variable slopes, the power system, traveling system, and leveling system of agricultural machinery chassis constitute essential technological pillars for ensuring operational safety and efficiency. Through improvements in terrain adaptability and energy utilization efficiency, coordinated innovation among these systems provides crucial technical support for advancing agricultural modernization in challenging environments. Based on multi-scenario application requirements and aligned with the strategic objectives of China’s Dual Carbon goals, this paper presents a systematic review of the developmental status of agricultural machinery chassis tailored for hilly–mountainous regions. It conducts a comprehensive analysis of the technical characteristics, working principles, and scenario-specific adaptability of key subsystems—namely, power transmission, terrain navigation, and dynamic stabilization. To meet the unique operational demands of these terrains, strategic directions are proposed for the development of intelligent chassis technologies, focusing on three core areas: innovative multi-mode locomotion mechanisms, optimized collaborative control algorithms, and integrated intelligent perception systems. These directions aim to provide both theoretical foundations and practical technical pathways for the design and implementation of next-generation terrain-adaptive intelligent chassis systems.

2. Power System of Agricultural Machinery Chassis

The power system constitutes a fundamental subsystem that ensures the normal operation and productivity of agricultural machinery. Based on drive types, it can be classified into four categories: mechanical, hydrodynamic, electric, and hybrid power systems [17,18,19,20,21]. Conventional mechanical power systems rely on rigid mechanical transmission chains for energy transfer, enabling robust performance in high-traction load operations. Hydrodynamic systems utilize fluid media to transmit or convert energy, providing superior operational smoothness and terrain adaptability. Electric power systems employ single-motor or multi-wheel-hub motor direct-drive configurations, which are characterized by high energy feedback efficiency. Hybrid systems integrate hydraulic-electric or hydroelectric strategies to improve overall energy utilization efficiency. In hilly and mountainous regions with complex operational environments, power systems must meet three essential requirements: high power density, low energy consumption, and enhanced environmental adaptability. Establishing chassis dynamic models using the Newton–Euler method (direct force analysis) and the Lagrangian method (energy and generalized force analysis) enables systematic characterization, prediction, and optimization of kinematic behaviors (e.g., posture and trajectory), dynamic loads (e.g., wheel–terrain interaction forces and pitch/roll moments), and energy-transfer efficiency under such demanding conditions. These modeling approaches serve as indispensable quantitative tools and foundational elements for addressing the core technical requirements. Figure 1 illustrates agricultural machinery equipped with different drive configurations.

Table 1. Working principles, characteristics, and applications of agricultural machinery powertrain systems.

Type	Working Principle	Core Characteristics and Performance Metrics	Typical Agricultural Application Scenarios
Mechanical	Gear/chain direct power transmission	Max climbing gradient: 25–30° Energy consumption per unit mass: 2.8–3.5 kJ/kg·km Torque density: ≈50 N·m/kg System efficiency: 60–75%	Small tractors (<50 HP), threshers, conventional harvesters, and other low-complexity equipment
Hydraulic	Pump–valve–cylinder energy conversion	Max climbing gradient: 35–40° Energy consumption per unit mass: 2.0–2.5 kJ/kg·km Torque density: >120 N·m/kg System efficiency: 70–80%	Large combine harvesters (e.g., CLAAS LEXION), sugarcane harvesters, orchard lifting platforms, and other high-torque/precision control scenario
Electric	Battery–motor direct drive	Max climbing gradient: 40°+ Energy consumption per unit mass: 0.9–1.4 kJ/kg·km Torque density: >200 N·m/kg System efficiency: >90%	Greenhouse operation robots, electric plant protection UAVs, and small electric tractors (e.g., Kubota X tractor)
Hybrid	Internal combustion engine + electric motor coordination	Max climbing gradient: 38–42° Energy consumption per unit mass: 1.6–2.0 kJ/kg·km Torque density: ≈150 N·m/kg System efficiency: 80–88%	Large intelligent tractors (e.g., John Deere 6R), silage harvesters, and hilly–mountainous multi-functional platforms

presents the operational principles, technical characteristics, and application scenarios of various power systems.

2.1. Mechanical Power System

The power transmission topology of a mechanical chassis follows the conventional internal-combustion-engine-driven architecture. A fuel-powered engine serves as the primary power source, while a clutch enables controlled coupling and disengagement of the power flow. The generated torque is sequentially transmitted through the gear transmission system, universal drive shaft, and drive axle assembly. Following this, the main reducer amplifies the torque and redirects the power flow direction, after which the differential distributes torque to the individual wheels. Finally, the half-shafts deliver rotational power to the drive wheel assemblies. The working principle is illustrated in Figure 2. This transmission design demonstrates high mechanical efficiency and strong terrain adaptability, which has established it as the dominant configuration in tractor applications [22,23,24]. It is particularly well suited for medium- to low-complexity hilly terrains with inclines below 25°.

Western nations, capitalizing on their advantageous geographical conditions and advanced industrial technological heritage, have conducted earlier and more mature research into tractor chassis designed for hilly and mountainous terrains. The Terratrac series of mountain tractors manufactured by Switzerland’s Aebi Company (Figure 3) feature a low-profile, wide-body design with a low center of gravity and agile steering, providing exceptional stability and traction adhesion in complex terrains such as hilly and mountainous regions. Figure 4 illustrates the MACH4 tractor developed by Karola AG [25], which utilizes a four-wheel drive system integrated with a quad-track chassis to achieve superior stability performance.

Traditional mechanical power systems exhibit commendable obstacle navigation capability, traction performance, and slope stability. However, they are inherently constrained by high energy consumption, low transmission efficiency, complex operational procedures, and inadequate load adaptability. These limitations become especially critical under abrupt slope change conditions, where inertial load impacts on the transmission system often lead to power interruptions and component fatigue failure, thereby intensifying energy efficiency degradation. Ahn et al. [26] proposed an integrated control algorithm that incorporates engine-transmission efficiency characteristics. Through experimental validation, they identified an optimal operational line that significantly enhanced overall system efficiency. Nevertheless, this approach did not account for the time-varying influence of dynamic load impacts on transmission efficiency, which restricts its effectiveness in real-world applications. To address the issues of low accuracy and high noise levels in hydro-mechanical continuously variable transmission (HMCVT) efficiency prediction, Lu et al. [27] introduced a hybrid methodology combining variational mode decomposition (VMD), particle swarm optimization (PSO), and backpropagation neural networks. Despite its advantages, this method fails to differentiate between meaningful signals induced by load fluctuations and random noise. This limitation underscores the need to integrate chassis dynamics models—such as Bekker’s equation—to reconstruct temporal load characteristics for more accurate efficiency modeling. Xia et al. [28] introduced an innovative design methodology for power-chained hydro-mechanical continuously variable transmission (PCHMCVT) systems, with the key advancement being the deep integration of chassis dynamics modeling into transmission system optimization. The study developed a tractor dynamics model based on traction force equilibrium equations to quantitatively assess the influence of rolling resistance, grade resistance, acceleration resistance, and tillage resistance on transmission loading. This approach effectively improved transmission efficiency across multiple drive modes. Neto et al. [29] performed a comparative evaluation of continuously variable transmission (CVT) and fixed-power split (FPS) transmissions by analyzing their impacts on transmission efficiency and fuel consumption. Through systematic measurements of the slip rate ($S_I$), drawbar power ($D{B}_p$), and other dynamic output parameters, they demonstrated that FPS transmissions offer superior energy efficiency at the expense of reduced operational speed. Notably, the observed traction force fluctuations under slope conditions highlighted the critical importance of incorporating chassis dynamics modeling for accurate prediction of transmission system robustness. Md-Tahir et al. [30] investigated the mechanism through which rigid lug wheels (RLW) enhance tractor traction performance by employing a wheel–soil interaction dynamics model. Utilizing Bekker’s pressure–sinkage and shear theory, the researchers quantified the soil shear reinforcement effects induced by RLW lugs. By solving Newtonian dynamic equations to determine the slip rate (TR) and power loss, they achieved significant improvements in traction coefficient and power conservation performance. Zhao et al. [31] conducted a comprehensive dynamics analysis that converted implement resistance and rolling resistance into power demand metrics. Their modeling framework incorporated transmission chain speed/torque equations, planetary gear power flow analysis, and chassis traction resistance calculations. This work revealed the essential trade-off between structural simplification of hydro-mechanical transmission (HMT) systems and energy consumption optimization. Xiao et al. [32] designed a wheeled tractor incorporating a novel hydro-mechanical continuously variable transmission (HMCVT) based on an in-depth analysis of fuel economy evaluation metrics, achieving substantial improvements in fuel efficiency. Ye et al. [33] applied fuzzy mathematics, reliability theory, and multi-objective optimization techniques to optimize the design of single-stage external spur gears, thereby enhancing transmission efficiency. Liang et al. [34] proposed a control strategy that utilizes multi-parameter adjustments to jointly optimize engine operating points and HMCVT working conditions. The objective function was defined as the ratio of engine fuel consumption rate to transmission efficiency, providing a theoretical basis for improving both tractor power performance and fuel economy. Wu et al. [35] developed a minimum-jerk polynomial-based local speed planning algorithm rooted in Bellman’s optimality principle. This approach effectively addresses the issues of poor operational stability and low fuel efficiency caused by frequent acceleration and deceleration events during tractor operation. It ensures smooth speed transitions in continuously variable transmission (CVT) systems while significantly improving fuel economy. Fu et al. [36,37] introduced two key innovations: (1) a gear-shifting control strategy with stochastic load adaptive correction to mitigate frequent random shifting induced by load fluctuations, and (2) a transmission system matching optimization method based on an improved non-dominated sorting genetic algorithm (NSGA). These approaches enable simultaneous enhancement of vehicle dynamic performance and fuel efficiency.

Current research on agricultural machinery power systems primarily centers on traditional internal-combustion-engine-driven architectures. These systems employ mechanical components such as clutches, gear transmission systems, final reducers, and differentials to achieve effective power distribution. Internationally, well-established solutions built upon industrial expertise have redirected research efforts toward optimizing transmission efficiency and integrating intelligent control algorithms. Domestically, studies focus on algorithm-driven reconstruction of power systems, utilizing fuzzy mathematics and multi-objective optimization techniques to enhance the efficiency of gear transmission.

However, existing chassis dynamics models largely rely on simplified assumptions—such as ideal terrain conditions and steady-state operating scenarios—which result in insufficient characterization of critical factors including tire–soil interaction, vehicle posture dynamics, and load response under complex field environments (e.g., alternating soft/hard soils, steep gradients, and asymmetric loading). This limitation compromises the robustness of control algorithms. Although certain advancements have notably improved fuel economy and achieved high transmission efficiency with moderate terrain adaptability in tractors, mechanical transmission systems still encounter challenges related to reliability and durability under high-speed, heavy-load, and highly dynamic operating conditions. Compared to hydraulic and electric drive systems, the intelligentization of mechanical systems lags behind, mainly due to the lack of real-time precision in dynamic models across full operational ranges, which hinders the integration of advanced control strategies.

Contemporary development of mechanical power systems is marked by three key characteristics: (1) the parallel advancement of transmission structural innovation and lightweight design; (2) the deep integration of data-driven methodologies with classical control theory; (3) dual optimization of power performance and energy efficiency through enhanced fuel economy. While data-driven methods, such as deep learning, offer new modeling opportunities, their reliance on extensive scenario-specific datasets and limited generalization capabilities constrain their applicability in addressing the extreme diversity and uncertainty inherent in agricultural operations. Future research should prioritize addressing critical bottlenecks in three key areas: the development of adaptive transmission systems for highly complex terrains, the realization of deep mechatronic-hydraulic collaborative control, and the integration of novel hybrid powertrain architectures. To support these advancements, it is essential to develop high-fidelity yet computationally efficient multi-physics dynamics models that incorporate mechanical components, soil properties, and crop interactions. Such models will significantly improve predictive accuracy and generalization capabilities under complex and variable boundary conditions. Furthermore, model verification and control system development must be strengthened through the application of digital twin and hardware-in-the-loop (HIL) technologies. These approaches can effectively shorten development cycles while enhancing system reliability and performance validation. In parallel, research efforts should focus on self-adaptive model updating mechanisms and transfer learning techniques. These methodologies will empower control systems to dynamically adjust to diverse agricultural implements, varying soil conditions, and different operational tasks. Collectively, these technological advances will facilitate the intelligentization and low-carbon transformation of agricultural machinery, aligning with broader goals of sustainable and precision agriculture.

2.2. Hydraulic Power System

The hydraulic power system of the chassis operates based on fluid dynamics and hydrostatic transmission principles, transmitting power through pressurized fluid to achieve energy conversion and precise control. As depicted in Figure 5, this system provides superior compliance, high torque output, and enhanced adaptability to complex operational conditions compared to traditional mechanical power systems. These advantages have facilitated its extensive adoption in agricultural machinery, including high-clearance crop protection equipment, harvesters, and specialized tractors.

Western countries initiated the application of hydraulic drive systems in agricultural machinery as early as the 1970s, establishing substantial technological leadership compared to China. As shown in Figure 6, representative high-clearance self-propelled crop protection machinery from international manufacturers utilizes fully hydraulic-driven chassis to achieve high operational efficiency, providing superior maneuverability and stability. However, their large-scale chassis designs and high costs limit their suitability for China’s hilly and mountainous terrains.

In recent years, domestic researchers have made substantial advancements in the development of hydraulic-driven chassis systems. These efforts have significantly narrowed the technological gap between China and Western countries in agricultural machinery design and implementation. Xiao et al. [38] designed a specialized hydraulic-driven chassis tailored for paddy field machinery to address challenges related to operational stability and insufficient power output. Their research focused on the hydraulic behavior of critical components during system operation. The experimental results demonstrated that the chassis exhibits low deviation characteristics and enhanced anti-slip performance under sloped or rugged paddy field conditions. Yang et al. [39] proposed an optimized design method for hydro-pneumatic suspension systems in high-clearance self-propelled sprayers based on an improved multi-objective particle swarm optimization (MOPSO) algorithm. This approach significantly enhanced vibration damping performance, demonstrating superior overall characteristics compared to conventional suspension systems. Zhang et al. [40] developed a hydrostatic chassis drive system for large-scale plant protection machinery. The system integrates a single-pump four-motor closed-circuit configuration with a flow divider valve to achieve anti-slip functionality, thereby improving operational stability. Hu et al. [41] designed a novel hydraulic chassis for rice transplanters using a single-pump four-motor drive architecture. An anti-slip valve block was incorporated to effectively prevent slippage in the distributed hydraulic motors during operation. Sun et al. [42] introduced a slip control strategy for hydrostatic drive systems in high-clearance sprayers. By establishing a bilinear relationship between slip rate and adhesion coefficient, this method effectively mitigates the negative effects of uneven terrain on driving stability. Chen et al. [43] presented a hydrostatic drive system for high-clearance corn detasseling machines, incorporating a four-wheel independent suspension structure. Simulation and experimental studies confirmed that the suspension system can effectively reduce ground impact forces and improve the performance of the drive system. Li et al. [44] developed a fully hydraulic remote-controlled power chassis. Field tests demonstrated stable performance during uphill/downhill movement, ridge crossing, and ditch traversal, meeting the mobility requirements for hilly and mountainous terrains. Zhang et al. [45] proposed a fully hydrostatic-driven intelligent spraying chassis with flexible control capabilities. Their research provides theoretical support for straight-line control of four-wheel independent drive systems and offers a reference for designing high-precision, reliable, and cost-efficient hydraulic synchronization control systems.

In conclusion, China has achieved significant advancements in hydrostatic drive technology and intelligent control systems for hydraulic-driven chassis. Nevertheless, several critical challenges remain, including dependence on imported core components, the inherent trade-off between system efficiency and energy consumption, and insufficient electro-hydraulic-mechanical coordinated control. Particularly, the lack of comprehensive system energy equations that integrate hydraulic actuators and suspension systems hinders in-depth analysis of fluid–mechanical–terrain coupling mechanisms and vibration energy transmission and dissipation dynamics. These limitations, combined with inadequate intelligentization, constrain the development of next-generation agricultural hydraulic power systems. To address these challenges, three strategic breakthroughs are essential: (1) The development of high-power-density hydraulic pumps and motors, along with low-leakage valves, should be prioritized. System topology should be optimized through innovations in hydrostatic drive technology to reduce energy consumption while simultaneously advancing the domestic production of core components. (2) A multi-objective optimization framework integrating terrain characteristics, load conditions, and energy consumption should be established using model predictive control (MPC) and reinforcement learning (RL). This will enable coordinated anti-slip control and optimal power distribution. (3) Multi-sensory integration—combining LiDAR, inertial navigation systems (INSs), and tactile feedback sensors—should be implemented alongside digital twin platforms to minimize hydraulic system response latency. Nonlinear disturbance observers should also be employed to compensate for dynamic uncertainties, thereby achieving multi-modal perception and electro-hydraulic-mechanical synergy. Additionally, lightweight design strategies must be emphasized to meet the operational demands of hilly and mountainous terrains. With strong national policy support and continuous technological innovation, hydraulic drive technology has become a key enabler for the modernization of agricultural machinery. Through component-level breakthroughs, algorithm-driven optimization, and system-level integration and validation, advanced hydraulic power systems are expected to play a pivotal role in promoting efficient, intelligent, and environmentally sustainable agricultural mechanization.

2.3. Electric Power System

Under the overarching goal of achieving the “Dual Carbon” targets (carbon peaking and carbon neutrality), traditional diesel-powered agricultural machinery is increasingly challenged by its high-pollution emissions and complex mechanical architectures. Electrified agricultural machinery, characterized by structural simplification through electrification integration, zero-emission operation, and low-noise performance, has emerged as a key pathway for agricultural equipment modernization. Its inherent compatibility with intelligent farming operations further solidifies its position as a transformative solution [46]. Compared to electric vehicles, electric agricultural machinery faces distinct challenges: high-torque output requirements for heavy-duty tasks; seasonal-intensive operational demands aligned with agricultural cycles; adaptability to complex and heterogeneous field environments. As a result, its power system requires deep integration with implements to ensure functional synergy [47]. The operational principles of electric drive systems are depicted in Figure 7.

Baek et al. [48] conducted a comprehensive evaluation of the agricultural and traction performance of electric all-wheel-drive (AWD) tractors, offering valuable insights for the optimization of distributed drive systems in agricultural machinery. Mao et al. [49] proposed a modeling methodology for electric drive systems using the Modelica platform. They developed individual component models for the energy system, motor system, and mechanical subsystems, which were subsequently integrated into a comprehensive simulation model of the electric tractor drive system. The reliability of the model was validated through experimental traction performance tests. This study quantified the efficiency of energy transmission from electrical to mechanical energy (${\eta }_{\tau }={\eta }_{mc}{\eta }_m{\eta }_{\delta }{\eta }_f$, ${\eta }_{\tau }$ is the traction efficiency of the electric tractor, ${\eta }_{mc}$ is the transmission efficiency of the drive system, ${\eta }_{\delta }$ is the slip efficiency of the drive wheel, ${\eta }_f$ is the rolling efficiency, and ${\eta }_m$ is the efficiency of the drive motor and controller), providing essential parameters for the design and optimization of electric drive systems. However, the current model does not incorporate vehicle dynamics equations based on either the Newton–Euler formulation or the Lagrangian approach that account for slope-induced potential energy. As a result, it lacks the capability to compute wheel–ground contact forces, estimate posture and slip states, or optimize torque distribution under complex terrain conditions. Future research should focus on advancing terrain-coupled dynamic modeling to enhance adaptive control performance in variable agricultural environments. To address the challenges of low traction efficiency and poor stability in distributed electric drive tractors, Yuan et al. [50] proposed a control strategy based on a Barrier Lyapunov Function (BLF)-enabled multi-constrained backstepping algorithm. By constructing component load models and designing constrained Lyapunov functions, this method achieves robust control performance under variable feeding disturbances, thereby significantly enhancing overall system stability. Deng et al. [51] developed a distributed drive electric tractor with a four-wheel independent drive architecture. Through static parameter matching verification, they demonstrated favorable dynamic performance under flat-road transportation conditions. Yu et al. [52] introduced a torque allocation strategy that divides total torque into base torque and compensation torque components. This approach integrates particle swarm optimization (PSO) for optimal torque positioning with battery state-of-charge (SOC)-based fuzzy rules to distribute torque between dual motors. The method effectively mitigates power–demand mismatches and improves operational performance during plowing and rotary tillage operations. An et al. [53] proposed a hierarchical control framework incorporating upper-layer sliding mode control for yaw moment regulation. Based on a comprehensive chassis model that includes nonlinear tire forces and vehicle dynamics, combined with real-time feedback of yaw rate and wheel speeds, the strategy enables optimal torque distribution by minimizing tire load rates and achieving efficient posture tracking. This significantly enhances adhesion margin and anti-slip capability in complex terrain conditions. Xie and Wen et al. [54,55] designed a novel dual-motor coupled transmission system for electric tractors, aiming to improve load capacity and operational efficiency. Chen et al. [56] presented a parameter matching and optimization methodology for dual-motor coupled powertrains, focusing on balancing traction performance with energy economy in electric agricultural machinery. Du et al. [57] and Wu et al. [58] explored advanced motor technologies, proposing the use of dual-salient pole-changing permanent magnet (PM) motors and variable leakage flux PM motors, respectively. These motor configurations provide theoretical support for achieving high-efficiency multi-mode operation, smooth pole-switching transitions, and stable performance under complex operating disturbances.

Electric agricultural machinery faces significant development challenges in hilly and mountainous regions due to complex operational environments and limited resource availability, particularly concerning endurance and energy efficiency. Li et al. [59] developed a real-time energy management strategy that integrates stochastic dynamic programming with extremum seeking algorithms, based on longitudinal vehicle dynamics. By dynamically optimizing dual-motor power distribution ratios and drive mode switching, the strategy enables adaptive torque allocation and real-time system efficiency optimization under complex terrain conditions, thereby improving overall drive efficiency and reducing average energy consumption. This highlights the essential role of chassis dynamics modeling in enhancing terrain adaptability. Li et al. [60] proposed an adaptive discharge threshold adjustment strategy using fuzzy logic combined with capacity allocation for hybrid energy storage systems, resulting in improved energy efficiency. He et al. [61] introduced an adaptive power management strategy that integrates operation condition recognition (OCR) with deep reinforcement learning (DRL) for hybrid energy systems. Through vehicle dynamics modeling that quantifies the influence of terrain parameters—specifically slope angle α and soil resistance k—on traction force $F_t$, and by constructing OCR modules using actual tillage data, this approach significantly reduces battery capacity degradation and power loss costs. Wang et al. [62] established a co-optimized energy management strategy based on longitudinal dynamics equations and wheel–soil interaction models. This method quantifies the effects of soil resistance “DF” and slip ratio “s” on torque demand, leading to enhanced operational efficiency and stability. Sun et al. [63] designed a fuel-cell-dominant hybrid power system supported by battery assistance, proposing rule-based energy management strategies that effectively delay fuel cell degradation while extending system endurance. Melo et al. [64] presented an automatic slip control solution for two-wheel-drive electric tractors, which improves both system stability and energy utilization. Fu et al. [65] developed an electric four-wheel-drive chassis and demonstrated that optimized suspension damping parameters can significantly enhance ride comfort. Liu et al. [66] proposed a tracked electric tractor with an optimized powertrain configuration, showing superior economic performance in field tests. Han et al. [67] developed a dual-motor electric tractor capable of delivering a maximum output power of 13.9 kW and continuous operation for up to 4.5 h, meeting the requirements of orchard operations. To advance intelligent mechanized harvesting for small green vegetables, Tang et al. [68] designed a self-propelled intelligent combine harvester capable of performing cutting, clamping, conveying, and collecting operations simultaneously.

In summary, current research on electric powertrain systems for agricultural machinery primarily focuses on achieving high torque output, enhancing adaptability to complex operational conditions, and optimizing energy efficiency. These efforts are characterized by the technical feature of “algorithm-driven hardware coordination.” In China, the development of electric agricultural machinery remains at an early exploratory stage, with most studies currently centered on theoretical validation and prototype testing. Several critical challenges persist: (1) insufficient electromechanical coupling optimization in powertrain systems, resulting in low energy conversion efficiency; (2) a lack of endurance prediction models that can support continuous operation under complex working conditions; (3) inadequate chassis dynamics modeling, which limits terrain adaptability; and (4) immature multi-mechanism cooperative control algorithms that negatively impact operational precision. To address these issues, breakthroughs in core technologies must be pursued through a mechatronic–hydraulic-integrated design—such as terrain–motor torque mapping based on comprehensive chassis dynamics models—as well as adaptive control algorithms tailored for variable working conditions and decision-making frameworks leveraging multi-source information fusion. Guided by the strategic directions of “intelligent electromechanical coupling, refined energy management, and precise scenario adaptation,” these technological advancements aim to overcome the practical implementation barriers faced by electric agricultural machinery operating in hilly and mountainous regions.

2.4. Hybrid Power System

The hybrid power system realizes dynamic coordination between diesel engines and electric motors through an innovative electromechanical coupling architecture. Its core features are manifested in three fundamental aspects: Firstly, a dual-mode power coupling mechanism based on planetary gear sets and intelligent control algorithms enables seamless transition between series and parallel operation modes, ensuring real-time torque compensation across diverse working conditions. Secondly, a multi-dimensional energy management strategy integrates braking energy recovery with power take-off (PTO) shaft reverse-drag power generation. This strategy, combined with state-of-charge (SOC) threshold control protocols, significantly enhances the duration of field operations. Thirdly, an intelligent redundancy design incorporates fault diagnosis systems to guarantee uninterrupted transitions between power sources under varying load demands. By addressing the operational challenges specific to hilly and mountainous terrain, the system effectively overcomes the limited endurance of purely electric agricultural machinery while reducing carbon emissions from conventional diesel-powered equipment. Furthermore, its design aligns with China’s strategic development requirements for next-generation agricultural machinery. The operational principle of this oil–electric hybrid system is illustrated in Figure 8.

Leading domestic and international enterprises have introduced hybrid tractors that achieve energy consumption reductions of 5–10% and carbon emission mitigations of 40–60% while effectively addressing adaptability challenges in hilly terrain. Mocera et al. [69,70,71] conducted comparative analyses between conventional and hybrid power systems, focusing on the performance characteristics of hybrid tractors equipped with electric continuously variable transmission (e-CVT) functionality. Using simulation techniques, they modeled and evaluated two parallel configurations, two series configurations, and one electro-hydraulic hybrid configuration. The experimental results indicated that the parallel configuration demonstrated superior peak power performance, whereas the series configuration achieved the highest fuel efficiency. The electro-hydraulic configuration was recommended as a viable alternative for engine downsizing. Li et al. [72] proposed a distributed hybrid electric tractor (DHET) architecture based on an AD–Extenics coupling design methodology. This approach enables intelligent electromechanical integration through powertrain hierarchical decomposition matrices and efficient operating range extension models, thereby establishing a theoretical foundation for future DHET powertrain innovations. Zhu et al. [73] developed a novel series-hybrid combine harvester by integrating series hybrid technology with distributed electric drive systems. They constructed mathematical models for key operational components, including chassis propulsion modules, which provide accurate references for selecting electric drive motors and matching hybrid system parameters. This approach simplifies the drivetrain structure while enhancing power transmission efficiency, offering new insights into improving the operational performance of combine harvesters. Sun et al. [74] introduced a novel methodology combining multi-objective particle swarm optimization (MOPSO) and wavelet decomposition algorithms to overcome the limitations of hybrid tractor crash testing. This approach significantly enhances the reconstruction of impact vibration signals, offering substantial value in the design and evaluation of tractor safety performance. Zhu et al. [75,76,77] developed a hybrid electro-mechano-hydraulic composite powertrain system for tractors to address the challenges of high fuel consumption and poor operational stability under complex and harsh working conditions. Additionally, they proposed an adaptive equivalent fuel consumption minimization strategy integrated with real-time operation mode prediction to further improve system efficiency. Yang et al. [78] designed a hydrogen fuel-cell-dominant hybrid power system supported by battery energy storage, aiming to mitigate both the environmental pollution caused by conventional diesel agricultural machinery and the limited operational duration of pure electric alternatives. The system incorporates a comprehensive energy management strategy (EMS) that integrates thermostat control, power tracking, and fuzzy logic control mechanisms. Ghobadpour et al. [79] implemented a two-layer energy management strategy (EMS) for plug-in hybrid electric tractors to regulate power distribution among multiple energy sources. This strategy effectively reduces fuel consumption while simultaneously extending vehicle range. Radrizzani et al. [80] constructed a predictive vehicle model that integrates chassis longitudinal dynamics, rolling resistance modeling, drivetrain efficiency losses, and wheel speed–torque relationships. The proposed energy management strategy (EMS) is designed to optimize fuel consumption while maintaining the battery state-of-charge (SOC) within a desired range. Zhang et al. [81] classified the power system into two distinct components—power sources and energy storage units—and proposed a dual-source collaborative optimization-based EMS (DSC-EMS). This strategy improves fuel cell efficiency, reduces ammonia consumption, and extends battery service life. Xu et al. [82] addressed the issue of insufficient battery discharge utilization in extended-range electric tractors by developing an EMS based on control parameter adjustment algorithms. The core innovation involves establishing a plowing-specific chassis dynamics energy consumption model, which segments the operational cycle into uniform acceleration, constant velocity, uniform deceleration, and turning phases. This model quantifies the energy required to overcome plowing resistance, rolling resistance, and gradient resistance, thereby significantly improving battery energy utilization efficiency and reducing overall fuel consumption. Wu et al. [83] formulated an energy management strategy for charge-depleting charge-sustaining (CD-CS) extended-range electric tractors to ensure continuous operation and meet range requirements. By constructing tire–soil interaction models and coulter–soil engagement models, they quantified the influence of different soil types on rolling resistance and determined optimal soil-entry angles, leading to a notable extension of pure-electric driving duration. Han et al. [84] proposed an energy-efficient hydro-electric hybrid tracked chassis that integrates independent hydraulic and electric drive systems. Compared to traditional single-drive agricultural equipment, this configuration demonstrates advantages in terms of lower power consumption, faster response times, and extended operational endurance. To enhance energy utilization efficiency in extended-range electric tractors, Wang et al. [85] established and experimentally validated a dual-motor independent drive energy management model. Their findings confirmed the theoretical soundness and practical feasibility of the proposed system.

Hybrid agricultural machinery is increasingly integrated into modern farming practices, marking a transition of China’s research from theoretical exploration to practical technological implementation. At the technical level, powertrain systems exhibit diversified development trajectories, while energy management strategies (EMSs) emphasize multi-objective collaborative optimization approaches. These include adaptive equivalent fuel consumption minimization strategies, fuzzy logic control algorithms, dual-source collaborative optimization frameworks, and hierarchical control architectures—each designed to balance fuel efficiency, operational range, battery durability, and emission constraints. Nevertheless, significant challenges persist, such as limited dynamic prediction capabilities for complex field conditions, difficulties in achieving optimal trade-offs among multiple objectives, and the high costs and reliability concerns associated with fuel cell and plug-in hybrid technologies. Future development directions highlight intelligent EMS optimization, multi-energy system integration, lightweight structural design, and modular system architecture. To fully enhance the sustainability and operational adaptability of agricultural machinery, it is essential to overcome key technological bottlenecks and reduce cost barriers.

3. Chassis Travel Mechanism System of Agricultural Machinery

The travel system constitutes a critical actuating component for agricultural machinery locomotion. In hilly and mountainous regions characterized by fragmented terrain, steep gradients, and heterogeneous soil compositions, these systems require enhanced stability, trafficability, and maneuverability [86,87,88,89,90]. To fulfill these performance requirements and enable accurate prediction of dynamic behaviors on complex terrains—including tractive force generation, anti-slip performance, and climbing capability—the application of precise wheel/terrain interaction models (e.g., Coulomb friction models, nonlinear slip-dependent models, or advanced soil mechanics-based models) is essential for optimizing chassis mobility systems. These mathematical models quantify mechanical interactions at the ground interface, thereby providing a quantitative basis for parametric chassis design (e.g., tire/track tread pattern configuration and contact pressure distribution optimization) and control strategy development (e.g., traction force allocation, slip ratio threshold setting, and slope anti-slip logic implementation). Through such systematic methodologies, the operational stability, terrain trafficability, and environmental adaptability of travel systems in fragmented mountainous environments are significantly improved. Structurally, travel systems are classified into wheeled and tracked configurations, with their comparative characteristics summarized in

Table 2. Principles, characteristics, and applications of agricultural machinery chassis travel systems.

Type	Working Principle	Structural Features	Key Parameters	Performance Indicators	Typical Agricultural Application Scenarios
Wheeled	Frictional drive through tire–ground interaction with steering mechanism for directional control	Simple structure; diverse tire configurations; optional independent suspension	Ground pressure: 50–150 kPa; minimum turning radius: 3–5 m; speed range: 0–40 km/h	Max climbing angle: 25–32; energy consumption per unit mass: 2.5–3.8 kJ/kg·km; traction efficiency: 60–75%; slip rate: 15–30%	Dryland farming in plains (e.g., John Deere 8R tractor), transport vehicles, and orchard management machines (narrow wheelbase design)
Tracked	Continuous ground contact via driven sprockets and track plates for pressure distribution	Multi-material track systems; Bogie wheel load-bearing; tension adjustment device	Ground pressure: 15–30 kPa; minimum turning radius: 1.5–3 m; speed range: 0–15 km/h	Max climbing angle: 35–42°; energy consumption per unit mass: 1.8–2.5 kJ/kg·km; traction efficiency: 75–88%; slip rate: 5–15%	Paddy field operations (e.g., Kubota PRO988Q harvester), wetland farming, and slope orchards (slip resistance > 85%)
Half-track	Electro-hydraulic coupling system combining front-wheel steering and rear-track drive	Articulated chassis; hydraulic speed regulation; modular configuration switching	Ground pressure: 30–50 kPa; minimum turning radius: 2–4 m; speed range: 0–25 km/h	Max climbing angle: 32–38°; energy consumption per unit mass: 2.0–3.0 kJ/kg·km; traction efficiency: 68–82%; slip rate: 8–20%	Muddy terrain operations (e.g., Case Steiger Quadtrac), hilly/mountainous transport, and silage harvesting (40% traction enhancement)

.

3.1. Wheeled Travel Mechanism

Wheeled chassis provides smooth mobility, low energy consumption, and precise control over speed and direction, facilitating adaptation to diverse operational environments. These advantages not only reduce operator fatigue but also enhance operational precision and efficiency. Based on drive configurations, wheeled chassis can be categorized into: front-wheel drive (FWD), rear-wheel drive (RWD), and four-wheel drive (4WD). These configurations exhibit excellent gradeability and tractive performance, thereby meeting the operational requirements of agricultural machinery in most environments.

Wheeled agricultural machinery predominantly utilizes pneumatic rubber tires, which offer exceptional damping and vibration absorption capabilities. However, when operating on unpaved surfaces, inadequate traction at the wheel–ground interface often results in wheel slip, significantly undermining obstacle-crossing performance and terrain adaptability. This issue becomes particularly critical in hilly and mountainous regions, where unstable traction heightens the risk of lateral rollover. Ben Hazem et al. [91] proposed a velocity–acceleration bivariate neural fuzzy friction model (NFFEM) that integrates fuzzy logic rules with radial basis function (RBF) neural networks. This approach overcomes the limitations of conventional single-variable velocity models, achieving a 94.55% reduction in positioning error for nonlinear friction models (NFNLFMs). The study introduces a novel paradigm for agricultural chassis development: (1) the dual-variable framework accurately captures wheel slip dynamics; (2) high-precision friction estimation optimizes traction control and energy consumption; and (3) the adaptive neural fuzzy mechanism enhances chassis trafficability and energy efficiency in complex farmland environments, providing an algorithmic foundation for intelligent anti-slip control systems in agricultural machinery. Ahamdi et al. [92] developed a dynamic model to address tractor rollover and stability issues in hilly terrain. By incorporating key parameters such as forward velocity ($V_f$), terrain slope ($\theta$), and wheel–ground friction coefficient ($\mu$), the research quantitatively analyzed the sensitivity of μ to both slip stability indices and overall stability indices under varying slope and speed conditions. The results indicate that, in both dry and wet grassland conditions, slip-induced instability has a greater impact on overall stability than rollover instability, particularly under low-friction scenarios where variations in slope significantly increase the susceptibility to slip-induced instability. Franceschetti et al. [93] established a dynamics model of articulated tractors by employing a virtual pivot point to simulate the relative rotation between front and rear wheels, thereby deriving the conditions for lateral stability. Li et al. [94] developed a three-dimensional nonlinear dynamic model that incorporates critical rollover criteria based on abrupt changes in tire–ground contact forces, revealing how operational speed affects both working performance and lateral stability. Li et al. [95,96,97] conducted a systematic investigation into the influence of front tire configuration, counterweight distribution, and rear track width on lateral stability mechanisms. Their results demonstrated that an increased track width ratio enhances tire–ground contact stability, while a greater wheelbase ratio improves disturbance recovery capability, with no significant impact on vibration characteristics. Qin et al. [98,99] integrated a single-axis momentum flywheel system with an active steering system to dynamically adjust tractor posture, effectively reducing rollover risks. Wang et al. [100,101] proposed an active anti-roll control strategy based on single-gimbal control moment gyroscopes (SGCMGs) for slope steering scenarios. This method reduced the peak roll angle by 27% and shortened the response time by 35%, significantly expanding the vehicle’s stability domain to meet lateral stability requirements. He et al. [102] introduced an active steering control strategy incorporating a modified critical position-based lateral stability index (RCP) for compact tractors operating in complex terrain. In off-road conditions involving 20° slopes with soil subsidence and obstacles, vertical load redistribution was used to characterize frictional constraints, resulting in a 35.98% reduction in lateral instability duration and contributing to the theoretical foundation for active safety technologies in agricultural equipment. Song et al. [103] implemented an active steering system based on sliding mode control. The control algorithm was designed using variable structure theory, incorporating dynamic posture-dependent lateral stability indices to achieve adaptive torque distribution.

Anti-rollover technologies for wheeled agricultural machinery mainly focus on four key areas: dynamic modeling, structural optimization, active and passive safety systems, and verification methodologies. Dynamic models allow for the quantitative assessment of stability boundaries, while parametric structural optimization improves the inherent anti-overturning capability. Furthermore, active control systems enable real-time regulation of stability performance. Existing research indicates that the integration of wheel–terrain interaction friction prediction algorithms with multi-modal control strategies can significantly reduce critical rollover risks. As smart agriculture continues to advance and autonomous driving technologies become more widely adopted, new challenges arise in the development of anti-rollover control systems. These challenges include enhancing motion control accuracy under complex wheel–terrain friction conditions, maintaining dynamic stability, and ensuring operational comfort during high-precision tasks.

3.2. Tracked Chassis Systems

Tracked chassis exhibit superior terrain adaptability compared to conventional wheeled chassis, primarily due to their larger contact area and reduced ground pressure. These characteristics effectively prevent vehicle slippage and subsidence in soft soil conditions while enabling the successful traversal of ditches and rocky obstacles, thereby minimizing operational downtime. Furthermore, the compact structure and lower center of gravity inherent in tracked chassis significantly reduce the risk of rollover, making them particularly suitable for lateral slope operations and uneven terrains commonly found in hilly regions. As a result, tracked chassis have been widely adopted in agricultural machinery and orchard management equipment.

International enterprises and research institutions maintain a technological lead in the development of tracked chassis. Through continuous advancements in suspension system design, modular power configurations, and intelligent control technologies, their tracked tractors and combine harvesters demonstrate excellent ground traction and terrain adaptability. As illustrated in Figure 9a, John Deere’s 9470RX quad-track tractor incorporates an enhanced drivetrain and optimized chassis design that improves traction performance on moist or loose soils while reducing soil compaction. Similarly, as shown in Figure 9b, Case IH’s Steiger series quad-track tractor features articulated steering and a continuously variable transmission (CVT), supported by a robust powertrain and high-strength steel frame, ensuring reliable performance during continuous heavy-load operations across diverse terrains. However, these large-scale and high-cost systems are not well suited for China’s fragmented farmlands in hilly areas with steep topography.

The development of tracked agricultural machinery in China is currently driven by a combination of technological advancements, market expansion, and supportive government policies. The domestic industry is primarily led by major enterprises such as Weichai Lovol, Dongfanghong, and Wode Agricultural Equipment, with additional contributions from innovative small and medium-sized enterprises (SMEs), including Liying Machinery and O’Power Dynamics. The product range covers a variety of machinery types, including tractors, rotary tillers, and transport vehicles, with ongoing technological improvements focusing on intelligent systems and eco-friendly innovations. Representative models of Chinese-developed tracked agricultural machinery are illustrated in Figure 10.

Although tracked chassis technology has been widely implemented in modern agricultural machinery for plowing, planting, crop management, and harvesting operations, research on tracked agricultural systems remains active, particularly in the areas of chassis stability enhancement, soil compaction mitigation, and turning radius optimization. To address issues such as slippage, subsidence, and instability under wet conditions, Yuan et al. [104] introduced bionic principles into ground contact mechanisms by mimicking ostrich foot cavity structures to enhance mechanical interlocking with soil. Their track–soil thrust model and adhesion regression equation established a cross-scale analytical framework for wheel–soil friction modeling, revealing the interaction effects between soil moisture content and structural parameters. This work laid a theoretical foundation for improving traction performance on wet terrain. Inoue et al. [105] investigated the nonlinear dynamic interactions between rollers and rubber tracks. They found that high-frequency excitation increases dynamic spring constants while decreasing viscous damping coefficients, thereby confirming the necessity of incorporating nonlinear characteristics for accurate system dynamics analysis. Tang et al. [106] developed a standardized driver posture model to evaluate the structural integrity and load-bearing capacity of tracked harvesters. The experimental results indicated stable operation under 3.5-ton loads across speed ranges of 1 m/s, 1.5 m/s, and 2 m/s. Keller et al. [107] proposed a vertical stress distribution model at the interface between rubber tracks and soil, which improved predictions of soil stress and compaction risks for tracked agricultural vehicles. Kormanek et al. [108] conducted field measurements and simulation analyses of the MHT 8002HV tracked harvester, identifying that average ground pressure remained ≤70 kPa. Dynamic load, effective contact length, and track tension were identified as key influencing factors, offering insights for balancing chassis stability with soil protection. Tang et al. [109] explored both straight-line and in situ steering principles of conventional single-side braking gearboxes to reduce soil compaction and residue accumulation. Based on this analysis, they designed and manufactured a bidirectional steering gearbox prototype, which demonstrated smooth steering performance without soil buildup. Liu et al. [110] analyzed the mechanisms and influencing factors of soil compaction caused by tracked combine harvesters operating in waterlogged paddy fields. Their findings indicated that increasing travel speed effectively reduces compaction risks in rice stubble soils. Jing et al. [111] conducted multi-speed compaction tests and revealed that self-excited vibrations from tracked harvesters induce soil “bouncing phenomena,” leading to increased structural damage in subsurface soil layers. Wang et al. [112] proposed a novel integrated driveline-steering chassis design for tracked harvesters to overcome the limitations of conventional hydrostatic transmission systems, such as low transmission efficiency and large turning radii. They established a track–ground slippage dynamics model (${\sigma }_1={\displaystyle \frac{S_S-S}{S_S}},{\sigma }_2={\displaystyle \frac{{S-S}_S}{S}}$, ${\sigma }_1$, ${\sigma }_2$ denote the slip ratios of the low-speed side and high-speed side tracks, respectively. $S_S$ represents the actual travel distance (m) of the tracks, while $S$ indicates the theoretical travel distance (m) of the tracks) combined with an empirical steering parameter correction model. By quantifying actual displacement deviations between the inner and outer tracks, their study revealed the direct influence of soil shear deformation on steering accuracy. Yu et al. [113] developed a self-propelled tracked broccoli harvester that exhibits strong terrain adaptability and superior obstacle-crossing performance during field operations.

In addition to conventional full-track chassis, the wheel–track hybrid chassis has emerged as a distinctive locomotion system that integrates the advantages of both tracked and wheeled configurations, playing an increasingly important role in modern agricultural production. This hybrid design typically incorporates traditional wheels on the front axle and track assemblies on the rear axle. Such a configuration enables agricultural machinery to maintain sufficient mobility for traversing soft, muddy, and uneven terrains while preserving the steering agility and operational convenience characteristic of wheeled systems. Neto et al. [114] evaluated the seeding performance of conventional dual-wheel tractors compared with semi-track configurations. Field experiments demonstrated that the rear semi-track system (front wheels + rear tracks) offered significant improvements over the traditional dual-wheel setup (tires on both axles) during soybean planting operations. Grazioso et al. [115] proposed a multi-body dynamics modeling approach to analyze the interaction between flexible tracks and soft soil. Their method quantifies the collision forces between track links and mechanical components using spring-damper elements, with a focus on developing soil-friendly designs for reconfigurable locomotion systems capable of switching between wheeled and semi-tracked modes. This research established a numerical simulation platform for parametric design of locomotion systems. Zhou et al. [116] constructed a high-precision multi-body dynamics model for articulated-steering semi-tracked tractors. For the rear-mounted track system, they formulated a ground pressure distribution model ($P_i={\displaystyle \frac{G_i}{dxdy}}={\displaystyle \frac{m_{2}g}{2dxdy}}$, $P_i$ is the grounding pressure on the tracked micro elements, $\mathrm{N}/{\mathrm{m}}^2$; $G_i$ is the weight of each half-track walking mechanism, $N$; g is acceleration of gravity, $\mathrm{N}/\mathrm{k}\mathrm{g}$; $m_2$ is the mass of the rear half car body, $\mathrm{k}\mathrm{g}$; dx is the micro element in the length direction of the part of the track contacting the ground, m; dy is the micro element in the width direction of the part of the track contacting the ground, m) and a micro-element friction resistance model, enabling quantitative evaluation of steering resistance torque. With respect to front wheel and track–ground interactions, contact-friction constraints were defined along with specified road surface friction coefficients, thereby establishing a comprehensive framework for quantitative dynamic analysis under complex operating conditions. Importantly, their methodology for road excitation reconstruction and parameter sensitivity analysis exhibits strong adaptability and can be extended to the development of other specialized agricultural machinery systems.

The effective implementation of tracked undercarriage systems holds significant potential for mitigating soil compaction induced by agricultural machinery, serving as a key technological approach to promote high-standard farmland development, ensure food security, and enhance farmers’ economic benefits. In recent years, China has developed a foundational research capacity in tracked undercarriage technologies, with relevant institutions proposing systematic strategies for structural optimization, scenario-specific adaptation, and large-scale deployment. Nevertheless, current tracked agricultural machinery still encounters major technical challenges across cultivation, planting, management, and harvesting operations: Firstly, the adaptability to complex field conditions—such as waterlogged clay soils and sloped terrain—remains insufficient. This limitation arises from inadequate modeling of chassis stability, wheel–soil interaction dynamics, and real-time soil response mechanisms, which calls for improvements in operational reliability and system robustness. Secondly, the integration level of intelligent control systems is relatively low, particularly concerning the synergistic integration of steering dynamics, accurate slippage compensation algorithms, and nonlinear friction damping regulation, thereby limiting precision operation performance and industrial scalability. Future research on critical tracked undercarriage technologies should focus on four core areas: (1) stability theory and autonomous control systems; (2) powertrain optimization and advanced steering mechanisms; (3) autonomous navigation combined with intelligent management systems; and (4) track–soil interaction modeling. This multi-disciplinary strategy aims to establish a comprehensive technology framework tailored to complex farmland environments, offering theoretical foundations and technical paradigms to support the intelligent transformation of modern agricultural equipment.

4. Chassis Leveling Systems of Agricultural Machinery

In agricultural production environments, farmland terrain often presents complex and variable topographies. Agricultural machinery, including combine harvesters, sprayers, and seeders, frequently experiences tilt during field operations due to uneven ground surfaces, slope gradients, and varying degrees of soil compaction. This tilting not only degrades operational performance—evidenced by uneven seeding, incomplete spray coverage, and increased harvesting losses—but also accelerates mechanical wear, ultimately shortening equipment service life. As a result, extensive research has been conducted globally on active leveling chassis systems, yielding significant advancements in the areas of leveling mechanism design, suspension system optimization, control strategy development, and scenario adaptability [117,118,119,120,121]. A standard active leveling chassis system typically consists of three key components: sensor networks, hydraulic actuators, and an intelligent control module. The system continuously monitors vehicle posture using inclination sensors and dynamically adjusts hydraulic cylinder displacements through an intelligent control unit. This unit integrates a chassis dynamics model that accurately captures the interactions between the machinery, terrain, and load under multi-degree-of-freedom motion, thereby enabling precise regulation of both lateral and longitudinal balance. Currently, such systems are predominantly applied in harvesting machinery, crop protection equipment, and orchard-specific vehicles, showing strong performance in challenging environments such as hilly and mountainous regions and orchard settings.

4.1. Mechanisms and Methods for Leveling

Leveling mechanisms for agricultural machinery can be categorized into five primary types based on their operational principles: hydraulic differential-height type, parallel four-bar linkage type, adjustable center-of-gravity (CoG) type, articulated frame steering type, and omnidirectional leveling type. These mechanisms enable lateral, longitudinal, and full-range adjustments in vehicle posture.

Table 3. Comparison of leveling mechanisms in agricultural machinery chassis.

Leveling Mechanism	Core Advantages	Performance Limitations	Performance Indicators	Suitable Application Scenarios
Hydraulic Differential Height Type	Excellent contour operation performance; simple structural principle; good slope adaptability	Poor lateral stability on uphill slopes; ground adhesion loss > 15%	Maximum climbing angle: 20–25°; energy consumption per unit mass: 3.2–4.0 kJ/kg·km; leveling error: <1.5°; response time: 0.8–1.2 s	Sloping terrain with gentle undulations; contouring on sloping terrain
Parallel Four-Bar Linkage Type	Simple structure; low failure rate	Limited leveling freedom (single axis); slow dynamic response	Maximum climbing angle: 15–20°; energy consumption per unit mass: 2.5–3.2 kJ/kg·km; leveling error: 2–3°; response time: >2 s;	Sloping terrain with gentle undulations; contouring on sloping terrain
Adjustable Center of Gravity Type	High traction efficiency; good uphill stability; excellent slope adaptability	Complex structure; requires multi-layer chassis	maximum climbing angle: 35–40°; energy consumption per unit mass: 2.0–2.8 kJ/kg·km; leveling error: <0.8°; response time: 0.5–0.9 s	Steep slope ascent/descent operations; steep slope surface undulations
Articulated-Torsion Type	High adaptability to rugged terrain; small turning radius; high flexibility; compact structure	Unsuitable for large-gradient slopes; high cost; complex structure; requires skilled operators	Maximum climbing angle: 25–30°; energy consumption per unit mass: 2.8–3.5 kJ/kg·km; leveling error: 1.2–2°; response time: 0.4–0.7 s	Rugged slope surfaces; small fragmented plots; gentle slope surface undulations
Omnidirectional Leveling Type	High traction efficiency; superior uphill stability; outstanding slope adaptability	High cost; complex maintenance;	Maximum climbing angle: 40°+; energy consumption per unit mass: 1.8–2.4 kJ/kg·km; leveling error: <0.5°; response time: <0.3 s	Steep slope ascent/descent operations; steep slope surface undulations; rugged slope surfaces

provides a comparative overview of the key characteristics among these different leveling systems.

Internationally developed leveling technologies continue to lead in terms of maturity and application scope compared to domestic Chinese systems, particularly in large-scale agricultural equipment. The VTT Technical Research Centre of Finland has developed a closed-loop leveling system utilizing servo proportional valves. This system achieves synchronized lateral and longitudinal leveling through four-stage linked hydraulic cylinders, with a maximum leveling angle of ±15° and a response time under 0.6 s, resulting in a 12% improvement in operational efficiency. John Deere’s 8R-3004 tractor (as shown in Figure 11a) integrates a distributed hydraulic suspension system controlled by load-sensitive pumps and pressure-compensated valves, effectively mitigating header vibration caused by uneven terrain. New Holland’s CR9070 mountain tractor (Figure 11b) employs a dynamic lateral leveling system that combines dual rack-and-pinion mechanisms with worm gear self-locking devices, enabling millimeter-level height adjustment during operation. However, its relatively large structural dimensions limit its suitability for hilly and mountainous regions in China. Ferrari’s K10s mountain tractor (Figure 11c) features a low-center-of-gravity design (with a center of gravity height ≤ 650 mm), incorporating multi-degree-of-freedom lateral swing mechanisms on the front axle. It achieves ± 15° dynamic tilt compensation through a hydraulic servo system. Balchanowski et al. [122] proposed a self-leveling solution for legged-wheel robots operating in complex terrains using a 22-degree-of-freedom multi-body dynamics model (six degrees of freedom for the chassis and four for each suspension unit). Their approach integrates linear actuators capable of ±300 mm vertical wheel adjustment with real-time inclination feedback, laying the foundation for robotic mechanism optimization and control algorithm development. Dettù et al. [123] introduced a dual cascaded control architecture for combine harvesters based on gray-box modeling. By applying third-order transfer functions to describe the asymmetric dynamics of hydraulic actuators and establishing posture-stroke mapping, they revealed geometric coupling mechanisms between roll angle (dependent on stroke difference) and pitch angle (dependent on stroke sum), enabling coordinated stability control. Pijuan et al. [124] designed a multi-degree-of-freedom suspension topology system that demonstrates strong terrain adaptability even under imprecise leveling constraints.

Domestic research on leveling machinery has made significant progress in structural design and optimization. As illustrated in Figure 12a, the crawler-type leveling tractor developed by Chongqing Xinyuan Agricultural Machinery incorporates a four-point hydraulic leveling mechanism that dynamically balances the vehicle by adjusting the contact area of the tracks in real time. Zoomlion’s PL2304 paddy field tractor (Figure 12b) utilizes coordinated servo motor and planetary roller screw actuation, achieving a leveling response time of ≤0.8 s. This system enables dynamic adjustment of support leg extensions while maintaining a maximum leveling error of ≤0.8° during paddy field operations. The DF404-5A orchard management machine developed by Changzhou Dongfeng Agricultural Machinery employs a bionic quadruped hydraulic leveling chassis, which realizes vertical compensation based on pressure sensor data and vehicle mass distribution models. Shandong Wuzheng Group, in collaboration with Shandong Agricultural University, has developed a torsion-frame mountain tractor (Figure 13) that demonstrates enhanced adaptability to sloped terrain through frame torsional articulation [125]. Meanwhile, Professor Yang Fuzeng’s research team at Northwest A&F University has developed an omnidirectional-leveling unmanned crawler tractor for use in mountainous areas. This system integrates parallel four-bar linkages with dual-frame mechanisms to achieve full-range leveling, significantly improving operational stability and safety on inclined surfaces. The working principle is schematically presented in Figure 14 [126].

Domestic scholars have made significant advancements in the design of leveling mechanisms. To address issues such as excessive machine tilt, reduced operational quality, and safety hazards during operations in hilly and mountainous terrain, Wang et al. [127] developed an omnidirectional automatic leveling system for tracked machinery based on a three-layer articulated frame structure (Figure 15). This system enables posture adjustment across pitch, roll, and compound inclination conditions. It also incorporates a sliding mode synchronous control method enhanced with disturbance observers to minimize synchronization errors among hydraulic cylinders and improve leveling accuracy. Sun et al. [128] proposed an omnidirectional leveling approach using a four-point adjustable hydraulic chassis for tracked combine harvesters operating in complex environments. Through multi-body dynamics simulations conducted on the ADAMS virtual prototype platform under slope climbing, lateral tilt, and compound posture scenarios, the system demonstrates a notable reduction in rollover risks caused by center-of-gravity (CoG) shifts, achieved through coordinated control strategies involving four hydraulic cylinders. Chai et al. [129] designed a legged-wheel unmanned chassis inspired by locust biomechanics, incorporating pneumatic–hydraulic hybrid actuation. The leveling control system integrates dual-axis (pitch/roll) posture sensing, virtual leg coordination control, and center-of-gravity dynamic compensation algorithms. As a result, the system achieves significantly improved leveling speed and precision, outperforming conventional systems. Tan et al. [130] addressed the risk of tipping during slope operations of self-propelled corn harvesters by developing a dynamic balancing control strategy based on a dual-degree-of-freedom (DoF) leveling mechanism. The system integrates lateral (−10–17°) and longitudinal (±43.1°) hydraulic modules into a multi-body structure containing 24 kinematic pairs. A geometric kinematic model was established to correlate cylinder displacement with machine posture, enabling full-range compensation under four-quadrant compound tilt conditions. Sun et al. [131] introduced an active leveling system based on parallel four-bar linkages. Through modeling of tractor lateral slip, they identified imbalanced lateral torque and non-uniform ground pressure distribution as the primary causes of sideslip. Gao et al. [132] developed an H-configuration variable-geometry power chassis tailored for mountainous agricultural applications. This system integrates balance rocker suspensions, hydraulic adjustment systems, and dual-drive axles to enable real-time adaptation of ground clearance, track width, and chassis posture.

In summary, continuous design and optimization of chassis leveling mechanisms can significantly improve the adaptability of agricultural machinery in complex terrains such as hilly and mountainous regions. Although considerable progress has been made in existing research and widely applied in operational equipment, current systems still face challenges related to insufficient adaptability, limited intelligent capabilities, and inadequate reliability under complex working conditions—particularly in areas such as multi-dimensional coordinated leveling, lightweight structural design, and high-precision dynamic response. Future research should focus on three key directions: (1) conducting strength analysis and structural optimization of core components; (2) developing modularized lightweight chassis systems that consider dynamic operational characteristics; and (3) designing intelligent leveling mechanisms with adaptive compensation functions. These improvements require optimizing the kinematic models and control logic of actuation systems to shorten posture adjustment cycles and achieve millisecond-level real-time feedback control. Moreover, it is essential to establish a comprehensive multi-dimensional testing framework that covers typical agricultural scenarios. This framework should rigorously assess system performance in terms of durability during continuous operation, stability under extreme loads, and human–machine interaction ergonomics, thereby ensuring the reliability necessary for the engineering application of leveling systems.

4.2. Control Systems of Leveling

In the field of slope-adaptive leveling technology for agricultural machinery, the dynamic stability and response speed of leveling control systems are recognized as key performance evaluation metrics.

Compared with domestic developments, international automatic leveling systems—particularly those in the United States, Western Europe, and Japan—have transitioned from purely mechanical adjustments to intelligent, high-precision systems through the integration of satellite positioning, sensor networks, and automated control algorithms. These technological advancements have significantly improved operational accuracy, efficiency, and terrain adaptability. CLAAS implemented a model predictive controller (MPC) in the LEXION 8900 harvester (Figure 16a), which generates real-time leveling commands by solving vehicle dynamics equations. This system is further enhanced by crawler-mounted spread spectrum satellite corrections, enabling it to operate effectively on slopes up to 30°, with an 18% reduction in energy consumption compared to traditional PID controllers. The John Deere S790 combine harvester (Figure 16b) integrates dual-axis tilt sensors and ground radar elevation scanners to dynamically detect terrain variations. Using model predictive control (MPC), the system adjusts header inclination based on forward travel speed, achieving a leveling response time of less than 0.5 s. The Case IH Axial-Flow 250 series harvester (Figure 16c) features a maximum leveling range of ±6°, capable of correcting body tilt within 0.8 s with an error margin of less than 0.3°. AGCO’s Fendt IDEAL 10T combines electromagnetic active suspension with synchronized adjustment of threshing drum tilt and header height, reducing vibration amplitude by 70% and completing leveling operations within 1.2 s (Figure 16d). Denis et al. [133] proposed a dynamic stability monitoring method based on an online adaptive observer, effectively addressing the limitations of conventional road dynamics approaches when applied to unstructured terrains. This approach provides a novel technical pathway for assessing rollover risks during off-road agricultural operations. Gonzalez et al. [134] developed HydraROPS, a dual-mode automatic rollover protection system that employs a hydraulically driven mechatronic architecture to regulate lateral posture and enhance safety during complex terrain operations.

Current control strategies for agricultural machinery leveling in China primarily rely on PID-based frameworks, encompassing classical PID, incremental PID, and various enhanced variants. Representative approaches include fuzzy PID (dynamically adjusting parameters via fuzzy logic to address nonlinear disturbances), dual-loop fuzzy PID (improving disturbance rejection through position-velocity dual feedback), and BP neural network PID (optimizing control parameters through neural network self-learning). A limited number of studies have explored advanced methods such as sliding mode control and model predictive control (MPC).

Table 4. Comparison of agricultural machinery chassis leveling control strategies.

Control Strategy	Core Advantages	Performance Defects	Performance Indicators
Classic PID Control	Simple implementation; flexible adjustment	Poor nonlinear adaptability; weak disturbance rejection	Max slope: 20–25°; energy consumption: 3.0–4.2 kJ/kg·km; response time: 300–500 ms; overshoot: 15–25%
Incremental PID Control	Strong anti-interference ability; easy to switch between automatic and manual modes	Poor adaptability to time-varying systems; high computational burden	Maximum slope: 22–27°; energy consumption per unit mass: 2.8–3.8 kJ/kg·km; response time: 250–400 ms; overshoot: 10–20%
Fuzzy PID Control	Strong adaptability; good robustness; high control accuracy	Complex rule-based design; limited real-time performance	Maximum slope: 28–33°; energy consumption per unit mass: 2.3–3.0 kJ/kg·km; response time: 150–250 ms; overshoot: 5–12%
Dual-Loop Fuzzy PID Control	Excellent stability; rapid response speed	Difficult to design parameters; over-regulation problem; steady-state error problem	Maximum slope: 30–35°; energy consumption per unit mass: 2.0–2.7 kJ/kg·km; response time: 80–150 ms; overshoot: 3–8%
Model Predictive Control (MPC)	Multi-objective optimization; strong constraint handling	High computational complexity; sensitive to model	Maximum slope: 35–40°; energy consumption per unit mass: 1.7–2.3 kJ/kg·km; response time: 200–350 ms; overshoot: 1–5%
Sliding Mode Control (SMC)	Fast response; strong robustness; strong anti-interference ability	High-frequency chattering	Maximum slope: 38°+; energy consumption per unit mass: 1.8–2.5 kJ/kg·km; response time: 40–100 ms; overshoot: 0–3%

presents a comparative analysis of the advantages and disadvantages of different PID control strategies. Qi et al. [135] proposed an automatic leveling control method based on a dual-loop fuzzy PID algorithm ($u\left(t\right)=k_{p}e\left(t\right)+k_i{\int }_0^{t}e\left(t\right)dt+k_d{\displaystyle \frac{de\left(t\right)}{dt}}, k_p=k_{p0}+∆k_p,k_i=k_{i0}+∆k_i,k_d=k_{d0}+∆k_d$ In the formula, $u\left(t\right)$ represents the output of the controller; $e\left(t\right)$ represents the input of the controller; $k_p$ is the proportional gain; $k_i$ is the integral gain; $k_d$ is the derivative gain; $k_{p0}{,k}_{i0},$ and $k_{d0}$ are the initial values of the PID parameters) to adjust the vehicle posture (Figure 17). Under identical PID parameters, this approach outperformed conventional dual-loop PID control by effectively reducing overshoot and leveling time. Yang et al. [136] developed a mountain crawler tractor that applied separate PID and fuzzy PID control algorithms for chassis and implement leveling. The Simulink-based simulation model demonstrated rapid response and stable coordinated control, with its architecture illustrated in Figure 18. Jiang et al. [137] introduced the QBP–PID hybrid algorithm, which integrates Q-learning reinforcement mechanisms, BP neural networks, and PID control architecture to achieve online neural network weight optimization and real-time parameter tuning. Chen et al. [138] proposed a single-neuron adaptive PID control algorithm that reduced response time by 32% compared to conventional PID, effectively resolving height instability during abrupt load changes in agricultural vehicles. Wang et al. [139] presented a BP neural-network-based automatic leveling method (Figure 19), demonstrating superior performance in stabilization time, maximum tilt overshoot, and steady-state error compared to PID control. Zhao et al. [140] proposed an adaptive dual-loop leveling system incorporating active suspension, employing composite dual-loop control and optimized adaptive genetic algorithms to enhance chassis leveling performance across diverse road conditions. Peng et al. [141] developed a sliding mode variable structure control algorithm with fuzzy switching gain regulation for four-point leveling mechanisms, achieving real-time dynamic leveling with improved precision and stability.

In summary, significant progress has been achieved in the research on agricultural machinery leveling control systems both domestically and internationally; however, distinct developmental patterns are evident between regions. Overseas technologies, particularly those in Europe, the United States, and Japan, have progressed to intelligent and high-precision stages. These systems demonstrate improved dynamic response, terrain adaptability, and energy efficiency through multi-sensor fusion, satellite positioning corrections, and advanced control algorithms. In contrast, domestic research primarily relies on PID-based frameworks, focusing on algorithmic enhancements to improve disturbance rejection and response speed. Although advancements have been made in overshoot suppression and parameter self-adaptation, the application of advanced strategies such as model predictive control (MPC) and sliding mode control remains at an exploratory stage, with limited integration of hardware systems like satellite positioning and active suspension. Globally, scholars have conducted pioneering research on intelligent algorithm integration and dynamic stability monitoring for unstructured terrains. However, China still faces notable technological gaps in engineering implementation and coordinated electro-hydraulic control. As operational requirements in complex terrains continue to grow, emerging intelligent control strategies—including sliding mode variable structure control, model predictive control (MPC), reinforcement learning control, and adaptive robust control—are anticipated to become key research areas within domestic studies. Furthermore, multi-modal control architectures—such as hybrid decision layers combining MPC with reinforcement learning and execution layers integrating adaptive robust control with fuzzy logic—will drive future technological convergence. Simultaneously, the development of high-precision domestic MEMS attitude sensors and low-latency hydraulic servo valves is crucial to support the practical implementation of advanced algorithms. This will gradually reduce the performance gap with international standards in comprehensive leveling capabilities under complex operating conditions.

5. Summary and Future Perspectives

The advancement of agricultural mechanization in China’s hilly and mountainous regions constitutes a critical foundation for achieving agricultural and rural modernization. In recent years, driven by national policy incentives and research initiatives, significant breakthroughs have been made in the development of intelligent agricultural machinery tailored for these challenging terrains. Currently, the technology underlying agricultural machinery chassis designed for hilly areas is transitioning from traditional mechanical drives to intelligent systems characterized by high adaptability. Core challenges include addressing coupling issues related to power, structure, and control within complex terrain environments.

Existing equipment and operational techniques remain largely experimental, indicating substantial opportunities for improvement in constructing reliable, comfortable, safe, and universally applicable chassis systems. To establish an intelligent chassis framework that can adapt to small plots, steep slopes (exceeding 25°), and obstacle-rich terrains while facilitating efficient low-carbon agricultural mechanization, future research should prioritize three key directions:

(1)

Innovative Multi-Modal Mobility Mechanisms: Enhance adaptability to wet and cohesive soils through terrain-adaptive structural optimization coupled with soil-friendly mobility design.

(2)

Cooperative Control Algorithm Optimization: Improve power distribution efficiency by 15–20% through energy-driven approaches alongside energy efficiency optimization.

(3)

Integrated Intelligent Sensing Systems: Upgrade leveling systems utilizing dual-loop fuzzy PID controllers combined with sliding mode observer technologies to achieve high precision and adaptability; develop specialized MEMS tilt sensors along with multi-source fusion positioning modules specifically designed for hilly terrain; establish real-time mapping models of terrain–load–energy consumption aimed at mitigating chassis instability on slopes exceeding 25°.

By integrating structural optimization with cooperative control algorithms, an intelligent platform will be developed to create efficient, smart, and low-carbon chassis systems. This initiative aims to accelerate the modernization of agricultural machinery in hilly and mountainous regions. The detailed research framework is illustrated in Figure 20.

5.1. Innovative Multi-Modal Mobility Mechanisms

Complex terrains in hilly and mountainous regions are characterized by significant slope differentials, irregular surface morphology, and high-adhesion soil media properties. These factors create systemic challenges for conventional rigid chassis systems, including limited terrain adaptability, compromised dynamic stability, and reduced operational precision. Therefore, the development of agricultural machinery specifically designed for these unique agroecological contexts must go beyond traditional architectural constraints. This necessitates the establishment of an adaptive chassis technology framework that focuses on innovations in multi-modal mobility mechanisms.

5.1.1. Terrain-Adaptive Mechanism Optimization

The complex operating conditions in hilly and mountainous regions—characterized by undulating terrain and heterogeneous soft soil—necessitate the optimized design of adaptive mechanisms for agricultural machinery. To address environmental uncertainties, a hybrid modeling approach can be employed that integrates physical dynamics models (e.g., track–soil thrust equations and tire slip ratio models) with empirical terrain data (e.g., 3D point cloud elevation maps and soil moisture distribution). This enables the construction of high-fidelity digital terrain fields, which support predictive decision-making for multi-degree-of-freedom suspension adjustments.

By developing a multi-degree-of-freedom suspension system that incorporates independently adjustable balance mechanisms and wheel–track composite structures, the chassis is capable of dynamically adjusting ground clearance and posture to accommodate variations in slope and gully topography. Slope prediction driven by elevation maps, combined with surface friction models (e.g., adhesion coefficient mapping based on the Mohr–Coulomb criterion), allows for proactive planning of center-of-mass trajectories, thereby reducing rollover risks. Furthermore, real-time fusion of IMU attitude data with digital terrain fields enables closed-loop control for mass-center stabilization, ensuring operational safety on challenging terrains such as slopes and gullies.

To meet both operational requirements and ecological preservation goals in mountainous areas, bionic mechanism optimization is essential. Inspired by biological locomotion principles and integrating soil bearing pressure distribution data with mechanical kinematic characteristics, this approach enhances obstacle negotiation capabilities while minimizing soil compaction damage. The integration of these technologies not only improves operational accuracy and safety but also provides highly adaptable solutions for agricultural mechanization in hilly regions, promoting terrain-adaptive retrofitting and advancing sustainable agricultural development.

5.1.2. Intelligent Mobility Mechanisms and Soil-Friendly Design

The thin and erosion-prone soils commonly found in hilly and mountainous regions necessitate the development of intelligent locomotion mechanisms as a fundamental component of soil-conservative agricultural designs. By integrating multi-source terrain modeling—specifically LiDAR elevation mapping combined with multi-spectral soil hardness detection—a dynamic surface friction model is established to enable autonomous switching strategies for wheel–track chassis systems. This dual-mode system operates in wheeled configuration on low-slope rigid surfaces (friction coefficient $\mu$ > 0.6) to maintain operational efficiency while automatically transitioning to tracked mode on steep clay slopes ($\mu$ < 0.3) to suppress slippage. This approach significantly improves terrain adaptability while simultaneously reducing the risk of soil compaction associated with conventional wheeled systems. Based on this framework, a hybrid-model-driven contact pressure optimization system has been developed. The system integrates physical models—particularly Keller’s track–soil vertical stress distribution model—with real-time soil moisture data to dynamically predict areas at high risk of soil compaction. Using model predictive control (MPC), the hydraulic suspension output is adjusted in real time. The iterative coordination between terrain modeling and physical-based mechanisms further refines the control parameters of the locomotion system, thereby enhancing both operational stability on slopes and energy efficiency. As a result, this integrated methodology effectively achieves the dual objectives of high-efficiency machinery operation and sustainable soil ecosystem conservation.

5.2. Cooperative Control Algorithm Optimization

The optimization of cooperative control algorithms for agricultural machinery operating in hilly and mountainous regions—particularly through energy-driven cooperative control and the deep integration with powertrains to enhance energy efficiency—represents a critical technical pathway for addressing the challenges posed by complex terrain and improving overall performance. In such terrains, frequent variations in slope and fluctuations in load lead to low energy utilization efficiency within conventional powertrains. Traditional mechanical transmissions encounter issues such as gear-shifting impacts and power interruptions when navigating variable-gradient landscapes. Furthermore, the nonlinear disturbances inherent to hilly terrains necessitate control systems that exhibit high robustness. Consequently, there is an urgent need for comprehensive research into cooperative control algorithms to tackle: dynamic energy management under complex operating conditions; multi-physics-coupled powertrain optimization; and terrain- and load-adaptive global optimal control.

5.2.1. Energy-Drive Cooperative Control

Hilly and mountainous regions are characterized by steep terrain gradients, variable soil adhesion, and high operational loads. Traditional agricultural machinery, which typically employs a single-power allocation mechanism and energy source, often experiences wheel slip or power interruptions during operation. To address these challenges, it is essential to integrate power distribution optimization, hybrid energy management systems, and hydraulic–electric hybrid drive systems into cooperative control strategies for energy drives. This integration enhances dynamic response and operational stability while balancing endurance with energy efficiency; it also achieves high precision and strong adaptability.

Currently, significant progress has been made domestically in the application of model predictive control (MPC) algorithms and the development of hybrid powertrains. However, challenges remain, including reliance on imported high-precision sensors and hydraulic components as well as insufficient generalization of algorithms. Looking ahead, breakthroughs in digital twin technology, lightweight edge computing solutions, and domestically developed core components—coupled with policy support and collaboration across the industrial chain—are expected to propel energy-drive cooperative control forward. This advancement will facilitate the transition of agricultural machinery in hilly regions toward full autonomy while enhancing energy efficiency. Ultimately, this evolution aims to improve both agricultural productivity and sustainability.

5.2.2. Deep Integration of Powertrains and Energy-Efficiency Optimization

Agricultural operations in hilly and mountainous regions encounter high energy consumption and significant fluctuations in power demand, rendering conventional single-power-source systems inadequate for meeting energy-efficiency requirements. Current technologies predominantly rely on hybrid powertrains (petrol–electric), which dynamically allocate power between engines and motors through algorithms that recognize operating conditions. However, existing systems still suffer from suboptimal energy management strategies and inefficient power coupling.

Future solutions may involve the development of a multi-level energy system that integrates fuel cells, lithium batteries, and supercapacitors, combined with dynamic energy management strategies to provide instantaneous power compensation under high-torque steep slope conditions. This approach aims to enhance operational endurance. Furthermore, the adoption of advanced materials such as carbon fiber composites can reduce overall machine weight while maintaining structural integrity, thereby improving energy utilization efficiency and increasing the dynamic response speed of control systems—ultimately bolstering stability during slope operations.

With advancements in domestically developed high-precision actuators and ongoing optimization of intelligent algorithms, deeply integrated powertrain technologies are poised to offer robust technical support for agricultural mechanization in hilly regions. This progress will drive the industry toward greater efficiency, intelligence, and sustainability.

5.3. Integrated Intelligent Perception Systems

Integrated intelligent perception systems amalgamate multi-source sensors, edge computing, and artificial intelligence algorithms to establish autonomous decision-making capabilities within smart systems. In the context of agricultural machinery designed for hilly and mountainous regions, these systems are predominantly applied in three key areas:

(1)

Adaptive Leveling Systems: Employing model predictive control (MPC) algorithms to facilitate rapid responses and enhance leveling precision during operations on slopes;

(2)

Intelligent Navigation Systems: Integrating three-dimensional path planning to achieve centimeter-level operational accuracy on sloped terrains;

(3)

Load-Power Matching Systems: Optimizing fuel efficiency in hybrid machinery through adaptive adjustments.

Currently, critical technologies such as LiDAR-based navigation have reached a stage of industrialization, while advanced methodologies like swarm intelligence coordination are still under exploration. This technological framework is propelling the transition of agricultural machinery in hilly regions from traditional mechanization towards intelligent and interconnected operations, thereby providing essential solutions to address agricultural challenges posed by complex terrains.

5.3.1. High-Precision and Adaptive Upgrades for Leveling Systems

Leveling systems are essential components for ensuring the safe operation of agricultural machinery in hilly and mountainous terrains. Their performance has a direct impact on machine stability, operational accuracy, and safety within complex topographies. In environments characterized by steep slope variations and diverse soil conditions, traditional leveling systems often demonstrate slow response times and low precision, which can result in machinery tilt, diminished operational quality, or even rollover accidents.

To address these challenges, research into high-precision adaptive leveling systems emphasizes the optimization of attitude adjustment mechanisms and control algorithms. By incorporating technological innovations such as multi-sensor fusion and advanced control strategies, these enhanced systems achieve millisecond-level response times and ultra-high leveling precision. This significantly improves machinery stability, operational safety, and travel performance in hilly regions.

Such advancements not only ensure operational safety but also enhance the mechanization level and economic efficiency of agricultural production in these areas, thereby laying a foundation for the intelligent development of modern agricultural equipment. However, domestic research in this field remains at an early stage. A comprehensive systematic analysis of machinery dynamics—integrating mathematical principles with mechanical and kinematic theories—is necessary to further advance the development of intelligent and reliable leveling control systems tailored to meet the unique demands posed by hilly terrain.

5.3.2. Construction of Multi-Modal Intelligent Operation Platforms

The construction of multi-modal intelligent operation platforms necessitates the systematic integration of heterogeneous sensors, which include but are not limited to visual sensing systems, acoustic monitoring arrays, tactile feedback networks, and high-precision positioning modules. This integration is complemented by deep learning algorithms and distributed edge computing architectures to establish intelligent agricultural machinery control hubs endowed with autonomous cognitive decision-making capabilities.

In hilly and mountainous agricultural production environments characterized by fragmented operational scenarios and diverse agronomic requirements, traditional single-function machinery systems demonstrate significant limitations in terms of operational adaptability and functional scalability. As a result, the development of integrated platforms that possess multi-modal perception and intelligent decision-making capabilities has emerged as a critical technological pathway for advancing smart agriculture.

Current technical solutions predominantly rely on modular quick-attach architectures; however, they continue to encounter bottlenecks such as insufficient flexibility in functional expansion and suboptimal real-time performance in intelligent decision-making processes. Addressing these challenges necessitates urgent breakthroughs through cross-modal data fusion and comprehensive optimization of adaptive control algorithms.