Search Results (79)

This research presents the design, numerical analysis, and experimental validation of a compact radial turbine intended for mini-turbocharger applications in UAV power systems. To meet the stringent requirements of UAV propulsion—such as lightweight construction, high efficiency at small scales, and stable performance across varying operating altitudes—a test rig was constructed to experimentally estimate turbine torque and shaft power across selected operating conditions. Complementary CFD simulations were performed to evaluate aerodynamic behavior, including flow distribution, torque generation, and power output at multiple rotational speeds matched to experimental mass-flow rates. Additional high-speed CFD simulations were conducted to predict turbine performance in operational regimes typical of UAV engines, where experimental testing is challenging. The combined CFD–experimental methodology provides accurate performance prediction for micro-scale radial turbines across different volute geometries and operating conditions. The results contribute essential insights for the development of next-generation miniaturized turbochargers aimed at enhancing UAV engine efficiency, high-altitude capability, and overall flight endurance. Full article

Precise electromagnetic torque estimation for permanent magnet synchronous motors (PMSMs) is crucial for enhancing the dynamic performance and energy efficiency of electric vehicles. To address the dynamic variations in dq-axis inductance caused by magnetic cross-coupling and saturation effects during motor operation—which lead to significant torque estimation errors in traditional fixed-parameter models under variable torque and speed conditions—this paper proposes a dynamic torque estimation method that integrates online dq-axis inductance identification based on a variable-step adaptive linear neural network (ADALINE) with an extended flux observer. The online identified inductance values are embedded into the extended flux observer in real time, forming a closed-loop torque estimation system with adaptive parameter updating. Experimental results demonstrate that, under complex operating conditions with varying torque and speed, the proposed method maintains electromagnetic torque estimation errors within ±3%, with a convergence time of less than 20 ms, while achieving inductance identification accuracy also within ±3%. These results significantly outperform conventional methods that do not incorporate inductance identification. This study provides a highly adaptive and engineering-practical solution for high-precision torque control of interior permanent magnet synchronous motors (IPMSMs) in automotive applications. Full article

The accurate estimation of helicopter engine power is crucial for ensuring operational performance and maintaining safety. Current methods, such as Maximum Power Checks (MPCs), are effective but resource-intensive and infrequent. This paper presents a novel machine learning-based framework tailored for operational helicopter fleets to estimate Engine Torque Factor (ETF) values from routine flight data obtained via Health and Usage Monitoring Systems (HUMS). The novelty lies in combining a statistically validated labeling strategy that links MPC-derived ETF values to regular flights with a dual-stage preprocessing pipeline, consisting of steady-state filtering and data consolidation, which is designed to produce high-quality, representative training data from noisy operational logs. Regression models, including XGBoost, CatBoost, and Random Forest, were trained and evaluated using HUMS data from AH-64A helicopters. Results demonstrate that focusing on specific ETF ranges significantly improves model performance, achieving R² values of up to 0.94. While the current implementation operates post-flight, the approach enables continuous monitoring between scheduled MPCs, potentially reducing unnecessary checks and providing earlier indications of power degradation. Full article

To meet the safety and compliance requirements pertaining to robots when interacting physically with humans or the environment in unstructured settings such as households and factories, in this study, we focus on methods for the design and control of a flexible robotic magneto-rheological actuator (MRA). Firstly, for the magneto-rheological fluid clutch (MRC), which is the core component of the MRA, an equivalent magnetic circuit model was established to accurately calculate the magnetic field inside the clutch, and a thermal circuit model was constructed to analytically determine the operating temperature of each component. Considering practical engineering constraints, including mechanical structure, magnetic saturation, maximum current, and maximum temperature, a genetic algorithm was used to optimize parameters of the MRC. Secondly, based on the dynamic characteristics of the MRA, a dynamic model incorporating the motor, reducer, MRC, and load link was established. Given scenarios where torque sensors cannot be installed due to cost and structural space limitations, a model reference PID feedforward control strategy was designed. Torque was estimated using input current. Finally, an experimental platform was built, and static and dynamic torque output experiments were conducted. These experiments verified the excellent torque tracking performance of the designed MRA. Through multi-physics modeling, parameter optimization, and control strategy design, this paper provides a solution for flexible robotic joints that integrates high torque, high compliance, and safety. Full article

Accurate estimation of tractor performance under various soil conditions is essential for enhancing operational efficiency in precision agriculture. This study developed machine learning models to estimate tractor performance based on key soil physical properties. Three algorithms—decision tree (DT), CatBoost, and LightGBM—were employed to capture nonlinear relationships between soil parameters and tractor performance indicators. The input variables included soil moisture content, cone index, and particle composition, while the output variables were engine torque, power, slip ratio, and axle power. The models in this study were trained and validated using field data collected from eight paddy fields in Chungcheongnam-do (two in Seosan, two in Cheongyang, and four in Dangjin) and two paddy fields in Gyeonggi-do (Anseong), Republic of Korea. Results showed that models using multiple soil variables significantly outperformed those using single variables. In Model D, CatBoost demonstrated superior performance in predicting engine torque, engine power, slip ratio, and axle power, achieving R² values that were 7.0–14.2% higher than those of DT and 1.6–3.8% higher than those of LightGBM. These findings demonstrate the feasibility of using machine learning with minimal input data to estimate tractor performance, potentially reducing the reliance on extensive physical testing. Full article

With the rapid development of deep-sea resource exploration and marine scientific research, wheeled remotely operated vehicles (ROVs) have become crucial for seabed operations. However, under complex seabed conditions, traditional ROV control systems suffer from insufficient trajectory tracking accuracy, poor disturbance rejection capability, and low dynamic torque distribution efficiency. These issues lead to poor motion stability and high energy consumption on sloped terrains and soft substrates, which limits the effectiveness of deep-sea engineering. To address this, we proposed a comprehensive motion control solution for deep-sea wheeled ROVs. To improve modeling accuracy, a coupled kinematic and dynamic model was developed, together with a body-to-terrain coordinate frame transformation. Based on rigid-body kinematics, three-degree-of-freedom kinematic equations incorporating the slip ratio and sideslip angle were derived. By integrating hydrodynamic effects, seabed reaction forces, the Janosi soil model, and the impact of subsidence depth, a dynamic model that reflects nonlinear wheel–seabed interactions was established. For optimizing disturbance rejection and trajectory tracking, a hierarchical control method was designed. At the kinematic level, an improved model predictive control framework with terminal constraints and quadratic programming was adopted. At the dynamic level, non-singular fast terminal sliding mode control combined with a fixed-time nonlinear observer enabled rapid disturbance estimation. Additionally, a dynamic torque distribution algorithm enhanced traction performance and trajectory tracking accuracy. Full article

A key trade-off persists in the control of permanent magnet synchronous motors (PMSMs): achieving fast finite-time convergence often exacerbates control chattering, while conventional chattering-suppression methods can compromise the system’s dynamic response. The existing literature often addresses these challenges in isolation. The core original contribution of this research lies in proposing a novel robust fuzzy adaptive control scheme that effectively resolves this trade-off through a synergistic design. The contributions are as follows: (1) A novel reaching law is formulated to significantly accelerate error convergence, achieving finite-time stability and improving upon conventional reaching law designs. (2) A super-twisting sliding mode observer is integrated into the control loop, providing accurate real-time estimation of load torque disturbances, which is used for feedforward compensation to drastically improve the system’s disturbance rejection capability. (3) A fuzzy adaptive mechanism is developed to dynamically tune key gains in the sliding mode law. This approach effectively suppresses chattering without sacrificing response speed, enhancing system robustness. (4) The stability and convergence of the proposed controller are rigorously analyzed. Simulations, comparing the proposed method with conventional adaptive sliding mode control (ASMC), demonstrate its marked superiority in control accuracy, transient behavior, and disturbance rejection. This work provides an integrated solution that balances rapidity and smoothness for high-performance motor control, offering significant theoretical and engineering value. Full article

To improve the prediction accuracy of soot load in gasoline particulate filters (GPFs) and the control accuracy during GPF regeneration, this study developed a prediction model to predict the soot mass concentration at the GPF inlet of gasoline direct injection (GDI) engines using advanced machine learning methods. Three machine learning approaches, namely, support vector regression (SVR), deep neural network (DNN), and a Stacking integration model of SVR and DNN, were employed, respectively, to predict the soot mass concentration at the GPF inlet. The input data includes engine speed, torque, ignition timing, throttle valve opening angle, fuel injection pressure, and pulse width. Exhaust gas soot mass concentration at the three-way catalyst (TWC) outlet is obtained by an engine bench test. The results show that the correlation coefficients (R²) of SVR, DNN, and Stacking integration model of SVR and DNN are 0.937, 0.984, and 0.992, respectively, and the prediction ranges of soot mass concentration are 0–0.038 mg/s, 0–0.030 mg/s, and 0–0.07 mg/s, respectively. The distribution, median, and data density of prediction results obtained by the three machine learning approaches fit well with the test results. However, the prediction result of the SVR model is poor when the soot mass concentration exceeds 0.038 mg/s. The median of the prediction result obtained by the DNN model is closer to the test result, specifically for data points in the 25–75% range. However, there are a few negative prediction results in the test dataset due to overfitting. Integrating SVR and DNN models through stacked models extends the predictive range of a single SVR or DNN model while mitigating the overfitting of DNN models. The results of the study can serve as a reference for the development of accurate prediction algorithms to estimate soot loads in GPFs, which in turn can provide some basis for the control of the particulate mass and particle number (PN) emitted from GDI engines. Full article

The series–parallel hybrid powertrain combines the advantages of both series and parallel configurations, offering optimal power performance and fuel efficiency. However, the presence of multiple excitation sources significantly complicates the torsional vibration behavior during engine startup. To accurately identify and analyze the torsional vibration characteristics induced by shaft resonance in this process, a torsional vibration feature identification algorithm based on disturbance observation and parameter partition learning is proposed. A simplified model of the drivetrain shaft system is first established, and an extended state Kalman filter (ESKF) is designed to accurately estimate the torque of the torsional damper. The inclusion of extended disturbance states enhances the model’s robustness against system uncertainties. Subsequently, continuous wavelet transform (CWT) is employed to identify the resonance characteristics in the torsional vibration process from the torque signal. Combined with the parameter partition learning strategy, resonance frequencies are utilized to infer key system parameters. The results demonstrate that, under a 20% perturbation of structural parameters, the observer model with fixed parameters yields a root mean square error (RMSE) of 10.16 N·m for the torsional damper torque. In contrast, incorporating the parameter self-learning algorithm reduces the RMSE to 2.36 N·m, representing an 85.2% improvement in estimation accuracy. Using the Morlet wavelet with a frequency resolution parameter (VPO) of 15 at a 50 Hz sampling rate, the identified resonance frequency was 14.698 Hz, showing a 1.1% deviation from the actual natural frequency of 14.53 Hz. Full article

Aiming at the control challenges faced by unmanned surface vessels (USVs) in complex environments, such as nonlinearities, parameter uncertainties, and environmental perturbations, we propose a non-singular terminal integral sliding mode control strategy based on an extended state observer (ESO). The strategy first employs a third-order linear extended state observer to estimate the total disturbances of the USV system, encompassing both external disturbances and internal nonlinearities. Subsequently, a backstepping sliding mode controller based on the Lyapunov theory is designed to generate the steering torque control commands for the USV. To further enhance the tracking performance of the system, we introduce a non-singular terminal integral sliding mode surface with a double power convergence law and redesign the backstepping sliding mode controller for the USV heading control. Meanwhile, to circumvent the differential explosion issue in traditional backstepping control, we simplify the controller design by utilizing a second-order sliding mode filter to accurately estimate the differential signals of the virtual control quantities. Theoretical analysis and simulation results demonstrate that the proposed control algorithm improves the convergence speed, adaptive ability, and anti-interference ability in complex environments compared to traditional linear backstepping sliding mode control, thereby enhancing its engineering practicability. This research offers a more efficient and reliable method for precise heading control and path tracking of USVs in complex and dynamic environments. Full article

Parameter identification in dynamical systems is a well-known problem with many applications in control design, system monitoring, and fault detection. As these systems are increasingly integrated into complex and demanding environments, challenges such as rapid response, uncertainty handling, and disturbance rejection must be addressed. This paper presents a real-time estimation technique for parameters and load torque in brushless DC (BLDC) motors. These electrical machines are extensively used in engineering applications and often operate under hard conditions. The proposed method is based on algebraic identification, known for its robust performance in both linear and nonlinear systems. In utilizing the mathematical model of a BLDC motor, a set of equations is derived to enable parameter estimation, assuming the availability of input and output measurements in open loop. Moreover, unknown load torque is estimated by approximating the disturbance over a short time window using Taylor series expansion polynomials. The theoretical contribution is analytically validated and is also verified through numerical evaluations revealing the effectiveness of the proposed technique for real-time parameter and disturbance estimation in BLDC motors over other important techniques. Additionally, to address potential peaks in the estimation process, a modification involving an exponent is introduced to mitigate these issues. Full article

Various model-based mass estimation approaches have been discussed for a long time. However, estimation performance often deteriorates in some driving situations and, in particular, slow convergence and excessive overshoot of estimates are a major issue for model-based approaches. Meanwhile, mass estimation approaches using ANN models have recently emerged to propose better solutions, but their usefulness has not been fully investigated. Therefore, this paper presents a vehicle mass estimation strategy using practical supervisory artificial neural networks to achieve more accurate results with better convergence. Here, the perturbed engine torque and vehicle longitudinal acceleration are selected as the inputs of the ANN (instead of the original engine torque and vehicle acceleration), which allows for the faster convergence of estimates with high accuracy; these inputs are existing sensor values already available in the vehicle system. The effectiveness of the proposed ANN approach was verified using simulation and software-in-the-loop simulation (SILS) with field test data, and it was found that the convergence speed of the proposed ANN is almost twice as fast as that of the model-based approach, the accuracy is much better, and the estimation quality is constantly stable without any excessive transient responses. This study will provide further insights into mass estimation using the ANN approach. Full article

Deep neural networks (DNNs) and transfer learning (TL) have been used to improve surface electromyogram (sEMG)-based force estimation. However, prior studies focused mostly on applying TL within one joint, which limits dataset size and diversity. Herein, we investigated cross-joint TL between two upper-limb joints with four DNN architectures using sliding windows. We used two feedforward and two recurrent DNN models with feature engineering and feature learning, respectively. We found that the dependencies between sEMG and force are short-term (<400 ms) and that sliding windows are sufficient to capture them, suggesting that more complicated recurrent structures may not be necessary. Also, using DNN architectures reduced the required sliding window length. A model pre-trained on elbow data was fine-tuned on hand–wrist data, improving force estimation accuracy and reducing the required training data amount. A convolutional neural network with a 391 ms sliding window fine-tuned using 20 s of training data had an error of 6.03 ± 0.49% maximum voluntary torque, which is statistically lower than both our multilayer perceptron model with TL and a linear regression model using 40 s of training data. The success of TL between two distinct joints could help enrich the data available for future deep learning-related studies. Full article

The engineering of actuation systems for active exoskeletons presents a significant challenge due to the stringent demands for mass reduction and compactness, coupled with complex specifications for actuator dynamics and stroke length. This challenge is met with a model-based methodology. Models for human body, exoskeleton and parametric actuation systems are derived and coupled. Beginning with an inverse dynamics human body simulation, loads in human joints are estimated, and the corresponding support torques are derived. Under the assumption of a control law ensuring these support torques, an optimization problem is stated to determine actuation system parameters such as the number of stator coils and number of battery cells. Lastly, results from the optimization are validated using sophisticated models. The methodology is applied to an exemplary exoskeleton and compared to an approach derived from previous studies. Full article

The exhaust gas recirculation (EGR) valve plays an important role in improving engine fuel economy and reducing emissions. In order to improve the positioning accuracy and robustness of the EGR valve under uncertain dynamics and external disturbances, this paper proposes a positioning servo system design for an electromotive (EM) EGR valve based on the Kalman filter. Taking a novel valve driven by a central winding limited angle torque motor (LATM) as the object, we have fully considered the influence of the motor rotor position and load current, as well as the magnetic field saturation and cogging effect, improved the existing LTAM model, and derived accurate torque expression. The parameter uncertainty of the above internal model and the external stochastic disturbance were unified as “total disturbance”, and a Kalman filter-based observer was designed for disturbance estimations and real-time feed-forward compensation. Furthermore, using non-contact magnetic angle measurements to obtain accurate valve position information, a position control model with real-time response and high accuracy was established. Numerous simulated and experimental data show that in the presence of ± 25% plant model parameter fluctuations and random shock-type disturbances, the servo system scheme proposed in this paper achieves a maximum position deviation of 0.3 mm, a repeatability of positioning accuracy after disturbances of 0.01 mm, and a disturbance recovery time of not more than 250 ms. In addition, the above performance is insensitive to the duration of the disturbance, which demonstrates the strong robustness, high accuracy, and excellent dynamic response capability of the proposed design. Full article