JVC-02 Teleoperated Robot: Design, Implementation, and Validation for Assistance in Real Explosive Ordnance Disposal Missions

: Explosive ordnance disposal (EOD) operations are hazardous due to the volatile and sensitive nature of these devices. EOD robots have improved these tasks, but their high cost limits accessibility for security institutions that do not have sufficient funds. This article presents the design, implementation, and validation of a low-cost EOD robot named JVC-02, specifically designed for use in explosive hazardous environments to safeguard the safety of police officers of the Explosives Disposal Unit (UDEX) of Arequipa, Peru. To achieve this goal, the essential requirements for this type of robot were compiled, referencing the capabilities of Rescue Robots from RoboCup. Additionally, the Quality Function Deployment (QFD) methodology was used to identify the needs and requirements of UDEX police officers. Based on this information, a modular approach to robot design was developed, utilizing commercial off-the-shelf components to facilitate maintenance and repair. The JVC-02 was integrated with a 5-DoF manipulator and a two-finger mechanical gripper to perform dexterity tasks, along with a tracked locomotion mechanism, which enables effective movement, and a three-camera vision system to facilitate exploration tasks. Finally, field tests were conducted in real scenarios to evaluate and experimentally validate the capabilities of the JVC-02 robot, assessing its mobility, dexterity, and exploration skills. Additionally, real EOD missions were carried out in which UDEX agents intervened and controlled the robot. The results demonstrate that the JVC-02 robot possesses strong capabilities for real EOD applications, excelling in intuitive operation, low cost, and ease of maintenance.


Introduction 1.Context and Motivations
The handling of explosive devices and suspicious packages has gained significant relevance in recent years due to the ease of obtaining these materials and supplies [1], coupled with the lack of regulations and standards for their safe handling [2][3][4].According to statistics managed by the Explosives Disposal Unit (UDEX) of the Peruvian National Police, the city of Arequipa reported 223 cases of suspicious packages between 2013 and 2020 [3], of which 163 involved objects such as grenades, projectiles, bombs, pyrotechnic products, and deflagration products.Incorrect manipulation of these devices could lead to fatal incidents [5][6][7].Therefore, robotic technologies are critically needed to minimize human exposure to these potentially hazardous situations.
Explosive ordnance disposal (EOD) robots are teleoperated unmanned vehicles that have revolutionized EOD missions, enabling operators to inspect, recognize, handle, and safely transport various dangerous explosive devices or suspicious packages from a safe distance without direct human intervention [8][9][10].These EOD robots are equipped with various tools, including sensors and cameras, and they possess a variety of locomotion mechanisms designed to navigate diverse terrains [11,12] and feature varied configurations in their robotic arms [13,14].Extensive research has been conducted in the development of EOD robot prototypes, resulting in a variety of designs grouped into three categories based on their size and transport capability, following the rescue robot (RR) approach [15][16][17]: man-packable, man-portable, and maxi-sized.

State of the Art of EOD Robots
Man-packable EOD robots weigh less than 20 kg and can be carried by a single agent using a backpack or carry-on suitcase.In this category, unmanned aerial vehicles (UAVs) are employed for EOD activities [18][19][20].However, despite their flexible mobility and quick inspections [21], UAVs have limitations in real EOD operations, including limited payload capacity, short battery life, and vulnerability to weather conditions [22].
Maxi-sized EOD robots weigh over 80 kg and typically require complex transport systems like trailers or external vehicles.Also classified as UGVs, notable examples in this category include KNIGHT manufactured by Transcend Robotics, San Francisco, CA, USA [28] (200 kg) and ANDROS manufactured by Remotec, Clinton, TN, USA [29] (220 kg), both equipped with a 7-DoF manipulator and employing wheel-based locomotion.Trackbased locomotion maxi-sized robots, such as CALIBER manufactured by ICOR Technology, Ottawa, ON, Canada [30] (89 kg with a manipulator featuring two robotic arms), TALON manufactured by QinetiQ North America, Waltham, MA, USA [31] (81 kg), and HARRIS T7 manufactured by L3Harris Technologies, Melbourne, FL, MA, USA [32] (322 kg) with a 7-DoF manipulator, are also prominent.Notably, HARRIS T7 features a haptic control system providing tactile feedback to operators.Despite their capabilities [28][29][30][31][32], the high market cost of maxi-sized EOD robots restricts accessibility for many security institutions, a crucial factor to consider when evaluating adoption in real-world environments.
Table 1 presents a detailed comparison of the different EOD robotic systems mentioned previously, highlighting their main features and approximate prices.  1 shows the various robots with oUSAding performance in carrying out EOD missions.However, it is crucial to highlight that access to these robots is restricted due to their considerable costs, which could hinder their adoption by security institutions with limited funds.In addition to the high costs, the complexity associated with repairing these robots must be considered.In the event of failure or damage during an EOD mission, the robotic system must be sent to the manufacturer for repair, which can also involve considerable costs and prolonged downtime.

Related Works on Low-Cost EOD Robots
Research has been conducted to address the design of low-cost EOD robots [33][34][35][36][37].A notable feature of these studies is the lightweight nature of their prototypes, facilitating mobility on flat surfaces.Furthermore, their designs prioritize ease of repair and cost less than USD 10, 000.Table 2 presents these EOD robots and their main features.It is crucial to emphasize that these low-cost EOD robot prototypes were primarily designed for research purposes and were not intended for real EOD missions.In other words, these robots were only tested in laboratory settings and were not experimentally validated in real scenarios, except for the EOD robots mentioned in [36,37].Additionally, they did not receive approval from EOD police officers in their respective countries for use in explosive ordnance disposal missions.1) and the limitations present in low-cost EOD robots (as presented in Table 2), the main objective of this research is the design, development, and implementation of a low-cost EOD robot prototype named JVC-02, which stands out for its affordability while satisfying the essential requirements to assist UDEX agents in their EOD work.Although this robot does not encompass all the features detailed in Table 1, it has all the essential functionalities needed to perform real EOD missions while maintaining a cost-effective approach.In addition, its modular design utilizes commercially available components, facilitating efficient maintenance and robot repairs.

Contributions
The main contributions of this research are the following:

•
A comprehensive analysis was conducted to define the design requirements of the JVC-02 robot for real EOD missions.To achieve this, we adopted the Quality Function Deployment (QFD) methodology [38] to identify the essential needs and requirements of UDEX agents.Additionally, we drew inspiration from the exceptional capabilities demonstrated by rescue robots in the RoboCup competition.This combined approach enabled us to rigorously and accurately obtain the necessary design requirements for developing the JVC-02 prototype.

•
Unlike low-cost EOD robots from previous works [33][34][35][36][37] and recent EOD robots from other research [39,40], our JVC-02 prototype has been tested in real EOD missions with the direct involvement of police officers from the UDEX in Peru.Additionally, field tests were conducted in real scenarios inspired by RoboCup challenges, including mobility, dexterity, and exploration assessments.This validation in real scenarios has allowed us to confirm the effectiveness and functionality of our design, demonstrating its capability to carry out real EOD missions.

Article Organization
Section 2 describes the first JVC-01 prototype designed, including its main features.The design requirements for the JVC-02 prototype are detailed in Section 3. In Section 4, the design of the entire JVC-02 robotic system is developed, detailing its design, mechanical system, electronic and control system, and human-machine interface (HMI) system.The cost analysis is discussed in Section 5, and the list of robot specifications is detailed in Section 6. Maintenance is discussed in Section 7. Field tests in real scenarios, such as actual EOD mission tests, are presented in Section 8.In Section 9, a comparison analysis is presented, and Section 10 describes the problems encountered in this research.Finally, conclusions and future work are discussed in Section 11.

Previous Works
In 2021, the Universidad Nacional de San Agustín de Arequipa, with the support of UDEX of Arequipa, Peru, started the development of the first EOD robot prototype in the region.This EOD robot, called JVC-01 [41], as shown in Figure 1, has a tracked locomotion system, a four-degree-of-freedom robotic manipulator, and a three-finger gripper, in addition to being integrated with a localization system.The tracked system with independent motors allows the robot to rotate on its axis.The chassis comprises 12V DC motors and gearboxes, each with a power output of 1.2 kW, allowing the JVC-01 to support up to 400 kg of additional weight on the chassis and tow vehicles up to 1301 kg on flat surfaces.Figure 2 shows the trailer of a UDEX pickup truck.Thanks to an interlocking system, the DC servomotors used in the robotic manipulator are equipped with gearboxes to control speed, provide high torque, and keep the arm fixed without excessive current consumption.The robot arm has an anthropomorphic design with three fingers and four degrees of freedom, enabling efficient manipulation of objects, as shown in Figure 3.This first EOD robot prototype design demonstrated its efficiency in tasks requiring power and stability.However, UDEX police officers commented that "The prototype is unnecessarily heavy, and the task of picking up objects is complicated".In certain situations, its high-power capacity is unsuitable for an EOD robot's requirements.Despite these observations, its tracked design with independent motors has been demonstrated to provide great maneuverability.

Design Requirements
This section presents the design requirements for the development of the JVC-02 robot.These requirements combine two design approaches: the capabilities of the RoboCup rescue robots and the needs and requirements of the UDEX police officers.

RR-RoboCup
The RoboCup is an international competition that has been held every year since 1997.It allows engineers and researchers in robotics to test their robotic systems and compare their performance [42,43].In this competition, robots dedicated to rescue tasks, also known as rescue robots, are evaluated in terms of their mobility, dexterity, and exploration capabilities [15][16][17].In [44], it was stated that these characteristics are the ones that an EOD robot must possess to be a useful assistance system for police officers.In that sense, the design requirements for the JVC-02 robot are inspired by the capabilities of the RoboCup rescue robots.The international standards of the National Institute of Standards and Technology (NIST) in conjunction with the American Society for Testing and Materials (ASTM) are used to evaluate the capabilities of the JVC-02 EOD robot.

UDEX-AQP
UDEX performs explosive ordnance disposal operations based on the experience and skills of its agents.However, there is a recognized need for a robotic system to assist in this type of operation, aiming to safeguard the integrity of officers and civilians.For this reason, the first robot prototype, JVC-01 [41], was designed, as detailed in the previous section.However, tests performed with the JVC-01 robot and UDEX agents revealed that the design was unnecessarily heavy, and the heavy-weight towing and carrying capabilities could have been more useful for an EOD robot.Therefore, the next robotic prototype, i.e., the JVC-02, should consider improvements based on this feedback.
The Quality Function Deployment (QFD) methodology is used in robot design to transform customer needs into design requirements [45][46][47].In this context, we employed this methodology to identify the specific needs and requirements of the UDEX agents for the optimal design of the JVC-02 robot in EOD missions.To achieve this, we first conducted individual interviews with each police officer to obtain a wide range of information about their experiences in EOD missions.Subsequently, a survey was administered to the whole UDEX police team.Based on the information obtained, the main needs of the UDEX officers were established, as detailed below: Once the needs of the UDEX agents were identified, the QFD methodology was applied.This methodology allowed us to translate the needs of police officers into measurable attributes in the robot prototype.Figure 4 shows the QFD matrix used for the optimal design of the JVC-02 robot, where the needs and requirements are presented with their relative importance based on the survey of UDEX agents, along with the technical requirements that must be met to satisfy the needs of UDEX agents and the evaluation of different EOD robotic prototypes.The following EOD robots were selected for this evaluation: JVC-01, MK2, LT2-F, HD2, TALON, and CALIBER.The selection of these robots was based on the fact that they feature a tracked locomotion mechanism, which is a feature that we also wish to incorporate in our prototype.The aim of this selection was to facilitate a detailed comparative analysis, allowing us to identify how these robots address the needs identified by the UDEX agents.
From the statistical data presented in the QFD matrix, it can be seen that a stable robotic mechanism presents a higher weight/importance relationship for UDEX agents in terms of the safe handling of explosive ordnance.Next, in decreasing order of weight/importance are the good design of a robotic manipulator, the mounting of cameras, an efficient locomotion mechanism, a well-implemented communication system, and finally, an intuitive graphical interface design for easy control of the JVC-02 robot.
With the information gathered from the previously discussed design approaches, encompassing the requirements inspired by the capabilities of rescue robots and the exhaustive collection of the needs and requirements of the UDEX agents, we finally arrive at the design requirements of the prototype robot.Table 3   x ▲

Use of commercial components
x

Design of the EOD Robot JVC-02
In this section, we present the integral design of the JVC-02 prototype, referencing the requirements detailed in Table 3 in the previous section.First, we focus on the design of the mechanical system, which consists of both the mobile platform and the robotic manipulator.Then, we detail the design of the electronic and control systems, highlighting the electronic components used, the power system, and the communication system.Finally, we present the operating system's design, which encompasses the control station and the humanmachine interaction (HMI) that enables effective communication between the operator and the JVC-02 robot.Figure 5 provides a comprehensive view of the JVC-02 robot, presenting the final mechanical design elaborated in SolidWorks version 2022 (Figure 5a) and the real prototype, which is composed of a robotic manipulator, cameras, antenna, etc. (Figure 5b).

Mechanical System
In this subsection, we start designing the JVC-02 robotic system by addressing the mechanical aspect.We present and analyze in detail the design of the mobile platform and the robotic manipulator, including a kinematic analysis to define its workspace.More detail can be seen in a previous work dealing exclusively with the mechanical design of the JVC-02 robot [48].

Mobile Platform Design
Commercial materials such as structural profiles, aluminum, nylon, rubber tracks, and automobile windshield wiper motors were used to design the mobile platform of the robot, significantly reducing its weight and lowering the manufacturing cost.The design of the mobile platform was developed using SolidWorks software, (2022 version) as shown in Figure 6.
External loads generated by traversing rough environments, such as slopes, stairs, and obstacles, were considered to ensure mobility and weight reduction.ASTM A36 structural sections were used in the design to support these loads and protect the control electronics, batteries, and motors.The locomotion system comprises two DC 12V wiper motors (one for each drive track); two speed reducers of catalytic converters and chains, each with a reduction ratio of Rt = 2.73; two AISI SAE 1045 steel axles; six aluminum wheels; ten nylon wheels; and two tracks (toothed belts with vulcanized rubber wishbones).The motors have a worm gear reduction gearbox that acts as a mechanical brake on slopes.The tensioning system is shared and adjusted by screws, and the tracks have a scalene trapezoidal configuration.This mechanical design of the mobile platform achieves considerable weight reduction and increased mobility in complex and rough environments, allowing the robot to climb stairs.In addition, it addresses the mechanical problems present in the previous JVC-0.1 prototype moving platform.

Robotic Arm Design
The robotic manipulator was designed to reduce weight using topologically optimized steel structural sections.Low-cost wiper motors, drill motors, and sprocket worm gearboxes were used.This arm provides explosive ordnance handling capability.Like the mobile platform, the manipulator design was created using SolidWorks, as shown in Figure 7.The structure of the robotic arm was fabricated with ASTM A 36 steel structural sections, mainly due to its accessibility in the local market.Topological optimization with laser cutting was applied to reduce its weight.The robotic arm has five degrees of freedom (base, shoulder, elbow, wrist, and gripper swing), and the gripper has two fingers, 12 joints, and one additional degree of freedom, allowing the gripper to open and close.
The base, shoulder, elbow, and wrist are driven by a worm gear mechanism in the robotic arm with a 12 V wiper DC worm gear motor.In addition to providing high transmission ratios, this worm gear offers mechanical safety in motor de-energization.The rotation of the gripper is driven by a chain reduction mechanism with a 12 V DC planetary geared motor.Finally, the opening and closing of the gripper, which extends up to 150 mm, is driven by a power screw mechanism with a 12 V DC planetary geared motor.The gear ratios of the gearboxes allow the robotic arm to lift loads of up to 8 kg.The gear ratios for each joint are as follows: Rt = 40 at the base, Rt = 70 at the shoulder, Rt = 40 at the elbow, Rt = 20 at the wrist, and Rt = 1.6 at the gripper swing.The critical joint is the shoulder, where the wiper geared motor has a maximum torque of 15 Nm.As mentioned above, the motor is coupled to a worm gear.For the shoulder, the Rt is 70, resulting in a torque of 1050 Nm at the output of the gearbox, which is higher than the required torque of 592 Nm to lift 8 kg.These calculations on motor torque are more precisely detailed in the following section.
Thanks to the design with topological optimization, a robotic arm with a significant weight reduction was possible.Furthermore, an acceptable level of precision has been achieved for the operators when handling explosive devices of up to 8 kg with the arm extended.

Motor Torque Calculation
For the power balance calculation, the torques required to move the mobile platform and the robotic arm of the EOD robot to manipulate and transport explosive devices across slopes and uneven terrain were determined as follows: • Calculation of the motor torque for the mobile platform: The critical scenario, where the highest motor power is required, was assumed to occur when the robot moves across slopes and stairs with an inclination of 45º.For the calculation, a robot weight of 100 kg and a friction coefficient of 0.8 were assumed, resulting in a maximum required force of 1248 N. Considering that the wheels have a diameter of 50 mm and that the robot uses two tracks, the torque required for locomotion is 31.2Nm for each motor.• Calculation of the torque for the robotic arm: A free body diagram (FBD) was developed, as shown in Figure 8.The robotic arm is in the critical extended arm position, where the greatest load is required to move and transport an explosive device with a maximum weight of 8 kg.The weights of each link and joint of the robotic arm were obtained from SolidWorks, where the robot was designed.These loads were included in the torque analysis.The results of the torques required for the motor of the joints

Kinematic Analysis
In order to determine the range of displacement of the end effector of the robotic manipulator during the manipulation of explosive devices, a kinematic analysis was conducted.First, a kinematic diagram of the manipulator was elaborated for the assignment of its reference systems in each joint, as shown in Figure 9.Using the Denavit-Hartenberg (D-H) method, the D-H parameters presented in Table 4 were obtained.
The kinematic equations necessary for calculating the end effector position as a function of the manipulator parameters were obtained using the D-H parameters.From these equations and parameters, a set of points representing the various positions achievable by the end effector of the robotic arm was obtained, as shown in Figure 10.In the top view (Figure 10a), the last joint of the manipulator has a reach of up to 1900 mm around the robot, while in the side view (Figure 10b), the manipulator can access heights of up to 1540 mm.Thanks to this information, the UDEX unit agents can have a complete view of all possible movements of the robotic arm, allowing them to execute the appropriate maneuvers to handle the explosive devices accurately and safely.

Stability Analysis
This section analyzes the stability of the EOD robot in critical situations it could face in real life.This analysis was carried out using SolidWorks simulations.The critical situations evaluated include manipulating an 8 kg object with the arm extended horizontally and displaced in an inclined plane.
Once the mechanical design of the EOD robot was completed, the SolidWorks simulations were prepared by placing an 8 kg weight in the gripper of the robotic arm with the arm extended.The centers of gravity were kept within the contact area of the tracks, ensuring stability in this situation (see Figure 11a).Subsequently, the EOD robot was tilted with the arm extended to the side until the vertical projection of the center of gravity reached the edge of the track contact area.The robot maintained stability up to a maximum tilt angle of 30°(see Figure 11b).

Electronic and Control System
In this section, we address the design of the JVC-02 robot from an electronic and control point of view.The selection of sensors, actuators, and controller boards is analyzed in detail to achieve optimal and low-cost control of the robotic prototype.Special attention is also dedicated to the power system and communication, ensuring fluid interaction between the components and an efficient connection between the robot and the operator.Figure 12 shows a general scheme of the whole electronic and control system of the JVC-02 robot, which is divided into three sets of blocks, corresponding to the control station, the manipulator, and the robot.Each of them contains blocks of different colors, where the light-blue blocks correspond to sensors, orange to controllers or processors, yellow to actuators, and green to intermediaries.The figure is followed by a description of each of the components, including the sensors, actuators, power system, controller boards, and communication system used in developing the electronic and control systems of the JVC-02.

Sensors
To effectively perform EOD tasks, the JVC-02 robot must be equipped with various sensors, including IP cameras, encoders, and limit switches.The following are short descriptions of these components: • Hikvision H.265 IP cameras (Hangzhou Hikvision Digital Technology Co., Ltd., Hangzhou, China), purchased from Amazon: To provide optimal visual feedback in real time to the operator and enable a complete view of the environment, three IP cameras were integrated into the robot: one on the turret, one on the front of the robot, and a third on the mechanical gripper of the robotic manipulator.These cameras are powered by 12V DC and offer a high resolution of 8 megapixels, with a fixed 2.8 mm lens and a maximum speed of 3840 × 2160 pixels at 20 fps.• AS5600 Rotary Encoders (AMS, Premstätten, Austria), purchased from Amazon: AS 5600 magnetic rotary position sensors, with a 12-bit resolution, a 3.3 V power supply, and an I2C output, were used to obtain real-time information about the position of each degree of freedom of the robotic manipulator, ensuring precise and efficient control in dexterity tasks.

•
Limit switches: These were used to avoid exceeding the mechanical limits of the manipulator arm structure.

Actuators
Actuators are essential in the robot's operation, allowing EOD tasks to be carried out effectively.Key actuators employed for the robot design include low-cost DC and stepper motors, which are described as follows: • DC motors: The JVC-02 robot is equipped with eight direct-current (DC) motors, of which six are low-cost wiper motors.Two wiper motors, with a voltage of 12 V, a torque of 15 Nm, and no-load oscillations of 41 RPM at low speed and 65 RPM at high speed, are designed to drive the chassis tracks.The remaining four wiper motors control the robotic arm with a voltage of 12 V, a torque of 20 Nm, and noload oscillations of 350 RPM.The two additional motors are located in the robot's mechanical gripper.

•
Stepper motors: To control the movement of the turret camera, two bipolar stepper motors-one for horizontal movement and one for vertical movement-are incorporated.This configuration allows the operator to precisely guide the camera to any point of interest, facilitating real-time environment visualization during EOD tasks.

Power System
The JVC-02 robot has a power system divided into two independent stages: the control stage and the power stage.This configuration was implemented to avoid possible failures in the control stage caused by voltage drops in the power stage due to the high and intermittent consumption of the motors.The power stages are as follows: • Control stage power supply: To ensure greater voltage stability and prevent the voltage from dropping below 12 V, two 12 V lead-acid batteries connected in series were used, allowing 24 V to be reached.Then, two regulators were used: one reduced the voltage to 5 V to power the Raspberry Pi board, while the other reduced the voltage to 12 V to power the controller boards, IP cameras, NVR, and access points.
In addition, step-down switching voltages and low-dropout-voltage regulator sources were used to achieve the 5 V and 3.3 V voltages needed to power the electronic circuits.

•
Power stage power supply: The power stage was powered by a 12 V 125 Ah Panasonic Chaos wet cell battery purchased from Amazon, which provided the power required for the drivers of the motors described as follows: -BTS7960 and Cytron20A Drivers purchased from Amazon: The BTS7960 (supply voltage: 5.5-27 V DC; control voltage: TTL 3.3 V/5 V; current capacity: 43 A with peaks up to 60 A) and Cytron20A (operating voltage: DC 6 V to 30 V; 3.3 V and 5 V logic level input; maximum motor current: 20 A with peaks up to 60 A) drivers are responsible for managing the power to control the DC motors through the signals emitted by the controllers.-Drivers L293D (Texas Instruments, purchased from Amazon): These drivers (logic supply voltage Vss: 4.5-36 V DC; power supply voltage Vs for motors: Vss-36 V DC; logic voltage: 3.3 V/5 V DC; channels: two full H-bridges (four half H-bridges); DC per channel: 600 mA; peak current per channel: 1.2 A) are responsible for managing the power for the stepper motors used for the turret camera movement.

Controller Boards
The JVC-02 robot control system has different controller boards that play crucial roles in coordinating and managing the robot's movements: • Raspberry Pi 3 (Raspberry Pi Foundation, Cambridge, UK) purchased from Amazon: This board establishes client-server communication with the control station computer, acting as a server.Through a socket connection, the Raspberry Pi 3 board receives the data string containing the speed and direction of each motor.It then sends this information via UART to the ESP32.• ESP-32 controller (Espressif Systems, Shanghai, China) purchased from Amazon: The ESP-32 board receives the data string from the Raspberry Pi 3 via UART and processes it.Subsequently, it uses the RS485 protocol to send each motor's specific speed and address.

•
ATmega328 controllers (Microchip Technology, Chandler, AZ, USA) purchased from Amazon: The JVC-02 robot has six motors for the arm and two for the tracks, each with an independent ATmega328 controller.This controller is capable of controlling the encoders to receive their absolute position and receiving the signal from the limit switch sensors that mark the limits of the movement of the links.

Communication System
Effective communication between the robot and the control station is essential during explosive ordnance disposal operations.It not only allows control of the robot but also provides critical information about the position of the manipulator's arm, the orientation of the robot, and images from the video cameras, crucial data for decision making in EOD situations.To ensure this reliable communication, the JVC-02 uses a CPE510 antenna that connects to an AC1750 access point.The socket protocol is implemented using TCP/IP in a client-server configuration, establishing two independent and bidirectional connections between the control station and the robot: one to transmit the images from the video cameras and the other to communicate with the control system.Data transmission is done through data frames, which contain information about each motor's speed and direction and the arm encoders' positions.

Operation System
In order to ensure intuitive and efficient operation between the UDEX police agent and the JVC-02 robot, the robot operation system was approached from two key aspects: the design of the operator control station and the design of the human-machine interface (HMI).The development of each of these elements is described below.

Control Station
The control station plays a fundamental role in the operation of the JVC-02 robot, as it allows the UDEX agents to control its movements, manipulate the robotic arm, and monitor the scenarios in real time.It is equipped with a water-and dust-resistant housing that ensures its operation in various environments.The station includes a computer laptop running the control software, a CPE510 antenna (TP-Link Technologies, Shenzhen, China) that facilitates communication with the robot, and an intuitive control joystick for maneuvering the robot.Figure 13a shows the control station housing, while Figure 13b presents the control station in its complete form.

Human-Machine Interface (HMI)
The appropriate design of the HMI is of utmost importance to optimize the efficiency and effectiveness of the JVC-02 robot in EOD missions.In order to fulfill the requirements set by the UDEX police officers, as detailed in Section and to ensure intuitive control of the robot during operations in hazardous situations, a user-friendly and easy-to-use graphical interface was developed using Python software (3.8. version) This graphical interface provides a real-time view of the environment through the cameras integrated into the robot, allowing zoom, brightness, and contrast adjustments to adapt to changing terrain conditions.In addition, it provides information on the overall status of the robot and the movement of the robotic arm joints, giving operators complete control.
Figure 14 shows the interaction between the operator and the JVC-02 robot during an exploration task.Figure 14a shows the operator operating the robot, while Figure 14b shows the graphical interface.In addition, Figure 14c shows the view provided by the end-effector camera, and Figure 14d

Cost Analysis
In this section, a cost analysis is carried out for the design of the JVC-02 EOD robot prototype.As mentioned at the beginning of this article, a modular design approach using commercial off-the-shelf components was chosen to facilitate the maintenance and repair of the JVC-02 robot.This section aims to identify and evaluate the main costs involved in the design process, which will provide us with an overall view of the total cost of the robot.Table 5 presents a summary of the approximate costs involved for each design system.

Systems Cost
Development USD 1600 Mechanical system and manufacturing USD 7220 Electronic and control system USD 3340 Operating system USD 1880 Approximate total cost USD 14,040 Table 5 shows that the total cost of developing the EOD JVC-02 robot amounts to USD 14,040, which is below the stated target of USD 15,000 presented in the "Design Requirements" section.Figure 15 shows a bar graph comparison of the man-portable and maxi-sized EOD robot costs from Table 1 concerning the total cost of the JVC-02 EOD prototype.The bar graphs in Figure 15, supported by Table 5, clearly show that the JVC-02 prototype costs significantly less than commercially available EOD robots.This fulfills our main objective in this research, which is the design of a low-cost EOD robot.Although our prototype does not include all the features of the robots presented in Table 1, it has all the essential functionalities needed to perform real EOD missions by UDEX police officers, which are experimentally validated through field tests and real EOD missions in the "Experimental Validation" section.

System Specifications
This section presents the specifications of the JVC-02 robot.These are evaluated in practice through field tests in real scenarios and real EOD mission tests, as discussed in the next section of this article.These tests aim to validate and verify the performance and operation of the robot in real situations, thus ensuring that it meets the requirements and objectives established for its operation.Table 6 presents the specifications of the JVC-02 robot.

Maintenance
• Routine check: The robot was checked before and after each deployment.A crucial aspect to consider is the wear of the tracks, as this can cause them to come off their axles.It is important to ensure that track tension is adequate to avoid this.

•
Overstressing and pressure on the axles: In case of overstressing, too much pressure may be exerted on the axles, transmitting the movement of the motors to the tracks.These shafts should be replaced when signs of wear are observed.

•
Checking the electronics and wiring: The electronics and wiring were checked, especially when the robot was operated on very rough terrain or received a shock or impact that could disconnect a component.

Experimental Validation
In this section, we present the experimental validation of the JVC-02 robot, which was carried out to determine its characteristics and capabilities.For this purpose, field tests were performed in real scenarios inspired by the RoboCup, following NIST/ASTM standards, covering aspects of mobility, dexterity, and exploration.In addition, real EOD mission tests were carried out, where the UDEX agents operated the robot during the whole mission.Subsequently, the results obtained were analyzed in relation to the requirements established at the beginning of the article.The mobility capabilities of the JVC-02 robot were evaluated through five field tests in real environments.These tests were classified into two categories: flat surfaces and inclined surfaces.Tests on flat surfaces involved performing straight 5-meter runs on grass and sandy terrain.On the other hand, tests on inclined surfaces involved going up and down stairs with slopes of 15°, 20°, and 30°.Each test was considered complete when the robot completed three repetitions on the test terrain.The JVC-02 robot successfully passed the tests on flat surfaces, demonstrating its ability to traverse rough terrain of grass and sand in up to three repetitions.As for tests on inclined surfaces, the robot successfully climbed stairs with 15°and 20°slopes in three repetitions.However, some drawbacks were observed in the tests with 30°slopes: while it completed the first repetition, in the second, it had difficulty completing the test, and in the last repetition, it failed to finish successfully, stopping during the process.Figure 16 shows scenes of these mobility tests.The conclusion regarding the mobility of the JVC-02 robot is that its performance on different terrains was successful, demonstrating that the robot can traverse both urban areas and natural environments in EOD missions.

Dexterity
The dexterity capabilities of the JVC-02 robot were evaluated through two field tests using the apparatuses proposed in the RoboCup RRL [49].The dexterity tests used the 'omni' apparatus and the 'linear' apparatus.These tests consisted of removing each object from each apparatus and placing it into a container next to the robot.Each test was considered complete when the robot smoothly extracted the object from the apparatus.The JVC-02 robot, with its 5-DoF manipulator, successfully completed the dexterity tests, as presented in Figure 17, demonstrating its dexterity in handling explosive ordnance in EOD missions.

Exploration
The exploration capabilities of the JVC-02 robot were evaluated through three tests of inspecting a house.These tests consisted of locating and recognizing three explosive devices previously placed in boxes at different locations within a house.Each inspection test was considered complete when the robot approached the designated target and transmitted information about the type of explosive device found to the operator.In this regard, the JVC-02 robot successfully inspected and recognized the three explosive devices inside the house, as illustrated in Figure 18.As a final note, the inspection tests of the JVC-02 robot demonstrated satisfactory performance in home exploration, which supports its ability to inspect homes, institutions, airports, etc., in real EOD missions.

Testing of Real EOD Missions
In this section, we present the real EOD mission tests with the direct involvement of UDEX police officers as part of the experimental validation process of the JVC-02 robot.During these tests, the officers assumed responsibility for all stages of the process, from the deployment of the JVC-02 robot to the final analysis and documentation of the operation.In order to guarantee the validity and realism of the tests, real intervention scenarios were designed where the UDEX agents faced EOD situations.In this regard, a test circuit was set up at the UDEX installations, where the explosive device was placed 50 m from the control station.In addition, a container was placed 30 m away from the explosive device for its transportation.The EOD mission tests involved UDEX police officers moving the robot to the target, identifying the explosive device, handling it, moving it to the container, and finally detonating it.Three tests were conducted, each with a different UDEX police officer controlling the robot.The test was considered complete when the UDEX agent successfully completed the mission, i.e., detonated the explosive device.The JVC-02 robot successfully passed all three real EOD mission tests, i.e., each UDEX police officer successfully completed the proposed mission.However, the UDEX agents provided comments and feedback on the performance of the JVC-02 robot.The most important observations were as follows: (1) Graphic interface: The camera system is suitable for allowing interaction without visual contact, but it is difficult to achieve depth perception.A side camera could improve this.(2) Increase in speed: The current speed of the robot is suitable for maneuvering when handling an explosive; however, when approaching an explosive, the speed may be too slow.Changing the motors to increase the revolutions per minute can address this problem.(3) Reduce the weight: The robot must be deployed by hand, and currently, at least four UDEX agents must be assigned for this purpose.However, only two agents should be needed to deploy the robot for a better distribution of tasks.The weight can be reduced by reducing the thickness of the platform and robotic arm.This feedback from the UDEX agents is of utmost importance, as it provides valuable information for improving and optimizing future designs.Figure 19 illustrates the EOD mission process, which begins with the arrival of the robot at the UDEX installation, as shown in Figure 19a.Figure 19b shows the deployment of the robot by the UDEX agents.Figure 19c shows the agents carrying out the EOD mission.Figure 19d shows the robot moving toward the target.Figure 19e shows the handling of the explosive device, and finally, Figure 19f shows the transfer of the device to the container for detonation.Throughout the mobility, exploration, and real EOD mission tests, constant monitoring of the quality of the wireless communication between the control station and the JVC-02 robot was carried out, ensuring its correct operation.During these tests, it was demonstrated that communication remains reliable in direct line-of-sight situations up to 450 m.In contrast, it extends up to 120 m in non-direct line-of-sight situations.These values corroborate the effectiveness of the wireless communication system used in the JVC-02 robot.The results of the experimental validation of the JVC-02 robot discussed previously are summarized in Table 7.

Comparison
In the QFD matrix analysis elaborated in this work, a previous comparison of robotic prototypes available in the market (JVC01, MK2, LT2-F, HD2, TALON, and CALIBER) was performed, and characteristics such as accessibility, manipulation, mechanical system, control, durability, vision system, ease of maintenance and low cost were compared.A scoring method from 1 to 5 was used, as shown in Table 8, where 1 is very bad and 5 is very good.The comparison of these characteristics can be summarized as follows: • Accessibility in different terrains: The JVC-02 demonstrates superiority over its predecessor, JVC-01, after locomotion tests, matching the capability of the prototypes available on the market.• Safe manipulation of explosive devices: In terms of manipulation, the JVC-02 shows an optimal grip superior to that of JVC-01 due to reducing one finger of the gripper, with two fingers being the optimal design, matching models such as LT2-F and TALON.However, due to the mechanical failures presented in the gripper, improvements are still required to match the manipulation characteristics of other robots in the market.• Robust mechanical system: Reducing the weight of the JVC-02 robot decreases its robustness, improving other features such as cost and rapid deployment.However, it is as effective as robots like the LT2-F despite this reduction in robustness.

•
Easy to control: Implementing the new control system, designed and presented in previous sections, undoubtedly improves and facilitates human-robot interactions over long distances in a safe way, far surpassing the JVC-01.
process, a specific maneuver was considered to attach the robotic arm to the chassis: first, the motor with the gearbox was inserted, then the robotic arm was rotated on its base axis, and finally, it was bolted to the chassis.• Detachment of the caterpillars: The caterpillars were detached during the mobility tests, mainly during turns.The track tensioning was corrected, and the optimum tensioning point was found, significantly reducing the number of times the track would dislodge.• Relocation of the lifting lugs: The lifting lugs, used by the UDEX agents to hold and transport the robot, were initially located at the corners of the main chassis structure.However, the tracks between the ears and the agents caused ergonomic problems.The front lifting ears were relocated to the extension of the main chassis structure.• Electronics protection: During the robot's locomotion testing, sharp impacts were observed that damaged the electronics.To address this, protective foam was installed inside the chassis, reducing damage to the electronics.

Conclusions and Future Work
In this article, we have presented the development of the EOD JVC-02 robot, a lowcost prototype designed to assist police officers from the UDEX in the city of Arequipa, Peru, in real explosive ordnance disposal missions.Our modular approach, involving commercially available components, has been fundamental in achieving a low-cost and highly functional robot that allows for easy maintenance and repair.
JVC-02 has demonstrated oUSAding mobility, dexterity, and exploration capabilities, which are essential for EOD operations.Although it does not incorporate all the features of other EOD robots available in the market, we focused on including the essential functionalities required in EOD missions, which have been validated through multiple tests in real scenarios, obtaining satisfactory results on their performance.However, we have identified some areas of improvement in the prototype that have provided us with valuable information for future designs: (1) in mobility tests, JVC-02 failed to complete the third repetition in the stair-climbing and -descending test with a 30°inclination; (2) in real EOD mission tests, agents suggested increasing the robot's speed, as they considered it slow for EOD missions; and (3) agents also recommended reducing the overall weight to allow for faster transfer and deployment.
In future work, our main objective will be to address the shortcomings encountered during the experimental validation of the JVC-02 to ensure optimal performance in real EOD missions.Currently, transporting and deploying the robot requires at least four people.To reduce this number to two, we will reduce the robot's weight by decreasing its structural thickness and using lower-density materials such as aluminum in some parts.The robot's current movement speed is adequate for maneuvering and interacting with explosive ordnance with precision.However, movement from the deployment point to the location of the explosive can be very slow over long distances.To address this, we will increase the rate of movement by changing the crawler motors to ones with more revolutions per minute.To improve the scanning system, we plan to increase the proximity sensors and add a flashlight to facilitate inspection in dark locations, allowing the robot to operate more effectively in low-visibility environments.Finally, we will improve control of the robotic arm using multimodal interfaces, allowing for more precise and intuitive control by the operator.
Funding: This research was funded by Universidad Nacional de San Agustín de Arequipa under contract numbers IBA-IB-27-2020-UNSA.
Informed Consent Statement: Informed consent was obtained from all subjects involved in this study.

Figure 8 .
Figure 8. Free body diagram of the robotic arm.

Figure 9 .
Figure 9. Reference system for each joint.

Table 4 .
D-H parameters of the 5-DoF manipulator. i

Figure 10 .
Figure 10.Working space of the robotic manipulator.(a) Top view.(b) Side view.

Figure 11 .
Figure 11.Stability analysis of the EOD robot in (a) horizontal plane; (b) inclined plane.

Figure 14 .
Figure 14.HMI implementation of the JVC-02 robot.(a) The Operator operating the robot on an EOD mission.(b) The Graphical user interface.(c) View of the end effector camera.(d) Field of operation of the JVC-02 robot.

Figure 18 .
Figure 18.Exploration tests in real scenarios.(a) The house to be explored.(b) Beginning of the test.(c) Entering the robot inside the house.(d) First object found.(e) Second object found.(f) Third object found.

Figure 19 .
Figure 19.Real EOD mission with the involvement of UDEX agents.(a) The arrival of the JVC-02 robot at the UDEX installations.(b) Deployment of the robot by UDEX agents.(c) UDEX agents carrying out the EOD mission.(d) Displacement of the robot to the target.(e) Manipulation of the explosive device.(f) Transfer of the device to the container for detonation.

Table 1 .
Comparison of EOD robots.
Based on the above considerations, which include the high cost of EOD robots (as detailed in Table presents the list of requirements for the EOD JVC-02 robot.
-Control station: 15 Kg Cost -Maximum cost for prototype design: USD 15,000

Table 7 .
Analysis of requirements and results.