Method for the posture control of bionic mechanical wheel-legged vehicles in hilly and mountainous areas

Kaoxin Pan, Qing Zhang, Zhenyu Wang, Sibo Wang, Aihong Zhou, Yong You, Decheng WangInternational Journal of Agricultural and Biological Engineering 2024 · DOI

Abstract

Traditional agricultural equipment chassis often suffer from poor environmental adaptability and inferior mobility in complex terrains. To address these weaknesses, this research presents a novel unmanned agricultural machinery chassis designed for stable and efficient operation in hilly and mountainous areas. Inspired by the motion principles of grasshopper hind legs, a new wheel-legged structure was developed using pneumatic-hydraulic linkage mechanisms. Kinematic analysis using the D-H parameter method revealed a significant end-effector travel range (up to 770 mm vertically), enabling features like independent four-wheel steering and automatic body leveling. A multi-objective posture control system based on the NSGA-II algorithm was implemented and validated through joint simulations and physical prototype testing. The results demonstrate superior leveling performance compared to traditional PID systems, with leveling errors maintained within approximately 1.2°, confirming the vehicle's feasibility for complex environment operations.

Introduction

Agricultural mechanization in hilly and mountainous regions faces significant hurdles. Traditional wheeled tractors, characterized by high centers of gravity, are prone to tipping on steep slopes and struggle with obstacles. While purely legged robots offer high adaptability, they often lack the speed and efficiency required for agricultural tasks. This paper proposes a hybrid wheel-legged chassis that combines the stability of legs with the efficiency of wheels, specifically focusing on the challenge of real-time posture control to maintain a level vehicle body on uneven terrain.

Bionic Design and Kinematics

The core of the vehicle is its bionic wheel-legged mechanism, which mimics the skeletal structure of a locust’s hind leg. By utilizing a parallel four-bar mechanism combined with pneumatic-hydraulic linkages, the design ensures that the wheels remain perpendicular to the ground during leg extension or retraction, maximizing traction and simplifying steering stability.

Application of bionics principle in the wheel-legged mechanism

Kinematic modeling using Denavit-Hartenberg (D-H) parameters established the workspace of the leg. The analysis confirmed that the end effector can travel 0-450 mm in the X-direction, 0-840 mm in the Y-direction, and 0-770 mm in the Z-direction. This wide range of motion provides the structural foundation for adjustable wheel tracks and significant ground clearance adjustments.

Advanced Posture Control Strategy

To manage the non-linear, highly coupled movements of the four independent legs, the researchers moved beyond standard PID controllers. They implemented the Non-dominated Sorting Genetic Algorithm II (NSGA-II) to optimize multiple objectives simultaneously: minimizing pitch and roll angles, managing the center of gravity height, and compensating for "weak legs" (legs losing ground contact).

Co-simulations using Adams for mechanical dynamics and Matlab/Simulink for control logic tested the vehicle against various virtual terrains, including 30° trapezoidal slopes and deep arc pits. A key feature of the control system is its biological reflex-inspired "weak leg" compensation, which automatically extends a leg if sensors detect a loss of ground support, preventing potential tip-overs.

Prototype Development and Experimental Results

A physical prototype was fabricated to validate the simulation findings. The assembly included wheel hub motors, hydraulic actuators, and high-precision inclination sensors.

CAD model of the wheel-legged unmanned chassis assembly

Key performance metrics from the study include:

  • Precision Leveling: The system achieved a maximum absolute leveling error of 1.08° for pitch and 1.19° for roll, significantly outperforming traditional PID control.
  • Stability: Standard deviations for leveling were remarkably low (0.216° for pitch and 0.176° for roll).
  • Response Time: The prototype achieved in-place leveling in an average of 1.2 seconds.
  • Terrain Adaptability: The vehicle demonstrated a ground clearance range from 24 mm to 574 mm and maintained a horizontal posture on side slopes up to 27°.

Conclusion and Future Work

The wheel-legged unmanned chassis successfully demonstrates that bionic principles combined with advanced optimization algorithms can solve stability issues in mountainous agricultural robotics. While the current study proves the effectiveness of stationary and low-speed leveling, future research will focus on dynamic leveling at higher speeds and optimizing the power efficiency of the hydraulic systems for extended autonomous operation.

Summary updated

Abstract

In response to the weaknesses of traditional agricultural equipment chassis with poor environmental adaptability and inferior mobility, a novel unmanned agricultural machinery chassis has been developed that can operate stably and efficiently under various complex terrain conditions. Initially, a new wheel-legged structure was designed by drawing inspiration from the motion principles of grasshopper hind legs and combining them with pneumatic-hydraulic linkage mechanisms. Kinematic analysis was conducted on this wheel-legged configuration by utilizing the D-H parameter method, which revealed that its end effector has a travel range of 0-450 mm in the X-direction, 0-840 mm in the Y-direction, and 0-770 mm in the Z-direction, thereby providing the structural foundation for features such as independent four-wheel steering, adjustable wheel track, automatic vehicle body elevation adjustment, and maintaining a level body posture on different slopes. Subsequently, theoretical analysis and structural parameter calculations were completed to design each subsystem of the unmanned chassis. Further, kinematic analysis of the wheel-legged unmanned chassis was carried out using RecurDyn, which substantiated the feasibility of achieving functions like slope leveling and autonomous obstacle negotiation. An omnidirectional leveling control system was also established, taking into account factors such as pitch angle, roll angle, virtual leg deployment, and center of gravity height. Joint simulations using Adams and Matlab were performed on the wheel-legged unmanned chassis, comparing its leveling performance with that of a PID control system. The results indicated that the maximum absolute value of leveling error was 1.08° for the pitch angle and 1.19° for the roll angle, while the standard deviations were 0.21647° for the pitch angle and 0.17622° for the roll angle, demonstrating that the wheel-legged unmanned chassis surpassed the PID control system in leveling performance, thus validating the correctness and feasibility of its full-directional body posture leveling control in complex environments. Finally, the wheel-legged unmanned chassis was fabricated, assembled, and subjected to in-place leveling and ground clearance adjustment tests. The experimental outcomes showed that the vehicle was capable of achieving in-place leveling with response speed and leveling accuracy meeting practical operational requirements under the action of the posture control system. Moreover, the adjustable ground clearance proved sufficient to meet the demands of actual obstacle crossing scenarios. Keywords: hilly areas, bionic machinery, wheel-legged vehicle, posture control, prototype testing DOI: 10.25165/j.ijabe.20241705.8383 Citation: Pan K X, Zhang Q, Wang Z Y, Wang S B, Zhou A B, You Y, et al. Method for the posture control of bionic mechanical wheel-legged vehicles in hilly and mountainous areas. Int J Agric & Biol Eng, 2024; 17(5): 151-162.
BibTeX
@article{pan2024method,
  author = {Kaoxin Pan and Qing Zhang and Zhenyu Wang and Sibo Wang and Aihong Zhou and Yong You and Decheng Wang},
  title = {{Method for the posture control of bionic mechanical wheel-legged vehicles in hilly and mountainous areas}},
  journal = {International Journal of Agricultural and Biological Engineering},
  year = {2024},
  doi = {10.25165/j.ijabe.20241705.8383},
}
Figure