Research on Accurate Motion Trajectory Control Method of Four-Wheel Steering AGV Based on Stanley-PID Control

Weijie Fu, Yan Liu, Xinming Zhang2023 · DOI

TL;DR

This paper proposes a novel Stanley-PID control method for improving the motion trajectory control of four-wheel steering Autonomous Guided Vehicles (AGVs). By combining the strengths of Stanley's lateral control and PID's longitudinal control, the method achieves more stable and precise path tracking, especially in complex environments.

What Problem Does This Paper Address?

The increasing adoption of Autonomous Ground Vehicles (AGVs) in various industries necessitates precise and stable motion control. While existing methods like PID and Stanley control have their merits, they often fall short in complex, dynamic environments. PID control, while simple, struggles with nonlinear systems and can suffer from saturation. Stanley control excels at path following but can deviate from the original path under unpredictable circumstances. This paper addresses the challenge of achieving robust and accurate motion control for four-wheel steering AGVs by developing a hybrid control strategy that leverages the advantages of both Stanley and PID controllers.

Methods

The core of this research is the development and simulation of a Stanley-PID control algorithm. This hybrid approach integrates:

  • Stanley Control: Primarily used for lateral control, it calculates the lateral error (distance from the desired path) and heading deviation (difference between the vehicle's orientation and the path's tangent). This information is used to determine steering angle adjustments.
  • PID Control: Applied for longitudinal control, it uses proportional, integral, and derivative components to adjust the vehicle's speed based on the error signal.

The Stanley algorithm's output (lateral error correction) is fed into the PID controller, which then generates the longitudinal commands. This combined approach aims to provide better adaptability to dynamic characteristics and disturbances, leading to more precise path-tracking control. The study also involves:

  • Kinematic Modeling: Developing a detailed kinematic model for a four-wheel steering AGV, considering its nonholonomic constraints.
  • Simulation: Using MATLAB 2021a to simulate the proposed Stanley-PID control algorithm and compare its performance against individual PID and Stanley controllers.

Datasets

Not discussed in the paper. The study relies on simulated data generated from the AGV's kinematic model and the proposed control algorithms.

Key Insights and Ideas

  • Synergy of Stanley and PID: The paper demonstrates that combining Stanley's lateral error correction with PID's longitudinal control significantly enhances AGV motion trajectory control. This hybrid approach addresses limitations of each individual controller.
  • Improved Stability and Accuracy: The Stanley-PID controller shows superior performance in terms of faster response, reduced fluctuations, and smoother variations compared to a single PID controller, indicating stronger system stability.
  • Adaptability to Complex Environments: The Stanley-PID method is proposed as a more adaptable solution for AGVs operating in complex road conditions, where traditional methods might struggle.
  • Addressing Nonholonomic Constraints: The kinematic modeling and control strategy specifically account for the nonholonomic nature of AGVs, which is crucial for accurate motion control.

Key Results

The simulation results, conducted using MATLAB 2021a, indicate that the Stanley-PID algorithm outperforms traditional PID control in terms of tracking accuracy and steady-state error. Specifically:

  • Response Time: The Stanley-PID algorithm achieved a response time of approximately 0.2 seconds, compared to around 0.3 seconds for the single PID algorithm.
  • Path Error: The Stanley-PID controller resulted in path errors initially fluctuating between 0 and -1.5, eventually stabilizing at around 0.4. The single PID controller showed fluctuations between 0 and -1.6.
  • Fluctuations and Stability: The Stanley-PID control exhibited smaller motion trajectory fluctuations and a more stable response compared to the PID control, which showed 3 to 4 fluctuations before converging.

Limitations and Future Work

The authors acknowledge a significant limitation: the lack of experimental validation. They emphasize that conducting real-world experiments on a four-wheel steering AGV is essential to fully evaluate its performance, including navigation accuracy, stability, and energy consumption, and to facilitate comparisons with other AGV types.

Future work directions suggested include:

  • Experimental Validation: Conducting a series of experiments to comprehensively assess the AGV's capabilities in various scenarios.
  • Path Planning Algorithms: Exploring new path planning algorithms and optimization methods that leverage the maneuverability of four-wheel steering AGVs.
  • Diverse Application Areas: Investigating the application of these AGVs in indoor navigation, smart warehousing, automated logistics, and service robotics.
  • Technology Integration: Integrating AGVs with sensor fusion, artificial intelligence, data analytics, and decision algorithms to enhance intelligence.
  • Anomaly Detection and Robust Control: Developing systems for automatic fault detection and resolution, along with robust control algorithms for uncertain environments.
  • Energy Optimization: Focusing on energy optimization methods to prolong battery life and improve energy utilization efficiency.

Summary updated

BibTeX
@misc{fu2023research,
  author = {Weijie Fu and Yan Liu and Xinming Zhang},
  title = {{Research on Accurate Motion Trajectory Control Method of Four-Wheel Steering AGV Based on Stanley-PID Control}},
  year = {2023},
  doi = {10.3390/s23167219},
}
Figure