Collision Avoidance Path Planning for Automated Vehicles Using Prediction Information and Artificial Potential Field

Area of Science:

• Autonomous Driving Systems
• Robotics and Control Theory
• Path Planning Algorithms

Background:

• Advancements in autonomous driving necessitate robust emergency avoidance path planning.
• Current artificial potential field (APF) methods face local minimum issues, hindering path completion.
• Integrating surrounding vehicle predictions is crucial for enhanced driving safety and stability.

Purpose of the Study:

• To develop an improved emergency avoidance path planning method for autonomous vehicles.
• To address the local minimum problem inherent in traditional APF algorithms.
• To enhance the efficiency and stability of autonomous driving through predictive path planning.

Main Methods:

• Integration of surrounding vehicle prediction data into the artificial potential field (APF) framework.
• Optimization of quintic Bézier curve control points using sequential quadratic programming.
• Simulation validation using IPG CarMaker and MATLAB/Simulink environments.

Main Results:

• The proposed method effectively integrates predictive data into APF for improved path planning.
• Optimization of Bézier curves enhances trajectory generation and stability.
• Simulations confirmed the method's validity and effectiveness in complex scenarios.

Conclusions:

• The novel approach successfully mitigates the local minimum problem in APF path planning.
• Integrating prediction data and optimizing Bézier curves leads to more efficient and stable autonomous driving.
• The validated method offers a significant advancement for autonomous vehicle safety systems.