This study presents a novel planning approach of lane changing for autonomous vehicles to ensure multiagent safety and comfortability in riding. First, a risk assessment model, Safety Field, combined with field theory and kinematics model was defined considering the passengers’ subjective feelings to identify the long-term safety trends. Based on the safety field, a new planning algorithm named S algorithm was developed, which was inspired by the A* algorithm. The special grid map, the lists, and the evaluation function designed in the S algorithm enabling its application to complex dynamic situations while considering riding safety and riding comfort. And the proposed path planning method has good scalability that we extend its application scenarios from straight roads to curved roads. Finally, the S algorithm was validated in virtual traffic environments on straight and curved roads, and the results from the test cases demonstrated the effectiveness and scalability of the algorithm.
This study presents a novel planning approach of lane changing for autonomous vehicles to ensure multiagent safety and comfortability in riding. First, a risk assessment model, Safety Field, combined with field theory and kinematics model was defined considering the passengers’ subjective feelings to identify the long-term safety trends. Based on the safety field, a new planning algorithm named S algorithm was developed, which was inspired by the A* algorithm. The special grid map, the lists, and the evaluation function designed in the S algorithm enabling its application to complex dynamic situations while considering riding safety and riding comfort. And the proposed path planning method has good scalability that we extend its application scenarios from straight roads to curved roads. Finally, the S algorithm was validated in virtual traffic environments on straight and curved roads, and the results from the test cases demonstrated the effectiveness and scalability of the algorithm.