Heuristic Path Planning Of Autonomous Driving In Social Scene

Robot path planning is one of the most basic functions of robots,but in some complex social scenarios,traditional obstacle avoidance navigation is far from enough.Different from obstacle avoidance navigation based on spatial information,social awareness navigation is mainly based on social information for path planning,which is a challenging navigation task in the field of robotics.Social-aware navigation aims at navigation performance,comfort,naturalness,and sociability.It can be considered that social-aware navigation is an interdisciplinary field of sociology,psychology,humancomputer interaction,and navigation.In this paper,by constructing a multi-layer cost map,including static layer,barrier layer,inflation layer,environment layer,and social layer,various social constraints are integrated into the cost map of each layer,and the A* algorithm is re-optimized so that it can be used in the cost map of each layer.Planning in line with social constraints in complex social scenarios.The environment layer cost map can automatically identify wide and narrow corridor scenes in the environment.For wide corridors,the strategy of driving on the right is adopted;for narrow corridors,the strategy of driving in the center is adopted.The social layer cost map takes special consideration of pedestrians,identifies pedestrians in the scene through pedestrian tracking,and models different personal space models according to the state of the person in the environment(static or walking),which can effectively avoid robots.Pedestrian discomfort caused by being too close to a person.At the same time,this paper groups pedestrians in the environment,divides pedestrians with similar positions,directions,and speeds into a group,and then models the newly established group as a whole,which can effectively prevent the robot from crossing the gap between humans.communication space.Then use the algorithm proposed in this paper and the traditional navigation obstacle avoidance algorithm to conduct a path comparison experiment to verify that the path obtained by the method in this paper is more humane.In order to make the robot travel more natural,this paper performs path smoothing and speed planning on the planned path.Path smoothing makes the robot’s driving path smoother and does not make surrounding pedestrians feel unsafe.Then,according to the maximum acceleration and deceleration limit of the robot,the maximum speed limit,and the different environment of the robot,the speed planning of the path is carried out,so that the robot can travel smoothly at a human-like speed without frequent acceleration and deceleration.At the same time,a comparison of various indicators,including running time,safety factors,etc.,was carried out on the speed-planned and non-speed-planned paths.Experiments show that speed planning and providing speed to the control system for reference can effectively improve the safety and efficiency of driving.