Toward Safety Navigation in Cluttered Dynamic Environment: A Robot Neural-Based Hybrid Autonomous Navigation and Obstacle Avoidance with Moving Target Tracking

In this paper, an autonomous navigation and obstacle avoidance strategy is proposed for an omnidirectional mobile robot. The robot plans a path, starting from an initial point going to a target point. A hybrid approach has been developed where a global approach has been applied to the motion along the desired path (DP) using 2nd order polynomial planning, while a local reactive approach is used to avoid collisions with static and/or dynamic obstacles based on the use of neural control. The neural controller design is based on the "sensing vector" and the "gap vector" concepts. The "sensing vector" is a binary vector which provides information about obstacles detection, while the "gap vector" provides information about a possible nearest gap the robot can pass through it. The proposed approach is extended to include the problem of moving target.