An effective vector-driven biologically-motivated neural network algorithm to real-time autonomous robot navigation

A novel biologically-motivated neural networks approach associated with developed vector-driven autonomous robot navigation is proposed in this paper. The biologically-motivated neural networks (BNN) algorithm is employed to guide an autonomous robot to reach goal with obstacle avoidance motivated by Grossberg's model for a biological neural system. As the robot plans its trajectory toward the goal, unreasonable path will be inevitably planned. A vector-based guidance paradigm is developed for guidance of the robot locally so as to plan more reasonable trajectories. In addition, square cell map representations are proposed for realtime autonomous robot navigation. The BNN based scheme demonstrates that the algorithms avoid the issue of local minima in path planning. In this paper, both simulation and comparison studies of an autonomous robot navigation demonstrate that the proposed model is capable of planning more reasonable and shorter collision-free paths in non-stationary and unstructured environments compared with other approaches.