Navigation of mobile robot in a grid-based environment using local and target weighted neural networks

The basic challenge for an autonomous mobile robot is to find a collision free path while navigating from a starting position to a target position. In real-life, autonomous mobile robots are useful in many fields including traffic planning for smart cities, military operations, warehouses applications etc. A ranging sensor-based collision-free navigation and mapping of an intelligent mobile robot using neural network is presented in this paper. A topologically grid-based map is utilized in the proposed neural dynamics. With the help of the ranging sensors the robot can sense only limited range of area. The next grid cell position of the robot is achieved by determining the maximum neural activity of its neighbouring neurons or the minimum target distance from the maximum activated neurons. Through simulation we have demonstrated the effectiveness of bio-inspired neural network and compare the effect of two weight calculation algorithms. The robot is successfully reaching to the target with the help of both the weight calculation algorithms.