Path Planning for a Mobile Robot using Neural Networks

To complete non-collision movement according to the environment information when it is in an unknown,complex,dynamic change condition,a reliable navigation of a mobile robot is demonstrated.The data about the states of the robot is fused using the framework of extended Kalman filter to localize a mobile robot more accurately.After the information of the dynamic workspace is constructed,an algorithm based on neural network is used.The navigation is implemented by adjusting the motion according to the tendency of the points set.Parameter optimization of the algorithm is investigated,and experiments and the simulations reveal that this method performs effectively without the information of the obstacle.