Neural control to simulate 4 autonomous navigation behaviors in a differential-drive mobile robot
Juan Carlos Vega Oliver, Pedro Freddy Huamaní Navarrete · 2017
This paper comprises the algorithmic design and simulation of an autonomous navigation system for a differential-drive mobile robot composed of 4 neural controllers using an intelligent control technique: neural networks. To do so, Matlab and the mobile robots simulator Sim.I.am were used. For the control design, each of the behaviors or controllers were modularly designed and tested following a reactive architecture. Subsequently, they were merged into a navigation system by the design of a finite-state machine, which selects the most appropriated behavior for the robot to follow at each time step. To test the approach, several simulations were carried out for each modular neural controller and for the overall system, which prove the validity and performance of the approach.