Time optimal trajectory planning for mobile robots by differential evolution algorithm and neural networks

A method is presented and tested for planning time optimal trajectories for a mobile robot with constraints by using an evolutionary technique with neural-networks components. The method establishes shortest time trajectories redefined to form a multi-constrained non-linear global optimization problem. The trajectory components such as the turning translational speeds of the mobile robot (i.e. the parameter vector of the problem) are found by using differential evolution algorithm (DE) to obtain the time optimally. DE is a floating-point genetic algorithm. Artificial neural networks learn kinematics structure and upper bound of the velocities on the trajectory. Experiments are successfully implemented on Nomad 2000 mobile robot.