Abstract
The paper makes an honest attempt to amalgamate two interesting open issues of path planning of mobile robots, such as learning automata-based planning and human in a loop using brain–computer interfaces. The learning automaton is employed to demarcate better actions at a state using a reward/penalty mechanism, whereas the brain–computer interface is used to identify the wrong actions taken by the robot at a given state and assist the robot to update the effect of penalty in its space of learning automaton. After convergence of the learning automaton, the robot utilizes it for determination of the next states from its current state by selection of the action with the highest reward function stored in the automaton. The proposed technique yields faster convergence than the one obtained by only learning automata-based system without any provisions for subjective feedback for error corrections. The choice of incremental reward/penalty value is also optimized with respect to minimum steps required to reach the goal during the planning phase.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have