LAUSR

Mobile robot path planning has attracted much attention as a key technology in robotics research. In this paper, a reformative bat algorithm (RBA) for mobile robot path planning is proposed, which is employed as the control mechanism of robots. The Doppler effect is applied to frequency update to ameliorate RBA. When the robot is in motion, the Doppler effect can be adaptively compensated to prevent the robot from prematurely converging. In the velocity update and position update, chaotic map and dynamic disturbance coefficient are introduced respectively to enrich the population diversity and weaken the limitation of local optimum. Furthermore, Q-learning is incorporated into RBA to reasonably choose the loudness attenuation coefficient and the pulse emission enhancement coefficient to reconcile the trade-off between exploration and exploitation, while improving the local search capability of RBA. The simulation experiments are carried out in two different environments, where the success rate of RBA is 93.33% and 90%, respectively. Moreover, in terms of the results of success rate, path length and number of iterations, RBA has better robustness and can plan the optimal path in a relatively short time compared with other algorithms in this field, thus illustrating its validity and reliability. Eventually, by the aid of the Robot Operating System (ROS), the experimental results of real-world robot navigation indicate that RBA has satisfactory real-time performance and path planning effect, which can be considered as a crucial choice for dealing with path planning problems.