Real-time path planning for autonomous vehicle based on teaching–learning-based optimization

Abstract This paper presents an online path planning approach for autonomous tracked vehicle in a cluttered environment based on teaching–learning-based optimization (TLBO), considering the smoothness, and potential collision with surrounding obstacles. In order to plan efficient that allows be autonomously navigated environments, problem is solved as multi-objective problem. First, perception fully achieved by means of inertial measurement unit (IMU), wheels odometry, light detection ranging (LiDAR). compensate sensors drift achieve more reliable data improve localization estimation corrections, fusion between outputs LiDAR, IMU made through extended Kalman filter (EKF). Then, TLBO proposed applied determine optimum path, where objectives are find shortest reach target destination, maximize while avoiding obstacles, taking into account dynamic algebraic constraints. To check performance algorithm, it compared simulation genetic algorithm (GA), particle swarm (PSO), hybrid GA–PSO algorithm. Finally, real-time experiments robot operating system (ROS) implementation conducted validate effectiveness