Optimal Pose Estimation with Particle Filters Using Unpowered Wheels

Accurate localization of the mobile robot is an essential element of autonomous vehicles that guarantees reliable navigation and secure passage along predetermined paths. Wheel odometry data from the powdered wheels and GPS-based location are the two main information sources for the pose estimation However, depending entirely on GPS might result in subpar performance owing to inherent errors and signal disturbances, particularly in metropolitan settings with tall buildings or tunnels. In addition, wheel odometry is used to increase precision and real-time responsiveness however, the reliability of the wheel odometry is reduced as the surface becomes slippery or the torque generation is very high. This resulted in the wheel slipping which is target as the problem statement in this research paper. In this proposed approach, the wheel odometry data is taken from the wheel encoder with unpowered wheel is taken that only change if the actual position is changed. As the unpowered wheel can only rotate when the mobile robot is moved however, the powered wheel may rotate even the robot is on stand still in the slippery surface. This proposed technique is implemented in various trails and the resulted are compared with various technique to check the reliability of the proposed technique. It has been experimentally obtained that the accuracy in the localization of the mobile robot is increased by 33.16% as compared to the conventional technique. This proposed technique may be implemented in path tracking, path planning, autonomous navigation, trajectory optimization, etc., to achieve accurate localization.