The path planning control of two wheeled mobile robot using extended Kalman Filter technique

Abstract

The project proposes a model of path planning control for two-wheeled mobile robot using the Extended Kalman Filter (EKF) technique. The main objective of this work is to minimize the error in calculation for the robot’s position, orientation, localization, and navigation. The results are compared to the past research using Linear Kalman Filter (LKF) technique. Due to the limitation of the LKF that is unable to handle non-linear dynamics measurement equations, makes the kinematics model of the robot is not accurate. The cause of uncertainty for the infinite precision of the distance between wheeled axes of the robot, noise from the sensor, and the slippage of the direction of motion in the perpendicular direction of the robot in the relative localization make it worse. This work is aimed to overcome the restraints of the path planning control of the robot by using a non-linear model with odometry and locomotion system using an optical encoder that can monitor the wheel position and speed. This is achieved by developing an estimator that recursively converges noise from sensor data with a model of the system dynamics of the robot. The distance between the robot and the desired landmark is set as input to EKF. To validate the accuracy of the proposed model, a simulation using MATLAB software is conducted. The initialization position is marked out at first, and the robot is programmed to navigate from point A to point B in a specific space. To verify the robot's position, prior knowledge of the environment is needed. The environment is set to be safe from obstacles therefore the robot doesn’t have a behaviour of obstacle avoidance using external sensors. The results of the proposed model are expected to be achieved successfully where the robot can course itself according to the equation of motion that has been set at the initialization state as path planning to the desired destination. The EKF technique will be verified to be the best method for the localization and navigation of the robot with reliable estimation of position and orientation of the robot compared to the standard Kalman Filter.