MODELING AND ANALYSIS OF TOUCH SCREEN BASED WIRELESS CONTROL OF FOUR MOTOR ROBOTIC VEHICLE EMPLOYING KNOWLEDGE-BASED SYSTEM AND ENSEMBLE MACHINE LEARNING

Robotic vehicles involve a wide range of steering activities. Most modern commercial utilities are now used and controlled by mobile applications. The modeling and analysis of touch screen based wireless control of four motor robotic vehicle employing skid steering mechanism is presented here. A 4 wheel- steering vehicle's trajectory motion for a given path is controlled using both kinematic and dynamic control laws. The instantaneous radius of curvatures at various time intervals are the sole input that is extracted from the given path needed to be traversed to control the entire vehicle automatically along that path employing back-stepping techniques and other supporting algorithms. The framework for touch screen layout for this application has been developed taking into consideration the response of the skid steering mechanism. This work aims at developing an algorithmic approach with which both the path and the speed at various segments of the path that the robot/drone is required to travel can be effectively and accurately traversed by finger drag gesture control input of the human controller on the touch screen employing Ensemble Machine Learning and Geometric Artificial Intelligence. Ensemble Machine Learning approach is used to develop a control model for any given plain ground/floor type on which the robotic vehicle has to move. This will have high utility for real-time robotic applications because ground to wheel resistance cannot be explicitly modeled for different types of grounds since those frictional dynamics properties are very difficult to formulate or calculate. Multi-sensor data fusion is employed for measuring the mechanical parameters pertaining to the vehicle. Ensemble Machine learning approach will measure the output of vehicle kinematics for various ranges of input controls and create a ground specific predictive control/drive model.