Static and Dynamic Obstacle Avoidance with Dynamic Analysis for Mobile Robots

Abstract

The self-driving car or robot is a key research area for the researchers due to increasing number of road accidents and erroneous driving. One of the major challenges in designing of a mobile robot is the autonomous travel of the robot. The key factor associated with the autonomous travel is the obstacle avoidance. The obstacle avoidance is classified as obstacle detection and avoidance control. Also, the dynamics of the mobile robot plays an important role during obstacle avoidance. A comprehensive literature review on mobile robot, obstacle detection and avoidance is done. Literature review on the overwhelming controller, physical model reduction, walking mobile robot, fault detection and isolation and bond graph methodology is also done. Accordingly, the objectives of the work are defined at the end of this literature review. The hybrid obstacle avoidance algorithm for static objects and fuzzy based obstacle avoidance algorithm for dynamic objects are proposed in this Thesis work. The model of bicycle vehicle model and four-wheel model of mobile robot are developed using bond graph theory. The effectiveness of the obstacle avoidance algorithm is tested on these models. The validation of the simulation results for obstacle detection and avoidance is also done with the experimentation-work. The obstacle avoidance for mobile robots is also done using overwhelming controller. Also, the quadruped robot based on Jansen linkage mechanism is developed and is used for obstacle avoidance application in a closed boundary-environment. The fuzzy logic controller with a set of rules is proposed to avoid the boundaries (static obstacle) of a closed environment. The simulation results are compared with the experimental result. The physical models for mobile robots are also reduced using eigenvalue sensitivity method to reduce the simulation time and system complexity. The response of the reduced model is compared with the full model.