Wheeled mobile robots (WMR) find increasing application in today’s market. The current trend shows that there is increase in need for intelligent autonomous vehicles in manufacturing as well as in other dominant fields such as health care, defense, transportation etc. Path planning, path tracking control and obstacle avoidance are some of the abilities which is essential for an autonomous WMR. Most of the path tracking controllers are designed for tracking a smooth reference path having continuous curvature. However, a path planner which is the upper layer in layered control architecture used for finding a feasible reference path, usually provide optimized waypoint path as the reference path. Such path consists of lines/links connecting the waypoints. The path thus generated is unsmooth in nature. The continuous tracking controllers consider the waypoint tracking problem as a point to point regulation problem rather than path tracking problem. Enough attention is not paid on waypoints tracking problem, especially the problem of controlling the robot through the line/link connecting the waypoints, which is actual reference path. Therefore, an effort has been made in this research work to address this waypoint tracking problem for a differential drive robot. A controller is developed for this purpose. It consists of a kinematic path tracking controller and a dynamic torque saturation controller. The kinematic path tracking controller enable path tracking while the dynamic torque saturation controller is utilized for avoiding the wheel slip. The dynamic controller uses a motor torque saturation strategy to limit the motor torque below the torque value above which the wheel slip occur. A detailed study is carried out in three stages. In the first stage, control of robot by utilizing the kinematic controller alone is studied and in second stage the dynamic torque saturation controller is added to avoid the wheel slip. Cross-track error and linear velocity tracking error is reduced in the second stage. However, the study showed a possibility of a modified control law for further reducing the cross-track error. Hence, in the third stage a modified control law is developed to further reduce the cross-track error. Simulation study as well as experimental study has been carried out to analyze the performance of the developed controller. For this study, the control gains of the developed controller are tuned with the help of multi-objective particle swarm optimization technique. An instrumented experimental setup with a measurement system is developed in-house to carry out the experimental study. The measurement system is used for obtaining the position of the mobile robot. The position information is further utilized to obtain the velocity of the robot by differentiation.The results of simulation as well as experimentation carried out suggest that the kinematic control law developed in the research work enables the waypoint tracking of differential drive robot. The dynamic controller reduces the cross-track error as well as linear velocity tracking error. The simulation results and experimental results of the dynamic controller are closer when compared with kinematic control result. The dynamic controller reduces the errors by avoiding the wheel slip. The modified control law developed in the third stage of the study is found to further reduce the cross-track error.Obstacle avoidance is another important area to be focused for an autonomous wheeled mobile robot as mentioned above. Obstacle avoidance of multi-robot system is a topic which is still in its developmental stage. This topic is addressed in this research work using two ways- (i) using a Bug-1 based strategy and (ii) using a reinforcement learning based strategy. The Bug-1 based strategy is developed by modifying the Bug-1 strategy developed by earlier researchers. The objective is to reduce the travel time and the average distance travelled by the robot for multi-robot obstacle avoidance. In multi-robot obstacle avoidance using the reinforcement learning, a technique called Q-learning is utilized to make the robot learn by itself - how to avoid the obstacle. It also enables the robot to move in formations as well as to change the formation as and when required. The system learns and improves the obstacle avoidance strategy by itself and also enables the progress towards artificial intelligence with respect to obstacle avoidance.Simulation study is carried out to verify the performance of the multi-robot obstacle avoidance strategies. From the simulation study, it is concluded that the Bug-1 based new strategy is a better strategy when compared with Bug-1 strategy. Experiments are carried out to verify the simulation results. The results of the simulation study on reinforcement learning based obstacle avoidance strategy suggest that this strategy is a feasible solution for multi-robot obstacle avoidance.