Multi – Objective Trajectory Tracking for an Autonomous Mobile Robot in Dynamic Environments Using Evolutionary Algorithms

Abstract

Mobile robots have got very significant role in today's modern life affecting nearly all aspects of human beings' lives; from military and agriculture to education, healthcare, social services, and tourism. Mobile robot can be an unmanned aerial vehicle (UAV), an unmanned submarine, or a waiter robot working at the restaurant, or an agent of delivery system. While talking about mobile robots, the first think sparks to think about is the navigation system and trajectory planning for them so that enables them to move in different terrains autonomously. Letting mobile robots run in an autonomous way needs an artificial system to be embedded on them to find the optimal route which is the shortest and the smoothest one with the lowest cost, exactly as a human operator would act, besides avoiding colliding the obstacles in the environment. This is called monitoring and surveillance of mobile robot's motions to improve their behavior in order to choose the best trajectories among the possible solutions. One of the most crucial areas of research in the field of mobile robotics is the development of the best methods for monitoring and routing trajectories for all types of moving unmanned robotic systems. Trajectory tracking is done in order to find a route without encountering environmental obstacles and hurdles. Although, lots of studies have been done for path planning, but most of them were done in static environments which is less like the real-world situation. In a real-life scenario, we may have multiple obstacles in the routing environment and the shapes and sizes of the obstacles may vary from each other. Furthermore, there may be some moving obstacle moving in different directions with different speed and characteristics. This makes the trajectory tracking more difficult for unmanned robotic systems. As the solution space for path planning problems is very wide, it is categorized as NP-Hard problems. The evolutionary algorithms having a heuristic approach to solve the problem have recorded a very significant trail in solving NP-hard problems till now and still there is a great appeal on improving them to be able to get better solutions. Depending on the path planning problem definition which induces us whether to take a discrete approach or a continuous approach to solve the problem, we can choose the appropriate one among the evolutionary algorithms to find the most appropriate solutions. In this Thesis, the investigation and comparison of two different points of view for solving the routing problem of moving robots in dynamic and unknown spaces with the presence of fixed and moving obstacles have been proposed. Our goal is to design and implement an algorithm according to which an optimal path is obtained from the starting point to the target point and also avoids any collision with static and dynamic obstacles during this path by the robot. First, the routing problem is solved with the ant colony optimization algorithm, and then routing is done in the same environment with the fuzzy logic method. The conducted investigations show the more appropriate efficiency of the fuzzy method in terms of simplicity in implementation in time-consuming processes and the results of the investigation of the time and length of the movement path in dynamic and static environments indicate the strength of this method compared to other evolutionary algorithms, especially in a dynamic environment.