Navigation and Control of Mobile Robots

The rapid development of robotics has benefited by more and more people putting their attention to it. In the 1920s, ‘Robota’, a similar concept, was first known to the world. It is proposed in Karel Capek’ s drama, Rossum’ s Universal Robots (RUR). From then on, numbers of automatic machines were created all over the world, which are known as the robots of the early periods. Gradually, the demand for robots is growing for the purpose of fulfilling tasks instead of humans. From industrial uses, to the military, to education and entertainment, di↵erent kinds of robots began to serve humans in various scenarios. Based on this, how to control the robot better is becoming a hot topic. For the topic of navigating and controlling mobile robots, number of related problems have been carried out. Obstacle avoidance, path planning, cooperative work of multi-robots. In this thesis, we focus on the first two problems, and mention the last one as a future direction in the last part. For obstacle avoidance, we proposed algorithms for both 2D planar environ- ments and 3D space environments. The example cases we raise are those that need to be addressed but have always been ignored. To be specific, the motion of the obstacles are not fixed, the shape of the obstacles are changeable, and the sensors that could be deployed for underwater environments are limited. We even put those problems together to solve them. The methods we proposed are based on the biologically inspired algorithm and Back Propagation Neural network (BPNN). In addition, we put e↵orts into trajectory planning for robots. The two scenarios we set are self-driving cars on the road and reconnaissance and surveillance of drones. The methods we deployed are the Convolutional Neural Network (CNN) method and the two-phase strategy, respectively. When we proposed the strategies, we gave a detailed description of the robot systems, the proposed algorithms. We showed the performance with simulation results to demonstrate the solutions proposed are feasible. For future expectations, there are some possible directions. When applying traditional navigation algorithms, for example, biologically inspired algorithms, we have to pay attention to the limitations of the environment. However, high-tech algorithms sometimes are not computationally friendly. How to combine them together so as to fulfill the tasks perfectly while the computational e ciency is not too high is a worthy topic. In addition, extending the obstacle avoidance al- gorithms to more competitive situations, such as applying to autonomous UAVs, is also being considered. Moreover, for cooperation among multi robots, which could be regarded as Network Control System (NCS), the issues, such as how to complete their respective tasks, how to choose the optimal routes for them are worth attention by researchers. All in all, there is still a long way to go for the development of navigation and control of mobile robots. Despite this, we believe we do not need to wait for too long time to see the revolution of robots