Research On Robot Path Planning Based On Memristive Neural Network

Intelligent car is not only usually used as a research platform to carry out relevant scientific research and gradually applied in all fields of social life. Because it has the advantages of simple structure, convenient operation, strong function and flexible and has a strong anti-interference to run in a variety of environments, especially can perform tasks in many environment where humans cannot reach or survival, has high research value.Based on the theoretical research of wheeled robot system, and widely referring to the current applications of mobile robot system, we analyze the key problems and research status of relevant technology in this article. On this basis, we used engineering technology, such as image processing, embedded control, software design, to build the experimental platform of mobile robot, and experimentally verified mobile robot navigation and path planning of theoretical research results. Two important control techniques are fuzzy control technology and neural network technology in the field of artificial intelligence. In this article, the neural network and fuzzy control are combined together, which were using their own advantages in the intelligent control field to construct a powerful and intelligent system. In the respect of neural network hardware realization, through continuous study on memristor in recent years, we gained some of its properties including:lower energy consumption and storage function and nanometer size. The appearance of memristor promoted the development of the artificial intelligence field, especially the progress in the research of artificial neural networks. The main contents of this paper are as follows:1. The software and hardware design of mobile robot system is established. The whole system is divided into two parts:the intelligent control and motion control. Intelligent control part is mainly used to achieve high level task like target recognition and path planning; the underlying control section is mainly used to control mobile robots performing a specific action. Among the rest, positioning navigation is based on the characteristics of the ultrasonic sensor and radio frequency module and the three side positioning principle. The communication between the intelligent control board and the motion control board is achieved with the USB mode, so that the control command after the intelligent control board can be quickly spread to the bottom motion control board.2. An ultrasonic positioning and navigation system based on the synchronization of radio frequency signals is adopted. The system consists of a reference node and a target node. In order to improve the real-time performance of the system, the RF module and the ultrasonic module are designed to be independent transceiver modules. Among them, the reference node is made by RF emission module and an ultrasonic transmitting module and fixed on the robot's location on the wall; target node is made by the RF receiving module and an ultrasonic receiver module, and it is installed on the robot. The radio frequency signal is used as the synchronization signal between the reference node and the target node, so that the transmission characteristics of the ultrasonic wave and the principle of the three side positioning are used to realize the real-time positioning of the robot.3. A new intelligent path planning strategy based on the combination of the neural network and fuzzy control algorithm is proposed. First, get the corresponding environmental information by means of image sensor on the robot; memristive neural network for mobile robot environment information was then used to identify and classify, and use a Bayesian algorithm to determine the optimal weights of network, so that we can correctly distinguish the obstacle and target; finally, based on fuzzy inference to fuzzify obstacle position information and target position information, and establish fuzzy rules and through the fuzzy solution produce accurate command driven, and at last make the mobile robot successfully arrive at the designated location. Simulation results show that the proposed path planning method can realize the accurate identification and classification of the unknown environment, and has high timeliness and reliability.The system hardware and system software are used with the modular structure; the system reliability is high, and function modules are clear. At the bottom of the intelligent vehicle is used with USB bus for message transmission, whose function can be conveniently expanded according to the requirement. Finally, the system simulation experiments were done on the location navigation and path planning algorithms.