Study On Autonomous Intelligent Car Based On Hierarchical Fuzzy Logic Control

With the development of science and technology, the application of intelligent body has already been a part of human’s daily life, including Smart Home, medical assistant, YuTu moon robot, Google robot AlphaGo, military robot, intelligent car, etc., which indicates that the AI has already put deep footprints on the human life. Intelligent car, as one of the scientific achievements of artificial intelligence, has got lots of technical improvements and plays a more and more important role in the scientific research and human life in the future.Based on the previous studies, the algorithm of hierarchical fuzzy logic control which is used in the obstacle avoidance of intelligent car has been studied in this paper. This algorithm divides all the fuzzy inputs into several layers according to their impacts on the consequence of intelligent system, decreasing the dimensionality of fuzzy rules number. In this paper, STM32 chip is chosen as the main controller and its application in the hardware control has been studied. Also Keil MDK coding tool is used to do the programming and code compilation.In the first chapter, the paper firstly describes the background and historical significance of the subject, and also cites some outstanding achievements in the field of intelligent car in academic circles at home and abroad. In second chapter, the mathematical model and mathematical expressions of intelligent car have been made based on knowledge of dynamics, paving the way to the calculating, programming and coding of algorithm in the future’s research. At the same time, the physical components of the car has been detailed, including power systems, motor drive module, commonly-used sensors, wired communication module, STM32 controller and obstacle detection module. These introductions help us better understand the structure and the working theory of car. The third chapter introduces the basic principle and algorithm of hierarchical fuzzy control which has successfully been simulated by the Matlab software and verified by experiment. The simulation result validates the effectiveness of double layer fuzzy algorithm for solving the "curse of dimensionality" problem. In the fourth chapter, the software development system and coding environment are well focused on. The Keil MDK software system is chosen as the coding tool, used to do the programming of all the components of the car, including obstacle detection module, motor drive module, speed detection module and CAN bus communication module. All these modules are relatively independent and the coding environment is friendly. Finally, the conclusion about this subject, the harvest I’ve got and all the work I’ve done is well made, as well as the wishes for future’s research.