基于FPGA的模糊控制自动垂直泊车系统的研究与实现

发布时间：2018-05-03 08:29

本文选题：自动垂直泊车 + 模糊控制　；参考：《大连理工大学》2015年硕士论文

【摘要】：自动泊车技术是目前汽车智能化趋势下的一个研究热点,它可以通过传感器和控制器探测环境信息并控制汽车自动行进到泊车位内,从而避免驾驶员在狭小泊车空间内进行复杂而需要经验的泊车操作,进而有效缓解泊车难的问题。针对现实生活中较为常见的垂直泊车情景,本文对自动垂直泊车系统的模糊控制策略进行了研究,并基于FPGA芯片对系统进行了软硬件协同的嵌入式设计。首先,从阿克曼转向几何原理出发,推导出前轮转向后轮驱动车辆的运动学方程,作为泊车过程的数学理论；并依据车体参数和环境信息计算出各个初始角度下的可行初始泊车区域和倒车过程中的防碰撞约束条件。其次,根据专家经验,在MATLAB模糊工具箱内,开发一种面向自动垂直泊车的角度、速度双输出模糊控制器,使车辆不仅可以调整舵机角度倒车,而且在向后倒车失败时,能够前后移动调整车身姿态,从而可以自适应的完成垂直泊车动作。为了验证算法有效性,在Simulink内搭建仿真模型,使用动画输出的方式检验模糊控制器的控制效果。再次,由于模糊控制器的设计存在较大主观性并需要大量的后续调试工作,本文使用混沌算法和遗传算法组合的优化策略对模糊控制器的隶属度函数参数和部分关键规则做优化。从Simulink动画输出易见,优化后模糊控制器输出的泊车路线更加合理,泊车动作更加简洁。最后,分析和设计超声波传感器在车体上的布局结构,搭建包含感知系统、决策系统和执行系统的自动垂直泊车智能车验证平台；同时,设计泊车系统的软件流程,并使用Xilinx公司的EDK开发平台将优化后的模糊控制器及车位寻找算法进行基于FPGA的软硬件协同嵌入式设计,将设计完成的智能车验证平台在垂直泊车的环境下进行实车验证,对实验结果进行分析。实验结果表明,本论文开发的模糊控制自动垂直泊车系统可以顺利的完成垂直泊车动作,进入停车位后车身姿态良好,达到了预期的控制效果。
[Abstract]:Automatic parking technology is a research hotspot in the current trend of intelligent automobile. It can detect environmental information through sensors and controllers and control the auto moving into parking space. In order to avoid the complex and experienced parking operation in the narrow parking space, the problem of parking difficulties can be effectively alleviated. Aiming at the common situation of vertical parking in real life, the fuzzy control strategy of automatic vertical parking system is studied in this paper, and the embedded design of hardware and software of the system is carried out based on FPGA chip. Firstly, the kinematics equation of the front wheel steering and rear wheel drive vehicle is derived from the Ackermann steering geometry principle, which is regarded as the mathematical theory of parking process. The feasible initial parking area and anti-collision constraint conditions in reversing process are calculated according to the parameters of the car body and the environment information. Secondly, according to the expert experience, in the MATLAB fuzzy toolbox, a kind of angle and speed double output fuzzy controller for automatic vertical parking is developed, so that the vehicle can not only adjust the steering gear angle reversing, but also when the backward reversing fails. The body can be adjusted forward and backward, so that the vertical parking can be completed adaptively. In order to verify the validity of the algorithm, the simulation model is built in Simulink, and the control effect of fuzzy controller is tested by animation output. Thirdly, the design of fuzzy controller is subjective and needs a lot of subsequent debugging work. In this paper, the parameters of membership function and some key rules of fuzzy controller are optimized by the combination of chaos algorithm and genetic algorithm. The output from Simulink animation is easy to see, and the output of the fuzzy controller is more reasonable, and the parking action is more concise. Finally, the layout structure of ultrasonic sensor on the car body is analyzed and designed, and the automatic vertical parking intelligent vehicle verification platform including sensing system, decision making system and execution system is built. At the same time, the software flow of parking system is designed. Using the EDK development platform of Xilinx Company, the optimized fuzzy controller and parking space search algorithm are used in the hardware and software co-embedded design based on FPGA, and the completed intelligent vehicle verification platform is verified under the environment of vertical parking. The experimental results are analyzed. The experimental results show that the fuzzy control automatic vertical parking system developed in this paper can successfully complete the vertical parking movement, and the body posture is good after entering the parking space, and the desired control effect is achieved.
【学位授予单位】：大连理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U463.6;TN791

【参考文献】