基于毫米波雷达定位的汽车三维防碰撞算法研究

发布时间：2019-01-15 07:29

【摘要】：近年来,随着互联网+汽车概念的提出,无人驾驶汽车、智能汽车的研发如火如荼。汽车防碰撞技术作为无人驾驶汽车、智能汽车的关键技术之一,也是避免交通事故的关键技术,成为国内外研究的热点。测量技术,是汽车防碰撞技术中的关键技术之一。经典的线性调频连续波(LFMCW)雷达测量方法用FFT变换进行数据处理,采样点增多时计算量显著增长,使测量系统实时性不理想。为解决此问题引入了互相关函数测量方法,结合互相关函数测量法测得的值,对采样信息的有效采样点部分进行FFT变换,可以快速高效地测出更精准的距离信息。汽车安全距离模型,是汽车防碰撞系统中的核心技术。现有的安全距离模型多以实际速度等实际信息为参考因素,没有合理考虑相对速度等相对信息的变化情况没有考虑道路状况、天气情况等复杂交通环境带来的影响。基于此,本文参考现有的安全距离模型的构造原理,基于车辆制动过程,从相对速度的角度,根据相对速度Δv等于0、大于0和小于0三种不同的情况建立了三种不同的安全距离模型。不仅合理考虑了相对速度对安全距离的影响,还考虑了不同材质的路面对安全距离模型的影响,同时考虑天气因素和驾驶员的驾驶习惯反应速度等因素对安全距离的影响。最后,利用防碰撞系统实时探测计算得到的交通信息,根据制定的汽车防碰撞预警策略,通过声音、灯光和液晶显示屏分等级进行报警,提醒驾驶员汽车当前行驶的安全状态,必要时启动汽车制动系统进行自动刹车,从而有效减少碰撞事故的发生,保证汽车出行的方便、安全、舒心。结合BP神经网络进行研究,有效解决了影响汽车安全距离因素多,汽车安全距离模型比较复杂,存在大量的非线性变化等问题,能够满足实际交通过程中需要不断更新模型参数,甚至不断更新模型的要求,使得汽车防碰撞系统能够根据交通状况实时更新安全距离模型及参数。
[Abstract]:In recent years, with the introduction of the concept of Internet vehicles, driverless vehicles, intelligent vehicles are in full swing. As one of the key technologies of driverless vehicles and intelligent vehicles, anti-collision technology is also the key technology to avoid traffic accidents, and has become a hot research topic at home and abroad. Measurement technology is one of the key technologies in automobile anti-collision technology. The classical linear frequency modulation continuous wave (LFMCW) radar measurement method uses FFT transform to process the data. When the number of sampling points increases, the amount of calculation increases significantly, which makes the real-time performance of the measurement system unsatisfactory. In order to solve this problem, the cross-correlation function measurement method is introduced. The effective sampling point part of sampling information can be transformed by FFT with the value measured by cross-correlation function measurement method, and more accurate distance information can be measured quickly and efficiently. Vehicle safety distance model is the core technology of automobile anti-collision system. Most of the existing safety distance models take the actual information such as actual speed as reference factors, and do not consider the change of relative information such as relative speed reasonably, and do not take into account the influence of complex traffic environment such as road condition, weather condition and so on. Based on this, this paper refers to the construction principle of the existing safety distance model, based on the vehicle braking process, from the angle of relative velocity, according to the relative velocity 螖 v = 0, Three different safety distance models are established for three different cases: greater than 0 and less than 0. Not only the influence of relative speed on safety distance is considered reasonably, but also the influence of different pavement materials on safety distance model is considered, and the influence of weather factors and driver's driving habit reaction speed on safety distance is also considered. Finally, using the anti-collision system to detect and calculate the traffic information in real time, according to the anti-collision warning strategy, alarm is made through the grade of sound, light and liquid crystal display to remind the driver of the current safety state of the vehicle. When necessary, the auto brake system can be started to reduce the collision accident and ensure the convenience, safety and comfort of the vehicle travel. Based on the research of BP neural network, the problems such as many factors affecting vehicle safety distance, complex vehicle safety distance model and a large number of nonlinear changes are effectively solved. It can meet the requirements of updating the model parameters and even updating the model constantly in the actual traffic process, so that the vehicle anti-collision system can update the safety distance model and parameters according to the traffic conditions in real time.
【学位授予单位】：兰州理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN958;U463.6

【参考文献】