复杂噪声环境下的ITS防碰撞技术研究

发布时间：2018-08-12 08:08

【摘要】：随着城市汽车化速度的加快,城市的交通拥挤、交通事故和交通污染等问题日益加剧。无论是发达国家还是发展中国家都受到这些交通问题的困扰。我国作为快速发展的发展中国家,此问题更是特别突出。追尾碰撞事故在城市交通事故中占据绝大部分,并且造成了严重的经济损失。因此,为了解决追尾碰撞事故问题,本文对复杂噪声环境下ITS(Intelligent Transportation System)防碰撞技术做了大量理论推导和实验研究,并提出了防碰撞方案。本文的ITS防碰撞方案由信号去噪、信号检测、动力学模型以及判决模型四个部分组成。本文方案所使用的信号为通用车载雷达信号,并且该信号会通过信号去噪进行预处理;去噪后的雷达信号将被输入检测器中进行信号检测,得到检测概率并且判断前方是否有车出现;同时动力学模型根据前后车的行驶状态计算安全距离;最后,判决模型将以上结果作为输入并设置相关门限进行判决,得到后车是否应该刹车的结果。本文完成的主要工作如下:第一,本文实现了基于概率密度函数的区域阈值EMD去噪算法,并且将其与小波去噪和其他基于EMD的去噪算法进行性能对比。实验证明前者去噪性能好于其他算法。第二,本文根据信号相关性的不同提出了时间相关性检测模型、空间相关性检测模型以及时空相关性检测模型。通过相同参数下的仿真实验证明基于GLRT的检测模型性能较好,特别是基于频域GLRT的检测模型性能表现最好。第三,本文结合动力学模型和判决模型提出了复杂噪声环境下的ITS防碰撞方案。和其他防碰撞方案相比本文方案具有优良性能的去噪算法和信号检测器,使得其在环境噪声较大的情况下仍然可以准确的检测到前方的车辆。其次,本文提出的动力学模型和判决模型结合了机动车的行驶过程,判决结果具有合理性和可靠性。最后,本文假设的防碰撞仿真实验场景充分考虑了交通中的真实场景,因此本文实验结果更具有真实性。通过仿真实验证明,本文提出的复杂噪声环境下的ITS防碰撞方案切实可行,并且能提高驾驶过程的安全性。
[Abstract]:With the acceleration of urban automobile, traffic congestion, traffic accidents and traffic pollution are becoming more and more serious. Both developed and developing countries are beset by these transport problems. China as a developing country with rapid development, this problem is especially prominent. Rear-end collision accidents occupy most of urban traffic accidents and cause serious economic losses. Therefore, in order to solve the problem of rear-end collision accident, this paper has done a lot of theoretical derivation and experimental research on anti-collision technology of ITS (Intelligent Transportation System) in complex noise environment, and put forward anti-collision scheme. The anti-collision scheme of ITS is composed of signal denoising, signal detection, dynamic model and decision model. The signal used in this scheme is a general vehicle radar signal, and the signal will be pre-processed by signal denoising; the de-noised radar signal will be input into the detector for signal detection. At the same time, the dynamic model calculates the safe distance according to the driving state of the vehicle. Finally, the decision model takes the above results as input and sets the relevant threshold to judge. Get the result of whether the rear car should brake. The main work of this paper is as follows: first, this paper implements a region threshold EMD denoising algorithm based on probability density function, and compares its performance with wavelet denoising and other EMD based denoising algorithms. Experiments show that the former is better than other algorithms. Secondly, time correlation detection model, spatial correlation detection model and spatio-temporal correlation detection model are proposed according to different signal correlation. The simulation results under the same parameters show that the performance of the detection model based on GLRT is better, especially the performance of the detection model based on frequency domain GLRT is the best. Thirdly, combining the dynamic model and the decision model, this paper proposes a collision prevention scheme for ITS in complex noise environment. Compared with other anti-collision schemes, the proposed scheme has excellent performance in denoising algorithm and signal detector, which makes it possible to detect the vehicle in front of the vehicle accurately under the circumstance of high ambient noise. Secondly, the dynamic model and the decision model are combined with the driving process of the vehicle, and the result is reasonable and reliable. Finally, the hypothetical anti-collision simulation scene takes full account of the real scene in traffic, so the experimental results in this paper are more authentic. The simulation results show that the proposed anti-collision scheme of ITS in complex noise environment is feasible and can improve the safety of driving process.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U495

【相似文献】