中文摘要：

为减少交通事件引起的交通延误,提出一种基于粒子群优化神经网络的交通事件检测算法;首先,利用车载激光测距仪和GPS设备作为实验平台,采集了反映路段车辆占有率及车辆运行速度特征的交通参数;其次,利用粒子群（PSO）算法训练随机产生的初始化数据,优化BP神经网络连接权值和阈值;最后,将PSO优化后的BP神经网络作为分类器进行交通事件的自动分类和检测;试验中比较了PSO神经网络算法、BP神经网络算法和经典算法对交通事件的检测效果,PSO神经网络算法在事件检测率（DR）、平均检测时间（MTTD）方面均优于其他目标算法;结果显示,粒子群优化的神经网络算法用于交通事件检测提高了检测性能。

英文摘要：

A new method was proposed for traffic incident detection based on particle swarm optimizer neural network. At first, the ve- hicular laser rangefinder and GPS equipment were used as the experimental platform, which collected the traffic parameters including the road vehicle occupancy rate and the vehicle running speed; Secondly, the particle swarm optimizer （PSO） was used to train the random initial da- ta to optimize the connection weights and thresholds of the back--propagation （BP） neural network; Finally The BP neural network after optimization was used to classify traffic incidents automatically. In the detection experiment, PSO neural network, BP neural network and traditional algorithms were compared in the same testing environment. PSO neural network was superior to the other objective algorithm in incident detection rate （DR） and mean time to detection （MTTD）. Results showed that particle swarm optimizer neural network brought a promising improvement in the detection capability for traffic incident detection.