现有的前向碰撞预警系统大多采用多个毫米波雷达叠加或毫米波雷达与视觉传感器融合等方式，存在成本高、算法受限等问题。在比较多种传感器的性能及应用优、缺点后，选择双目视觉传感器作为前向碰撞预警系统传感器。将改进后的碰撞时间(Time to Collision，TTC)算法与卡尔曼滤波融合，结合双目视觉传感器，比较TTC值与适应性阈值，评估风险等级，确保行车安全，降低事故率。在Matlab环境下，基于改进算法对两个不同行车场景进行仿真分析。结果表明，与传统的TTC算法相比，融合卡尔曼滤波TTC算法的碰撞时间预警响应及时性和可靠性显著提高。
Improvement of Forward Collision Warning Algorithm Based on Binocular Vision Sensor
The existing forward collision warning systems primarily utilize multiple millimeter-wave radars in combination or integrate millimeter-wave radar with visual sensors.However,these approaches often face challenges such as high costs and algorithmic limitations.After comparing the performance and application advantages and disadvantages of various sensors,a binocular vision sensor has been selected as the sensor for the forward collision warning system.This study integrates an improved Time to Collision(TTC)algorithm with Kalman filtering and utilizes the binocular vision sensor to compare the TTC values against an adaptive threshold to assess risk levels,thereby ensuring driving safety and reducing accident rates.Simulations of two different driving scenarios were conducted in a Matlab environment based on the improved algorithm.The results show that the TTC algorithm incorporating Kalman filtering significantly improves the collision time warning response timeliness and reliability compared to the traditional TTC algorithm.
国家科技期刊平台
NSTL
万方数据
