一种改进的毫米波雷达多目标检测算法

摘要

智能驾驶主要依据前视雷达进行目标探测，基于毫米波雷达的多目标检测算法是目前的研究热点。但现有的基于FMCW的雷达多目标检测算法存在大量虚警、漏检的问题，且无法获得目标的方位信息，很难应用到复杂多变的道路环境中。提出了一种改进的毫米波雷达多目标检测算法。该算法在传统FMCW算法的基础上，提出相位差和功率差联合匹配模型进行频率匹配的思想，较好地解决了FMCW算法中频率匹配差错、无法得到目标方位信息等问题。对该算法进行了多目标模拟仿真和实测道路试验，仿真和实验结果表明，基本上能正确检测出多目标的距离、速度、方位角，且性能可靠稳定。%Intelligent driving relies on forward-looking radar for target detection, and recently multi-target detection algorithm based on millimeter-wave radar becomes a research hotspot. However,certain deficien-cies still exists in the traditional FMCW radar target detection algorithm, induding false alarm, missed de-tection, and so on,thus being unable to obtain the azimuth information of the targets, so this traditional al-gorithm couldn’t be used in complex and variable road conditions. Based on the traditional FWCW algo-rithm,a modified millimeter-wave radar algorithm is proposed, and the idea of joint matching model by phase difference and power difference to match the frequency is given, thus fairly solving the problems in FMCW, such as frequency error matching and missing azimuth information of targets. Multi-target simula-tion and actual measurement indicate that this new algorithm,reliable and stable performance could correct-ly detect the distance, speed and azimuth of multiple targets.