基于惯性/仿生视觉/激光雷达的智能感知无人系统

INS/Bionic Vision/LiDAR Based Intelligent Perception Unmanned System

针对当前缺乏对集成自主导航、目标探测、环境感知技术为一体的无人系统的研究,提出一种基于惯性/仿生视觉/激光雷达的智能感知无人系统,将基于仿生偏振视觉/惯性/激光雷达的组合导航技术、基于YOLOv7的目标探测与环境感知技术应用在无人系统上,一方面可避免战时卫星导航易被反侦察与干扰的问题,提高无人系统在复杂环境下(野外环境与封闭未知环境)的隐蔽性与自主导航能力,增强组合导航系统的稳定性与鲁棒性;另一方面针对复杂环境利用视觉相机进行智能感知,提高无人系统的目标探测以及自主避障能力。利用搭建的无人系统进行实验验证。结果表明,该方法可以实现无人系统的自主导航以及目标识别功能。与传统方法相比,提高了自主避障能力。其导航精度在野外环境下误差在1°以内。

Aiming at the current lack of research on unmanned systems that integrate autonomous navigation,target detection,and environmental perception technologies,an INS/Bionic Vision/LiDAR based intelligent perception unmanned system is proposed.The integrated navigation technology based on bionic polarization vision/INS/LiDAR and the target detection and environmental perception technology based on YOLOv7 are applied to unmanned systems.On the one hand,it can avoid the problem that wartime satellite navigation is easy to be anti-reconnaissance and interference,improve the concealment and autonomous navigation ability of unmanned systems in complex environments(field environment and closed unknown environment),and enhance the stability and robustness of integrated navigation systems.On the other hand,visual cameras are used for intelligent perception in complex environments to improve the target detection and autonomous obstacle avoidance capabilities of unmanned systems.The built unmanned system is used for experimental verification.The results show that the method can realize the autonomous navigation and target recognition function of the unmanned system.Compared with the traditional method,the autonomous obstacle avoidance ability is improved.The navigation accuracy error is within 1°in the field environment.