基于传感器融合的矿井运输车辆环境感知研究  

Study on Environmental Perception of Mine Transportation Vehicles Based on Sensor Fusion

作　　者：蒲德全 高振刚 李鹏洲 PU Dequan;GAO Zhengang;LI Pengzhou(College of Mechanical Engineering,Inner Mongolia University of Technology,Hohhot 010051,China;College of Big Aeroplanes,Ordos Institute of Technology,Ordos 017000,China)

机构地区：[1]内蒙古工业大学机械工程学院,内蒙古呼和浩特010051 [2]鄂尔多斯应用技术学院大飞机学院,内蒙古鄂尔多斯017000

出　　处：《金属矿山》2024年第10期176-181,共6页Metal Mine

基　　金：内蒙古自治区关键技术攻关项目(编号:2020GG0200);内蒙古自治区自然科学基金项目(编号:2021MS05051);内蒙古自治区高等学校科学研究重点项目(编号:NJZZ21058)。

摘　　要：针对矿井环境的复杂性以及单一传感器在井下感知中精度较低等问题,为降低矿井运输车辆发生安全事故的概率,提出了基于激光雷达与超声波雷达的信息融合状态估计算法。算法基于改进的迭代扩展卡尔曼滤波(Iterative Extended Kalman Filter,IEKF)进行设计,充分利用了激光雷达检测精度高、不受光线突变影响及超声波雷达可补偿近距离盲区的特点。将2种雷达的坐标系标定到同一坐标系下,以超声波雷达数据为先验状态,再利用激光雷达数据更新先验状态,得到矿井运输车辆与障碍物、巷道之间的精确位置关系。采用PreScan与Simulink联合仿真验证算法的有效性。仿真结果表明:所设计的算法在基于PreScan模拟的矿井环境中,能较为准确地检测矿井运输车辆与障碍物的位置关系,与单一传感器检测结果相比,既降低了检测距离的平均绝对误差,又有效解决了激光雷达近距离检测的数据漂移问题,有效提升了矿井作业的安全性,为矿井运输车辆防碰撞预警及智能驾驶提供了精确的环境感知信息。In order to reduce the probability of safety accidents occurring in mine transportation vehicles,an information fusion state estimation algorithm based on LiDAR and ultrasonic radar is proposed to address the complexity of the mine environment and the low accuracy of a single sensor in mine sensing.The algorithm is designed based on the improved Iterative Extended Kalman Filter(IEKF),which makes full use of the features of high detection accuracy of Lidar,which is not affected by the sudden change of light,and the ultrasonic radar that can compensate for the blind zone in the near distance.The coordinate systems of the two radars are calibrated to the same coordinate system,and the ultrasonic radar data are used as the a priori state,and then the laser radar data are used to update the a priori state to obtain the precise positional relationship between the mine transportation vehicles and the obstacles and roadways.The validity of the algorithm is verified by joint simulation of PreScan and Simulink,and the simulation results show that the designed algorithm can more accurately detect the positional relationship between mine transportation vehicles and obstacles in the mine environment simulated based on PreScan,and compared with the detection results of a single sensor,it not only reduces the average absolute error of detection distance,but also effectively solves the data drift problem of the Lidar close range detection,and effectively solves the data drift problem of the Lidar close range detection.At the same time,it effectively solves the data drift problem of Lidar near distance detection,effectively improves the safety of mine operation,and provides accurate environment sensing information for anti-collision warning and intelligent driving of mine transportation vehicles.

关 键 词：传感器信息融合 矿井环境 卡尔曼滤波 激光雷达 超声波雷达 

分 类 号：TD67[矿业工程—矿山机电]