王浥晨, 王学文, 马凯,等.基于全向激光雷达的液压支架护帮板与采煤机防撞监测预警系统研究[J].河南理工大学学报（自然科学版）,2025,44(3):44-52.

WANG Y C WANG X W, MA K, et al. Research on the anti-collision monitoring and early warning system for hydraulic support panels and shearers based on omnidirectional laser radar[J]. Journal of Henan Polytechnic University(Natural Science) , 2025, 44(3): 44-52.

基于全向激光雷达的液压支架护帮板与采煤机防撞监测预警系统研究

王浥晨1,2, 王学文1,2,3, 马凯1,2, 李辉1,2, 倪强4, 谢嘉成1,2,3

1.太原理工大学 机械工程学院，山西 太原  030024；2.太原理工大学 煤矿综采装备山西省重点实验室，山西 太原  030024；3.太原理工大学 智能采矿装备技术全国重点实验室，山西 太原  030024；4.山西康伟集团有限公司，山西 长治 046006

摘要: 目的 针对采煤机运行过程中滚筒与液压支架护帮板易发生碰撞，而现有监测设备传感器数据融合处理复杂、精度受限且有滞后性等问题。  方法 提出一种基于全向激光雷达的液压支架护帮板与采煤机空间相对位姿解算方法，对液压支架护帮板与采煤机防撞监测进行研究。该方法可分为3部分：（1）将二维激光雷达和精密旋转云台组成的三维激光扫描系统安装在液压支架顶梁下方，对目标物进行数据采集，建立护帮板与采煤机滚筒位姿解算模型；（2）结合SIFT 算法提取特征点，对采煤机滚筒和液压支架护帮板的相对位姿进行实时解析，并建立行为数据库；（3）建立分级预警机制，通过Unity 3D可视化界面输出护帮板状态是否异常，并及时发出警报。  结果 对实验室10组液压支架护帮板与采煤机滚筒位姿信息多次解算并判断是否发生碰撞，结果表明：护帮板伸缩位移平均误差1.5 cm，护帮板倾角平均误差1.7 °时关键点坐标解算x轴相对误差为4.5%，y轴相对误差为3.6%，z轴相对误差为2.8%。5种工况下， x，y，z坐标轴方向绝对误差均值均<3.5 cm，通过位姿信息解算能够有效评估其截割干涉状态并确保评估结果与实际情况相符，满足应用需求。  结论 该方法以较低成本获取护帮板和滚筒在三维空间中的相对位姿信息，且精度较高，实时监测碰撞状态，有利于提高井下设备的安全性和综采工作面的稳定性。

关键词:采煤机滚筒;液压支架护帮板;二维激光雷达;三维扫描;防撞监测系统

doi:10.16186/j.cnki.1673-9787.2024080051

基金项目:国家自然科学基金资助项目（52474178）；中央引导地方科技发展项目（YDZJSX2022A014）；山西省留学人员科技活动择优重点资助项目（20230008）；山西省科技成果转化引导专项项目（202304021301036）

收稿日期:2024/08/29

修回日期:2024/12/11

出版日期:2025-04-18

Research on the anti-collision monitoring and early warning system for hydraulic support panels and shearers based on omnidirectional laser radar

WANG Yichen1,2, WANG Xuewen1,2,3, MA Kai1,2, LI Hui1,2, NI Qiang4, XIE Jiacheng1,2,3

1.College of Mechanical Engineering， Taiyuan University of Technology， Taiyuan  030024， Shanxi， China；2.Shanxi Key Laboratory of Fully-Mechanized Coal Mining Equipment， Taiyuan University of Technology， Taiyuan  030024， Shanxi， China；3.National Key Laboratory of Intelligent Mining Equipment Technology， Taiyuan University of Technology， Taiyuan  030024， Shanxi， China；4.Shanxi Kangwei Group Co.， Ltd.， Changzhi  046006， Shanxi， China

Abstract: Objectives Collisions between the shearer roller and the face guard of the hydraulic support are prone to occur during the operation of the shearer. However， existing monitoring methods are hindered by the complexity of sensor data fusion， limited accuracy， and inherent lag，  which results in poor monitoring precision. Methods A spatial relative pose estimation method based on omnidirectional LiDAR was proposed to monitor collisions between the face guard of the hydraulic support and the shearer roller. The method comprises three main components： （1） A 3D laser scanning system， consisting of a 2D laser radar and a precision rotating platform， was installed beneath the hydraulic support beam to collect data from the target object. A pose estimation model for the face guard and the shearer roller was developed. （2） The SIFT algorithm was employed to extract feature points， allowing real-time analysis of the relative pose between the shearer roller and the face guard， and the creation of a behavior database. （3） A hierarchical warning mechanism was implemented. The Unity 3d visualization interface was used to display the status of the support plates， indicating whether they were abnormal， and to promptly issue alerts.  Results The relative pose information of 10 sets of hydraulic support face guards and shearer rollers was calculated in the laboratory to determine the potential for collisions. The results showed an average displacement error of 1.5 cm for the extension of the support plate， and an average tilt angle error of 1.7°. The relative errors were 4.5% for the x-axis， 3.6% for the y-axis， and 2.8% for the z-axis. Under five different working conditions， the mean absolute errors for the x， y and z axes were all less than 3.5 cm. The pose information effectively evaluated the cutting interference state， with results consistent with the actual situation， meeting the application requirements.  Conclusions The method enables the acquisition of the relative pose information of the face guard and shearer roller in three-dimensional space at relatively low cost， with high accuracy. Real-time collision monitoring improves the safety of underground equipment and the stability of the fully mechanized mining face.

Key words: shearer roller; face guard of hydraulic support; 2D laser radar; 3D scanning; anti-collision monitoring system