统计显示，2003年至今，煤矿每年平均死亡4000余人，安全生产形势十分严峻，在生产过程中实时监控与管理已刻不容缓。若能实时地、较为精准地监测到井下人员的位置信息，矿难发生时就能更及时、有效地实施救援，减小损失，提高井下作业的安全性。因此，开展煤矿井下定位系统设计、提高定位精度以及定位误差分析具有重要的现实意义和研究价值。本课题是围绕东北某中小型煤矿展开的，论文研究内容也是基于该煤矿具体的井下环境进行的，主要开展了井下定位系统总体设计、模拟实验平台搭建与接收信号强度（RSSI）测量实验；RSSI测量值、基于RSSI得到的测距值的误差修正与分析；测距误差对最终定位误差的影响规律与实验验证等研究工作。取得的主要研究成果如下：1）调研对比了各种无线传感器网络（WSN）定位算法，确定了本课题基于RSSI定位算法开展研究。在基于测距技术的定位中，RSSI算法的测距精度虽不如其他方法好，但其系统所需硬件简单，且易与现有硬件匹配，通信和计算开销比较小，采用这种算法可减小整个网络的造价。 2）根据煤矿井下多为巷道，人员和机车也多在巷道中活动的特点，提出锚节点一维线性部署方案，介绍了煤矿井下定位系统的总体设计方案，搭建了定位系统模拟仿真平台，进行基于RSSI的测量实验。实验结果表明随着锚、盲节点间距增大，接收信号强度降低，符合RSSI路径-损耗模型，验证了该套定位系统的可靠性，为后续基于RSSI的测距做好了实验准备。3）从RSSI测量值和基于RSSI得到的测距值两个层面上开展了误差的修正与分析研究，以期减小误差累积，提高最终定位的精度。实验结果表明：数据筛选法对RSSI测量值的修正结果优于统计均值法；经最小二乘法修正，测距误差的最大值和平均值分别由2.73 m、0.97 m降低为1.62 m、0.54 m。尽管如此，测距误差仍然不可避免。4）理论分析并通过Matlab仿真测距误差对最终定位误差的影响规律，并进行相关实验验证。理论分析和数值计算结果表明，相同的测距误差在不同的定位区域表现出了非均匀的目标定位误差分布；测距误差会对盲节点最终的定位估计产生明显影响，实验结果也验证了这一点。该部分工作有望能够对煤矿井下巷道定位系统的建设和将来可能的事故救援工作提供参考依据。

Statistics show that since 2003, average more than 4000 people per year died of the coal mining disaster. The mining production safety situation is very grim. Real-time monitoring in the process of production and management are imperative. If more accurate monitoring location information of the mine personnel can be obtained in real time, mine accidents rescue will be more timely and effectively, and loss can be reduced. After that, the security of mine work will be improved. Therefore, it is harsh necessary for practical significance and research value to design a coal mine location system, analyze location errors, and improve location accuracy.This topic is carried out about a small and medium-sized coal mine in northeast. This paper research content is also based on the specific coal mine underground environment. There are three parts work mainly studied. The first is the overall design of the underground location system, simulation platform building and received signal strength (RSSI) experimental measurement. The second is the error correction and analysis for RSSI measured values and range values. The last is the influence law of ranging error on the final location error and experimental verification. The main research results are listed as follows:1) All kinds of wireless sensor network (WSN) localization algorithm in coal mine were researched and compared. After that, the RSSI localization algorithm was chosen as the study object in this paper. In distance-based measurement technology, the measured range precision of the RSSI method is not better than other methods. However, the RSSI localization algorithm has a lot of advantages, including simple hardware system, compatible with existing hardware, low communication and computing cost, and can reduce the cost of the entire network.2) One dimensional linear anchor node deployment scheme is put forward based on the situation that most of the coal mine is roadway and most of personnel and locomotive are activity in the roadway. In this study, the overall design scheme of the coal mine location system introduces. The location system simulation platform was set up. The measurement experiment based on RSSI was carried out. The experimental results show that with increasing the space between anchor nodes and blind nodes, the received signal strength is reduced. This phenomenon is in line with the RSSI path loss model and verifies the reliability of the proposed location system. This part work laid the foundation for the follow-up RSSI ranging experiment.3) To reduce the accumulation error and improve the accuracy of RSSI localization algorithm, the error correction and analysis research of two levels for the measured RSSI values and range values were carried out, respectively. The experimental results show that the corrected RSSI measured values by the data filtering method is superior to that by statistical average method. Modified by least-square method, the maximum range error and the average range error decreased from 2.73 m and 0.97 m to 1.62 m and 0.54 m respectively. However, the ranging error is still inevitable.4) The influence of range error on the final location error was theoretically studied with the help of Matlab code, and the relevant experimental verification was also done. Theoretical analysis and numerical calculation results reveal that the same distance measurement error in different location area shows heterogeneous target location error distribution. Range error will significantly impact on the final location estimation for blind node, and the experimental results have verified this result. This part work is expected to provide some references for the construction of the coal mine roadway location system and the possible accident rescue in the future.