煤矿井下通信系统是重要的信息传递通道，尤其在煤矿救援工作中具有极其重要的作用。然而，在矿难井下复杂环境中，实现通信系统的实时搭建以及在有限的通信系统中传输大量数据是十分困难的。本文主要研究煤矿救援机器人无线通信系统以及数据传输的关键技术问题，旨在构建井下救援人员与机器人之间的无线通信网络，实现大量现场探测数据的实时有效传输，为实施救援提供准确、可靠的事故现场信息。 研究了矿难非结构环境中搭建无线通信系统的需求和无线通信系统信道传播特性，提出了基于机器人、“不倒翁”中继和布放平台的煤矿救援机器人无线通信系统控制体系结构，为实现复杂环境中煤矿救援机器人无线通信系统的快速搭建，以及事故现场信息数据的实时传输提供了可靠的解决方案。 针对煤矿井下救援机器人无线通信系统应用环境的复杂性，归纳了常见事故巷道结构类型。借助Wi-Fi通信技术组建通信平台，基于麦克斯韦方程组和无线电磁波传播机理，分析了正常和灾后巷道结构下无线通信信道传播特性，建立了煤矿井下两类非结构巷道的无线通信信道传播特性模型，提出了适合煤矿救援机器人无线通信系统中继的布放策略，为煤矿救援无线通信系统的快速搭建提供可靠的理论依据。 研究了煤矿救援机器人无线通信系统特定信道条件下的环境信息数据实时传输问题，借助多尺度小波理论在数据压缩与重构中的应用，提出自衍生多尺度小波的改进霍夫曼数据压缩编码算法，该算法提高了环境信息数据的压缩效率及自适应能力。通过自匹配小波重构算法，实现了自衍生小波分解后数据的实时重构。该算法在有限通信环境中，提高了环境信息数据的压缩效率及压缩比。通过对某矿采集的瓦斯浓度数据压缩重构试验，由还原结果表明算法误差较小，保真度较高。 分析了煤矿救援过程中视频数据传输对无线信道性能的要求，研究了视频数据的实时、准确传输问题。基于小波边缘特征提取的图像选择和二维小波图像压缩理论，提出自适应多重小波压缩的改进H.264视频编解码算法，具有较高的压缩效率和自适应能力。结合自适应多重小波重构算法，实现了矿难井下非结构环境中视频数据的实时传输。利用实验巷道视频数据对算法进行验证，数据还原结果表明压缩算法误差可控，不仅提高了视频数据传输能力，而且降低了视频数据的信道占用率。 通过一系列的仿真分析和实测数据测试，结果表明本文提出的无线通信系统搭建方法及数据压缩算法能够保证在有限通信环境下的数据传输，为在矿难井下实时搭建煤矿救援机器人通信系统并实时传输机器人采集的环境和视频数据奠定了坚实的基础。

The communication system used in coal mine is an important information transmission channel, it plays an important role in coal mine rescue. As the coal mine environment is complicated, it is extremely difficult to construct the real-time communication system and to transmit a large amount of data using the system with limited functions. Aiming to the setting-up wireless communication network between rescue personnel and robot under the mine, the key technical issues of coal mine rescue robot wireless communication system and data transmission are studied and all these results are significant for implementation of the rescue and for providing accurate on-site information. By studying channel propagation characteristics of wireless communication systems in the demands of wireless communication system in non-structural environment, the coal mine rescue robot wireless communication system control structure are proposed based on robot, “tumbler” repeater, and laying platform. This lays a solid foundation for speedily building coal mine rescue robot wireless communication system and provides reliable solutions to the real-time transmission of on-site data. The tunnel structure of common types of accidents is summarized according to the extremely complicated communication environment in coal mine. It then uses Wi-Fi communication technology to set up a communication platform and analyzes the wireless communication channel characteristics of normal and post-disaster tunnel configuration based on Maxwell’s equations and the radio wave propagation mechanism. Moreover, the wireless communication channel propagation characteristics model of two non-structural tunnels in coal mine is built. The laying strategy of wireless emergency communication networks is proposed. The real-time transmission problems of environmental information data in specific channel conditions of coal mine rescue robot wireless communication system are studied. Using the application of multi-scale wavelet theory in data compression and reconstruction, the improved Hoffman lossless data compression coding algorithm of self-derived multi-scale wavelets is proposed, which is improving the compression efficiency and adaptability to environmental information data. Through self-matching wavelet reconstruction algorithm, the real-time reconstruction of data is achieved after self-derived wavelet. The algorithm improves the compression efficiency of environmental information and data compression ratio in the limited communication environment. Conducting the compression and reconstruction of gas concentration data collected from a certain coal mine and analyzing the restored results, the algorithm proposed in this thesis has few deviations but high data fidelity. Analyzing the demands and transmission requirements of video data in coal mine rescue process, the real-time and accurate transmission problems of video data are studies. Based on the image selection taken from edge feature of wavelets and two-dimension wavelet image compression algorithm, the improved H.264 video encoding and decoding algorithm based on self-adaptive multi-wavelet compression is provided in order to ensure higher capability of video data compression and self-adaptive capability. Combining the reconstruction algorithm of self-adaptive multi-wavelets, it can realize the real-time transmission of video data in non-structural environment. Conducting the algorithm verification of tested roadway video data, the deviation of the restored results is controllable; the overall data transmission has high quality; and the amount taken up by the data transmission can also be reduced. Through a series of simulation and real data tests, the proposed methods of wireless communication system setting-up and data compression algorithms ensure data transmission in a limited communication environment. This lays a solid foundation for the real-time setting-up of coal mine rescue robot wireless communication system and transmission of environment and video data collected by robots in the mine.