Research On The Location Algorithms Of Miners Based On Frequency-Domain Model Of Coal Mine And TDE Technology

Nowadays, the requirement of domestic coal mine safe production becomes higher and higher, so it is very important to set up reliable and practical miner tracking and locating systems for coal mine safe production management. Compared to the developed countries, the information level of our country is lagging. In fact, the miner tracking systems used widely in our country are only a kind of recording systems. They are not real miner tracking and locating systems and can' t report the accurate positions of miners. The precise miner locating technology remains in the theoretical stage home and abroad, and the scholars are focusing on how to improve the location accuracy.In this thesis, the precise miner locating algorithms based on the frequency-domain model of coal channel and TDE technology were investigated. First, the position reckoning method, approach method and wireless location method were analyzed according to the principle of miner location, then, the TDE method that was chosen for precise miner location, which is one of the distance-based location methods.Next, algorithms based on the frequency-domain model of coal channel were investigated. The important factors that affect the accuracy of TDE were analyzed. The IFT,DSSS/XCORR and the subspace algorithm based on singular value decomposition of signal-subspace correlation matrix were investigated.In this thesis, the Root-Min-Norm algorithm was applied in miner location. The statistical characteristics of the algorithm in different multi-path environments, and in different SNR and different bandwidths conditions by simulation were analyzed. The statistical results demonstrate that the Root-Min-Norm algorithm has high resolution, estimation accuracy, stability and anti-noise characteristics, but it is not good at resolving the correlated signals which affects its abilities of resolving the signals with small intervals. So the modified Root-Min-Norm algorithm was proposed for the correlated signals. The modified algorithm improves the accuracy of correlation matrix estimation in small sample conditions. The statistical results of simulations demonstrate that the modified Root-Min-Norm algorithm is an algorithm with high resolution, accuracy and stability, and it has great values in the actual applications.