Research Of Wireless Positioning Technology For Moving Target In Mine

The coal resource is important foundation energy in our country and plays animportant role in the development of national economy. But in the process of coalmining, the frequently happened accidents not only endanger the personnel’s livesand cause great economic loss, but also leave a bad impact on society. Realizing thepositioning of personnel and locomotive in mine can not only achieve scientificdispatching and real-time control, improve the mine productivity in dailyproduction, but also obtain their position information timely when accidenthappens, which helps to improve the rescue efficiency and ensure the safety of thepersonnel’s lives. Thus, realizing the positioning of personnel and locomotive inmine is of great significance to enhance the level of mine production safety.With the rapid development of wireless network and mobile communicationtechnology, indoor positioning using wireless local area network (WLAN) isbecoming a research hotspot. In this dissertation, WLAN-based wirelesspositioning technology is applied to the mine and its research hotspot—fingerprintpositioning algorithm is modified combining with one-dimensional linear feature ofmine laneway. To solve time-consuming and low-efficiency problem in the trainingphase of fingerprint positioning algorithm, a neural network interpolationalgorithm based on region division is proposed, which uses the signal strengthreceived by reference points in the same region to realize interpolation. Theinterpolation algorithm proposed in this dissertation ensures the accuracy ofinterpolation and improves the efficiency of training phase of fingerprintpositioning algorithm. And in the positioning phase, a wireless positioningalgorithm based on signal strength correctness judgment is proposed to deal withinaccurate signal strength that moving personnel and locomotive may receive. Inorder to ensure the positioning accuracy, the algorithm judges the signal strengthreceived by the target which is being positioned whether correct or not to excludeinaccurate data.Finally, the proposed interpolation algorithm and positioning algorithm in thisdissertation are simulated by Matlab. The simulation result shows that, thealgorithms proposed in this dissertation improve the efficiency of training phaseand the positioning accuracy compared with traditional fingerprint positioningalgorithm.