Research On Underground Personnel Location Algorithm For Wireless Sensor Network Based On RSSI

Accurate underground personnel positioning technology is an important scientific and technological means to improve the safety of the coal mines production management.At present,underground personnel positioning technology is mainly based on Received Signal Strength.This technology can only recognize the approximate area of the target,and it is lack of positioning accuracy.Wireless sensor network(WSN)is relatively mature and suitable for building underground personnel positioning systems with higher requirements.Based on the disadvantages of large location error and poor environmental adaptability of traditional WSN location algorithms,an improved algorithm based on the location fingerprint location model of limit learning machine(ELM)is proposed.This makes the model more accurate and suitable for the actual downhole environment application.The research contents are as follows:1.The improved whale optimization algorithm(IWOA)is employed to optimize the weights and thresholds of ELM,which can improve its convergence speed and generalization ability to a certain extent.The IWOA-ELM positioning model is established to make the algorithm robust and accurate.In order to set the relevant parameters of the positioning model well,the ray tracing method is used to simulate the downhole environment and obtain the relevant data required for the simulation experiment.The simulation results show that the positioning error of the proposed IWOA-ELM model is reduced by about 17% compared with ELM,and the position of the target to be located can be more accurately estimated.2.According to the characteristics of downhole environment,the further improvement measures have been promoted.The characteristics of underground electromagnetic propagation environment are analyzed.In view of the characteristics that the received signal strength indication(RSSI)is easy to jump and time-varying,a Gaussian filtering method is proposed to preprocess the RSSI.In view of the electromagnetic propagation environment changes in a long-time scale,an online sequential limit learning machine(OS-ELM)is introduced to modify the positioning model.It can overcome the errors caused by the change of the electromagnetic propagation environment.In order to overcome the problem that the unreasonable weight setting of OS-ELM causes the model correction not in place,a dynamic weight factor is added to OS-ELM to establish the IWOA-DOS-ELM positioning model.It is more adaptable to the changes of the environment and has stronger robustness.3.The underground personnel positioning system based on WSN is designed to simulate the underground scene in the narrow and long channel,and the feasibility and superiority of the algorithm in this paper are verified in practice.Compared with the unmodified IWOA-ELM,the experiment show that the improved algorithm is more reliable after the environment changes.The positioning accuracy can be better maintained.It has certain theoretical and practical significance.