Journal of System Simulation ›› 2018, Vol. 30 ›› Issue (9): 3586-3595.doi: 10.16182/j.issn1004731x.joss.201809046

Previous Articles    

An Improved Algorithm of VIRE Based on BP Neural Network

Kong Hongshan, Yu Bin   

  1. Information Engineering University, Zhengzhou 450001, China
  • Received:2017-01-09 Online:2018-09-10 Published:2019-01-08

Abstract: Aimed at the problem that the proper threshold affects the accuracy of positioning in VIRE algorithm, a RFID positioning method combining BP neural networks with VIRE algorithm is presented and is named BP-VIRE. BP-VIRE utilizes the same reader’s configuration and deployment as that of VIRE, and regards referenced tag RSSI and coordinate values as training samples, and constructs the positioning model based on neural networks with the help of BP algorithm which uses the method of error back propagation. According to the simulation results, compared with VIRE, BP-VIRE performance is better in the capability of positioning, esp. in the aspects of average deviation, maximum deviation, and boundary area deviation, so the improved method is more suitable for indoor positioning application in comparison with VIRE.

Key words: indoor positioning, BP neural networks, virtual reference, radio frequency identification, received signal strength indicator

CLC Number: