Indoor visible light fingerprint positioning based on dynamic weighted RSSI and LGSHHO-LSSVM

Abstract

To address the issues of poor stability and low positioning accuracy in traditional indoor visible light fingerprint positioning methods based on received signal strength indicator (RSSI), this paper proposes an indoor visible light positioning method based on dynamically weighted RSSI signals and an improved harris hawk optimization (HHO) algorithm for least squares support vector machine (LSSVM) (abbreviation DW + LGSHHO-LSSVM algorithm). First, a dynamic weighting algorithm is introduced to process the RSSI signals, considering both the geometric distance between the receiver and the light source as well as the signal fluctuation, dynamically adjusting the weights to reduce noise interference and enhance the positioning features of the signal. Second, the logistic chaotic map is employed to improve population diversity, and a golden sine strategy is used to enhance the search efficiency of the HHO algorithm, thereby improving the global search ability and convergence accuracy of the algorithm. Finally, the improved HHO algorithm is used to optimize the key parameters of the LSSVM model, establishing a high-accuracy indoor positioning prediction model. Experimental results show that, in a 3 m × 4 m × 3 m test environment, the average positioning error of the DW + LGSHHO-LSSVM algorithm is 4.7 cm, with errors mainly concentrated in the range of 0.05–0.1 m. Compared to the LSSVM, HHO-LSSVM, DW + HHO-LSSVM, LGSHHO-LSSVM and MPA-LSSVM algorithms, the positioning accuracy is improved by 53.92 %, 48.35 %, 27.69 %, 38.96 % and 51.55 %, respectively. These results demonstrate the significant advantages of the proposed method in terms of positioning accuracy and stability.