Efficient Localization Algorithm With UWB Ranging Error Correction Model Based on Genetic Algorithm-Ant Colony Optimization-Backpropagation Neural Network
Peipei Dai, Sen Wang, Tianhe Xu, Min Li, Fan Gao, Jianping Xing, Linghan Yao
Abstract
The development of wireless sensor network technology has extended the diverse range of tools available in location-based services (LBS). Indoor high-precision positioning is among the most popular topics in location tracking and positioning. Ultrawideband (UWB) double-sided two-way ranging (DS-TWR) is used widely because it provides a reliable ranging performance. In this study, the effect of the UWB DS-TWR ranging error was suppressed using a ranging error model to improve the reliability and accuracy of indoor positioning services. Based on the conditions described above, an optimized backpropagation neural network (BPNN) correction model that integrates both a genetic algorithm (GA) and the ant colony optimization (ACO) algorithm, forming the GA-ACO-BPNN model, is established and verified experimentally. In addition, under static and kinematic actual positioning conditions, improvements in the BPNN, GA-BPNN, ACO-BPNN, and GA-ACO-BPNN ranging error correction models in terms of their positioning performances are calculated and compared. The experimental results show that the proposed GA-ACO-BPNN model can reduce the impact of the ranging error on ranging and positioning effectively. The positioning accuracy and reliability of the UWB DS-TWR solutions are improved significantly after application of this model, which provides a reference point for solutions to subsequent fusion positioning problems, e.g., UWB and inertial measurement unit integration.