Biased Kalman Filter for Eliminating NLOS Errors in Wireless Localization

In wireless localization systems, non-line-of-sight (NLOS) errors occur when signals are obstructed by obstacles. To mitigate these errors, the biased Kalman filter serves as an effective solution. This program implements an NLOS detection module that identifies NLOS errors in wireless positioning. When NLOS errors are detected, the system employs a biased Kalman filter with modified innovation statistics to handle the positive bias characteristic of NLOS measurements. For line-of-sight (LOS) conditions where no NLOS errors are present, the program utilizes a standard Kalman filter with optimal Gaussian assumptions. The implementation features two parallel filtering paths with a mode-switching mechanism based on statistical tests (like chi-square tests or residual analysis) on measurement innovations. This dual-filter approach ensures improved localization accuracy and reliability by dynamically adapting to different propagation conditions, thereby enhancing the practical value of wireless positioning systems.