Mobile robot localization via outlier rejection in sonar range sensor data

Localization is an important ability for a mobile robot. The probabilistic localization method becomes more popular because of the ability of representing the uncertainties of the sensor measurements and inaccuracy environments, robust solutions for a wide perspective of localization problem. The particle filter is one of the Bayesian-based methods. In this study, data taken by sonar range sensor is used to localize mobile robot. Sonar range sensors suffer from wrong reflection effects which may cause outliers. Also, outliers may occur in the particle filter process. In this study, a new sensor model Repealing Range Sensor Model (R²SM) is proposed and integrated to particle filter to reduce the effects of outliers. In order to show the effectiveness of the proposed method, Grubbs´ T-Test, a well-known outlier rejection method, is implemented. Experiments show that results of the proposed approach are comparable to the results of the Grubbs´ T-Test in terms of Localization Success Ratio (LSR) and Number of Iterations (NOI) required for localization. The main advantage of the proposed R²SM is that it does not require any additional information such as critical value table. This provides more flexible outlier rejection approach.