A Non-Line-of-Sight mitigation localization algorithm for sensor networks using clustering analysis

When the number of corrupted measured distances is minor (e.g.: 3 corrupted from 10), all the pre-located estimators before MRDF processing distribute densely and coherently. Using MRDF processing accurate estimators can be got. When all measured distances are attached with the same NLOS errors ranging from 5% to 50% of each measured distance, MRDF algorithm has similar performances with LSE and performs better than RWGH and RMIN. [Display omitted] •Pre-located estimators are obtained using adopted subsets of the distances.•Utilize MRDF to recognize the pre-located estimators with high accuracy.•Localization accuracy can also be improved under random NLoS measurement error. Non-Line-of-Sight propagation of wireless signal has an impact on measured distances in range-based localization and will bias the final localization results. A new localization algorithm is proposed in this paper to mitigate Non-Line-of-Sight errors when there are more than enough anchor nodes deployed around the node to be located. This algorithm utilizes multi-round clustering analysis to filter the pre-located estimators which derive from all possible subsets of measured distances. In each round, the method density-based spatial clustering of applications with noise is adopted. Simulations show that the proposed algorithm can effectively improve localization accuracy not only when the measured distances with Non-Line-of-Sight error are minor but also under the condition that all of them suffer random Non-Line-of-Sight error.