An Asymmetrical Probabilistic Estimation using Structural Matching for Outdoor Localization ⋆

Location-Based Services (LBSs) for mobile communication are widely used in which location estimation is the prerequisite. We propose an Asymmetrical Probabilistic Estimation (APE) method based on the location fingerprinting. In proposed method, the signal intensity variations for each cell in each fingerprint are modeled by Gaussian distributions with asymmetrical probability density function due to the nonlinear relationship of the signal space and physical space. Moreover, to enhance the poisoning accuracy and decrease time-consuming, the candidate fingerprints for location estimation are filtered by proposed Structural Matching (SM) algorithm. We have implemented the localization method based on a real cellular network in an urban area. The results indicate that our proposed estimation method achieve higher positioning accuracy.