Study of fuzzy logic and particle swarm methods in map matching algorithm

Abstract

Navigation system is used in deploying the real-time position of vehicles on the map of route with the help of Global Positioning System (GPS). However, the imprecision is included in the map of route due to erroneous GPS signal (i.e., inclusion of errors related to receiver and/or propagation). Also, the current navigation system involves dense road network areas having higher probability of imprecise inputs. The map matching (MM) method requires capability of inherent tolerant to these imprecise inputs and is capable of suggesting certain vehicles more likely to run on one specific link than other links in navigation system. The fuzzy inference system (FIS)-based technique converting the qualitative terms into quantitative values is very effective to deal with human intervention and linguistic vagueness in MM algorithms. In case of highly unreliable positioning errors included in MM, the particle swarm optimization (PSO) is used as an option for continuously updating sensor information to find vehicle motion and respective route link on map. This has directed us to do the comparative study of both the above two methods with the other methods. Simulation results show that FIS- and PSO-based MM algorithm significantly outperforms the previous algorithms in terms of running time, accuracy and recall.