Particle filters for positioning, navigation, and tracking
IEEE Transactions on Signal Processing
Neural Networks for Improved Tracking
IEEE Transactions on Neural Networks
| Hi-index | 0.00 |
Accurate Mobile Geographic Information System (GIS) is a major building block of many applications, particularly in Intelligent Transportation Systems (ITS). In this context, GPS provides position information of each vehicle, while immediate surrounding information is gathered through the exchange of beacons. Yet, the ITS environment is characterized by frequent losses of GPS signal and beacons. Estimation/tracking based on Kalman or Particle filters could be an alternative to support the precision of the Mobile GIS, but both approaches are equally sensitive to missing and unreliable data. In this paper, we propose GSF, a Glowworm Swarm Optimization to particle filters, adding the bio-inspired capabilities of Glowworms to converge to multiple potential estimates, when unreliable mobile GIS lack precise updates. We first analyze the performance of GSF by considering perfect conditions. Second by considering GPS signal loss, packet loss and positioning errors. Simulation results show that our approach achieves its design goal of improving the precision of the mobile GIS. GSF performs better than standard particle filter scheme in terms of position accuracy, and this at a reduced complexity and fair convergence time.