Abstract
Accurate and reliable vehicle localization is a key component of Intelligent Transportation System (ITS) applications. Personalized travel related services and recommendation systems like collision avoidance rely principally on the accurate and reliable knowledge of vehicles’ positioning. In this paper we propose a cooperative multi-sensor multi-vehicle localization method with high accuracy for terrestrial consumer vehicles. Two streams of real-time data are assumed available. One in the form of GPS coordinates of nearby vehicles received from a vehicle-to-vehicle (V2V) network and the other in the form of inter-vehicle distance measurements from a range sensor. In real-world situations, these heterogeneous sources of information are noisy and could be unavailable during short intervals. To overcome the effect of noise, measurements from two sources are fused together to estimate the number and motion model parameters of the vehicles. The problem is formulated in the context of Bayesian framework and vehicle locations as well as their velocities are estimated via a Sequential Monte-Carlo Probability Hypothesis Density (SMC-PHD) filter. To test the effectiveness of the proposed approach, a simulated scenario based on a grid portion of downtown Calgary is designed. Traffic intensity values match real-world reported data for the selected test location. Results of the simulation indicate that the proposed method provides reliable estimation of motion model parameters that predict the future location of vehicles.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call