影像輔助之協同式車輛定位系統

Abstract

智慧型交通運輸系統(Intelligent Transportation System, ITS)在車載環境中提供了各種不同的服務。許多的ITS應用都需要利用到車輛準確的位置資訊。然而目前的一些定位方法無法擁有充足的準確性或是受限於佈建成本，並不能完全適用於目前的車載環境中。故在此論文中，我們提出了影像輔助之協同式車輛定位系統 (Video-Assisted Inter-Vehicle Positioning, VIP)。此系統整合了GPS接收器與行車紀錄器透過影像處理所摘取的感測資料，並經由車間通訊來交換這些資訊以提升定位的準確性。在此系統中，我們假設每台車輛都裝設GPS接收器以及WAVE/DSRC通訊模組；並且有部分的車輛裝設了行車紀錄器。裝設行車紀錄器的車輛可以透過影像辨識出的車道別，以及透過電子地圖找出GPS位置對應的車道別來驗證GPS位置的誤差程度。在協同式定位的過程中，這些資訊與附近車輛進行交換，區別出車輛GPS位置的權重並對自己的GPS位置進行修正。模擬的結果顯示我們的方法在一半的車輛裝設行車紀錄器的情形下，車輛位置的準確性即可達到百分之十至三十的提升。而在所有車輛都裝設了行車紀錄器的情形下，準確性提升了百分之三十至四十(修正到三公尺內)。

Intelligent Transportation System (ITS) provides various services for vehicular environments. Most of the ITS applications require accuracy position of vehicle to be developed. However, the current positioning techniques and devices may not satisfy the accuracy. In this thesis, we present a Video-Assisted Inter-Vehicle Positioning (VIP) system. The system integrates sensed data extracted from GPS receiver and driving video logger with image processing techniques, and then the sensed data is exchanged to nearby vehicles for improving the accuracy of cooperative positioning. In our system, each vehicle is equipped with a GPS receiver and WAVE/DSRC equipment, and some vehicles are equipped with driving video logger, called fully equipped vehicle. Each fully equipped vehicle will assess the difference between the lane recognized by the image processing module and the lane is referred to digital map module to validate the accuracy of its GPS position. Then, these information will be exchanged among vehicles to differentiate the weights of received positions during the cooperative positioning process. Simulation results show that our approach can achieve 10~30 percent of improvement in the position accuracy when only half of vehicles were fully equipped, and improve the accuracy by factor of 30~40 percent (within 3m) if all vehicles were fully equipped.