Trajectory Based Collaborative Navigation Service Framework: Car Navigation as a Telematic Service

Vehicle navigation applications are receiving more and more interests today. However, due to the frequently changed urban map, inaccuracy of GPS data, and the small screen size of the navigator, navigation result is not as good as we expect. In this paper, we propose a new idea called “collaborative navigation”, where experienced drivers are selected to transfer their “experiences” to naïve drivers during driving. The major obstacle to make the idea into reality is its business model, the accuracy to select the experienced driver, and the overall performance of the system. To address the above issue, we designed a collaborative navigation service framework (CNSF), and implemented it in J2EE and J2ME. In this framework, business model based on Telematics is suggested; the trajectories collected from GPS devices of experienced drivers are regarded as the basic “experiences”; algorithms are designed to mine the best route and best candidate for collaboration; software modules are implemented to facilitate the collaboration between experienced and naïve drivers. Experiments designed in the end of this paper show such kinds of services are feasible in current computing environments with our own trajectory storage structure and filtering algorithms.