| Summary: | 碩士 === 逢甲大學 === 資訊工程學系 === 103 === Taxi is becoming an intelligent traffic industry. Several studies proposed new carrying models, and improved the business efficiency of taxis. Zone-based queuing combines the concept of hunting and waiting. The taxis can hunt in the queuing zone when they are queuing in the queuing zone. However, taxi drivers easily gather in certain areas in the queuing zone according their experience. This may lead to the traffic congestion in these areas. To solve above problems, we propose an adaptive and hotspot aware taxi zone queuing system. Our system adaptively adjusts the service status of queuing line in the zone, and guides the taxis to different hotspots to wait for the queuing mission based on the mechanism and infrastructure of the zone-based queuing mode. This not only reduces the probability of the traffic congestion, but also reduces the waiting time of passengers after they call a taxi. First, we analyze the historical taxi GPS record data from Taiwan Taxi Inc., and discuss the temporal and spatial distributions of carrying records. We extract the areas that have the massive carrying records in the real environment, and deem the areas to be the hotspots in queuing zone. Then, we develop an adaptive zone queuing scheme based on M/M/1 queuing model, and extract the hotspots that have the shortest average waiting time depending on the historical data and real-time data. Moreover, we also implement a system prototype based on the above mechanism, and provide the navigating service using the taxi on-board unit.
|
|---|
