利用智慧型手機以詢問回覆機制收集即時交通路況之設計與評估

Abstract

傳統利用GPS探偵車收集交通資訊的方法，一般採用探偵車週期性回報，或經過特定地點時回報，也就是，交通路況在固定時間點或固定地點被探偵車取樣。這種取樣的方式並不能保證能偵測到交通路況的動態變化，因為交通路況同時在時間軸和空間面上動態變化。在此一計畫，我們提出創新的動態交通資訊收集方法。我們利用配備GPS接收器與加速度偵測器之智慧型手機偵測交通路況，並採用詢問與回覆機制收集即時動態交通路況。首先，我們將分析高速公路與平面道路的車流模式。其次，我們決定智慧型手機可偵測的交通路況與事件，並設計其演算法與開發程式。第三，我們設計以詢問與回覆機制收集動態交通資訊之演算法與開發程式。基本的構想是：交通資訊中心(Traffic Information Center, TIC)根據目前收集的交通路況，以區域廣播的方式要求各路段的探偵車所需偵測與回報的事件。如此，交通路況在時間與空間上的動態變化都能即時的被探偵車偵測到，並回報給TIC。為了驗證與評估此一方法，我們也將利用交通模擬軟體VISSIM產生高速公路與平面道路在各種合理的車流模式組合下之車輛軌跡與速度。我們以這些軌跡作為路面實況，將用以驗證我們的以詢問與回覆機制之動態交通資訊收集方法，並評估其效能。此外，我們也將在高速公路與平面道路進行實際路測，紀錄車型軌跡與加速度偵測器記錄。這些記錄可用以比對VISSIM軌跡之真實性，也可以用以驗證智慧型手機端之事件偵測程式。

Traditional GPS-based probe vehicles send traffic reports to the traffic information center (TIC) either periodically in time or when passing pre-determined locations, i.e., traffic conditions are sampled by the probe vehicles either at fixed times or at fixed locations. These kinds of sampling methods cannot always detect the dynamic behaviors of traffic conditions because traffic conditions vary both temporally and spatially. In this project, we propose a novel method to collect real-time dynamic traffic information from vehicles carrying smart phones equipped with GPS receivers and accelerometers. We use a client-server architecture and a query-and-response mechanism. First, we analyze the traffic models for both highways and surface roads, and determine the events that need to be detected by the probe smart phones. Second, we design algorithms and develop programs for the smart phones to detect the events. Third, based on the current collected traffic information, the TIC instructs the smart phones what events to detect and report. In this way, both the temporal and spatial changes in traffic conditions can be detected and reported to the TIC in a real-time manner. We will develop computer programs for the PIC to collect traffic information from the smart phones. To evaluate the feasibility of our methods, VISSIM will be used to generate vehicle trajectories and speeds for both highways and surface roads under various traffic conditions. These simulated vehicle trajectories and speeds will be used as ground truth to verify our traffic information collection method. In addition, road tests with smart phones on the test vehicles can provide realistic vehicle trajectories, and accelerometer records under various traffic conditions. The real vehicle trajectories and accelerometer records can also be used to verify the event detection algorithms on smart phones.