物聯網旅行計劃系統的比較和分析

Abstract

不確定性是乘客減少使用公共交通系統的一個關鍵因素。這是因為並無法保證公車能依時刻表運行。因此，它減少了使用者對系統的信心而導致使用率下降。即時旅行計劃系統可使客運公司盡可能的改善使用者的使用經驗也能有效減少用戶在旅行途中的不確定性。在出發前，它提供用戶從出發點到目的地的行程規劃。在確定行程後，即時旅行計劃系統會對行程可能發生延遲的用戶送出提醒。 雖然即時旅行計劃系統已有廣泛的相關研究，但大都著眼於整合多種交通工具，並提供多種標準讓用戶規劃自己的行程。然而，這樣的系統，其即時資訊的來源，現今仍依賴持續地監控他們的運輸車輛。到目前為止，就作者所知，只有很少的研究著眼於將監控系統分散至各運輸車輛，使其監控自己的行車狀態。 在這項研究中，我們提出一個物聯網即時旅行計劃系統，它提供車輛〈在這個例子中為公車〉監控它們自己狀態的能力。在我們的系統中，公車只有在運行時刻與表定時刻表有差異時，才會對系統發出通知。從而優化了網路流量，並減少監控系統因連續監測所需要的計算量。 為了比較我們提出的方法與傳統的監控系統，我們以交大巴士時刻表實作了兩個實驗原型。在此我們引進以M2M通用服務層作為公車與系統之間的中介平台的概念。這兩個實驗原型都建立在OpenMTC〈一個符合ETSI M2M標準的通用服務平台〉之上。 我們在實驗中以兩種方式測試其效能─穩定的使用者查詢以及突發性大量的使用者查詢，而在使用者查詢的同時，公車會將其狀態回傳給系統。結果顯示，在公車處於相同狀況下，我們提出的方法可以有效增進系統的流量率。

Uncertainty is a key factor that prevents commuters from using public transportation system. This is because there is no guarantee that public transportation vehicles will always follow their published schedules. It thus reduces a commuter’s confidence in utilizing the system. Real-time Trip Planning systems are a genuine effort from transportation agencies to increase commuter’s confidence. Real-Time Trip Planning systems help to reduce users’ uncertainty with timely information on every phase of their trips. On a pre-trip phase, they provide users with steps to arrive at their destination. And afterwards, on pre- and on- board phases, Real-time Trip Planning systems alert users of possible delays on their plan. Although research about real-time Trip Planning systems is extensive, it focus on incorporating multiple modes of transportation and offer multiple criteria for users to plan their trips. However for their source of real-time information, such systems, nowadays still rely on continuously status monitoring of their transportation vehicles. Until now little research, to the authors’ best knowledge, has focused on distributing status monitoring task to the vehicles themselves. In this study we propose a real-time Trip Planning system that provides vehicles, in this case buses, with capabilities to monitor their own status. In our approach, buses only notify their status when it differs from their published schedule. Thus optimizing network traffic, and reducing system overhead caused by continuous monitoring in the traditional approach. In order to compare our proposed approach against the traditional one, two prototypes were implemented for NCTU buses. We incorporate the concept an M2M common service layer as middleware between buses and the prototypes, using OpenMTC, an ETSI M2M standard compliant common service platform. Prototypes’ performance was measured under constant and burst load of user inquiries, while simulated buses simultaneously updated their status. It turned out our proposed system can achieve an improvement in system’s capacity in terms of throughput.