| 摘要: | 平安是回家唯一的路。即使全世界的政府都在大聲疾呼交通安全的重要性，每年道路交通事故依然頻傳，似乎難以避免。而這些事故中又以連環車禍最易發生慘劇，受傷的人數也最多。這樣的問題層出不窮，也亟待解決。 而隨著無線通訊技術日趨成熟，車載通訊的應用亦可藉此來交換或傳遞交通訊息，進而提升交通安全或是減少塞車。其中以利用廣播技術來傳遞緊急訊息給駕駛人，為最常被廣泛應用的一種方式，然而這樣的傳遞模式必須克服存在的無線傳輸問題，如廣播風暴、涵蓋盲點、訊號干擾或是封包碰撞等，而使得緊急訊息無法被送達或是延遲，因此現有的方式來傳送緊急訊息，仍無法有效地降低交通事故的發生。 本篇論文提出了一個新穎的廣播演算法，以事件驅動的方式來執行，並且採用以區域位置為基礎的廣播設計，除了能有效地解決廣播風暴的問題，同時能適用於高速公路雙向車道的防撞機制。主要的轉送封包條件是依據目前車輛及所收到前車的位置、方向及速率來做決定，並藉由一個調適等待時間的方法，使得對向車道能有效地協助封包轉送，並且能減少涵蓋盲點的問題。 模擬的結果顯示本演算法能夠有效率地傳遞緊急訊息。位於四?五公里內的百來輛車，僅需15.30毫秒及21個轉送點就能傳送完畢，其延遲時間至少比其他的演算法少了10%，轉送率則最多可少79%。而除了比其他的廣播演算法更能快速地傳達訊息外，同時能避免封包的遺失及提升傳送成功率，並且以最少的轉送點來完成訊息的傳遞，因此更能夠避免連環車禍事故的發生。此外對於演算法的模型，除了加以分析求得最佳的參數值外，並以簡明的轉送步驟來傳送封包，同時提供明確的封包格式，以利未來應用程式的擴充，及發展更多的車載服務。 Safety is the only way home. Even though governments around the world announce the importance of traffic safety, it seems extremely hard to avoid traffic accidents happening repeatedly every year. Rear-end crashes are the most frequently occurring for tragedy and resulting in a substantial number of injuries. People are eager to find clever solutions to such critical problems. With the techniques of wireless communications growing mature, Inter-Vehicle Communications provide an efficient way to improve traffic safety and prevent congestion through exchanging or disseminating the traffic conditions. Broadcasting techniques are the widely used applications of delivering the emergency messages to drivers; however, they often suffer the broadcast storm, hidden node, interference, and contention problems, which makes the emergency message undelivered or delayed. For this reason, using current techniques to forward emergency messages, it remains difficult to lower the incidence of traffic accidents efficiently. A novel broadcasting algorithm by event-driven was proposed in the thesis and is categorized as area-based method broadcast. Not only does it overcome the broadcast storm problem effectively, but it is also suitable for cooperative collision avoidance on two-way highway. Its message-relaying decisions are made by receivers based on the position, direction, and speed properties of re-transmitters and itself. Moreover, the time delay is adaptively adjusted in opposite directions to reduce the hidden node problem and result the delivery more efficiently and reliably. The simulation results show that the proposed algorithm disseminates the emergency messages most efficaciously and steadily, for instance, it can reduce the delay of compared protocols at least 10% and the retransmission rate at most 79% since the delivery on 100 vehicles within 4.5 km is only taking 15.30 ms and 21 retransmissions. Not merely does it have the lower latency and packet loss rate, but also it takes the fewer retransmissions than other flooding based protocols. Consequently, it can help prevent the rear-end crashes. In addition, a sophisticate data format is defined as well for future extensions and developments of vehicular services. |
