車載網路中透過RSU的規劃傳送多路徑影像串流

Abstract

車載網路是在近幾年來新興的議題，隨著科技的進步，車載網路上的應用也越來越豐富，在車載網路下傳遞文字訊息已經變得不敷使用。影像相較於文字較容易被人們所理解和接受，透過影像豐富的內容可以提供人們行車的安全和更便利的生活。然而在車載網路下傳遞影像是一個嚴峻的挑戰。在本篇論文，我們提出了一個在車載網路中提供影像串流的系統，並且導入使用者的路徑。根據之前多路徑的相關研究，可以知道在車載網路下，透過多路徑的備份機制可以提高封包成功傳遞率，然而透過實驗發現到影像串流透過多路徑傳遞，其傳輸頻道競爭的情形會很嚴重，會降低影像的品質，故我們的系統透過路側單元的協助來降低傳輸頻道競爭的情形。因此如何選取一組路側單元以提供多路徑的影像串流是本篇論文主要探討的部分，我們將路側單元選取與排程問題轉化成覆蓋問題以求得最佳解。並且我們設計一個系統和路由協定在車載網路下提供影像串流服務，影像傳輸路由協定則是每當影像傳遞到路口時，再透過優先權來做下一個傳輸路徑的選擇。根據實驗的結果，相較於單一路徑，透過路側單元協助的多路徑的影像串流傳遞會有較高的影像觀賞品質，並且系統透過我們提出的覆蓋演算法和單純用貪婪演算法來尋找規劃路側單元相比，我們提出的演算法會較容易找出一組完整的路側單元集合，使得系統在提供影像串流服務時，在相同的路側單元數目下可提供更多使用者的影像服務，讓本篇提出的影像系統在真實的車載網路下可以更適用。

With the rapid development of wireless communication technology and the rapid increase in demand for traffic information in the car, a Vehicular Ad-Hoc Network (VANET) is an emerging network technique that will be deployed in the future. A VANET is a technology that equips moving cars with wireless communication capability to create a mobile ad-hoc network. To cope with the highly partitioned nature of a VANET, some static Roadside Units (RSUs) can be deployed to improve the data dissemination. With the spreads of content-rich web-based intelligent transport system applications, the simple text data dissemination in VANET has become insufficient. In contrast, video is easier to be comprehended by people than text data. Such as an accident video on the road provides instant traffic safety information to the drivers. However, how to disseminate a video over VANET is a challenge. In this paper, we propose a system to provide video streaming services along with car’s route on the roads. According to the previous studies backup mechanism through multiple paths routing can improve the packet delivery ratio. Nevertheless, our experiments showed that the video streaming via multi-path transmission leads to video quality degradation due to severe channel contention. Therefore, in this paper, we focus on the problem of selecting RSUs to provide multi-path video streaming services. We find that the RSU scheduling problem can be transformed to the well-known Set-Covering-Problem. Our RSU-Scheduling algorithm finds a set of RSUs which support video servers to provide high quality video streaming service. Based on our experimental results, RSU support video streaming via multi-path routing has higher video quality than single path routing. Furthermore, our RSU-Scheduling algorithm can support more video users with the same number of RSUs comparing with greedy method.