在車載網路(VANETs)中，如何確保事件訊息的可信度在近幾年來一直獲得相當多的關注，希望可增加車輛的安全性與確保隱私防護機制不會影響到信任關係的建立。傳統相關研究的主要不足在於隱私保護、快速轉傳惡意訊息以及在共謀攻擊的情況下，嚴重的影響信任管理系統在事件訊息可信度決策的準確性，此外傳統決策判斷造的延遲時間更是影響車輛安全的重要因素。為了解決傳統研究上的不足，本研究提出一個利用路側單元(RSU)與信標(Beacon)為基礎的信任管理系統，可有效對抗快速轉傳的惡意訊息及共謀攻擊。而在隱私防護及確保信任關係的有效性上，本研究提出安全廣播認證協定，除了可以有效確保廣播訊息的隱私性外，亦可作為信任關係建立的基礎。而在縮短決策的延遲時間部份，本研究採用了調適性的決策模式，藉由車旁單位快速提供輔助資訊，可有效縮短決策的延遲時間。為了評估系統的可靠性與效率，本研究在不同假造訊息或變造訊息攻擊比例組合下及共謀攻擊發生時進行大量的模擬測試與比較，得到的結果，顯示本研究所提出的系統足以抵抗訊息丟棄攻擊、訊息修改攻擊及假訊息攻擊，並且表現較比較對象於準確度至少可超過23%以上，在決策延遲時間至少可縮短12%以上，在共謀攻擊的情況下在準確度與延遲時間仍可以有超過18%及14%領先比較對象。 In VANETs, how to determine the trustworthiness of event messages has received many attentions in recent years for improving the traffic safety and ensuring the location privacy of vehicles. In existing researches, the decision making accuracy on event message under privacy enhanced scheme, fast forwarded malicious message and collusion attacks are all the outstanding problems. In addition, the decision delay while making decision for distinguishing the message trustworthiness is also a critical issue and need to be resolved.In this paper, we propose a RSU and beacon-aided trust management system, called RaBTM, which aims to provide message opinion quickly and thwart internal attacker from sending or forwarding forged message in privacy-enhanced VANETs. In addition, a secure broadcast authentication protocol has also been proposed which can not only ensure the privacy of broadcast messages, but also can provide the linking ability of the trust relations of vehicles. On shorten the decision delay, in this paper, we propose an adaptive decision making model which makes decision quickly by the effective and quick assistance of the RSUs. This propose model can effective shorten the decision delay.To evaluate the detection accuracy and efficiency of our proposed system, we conducted a set of simulations under various trust attacks. Under non-collusion attacks, the simulation results show that the proposed RaBTM can highly resilient to adversarial attacks and performs at least 23% better on detection accuracy, and 12% shorter on decision delay than VEBAS with weighted vote (WV) scheme. Under collusion attacks, RaBTM can still perform better than comparison baseline more than 18% and 14% respectively.
