汽車智能化的道路：智能汽車、自動駕駛汽車安全監管研究

The Intelligentization of Automobiles: Smart-Cars, Robo-Cars and their Safety Governance

近年來，“功能汽車”正呈現出向“智能汽車”逐漸進化的革命性趨勢，其原因在於智能汽車對非結構化環境的高適應性。從安全監管角度來看，功能汽車的車用電子控制系統具有高度可預見性，因此功能汽車之ECU嵌入式控制單元對於環境的理解程度十分接近地面實況。然而，德國TUV安全認證機構的研究報告指出這一切將可能不再適用在智能汽車上面，特別是當駕駛輔助系統開始展現一些（半）自動行為以後。除“智能汽車”之外，無人介入的“自動駕駛汽車”亦將引發更多關於安全監管和道德倫理等問題，因此有必要以自動化伴隨的“開放組織風險”為中心，來貫穿這股科技變革的起點（“智能汽車”）和終點（“自動駕駛汽車”），並分析其潛在安全隱患以及治理之道。

Currently, we are witnessing the transition from feature- to smart-cars, and it will be evolutionary trend since smart-cars’ highly perception with unstructured environments. The control systems in feature cars are well understood and within our theoretical frameworks. Certain safety properties of stability control systemsfor example-can be formally proved. The underlying physical models are well understood and can be fitted with accurate sensor readings. The ECUs understanding of the environment is therefore close to the ground truth. However, as pointed out in this is no longer the case for smart-cars where driver assistance systems implement (semi)-autonomous behaviors. Except to smart cars, the other unmanned autonomous robo-cars could derive more safety and ethical problems. Thus, the authors would like to analyze potential safety gaps and also to provide our suggestions for safety governance regarding the implementations of smart/robo-cars.