Self-driving safety is one of the major concerns raised with regard to pushing the use of automated vehicles on roads. Fully automated vehicles are forced to make appropriate decisions in an uncertain environment where driverless and human-driven vehicles share the road together. This study proposes a new model that can be used to enhance the autonomous driving behavior at a non-signalized intersection considering the traffic safety guarantee, delay time, and smooth driving. The proposed model is called the responsibility-sensitive safety-based partially observable Markov decision process model for the decision-making mechanism of automated vehicles. The model not only increases traffic safety guarantee and smooth driving, but also reduces the delay time. First, we generate some specific driving scenarios using the automated driving toolbox in MATLAB. Second, the driving strategy of automated vehicles is optimized by the partially observable Markov decision process framework using the adaptive cruise control system. Finally, the responsibility-sensitive safety algorithm is implemented under adaptive model predictive control. The proposed model performs better than the classical adaptive model predictive control. In the best case, the proposed model took a 31.60 % reduction in braking time and a 51.20 % improvement in smoothing speed control.

无人驾驶安全是在道路上推动自动驾驶汽车使用的主要问题之一。全自动汽车被迫在不确定的环境中做出适当的决定，在这种环境中，无人驾驶和人力驱动的车辆将共同共享道路。这项研究提出了一种新模型，该模型可考虑交通安全保证，延迟时间和平稳驾驶，来增强非信号交叉路口的自动驾驶行为。该模型被称为自动驾驶汽车决策机制的基于责任敏感安全性的部分可观察马尔可夫决策过程模型。该模型不仅增加了行车安全保障和顺畅的驾驶，而且减少了延误时间。第一的，我们使用MATLAB中的自动驾驶工具箱生成了一些特定的驾驶场景。其次，通过使用自适应巡航控制系统的部分可观察到的马尔可夫决策过程框架来优化自动驾驶汽车的驾驶策略。最后，在自适应模型预测控制下实现了责任敏感的安全算法。与经典的自适应模型预测控制相比，该模型的性能更好。在最佳情况下，提出的模型在制动时间上减少了31.60％，在平滑速度控制方面提高了51.20％。在自适应模型预测控制下实现了责任敏感的安全算法。与经典的自适应模型预测控制相比，该模型的性能更好。在最佳情况下，提出的模型在制动时间上减少了31.60％，在平滑速度控制方面提高了51.20％。在自适应模型预测控制下实现了责任敏感的安全算法。与经典的自适应模型预测控制相比，该模型的性能更好。在最佳情况下，提出的模型在制动时间上减少了31.60％，在平滑速度控制方面提高了51.20％。