No abstract

No abstract

Speed judgment is a vital component of autonomous driving perception systems. Automobile drivers were able to evaluate their speed as a result of their driving experience. However, driverless automobiles cannot autonomously evaluate their speed suitability through external environmental factors such as the surrounding conditions and traffic flows. This study introduced the parameter of overtaking frequency (OTF) based on the state of the traffic flow on both sides of the lane to reflect the difference between the speed of a driverless automobile and its surrounding traffic to solve the above problem. In addition, a speed evaluation algorithm was proposed based on the long short-term memory (LSTM) model. To train the LSTM model, we extracted OTF as the first observation variable, and the characteristic parameters of the vehicle’s longitudinal motion and the comparison parameters with the leading vehicle were used as the second observation variables. The algorithm judged the velocity using a hierarchical method. We conducted a road test by using real vehicles and the algorithms verified the data, which showed the accuracy rate of the model is 93%. As a result, OTF is introduced as one of the observed variables that can support the accuracy of the algorithm used to judge speed.

Based on the NaSch cellular automaton traffic model, a modified single lane traffic model is proposed by considering the dynamic headway of successive vehicles, in which the complex characteristic and driving behavior difference between drivers are taken into account. The relationship between the flow rate and the traffic density is obtained by the numerical simulation, and it shows a two-dimensional region in the flow density plane. The three traffic phases, i.e., free flow, synchronized flow, and wide moving jams, are exhibited. It indicates that the synchronized flow and traffic jams can appear even if there is no traffic bottleneck. Besides, the high speed car-following phenomenon is indicated when the traffic is in the synchronized flow. The rate of the high speed car-following is in good agreement with the measured result.

The driving state of a self-driving vehicle represents an important component in the self-driving decision system. To ensure the safe and efficient driving state of a self-driving vehicle, the driving state of the self-driving vehicle needs to be evaluated quantitatively. In this paper, a driving state assessment method for the decision system of self-driving vehicles is proposed. First, a self-driving vehicle and surrounding vehicles are compared in terms of the overtaking frequency (OTF), and an OTF-based driving state evaluation algorithm is proposed considering the future driving efficiency. Next, a decision model based on the deep deterministic policy gradient (DDPG) algorithm and the proposed method is designed, and the driving state assessment method is integrated with the existing time-to-collision (TTC) and minimum safe distance. In addition, the reward function and multiple driving scenarios are designed so that the most efficient driving strategy at the current moment can be determined by optimal search under the condition of ensuring safety. Finally, the proposed decision model is verified by simulations in four three-lane highway scenarios. The simulation results show that the proposed decision model that integrates the self-driving vehicle driving state assessment method can help self-driving vehicles to drive safely and to maintain good maneuverability.