Improvement for vehicle following safety in mixed traffic flow: A cooperative guidance method for driving behavior decision-making

Under the vehicle following scenario that do not involve the lane changings in a relatively saturated traffic flow operating state, the driver’s driving behavior decision-makings mainly involve the compliant behavior of maintaining the existing following time and the illegal behavior of reducing the following time regardless of the risks. For the purpose of releasing the time pressure, the drivers will often choose the illegal behavior, which causes a sharp decline in the safety and the stability of the traffic flow operations. A method that can guide the drivers to actively choose the compliant behavior needs to be proposed urgently to improve the overall safety and gradually alleviate the efficiency crisis during the more orderly driving process, which is defined as the cooperative guidance. However, the cooperative guidance is not easy to establish because the underlying mechanism of the driving behavior decision-making is difficult to quantify. Based on this background, this paper proposes a cooperative guidance method that acts on the driver’s behavior based on the quantitative analysis of the driving behavior decision-making, which can improve the safety and the efficiency of the traffic flow operations by guiding the driver’s behavior to become more compliant and orderly during the vehicle following processes. In order to better utilize the safety improvement effect of the established cooperative guidance method, this paper sets the vehicle following scenario into a mixed traffic flow composed of the human-driven vehicles (HDVs) and the connected autonomous vehicles (CAVs). At the same time, in order to further exert the safety guarantee and auxiliary guidance role of the CAVs, this paper creates a mixing pattern that makes the CAVs into the CAV platoons (CPs) and the HDVs into the HDV platoons (HPs). By constructing a measuring architecture for different types of the following time in this mixing pattern, this paper proposes a decision-making quantification paradigm based on the comprehensive benefit that can consider both the safety as well as the time saving, and establishes a decision-making model based on the comprehensive benefit (CBDM) to quantitatively simulate the drivers’ benefit-oriented driving behavior decision-makings. Then, an intelligent transportation points system (ITPS) based on the Elo rating algorithm is involved to adjust the comprehensive benefit that can be obtained from different driving behaviors described by the CBDM. Finally, a mixed traffic flow micro-indicator system (M-TFMS) is proposed to update the traffic flow indicators of the constructed scenario and assess the safety improvement effect by using the cooperative guidance under the action of the CBDM and the ITPS. When the cooperative guidance comes into play, the overall safety level when the drivers choose the compliant behavior is up to 4.016% higher than when they choose illegal behavior. At the same time, on the premise of ensuring safety, drivers choosing compliant behavior will increase the lane capacity by 16.11% higher than when they choose illegal behavior. In other words, the cooperative guidance is expected to achieve a win-win situation of the safety and the efficiency.