Driver-pedestrian Interaction Analysis And Right-of-way Assignment Strategy In Typical Urban Road Scenarios

The development of artificial intelligence and computer vision promote the transformation and upgrading of the automotive industry,and the research of autonomous driving is now in the ascendant.In urban environment,mixed traffic is a challenging driving scenario for selfdriving vehicles due to its complex traffic composition and multiple interactive behaviors.Autonomous vehicles need to communicate with other road users to understand their intentions and give feedback as human drivers.Pedestrians,as one of the most vulnerable road users,are common in urban traffic.Therefore,it is of great significance to understand the process of interaction between pedestrians and human drivers and the behavioral characteristics of both.Based on the analysis of the interaction characteristics between pedestrians and human drivers in traditional mixed traffic,this paper establishes a pedestrian trajectory prediction model,and propose a right-of-way assignment strategy under the condition of pedestrian-vehicle interaction.Specifically,the research work of this paper mainly includes the following contents:(1)Analysis of pedestrian-vehicle interaction characteristics in typical scenarios.Roundabout,unsignalized/signalized intersection,and mid-block crosswalk are selected as the typical interactive scene of urban mixed traffic.According to the mixed traffic flow trajectory data from traffic video,the motion characteristic parameters of each road users are extracted,and the statistical distribution of relevant parameters of the trajectory data are obtained.On this basis,the detailed process of pedestrian vehicle interaction in typical scenes is given,and the motion characteristics and state of pedestrian crossing are extracted,and the correlation mechanism of interaction behavior decision-making is proposed.(2)A pedestrian trajectory prediction method based on pedestrian-vehicle interaction is constructed.On the basis of decision-making behavior of pedestrians,this paper analyzes the factors of pedestrian-vehicle interaction behavior,selects appropriate explanatory variables to describe the key indicators,and identifies the most important interactive indicators through the importance ranking process.Then,an LSTM-based pedestrian crossing trajectory prediction method in mixed traffic environment is developed.(3)Dynamic right-of-way assignment strategy for mixed traffic flow in typical scenarios.This paper analyzes the traffic characteristics and passing demand of road users,divides the different passing stages of vehicles and pedestrians,calculates the front and rear safety distance for each road user,and based on the analysis of conflict,determines the ownership of right-ofway,and gives the reasonable passing order of road users in each direction.A dynamic rightof-way assignment strategy is finally proposed to balance the passing efficiency and road safety in each scenario.