Optimization Of Traffic Sensors’ Layout On Highways With On/off Ramps Under Intelligent Connected Vehicle Environment

With the promotion of connected and Automated Vehicles(CAV),the highway will exist a mixed traffic flow of connected vehicles and human-driven vehicles(HV).The existing sensors on China’s highways are mainly for human-driven vehicles and have many problems,such as low detection accuracy and poor timeliness,which cannot meet the needs of mixed traffic flow scenarios.Based on this,this article mainly studies the existing sensors on the highway on-ramp system and the communicative detectors in the new highway ramp system.Taking the estimation accuracy of the travel time and the time-average velocity of the highway section as the main optimization goal,the existing sensors equipped with communication devices and the layout of new highway communicable sensors are optimized on two levels.Firstly,in the vehicle-to-vehicle(V2V)communication environment and vehicle-toinfrastructure(V2I)communication environment,the car-following and lane-changing models are constructed for CAV and HV,respectively.A mixed traffic flow system performance evaluation system was established from average speed,delay,speed standard deviation,and average absolute relative error of travel time estimation.The influence of sensor communication range,sensor number,and location on-road performance was analyzed.The results show that when the communication range of the sensor is between 500 and 1000 meters,the improvement effect of mixed traffic flow system performance is the best.As the number of sensors increases,the estimation accuracy of mixed traffic flow system performance will first increase and then converge to a constant value.Then,based on the Long short-term Memory Network(LSTM)information fusion,an optimized layout model for installing communicable devices is established for the highway ramp system with existing sensors.In the optimization model,the information collected by the fixed sensor and CAV is taken as the input of LSTM,the average road time speed after fusion is taken as the output,the estimation error of the middle road time speed is taken as the target,the location of the existing sensor with the communicable device is taken as the decision variable,and the installation cost is taken as the constraint.A genetic algorithm based on LSTM to calculate individual fitness was proposed to solve the problem.The influence of the maximum number of mobile devices and CAV penetration on the estimation error of mixed traffic flow system time average speed was analyzed.The case analysis shows that when the number of sensors installed with communication devices reaches the optimum,the velocity estimation error does not change.In addition,when the penetration is 40%,the estimation error of mixed traffic flow system average velocity is the smallest.When the penetration reaches a particular value,the estimation error of rate is no longer sensitive to penetration changes.Finally,considering the information fusion of moving sensors,communicable fixed sensor,and CAV,the optimal layout model of communicable sensors for the newly built highway ramp system is established.The model is established taking the minimum time average velocity estimation error as the optimization goal,with the number of the communicable fixed sensors and the proportion of moving sensors as constraints.Genetic algorithm is used to solve the model.The effect of the maximum number of distributions of communicable fixed sensors,the proportion of CAV penetration and the proportion of moving sensor on the calculation error are analyzed.The results show that the optimal proportion of moving sensors is 6%.It can improve the estimation accuracy of mixed traffic flow system state parameters,and the layout cost is negligible.