Vehicular Speed Control Of Eco-Driving Systems Based On Connected Vehicles

Eco-Driving Assistance Systems (EDAS) assist drivers to drive in an ecologicalway through collecting information of the road, traffic, driving behavior and vehiclestatus. Vehicle economy is related to many factors, among which, the complexity ofroad enviroments, diversity of driving behavior and variability of driving conditions,hamper the development of EDAS systems. In order to improve the adaptability ofEDAS in complex conditions, this paper proposes an Eco-Speed control algorithmbased on connected vehicles through telematics. Based on analysis of the fuelconsumption characteristic, combined with road and traffic flow conditions, the vehicledrives ecologically under the double-layer control. An Eco-Speed control algorithmconsidering road and traffic conditions is designed, and the effectiveness of thealgorithm is validated.Based on definition of the system control objectives, hierarchical control method isused to establish the architecture of the Eco-Speed control system. Difficulties in thekey technologies, fuel consumption analysis and modelling, design of control law andverification methods, along with their relations are clarified.After systematic analysis of traditional fuel consumption models, vehicle fuelconsumption characteristics are analized and polynomial fuel consumption model isselected. Akaike’s information criterion (AIC) is introduced to evaluate polynomial fuelconsumption models with different structures, and real vehicle fuel consumption dataare used to validate the best fuel consumption model under AIC, which provides supportfor the Eco-Speed control algorithm.Double-layer control law is designed, constrained by vehicle status, road andtraffic condition, to optimize vehicle speed. The first layer is the average eco-speedcontrol, within which, traffic light signals, traffic flow rate and speed limit of the roadsare abstracted to node constraints and link constraints in mathematical equations, fuelconsumption estimated from the proposed fuel consumption model is used as costfunction, and optimization methods are use to calculate the average eco-speed in eachsection of the road. When there are multiple node and link constraints, the difficulty inoptimization is increased, and thus a multi-intersection eco-speed control algorithm based on Dijkstra algorithm is proposed. The second layer is the real-time eco-speedcontrol, within which, eco-speed from the first layer control is added as constraint, andvehicle speed is smoothed to reduce fuel consumption when approaching an intersection.In the second layer of control, an eco-speed control based on dynamic programming isdesigned to save fuel for vehicle travelling through slopes. Meanwhile, switching logicamong these sub-algorithms is also designed.For system validation, simulation platform is built, and the sub-algorithms arecombined into8speed control algorithms. Fuel consumption, travelling time and stoptimes of the8speed control algorithms are compared through simulation. The resultsshow that the double-layer Eco-Speed control algorithm can reduce fuel consumptionwithout increasing travelling time and number of stops, which leads to the best overallperformance.