State Estimation And Road Identification For A Full Drive-by-wire Electric Vehicle

In the year of2013,84.27million cars were sold, which increased by3.8%comparedwith the year of2012, in which81.21million cars were sold. In China,21.93million carswere sold in2013, which increased by14.2%compared with the year of2012, in which19.21million cars were sold. Earlier in the year of2009, the production and salesnumbers of Chinese domestic car market both exceeded those of American domestic carmarket. As a result, Chinese car market replaced American car market and thus becamethe largest car market in the world. As the number of cars increasing year by year, trafficand environment problems are becoming more and more serious. In some cities of China,traffic is becoming more and more crowded, and the environment is becoming worse andworse. Take Beijing as example. There are more than5.2million cars in Beijing. Thetraffic in urban areas is very crowded, and is becoming more crowded. Haze weatheris also becoming more frequent. As a result, Beijing government has introduced a seriesof policies that limit the purchase and driving of cars. Some other cities in China havefollowed Beijing and introduced similar policies as well to alleviate theincreasingly serious traffic and environment problems.As a kind of new energy vehicles, electric vehicles can realize―Zero Emission‖andtheir use-cost is lower than that of traditional vehicles. As a result, electric vehicles areobtaining the widespread attention and researches in the world. Now in the worldwide,several electric vehicles have been put into production, such as Nissan Leaf and BYD e6.These vehicles have won the recognition of certain consumers. The full drive-by-wire electric vehicle is a new kind of electric vehicles in whose actuators X-by-wiretechnology is used. For a full drive-by-wire electric vehicle named UFEV (Urban FutureElectric Vehicle) which is designed and made by Professor Zong’s research group,X-by-wire technology is used in its steering, driving and braking systems. Meanwhile,UFEV can perform4wheel independent steering (4WIS),4wheel independent driving(4WID) and4wheel independent braking (4WIB).This paper is carried on with the full drive-by-wire electric vehicle UFEV as theobject of study. Based on the learning, analysis and summary of the domestic and foreignresearch achievements on vehicle state estimation and road identification, vehicle states(longitudinal velocity, lateral velocity and yaw rate) are estimated using EKF and UKFrespectively in this paper, and the two methods are compared. Road properties (peakvalue of road adhesion coefficient and optimal slip ratio) are identified using Mamdaniand T-S fuzzy controllers respectively in this paper, and the two methods are compared.The main research works of this paper are summarized as follows：First, based on the features of UFEV and the purpose of this paper, seven degrees offreedom vehicle dynamic model is built to describe the motion and turing modes of UFEV.The seven degrees of freedom are longitudinal velocity at the center of mass, lateralvelocity at the center of mass, yaw rate of the vehicle and the turning angles of fourwheels respectively. This vehicle dynamic model can describe any turning mode of UFEVas well as any motion in longitudinal, lateral or yaw direction. The vehicle model inCarSim is modified based on the features of UFEV in order to ensure that simulationusing Simulink and CarSim can be close to actual situations.Then using HSRI tire model, vehicle states are estimated using EKF and UKFrespectively in several simulation conditions, and the two methods are compared. Roadproperties are identified using Mamdani and T-S fuzzy controllers respectively in severalsimulation conditions, and the two methods are compared.At last, the combination of EKF and T-S is adopted in the vehicle road tests with thecomparison of various methods. Considering the several turning modes of UFEV, different road tests have been carried on. The combination of EKF and T-S is verified byoffline post-processing with vehicle tests data. The results of both the simulation and thevehicle road tests show that both EKF and UKF can estimate the vehicle states with agood accuracy, and EKF may be a little intuitively clearer so that debugging of EKF maybe a bit more convenient. As for road identification, both Mamdani and T-S fuzzycontrollers can identify the road properties with a satisfying accuracy, and T-S fuzzycontroller may be easier in structure and more efficient.