Algorithm Design And Experimental Verification For Lane Keeping Assisted Control Based On Electric Power Steering

Traditional automotive safety generally means passive safety, which reduces passengers’ damage during traffic accidents. Recently, automotive active safety is attracting more and more attention from specialists and consumers. DAS(Driver Assisted System) is one of the most important active safety techonology, including LKAS(Lane Keeping Assisted System), APAS(Automatic Parking Assisted System), PAS(Parking Assisted System), etc. During driving, a driver is in distraction because of sleepiness, making calls, talking with other passengers, etc. LKAS is designed to solve this problem, which could keep a vehicle in its target lane by active control to ensure the vehicle safe.This paper combines itself with the National Natural Science Fund Project "Research on the steering control mechanism and evaluation method of information system based on the characteristics of drivers by wire"(Number: 51575223). Based on learning researches and development related to LKAS both at home and abroad, this paper builds up warning and assisted control algorithm applied to control LKAS based on EPS(Electric Power Steering) System. Then the paper verifies the designed algorithm by simulated tests and hard-ware-inloop experiments. Main contents are as follows:(1) In order to improve LKAS warning accuracy, the paper sets up LKAS warning algorithm considering driving habit. The paper builds up multi-mode TLC(Time to Lane) algorithm and FOD(Future Offset Distance) algorithm to satisfy driver’s diversity. Each algorithm has its applicable area and different algorithms have different applicable areas, but there is no one algorithm covering all the employed conditions. In order to solve this problem, the paper proposes a kind of joint warning algorithm to ensure LKAS warning accuracy. In order to prove the proposed joint algorithm valid, this paper carrys out joint simulated experiments.(2) For the purpose of decreasing the nuisance alarm rate of LKAS, the paper studys about LKAS application situation and working principle for a long time to build driver steering state identification model. When a driver distracts his attention from controlling his car, LKAS needs to detect a vehicle departing from the original lane or not based on steering wheel torque and steering light. The paper judges driver’s aim to avoid nuisance alarm, which guarantees machine interaction friendliness.(3) The departing vehicle is still in the danger of departing from the target lane, while the driver does nothing druing LKAS warning time. In order to solve this problem, the paper builds LKAS assisted control algorithm based on EPS. By the deviation between the real and target lateral position, the driver model puts out the target steering wheel angle used to correct the departing motion of vehicle. EPS assisted motor drives the steering mechanism moving to correct the motion locus of the vehicle. For the purpose of verifying the validity of the designed algorithm, the paper conducts joint simulated experiments, which provides theoretical basis for the following hardware in the loop test bench experimrnts.(4) The paper remolds the exiting EPS hardware-in-loop test bench and does experiments on it for the sake of verifying LKAS warning algorithm and assisted algorithm valid. This test bench selects DS 1006 as real-time interface and uses Car Sim RT and Simulnik to build simulated environment. After finishing configuring hardware interface, the paper tests sensors and designs test conditions according to intelligent assisted driving testing criterion. Under the circumstance of Control Desk, it allows testers to adjust LKAS algorithm parameters realtimely. By tests on this test bench, the paper verifies LKAS warning and assisted control algorithm valid.