An active DoE method for the automated driving function ACC controller calibration. (25th March 2022)

• Record Type: Journal Article
• Authors: Zhou, Su; Wang, Rucai; Koegeler, Hans-Michael; Jin, Jie
• Is Part Of: International journal of powertrains. Volume 11:Number 1(2022)
• Page Start: 19
• Page End: 37
• Publication Date: 2022-03-25
• Subjects: advanced driver assist system -- ADAS -- design of experiment -- DoE -- key performance indicator model -- calibration -- simulation; Motor vehicles -- Power trains -- Periodicals; 629.25
• Languages: English
• ISSNs: 1742-4267

Abstract

Generally, the calibration of an advanced driver assist system (ADAS) controller is carried out by an experienced operator, who tunes the most relevant parameters based on his experience. However, the calibration results are highly subjective, and could easily end up with a local optimum. To solve this problem, this paper studies an active design of experiment (DoE) calibration method for an adaptive cruise control (ACC) system controller, which needs to be calibrated with the targets of improving fuel efficiency and driving comfort. A statistical method was applied to arrange the simulation points of the most significant parameters. With the simulation results, the key performance indicator (KPI) models of the ACC system were built, to find the global optimum. Finally, the ACC controller was successfully calibrated with the least possible efforts and achieved a 19% reduction of fuel consumption and 49% improvement of driving comfort comparing to the ACC controller with the base set.