文章基本信息

标题：Identification of the Driver Demand Torque Map in Engine ECU Based on Unscented Kalman Filter
本地全文：下载
作者：Diwang Ruan ; Julia Baldyga
期刊名称：IFAC PapersOnLine
印刷版ISSN：2405-8963
出版年度：2022
卷号：55
期号：1
页码：217-222
DOI：10.1016/j.ifacol.2022.04.036
语种：English
出版社：Elsevier
摘要：AbstractIn the electronic control unit (ECU) of a gasoline engine, there are more than a thousand of maps to define the nonlinear and difficult physically-modeled relation between the control input, output, and condition parameters. These maps are normally stored as 1D or 2D lookup tables in ECU after calibration and become part of the controlled object. To implement a high-quality up-layer controller, identification for the actual controlled plant is very necessary, including the inherently embedded control maps. In this paper, the driver demand torque map (DDT map) was taken as an example. At first, the DDT map was inverted and polynomial model was adopted to characterize the relation between torque demand, engine speed and pedal position, with minimum measurement data to initialize the polynomial model coefficients by regression fitting. After that, the fitted DDT map model was taken as a nonlinear system, the polynomial coefficients were taken as states and further recursively updated with unscented Kalman filter (UKF) once new measurements were obtained. Finally, the whole map was identified and compared with the original one. Two different kinds of DDT maps (economic-mode, sport-mode) were studied, and different orders of polynomial models and parameters fine-tuning were also addressed. Results confirm the feasibility and superiority of the UKF-based map identification.
关键词：KeywordsDriver demand torque mapMap identificationUnscented kalman filter(UKF)Gasoline engine