A Cost Effective System Test-Bed using Model Based Approach
Published April 5, 2016 by SAE International in United States
Annotation of this paper is available
The effort involved in automotive software test/calibration at road level is very high and cost involved is also commendable because of the involved proto level samples. Further the on-road test/calibration process is sensitive to external factors like drive pattern and environmental conditions. It is always a challenge for any OEM, to come up with an efficient process, which optimizes development cost, time and reliability of the product.
The model based test/calibration process is always a dream for any engineer to work on, as it has big advantage of cost, reproducibility and repeatability of test cases . But the challenge lies in achieving the closeness to reality with limited availability of crucial data for model parameterization.
Activity at test bed level bridges the gap between the on-road and model based test/calibration achieving high maturity level at optimal cost/time.
Current vehicle has many systems, which work in synergy to create an impact on end customer. The integration of all these systems at test bed level may not be viable in terms of cost and component availability. Further, the objective may be to test all the involved systems in the vehicle.
The paper addresses this problem, wherein the system under test will be real hardware, and the rest of the interacting systems will be emulated by a real-time capable model . The vehicle plant model is developed in AVL-CRUISE and MATLAB/SIMULINK environment. The real-time capable model runs in ETAS-RTPC environment simulating the vehicle dynamics synchronous with the real components. This paper also describes the challenges faced during the implementation/commissioning phase of the test bed. The application being the testing of actual component functionality or calibration at system level, in real-time.