Application of Robust Engineering Methods to Improve ECU Software Testing

Robust Engineering techniques developed by Taguchi have traditionally applied to the optimization of engineering designs. Robust Engineering methods also may be applied to software testing of ECU algorithms. The net result is an approach capable of improving the software algorithm in one of two ways. First the approach can identify the range of areas which prove problematic to the software such that a robust solution may be developed. Conversely, the approach can be used as a general strategy to verify that the software is robust over the range of inputs tested. The robust engineering methods applied to software testing utilize orthogonal array experiments to test software over a range of inputs. The actual software trials are best performed in the simulation environment and also via automated test hardware in the loop configurations in realtime. This paper outlines a process for applying Robust Engineering methods to software testing. Experiments derived from the robust methodologies are then implemented using both software simulation and hardware in the loop environments. A case study is presented where the process is applied to improve the robustness of an ECU software algorithm. Results are obtained from both using simulation and performing automated hardware in the loop tests.