HIL Component Test Benches for Boosting Testing Efficiency

Lecturers: Dr.-Ing. E. Sax, Dipl.-Ing. Stefan Abendroth, Dipl.-Ing. Alexander Radler, Dr.-Ing. S. Schmerler, Mercedes-Benz Technology GmbH

Constant reductions in development time, rapidly increasing software complexity, and at the same time tough requirements for the quality of E/E components, together make up one of the greatest challenges facing the automotive industry. To keep the cost of testing at a reasonable level despite these shifting variables, optimized and coordinated methods and processes must be used both for development methods and for validating and testing. Uniform design, reusability, and the automation of component and integration tests are major factors in achieving the necessary improvement in efficiency.

Particularly safety-critical applications such as restraint systems (airbag), and steering and braking assistants, must therefore be tested in a coordinated process parallel to the development process itself. This requires transparency of the tests performed from the early design phases, when the function is available as an executable model or as code on a prototype, right through to the electronic control unit (ECU) integrated into the vehicle.

Different testing equipment is used depending on the testing objective and the version of the object under test. If an individual ECU needs testing under hard real-time conditions, the obvious, and proven, solution is hardware-in-the-loop technology for component testing. Modeling the ECU’s environment enables its functionality to be tested in the development phase even before concrete actuators and sensors are available. The maturity of the ECU can be assessed at an early stage by skillfully varying the model parameters and combining constraints and events.

This paper focuses on project planning and the process-oriented use of HIL component test benches for restraint systems, based on our own experience in practical application. The challenges to implementation were the extremely short processing times of under 200 µs that usually elapse between the detection of a crash input and the triggering of an ignitor. The potential, however, is obvious: Being able to test highly safety-critical airbag systems not only in practice but also on the model in the lab provides enormous benefits. These include the reusability of equipment, the ability to combine and sequence the crash patterns freely to mimic a great variety of crash scenarios, and regression of test cases. The result is long-term flexibility and cost reduction.

dSPACE components were used to implement the test bench. Individual extensions complete the test equipment. The test process that was set up, and that is now a daily reality, is oriented to TEmb. The paper reports in detail on both.