Controlled time series generation for automotive software-in-the-loop testing using GANs
This work addresses the problem of insufficient test coverage for automotive mechatronic systems by enabling controlled and realistic synthetic input generation, though it is incremental as it applies existing GAN frameworks to a specific domain.
The paper tackled the challenge of generating realistic input stimuli for automotive software-in-the-loop testing by using Generative Adversarial Networks (GANs) to learn from unlabeled in-vehicle signal data and a metric-based linear interpolation algorithm for controlled generation, resulting in improved virtual test coverage and reduced reliance on expensive field tests.
Testing automotive mechatronic systems partly uses the software-in-the-loop approach, where systematically covering inputs of the system-under-test remains a major challenge. In current practice, there are two major techniques of input stimulation. One approach is to craft input sequences which eases control and feedback of the test process but falls short of exposing the system to realistic scenarios. The other is to replay sequences recorded from field operations which accounts for reality but requires collecting a well-labeled dataset of sufficient capacity for widespread use, which is expensive. This work applies the well-known unsupervised learning framework of Generative Adversarial Networks (GAN) to learn an unlabeled dataset of recorded in-vehicle signals and uses it for generation of synthetic input stimuli. Additionally, a metric-based linear interpolation algorithm is demonstrated, which guarantees that generated stimuli follow a customizable similarity relationship with specified references. This combination of techniques enables controlled generation of a rich range of meaningful and realistic input patterns, improving virtual test coverage and reducing the need for expensive field tests.