Autonomous Driving Validation with Model-Based Dictionary Clustering

Validation of autonomous driving systems remains one of the biggest challenges that car manufacturers must tackle in order to provide safe driverless cars. The complexity of this task stems from several factors: the multiplicity of vehicles, embedded systems, use cases, and the high level of reliability that is required for the driving system to be at least as safe as a human driver. In order to circumvent these issues, large scale simulation that reproduces physical conditions is intensively used to test driverless cars. Therefore, this validation step produces a massive amount of data that needs to be processed. In this paper, we present a new method applied to time-series produced by autonomous driving numerical simulations. It is a dictionary-based method that consists in three steps: automatic segmentation of each time-series, regime dictionary construction, and clustering of produced categorical sequences. We present the time-series specific structure and the proposed method's advantages for processing such data, compared to state-of-the-art reference methods.