Belief-Space Residual Risk for Automated Driving under Localization Uncertainty
For developers of automated driving safety systems, this addresses the practical problem of localization uncertainty in complex urban and adverse weather conditions, providing a more realistic risk assessment.
This work extends residual risk metrics for automated driving to account for localization uncertainty by modeling ego pose as a Gaussian distribution and reformulating risk as expected degradation-induced risk over the belief distribution. The method incorporates localization uncertainty into collision probability computation via covariance fusion.
Residual risk metrics have recently been introduced to assess the safety implications of automated driving systems. Existing approaches typically assume a deterministic ego pose and concentrate mainly on perception errors related to surrounding objects and latency effects. In practice, however, automated vehicles operate under considerable localization uncertainty, especially in complex urban settings and in adverse weather conditions. This work extends the spatial residual risk formulation to the belief space by explicitly modeling ego pose uncertainty as a Gaussian distribution. Residual risk is reformulated as the expected degradation-induced risk over the ego pose belief distribution. Within a particle-based risk estimation framework, localization uncertainty is incorporated into the computation of collision probabilities through covariance fusion of ego and object uncertainties.