Functional degradation strategy for localization failure of urban autonomous driving vehicles

Localization failure is a significant challenge faced by urban autonomous driving applications. To ensure vehicle safety in the event of localization failure, a functional degradation strategy is proposed from the perspective of the autonomous driving system. The strategy introduces a hierarchical four-stage degradation framework, enabling dynamic switching between localization-based path tracking, dead reckoning (DR)-based tracking, lane-keeping control, and manual takeover according to localization status, DR duration and lane lines quality. Specific degradation strategies are designed for typical urban scenarios—including straight and curved roads, intersections, and on-ramp scenarios. Hardware-in-the-loop (HIL) experiments demonstrate that the functional degradation strategy effectively addresses localization failures in real urban scenarios.