Scenario Taxonomy

6 categories, 27 sub-scenarios classifying all known L4+ robotaxi operational anomalies. Each maps to a recommended response layer (1/2/3).

Layer 1: AI Autonomous Response (70%)
Layer 2: AI-Assisted + Human Confirm (25%)
Layer 3: Remote Driving / On-Site (5%)
A

System-Wide Failure

系统性故障

Anomalies affecting multiple vehicles simultaneously due to infrastructure, cloud, or platform-level failures.

B

Perception/Decision Failure

感知/决策失效

Vehicle's perception system or decision-making logic fails to correctly interpret the driving environment.

C

Planning/Execution Anomaly

规划/执行异常

Vehicle's planning or control system produces dangerous or frozen output despite correct perception.

D

Vehicle Hardware Failure

车辆硬件故障

Anomalies caused by failure of vehicle hardware components — sensors, powertrain, battery, or actuators.

E

External Conflict

外部环境冲突

Anomalies caused by external events or road users that the ADS must respond to.

F

Passenger-Side Issue

乘客端异常

Anomalies originating from or primarily affecting passengers.