Autonomous Vehicle Planner Recovery Using Fallback Trajectories
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face issues with safely managing recoverable errors during autonomous driving, leading to potential stranding due to immediate shutdowns without adequate recovery time, which can result in unsafe conditions.
Innovation Solution
Implementing a system where computing devices assess errors against a recoverable error list, allowing the planner system a 'second chance' to generate a new trajectory within predetermined time limits, and using fallback trajectories to ensure safe vehicle control if recovery is not possible.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the planner system immediately shuts down when an error occurs, then safety is ensured by preventing continued operation with faulty systems, but the vehicle may be stranded unnecessarily due to recoverable errors
Solution Approach 1:
The system implements a buffer mechanism by introducing a recoverable error list and timeout period before shutdown. When an error occurs, instead of immediate shutdown, the system checks if the error is recoverable and allows a timeout period for recovery attempts. This cushioning mechanism prevents premature shutdowns for transient errors while still ensuring safety through eventual shutdown if recovery fails, thus resolving the contradiction between safety and vehicle availability.
2Productivity
If the planner system allows unlimited recovery attempts, then vehicle availability is maintained by keeping the system operational, but system stability deteriorates due to persistent errors that should trigger shutdown
Solution Approach 1:
The system implements dynamic error handling by introducing a timeout mechanism that adapts the system's response based on error persistence. The timeout period acts as a dynamic threshold: if recovery succeeds within the timeout, normal operation continues; if the timeout expires, the system transitions to shutdown mode. This dynamic approach allows flexible recovery attempts while maintaining system stability by enforcing a definitive shutdown condition if recovery fails, resolving the contradiction between vehicle availability and system stability.
3Reliability
If the system implements comprehensive error checking and recovery protocols, then reliability is improved by ensuring safe operation, but device complexity increases due to additional computing devices and protocols
Solution Approach 1:
The system segments the error handling process into distinct functional modules: error detection, error classification (recoverable vs. non-recoverable), timeout management, recovery attempt execution, and shutdown decision. Each module is implemented as a separate computing device or software component with specific responsibilities. This segmentation improves reliability by ensuring thorough error checking while managing complexity through modular design, where each segment can be independently developed, tested, and maintained.
Data Source
AI summary
Aspects of the disclosure relate to controlling a vehicle in an autonomous driving mode using trajectories. For instance, a trajectory may be received by one or more first computing devices from one or more second computing devices. While the first computing devices are controlling the vehicle in the autonomous driving mode based on the trajectory, an error may be generated by second computing devices. Whether the error is a recoverable error may be determined, and if so, the second computing devices attempt to generate a new trajectory. When the second computing devices generate the new trajectory, the vehicle may be controlled by the first computing devices according to the new trajectory.