Vehicle Feedback Control for Lane-Keeping and Collision Avoidance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in ensuring driving safety due to the complexity of driving environments, particularly in maintaining lane position and avoiding collisions.
Innovation Solution
A computer program product is implemented in an onboard vehicle computer system that receives command data, corrects driving commands to generate modified command data, and modifies the vehicle control system's operation to ensure the vehicle operates within an intended lane and maintains a safe distance from other vehicles, thereby avoiding potential collisions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If autonomous vehicle uses machine learning system to generate driving commands, then the vehicle can navigate autonomously, but driving safety cannot be ensured in complex scenarios
Solution Approach 1:
The patent introduces a feedback system as an intermediary component between the machine learning system and vehicle control system. This feedback system receives driving commands from the machine learning system, evaluates their safety using trained feedback models, and provides feedback signals to modify commands that fail safety checks. This mediator architecture allows the autonomous navigation capability to be maintained while adding a safety verification layer that ensures driving safety in complex scenarios.
2Productivity
If the vehicle strictly follows machine learning commands, then the navigation efficiency is maintained, but the vehicle may deviate from intended lane or collide with other vehicles
Solution Approach 1:
The patent implements a feedback mechanism where the feedback system continuously monitors driving commands and vehicle state, compares actual trajectory with intended trajectory using trained feedback models, and generates feedback signals to correct deviations. This closed-loop feedback approach maintains navigation efficiency by only intervening when necessary while preventing harmful outcomes through continuous safety verification and command adjustment.
3Reliability
If the vehicle modifies driving commands to ensure safety, then collision avoidance is achieved, but the navigation route may be disrupted
Solution Approach 1:
The patent applies partial action by selectively modifying only those driving commands that fail safety verification while leaving compliant commands unchanged. The feedback system performs minimal necessary interventions - adjusting steering angles, speeds, or trajectories only when safety risks are detected - thereby maintaining collision avoidance capability while preserving the overall navigation route consistency and minimizing disruptions to the planned path.
Data Source
AI summary
Embodiments for improving driving commands for vehicular end-to-end driving include a method for a connected vehicle that includes receiving command data describing one or more driving commands for the connected vehicle. The method includes correcting the one or more driving commands to generate modified command data that describes one or more modified driving commands. The one or more modified driving commands control an operation of the connected vehicle to satisfy a trajectory safety requirement. The method includes modifying an operation of a vehicle control system of the connected vehicle based on the modified command data so that the connected vehicle operates based on the one or more modified driving commands.