Autonomous Driving Control via Sensor Confidence-Based Action Gating
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High-order autonomous driving vehicles face safety issues when sensors fail, leading to sudden transfer of control rights to drivers, compromising driving safety and user experience.
Innovation Solution
An intelligent driving method that determines a confidence level of sensing sectors based on sensor faults and adjusts execution strategies for autonomous driving actions to prevent sudden control transfers, allowing or prohibiting actions based on confidence levels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If sensor fault detection is implemented with direct control right transfer to driver, then driving safety is compromised due to sudden handover, but maintaining autonomous driving function requires keeping control rights with system
Solution Approach 1:
The system performs preliminary actions by determining confidence levels of sensing sectors before executing autonomous driving actions. When sensor faults are detected, the system proactively adjusts execution strategies in advance, allowing ongoing actions to complete while preventing new actions that would require faulty sensing data. This preliminary assessment and strategic adjustment prevents sudden control handovers while maintaining autonomous driving functionality.
Solution Approach 2:
The system dynamically adjusts execution strategies based on real-time confidence level assessments. When sensor faults occur, the control strategy transitions from full autonomous execution to a restricted mode where actions are evaluated against confidence thresholds. This dynamic adaptation allows the system to maintain autonomous driving experience in safe conditions while ensuring driving safety when sensor reliability deteriorates.
2Extent of automation
If autonomous driving actions are executed based on faulty sensor data, then autonomous driving function is maintained, but driving safety deteriorates due to incorrect decisions
Solution Approach 1:
The system applies local quality by assessing confidence levels for specific sensing sectors rather than uniformly disabling autonomous driving. Each sensing sector is evaluated independently, and execution strategies are adjusted locally for actions dependent on specific sectors. This allows the autonomous driving function to continue at high level for actions with sufficient confidence while restricting only those actions affected by sensor faults.
Solution Approach 2:
The system implements feedback by continuously monitoring sensor status and confidence levels, then adjusting execution strategies accordingly. When sensor faults are detected, the feedback loop triggers strategic adjustments that prevent execution of actions dependent on unreliable sensing data. This closed-loop feedback mechanism maintains autonomous driving functionality while ensuring safety by adapting to real-time sensor conditions.
3Reliability
If sensor fault handling immediately terminates autonomous driving mode, then driving safety is improved, but autonomous driving experience deteriorates due to frequent mode switching
Solution Approach 1:
The system performs preliminary assessment of confidence levels before determining execution strategies. When sensor faults occur, rather than immediately terminating autonomous mode, the system proactively evaluates which actions can still be safely executed. This preliminary action allows the system to maintain autonomous driving mode for actions with sufficient confidence, extending autonomous driving duration while preserving safety.
Solution Approach 2:
The system dynamically adapts autonomous driving mode rather than statically terminating it. The execution strategy transitions from full autonomous execution to a confidence-based restricted mode, allowing the system to maintain high-level autonomous functionality for actions with adequate confidence while restricting actions dependent on faulty sensors. This dynamic adaptation extends autonomous driving duration without compromising safety.
Data Source
AI summary
An intelligent driving method and apparatus are related to the field of intelligent driving. The method includes determining a confidence level of a sensing sector based on a fault condition of a sensor included in the sensing sector. The method further includes determining, based on the confidence level, an execution strategy corresponding to an autonomous driving action. The method further includes executing the autonomous driving action according to the execution strategy.