Autonomous Vehicle Mode Switching Under Low-Confidence Conditions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in maintaining confidence in their operation, particularly when encountering uncertain or complex environments, leading to a need for assistance when their confidence level falls below a predetermined threshold.
Innovation Solution
The vehicle identifies situations where its confidence is low and sends a request for assistance to a remote operator, including sensor data and video streams, allowing the remote assistor to propose a new mode of operation, which the vehicle then switches to, using a control system configured to operate in different autonomous modes based on received guidance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If the autonomous vehicle operates independently without remote assistance, then the extent of automation is improved, but the reliability deteriorates in low-confidence situations
Solution Approach 1:
A remote operator serves as an intermediary between the autonomous vehicle and the control system. When the vehicle's confidence in its autonomous operation falls below a threshold, the system automatically connects to a remote operator who can view sensor data and provide guidance or take control, thus maintaining high automation while ensuring reliability through human intervention when needed.
2Reliability
If the vehicle switches to remote assistance mode, then the reliability is improved, but the extent of automation deteriorates
Solution Approach 1:
The system dynamically adjusts the extent of automation based on real-time confidence levels. The vehicle operates autonomously by default (high automation) but can switch to remote-assisted mode (lower automation) when confidence drops below a threshold. This dynamic adjustment allows the system to optimize between automation and reliability depending on the operational context.
3Reliability
If the vehicle requests remote assistance frequently, then the reliability is improved, but the loss of time increases due to communication delays
Solution Approach 1:
The system takes preliminary action by establishing a confidence threshold that triggers remote assistance requests. By proactively identifying low-confidence situations before they lead to failures, the system can seek assistance early, potentially preventing more time-critical issues from developing. The threshold-based approach ensures requests are made at appropriate moments rather than continuously.
Solution Approach 2:
The vehicle prepares for potential remote assistance by continuously monitoring its confidence level and having the communication infrastructure ready. When the confidence threshold is breached, the system has already identified the issue and can immediately initiate the connection to remote operators, reducing the overall response time compared to reactive assistance models.
Data Source
AI summary
Example systems and methods enable an autonomous vehicle to request assistance from a remote operator when the vehicle's confidence in operation is low. One example method includes operating an autonomous vehicle in a first autonomous mode. The method may also include identifying a situation where a level of confidence of an autonomous operation in the first autonomous mode is below a threshold level. The method may further include sending a request for assistance to a remote assistor, the request including sensor data representative of a portion of an environment of the autonomous vehicle. The method may additionally include receiving a response from the remote assistor, the response indicating a second autonomous mode of operation. The method may also include causing the autonomous vehicle to operate in the second autonomous mode of operation in accordance with the response from the remote assistor.