ML Maneuver Output Biasing for Adjustable Aerial Vehicle Risk
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
ML algorithms used in safety-critical applications like aerial vehicle flight control systems are not designed to be altered based on acceptable risk levels, limiting their ability to adjust maneuver recommendations to mitigate structural damage during adverse conditions.
Innovation Solution
The implementation of an output biasing system that adjusts the maneuver confidence threshold using risk tuning parameters, allowing the ML algorithm to increase or decrease structural risk based on specific mission parameters or conditions, thereby influencing the selection of maneuver recommendations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the maneuver confidence threshold is set to ensure high safety standards, then the structural damage risk is reduced, but the ability to complete missions in adverse conditions deteriorates
Solution Approach 1:
The patent applies dynamics by making the maneuver confidence threshold adjustable rather than fixed. The threshold can be dynamically modified based on risk tuning parameters and mission-critical factors, allowing the system to adapt between conservative (high threshold) and aggressive (low threshold) maneuver selection strategies depending on operational context
Solution Approach 2:
The patent changes the parameter of the confidence threshold based on risk tuning parameters. By modifying this key parameter, the system can shift between different risk tolerance levels, enabling mission completion when appropriate while maintaining safety standards when required
2Reliability
If the ML algorithm is made rigid to ensure consistent safety standards, then the reliability is improved, but the adaptability to different risk scenarios deteriorates
Solution Approach 1:
The system transitions from a static, rigid ML algorithm to a dynamic one that can adjust its behavior based on risk tuning parameters. The algorithm maintains its core safety logic while gaining the ability to adapt to different operational scenarios through parameter modification
Solution Approach 2:
The patent introduces risk tuning parameters that modify the ML algorithm's output without changing its fundamental structure. This allows the same algorithm to serve multiple risk tolerance levels by changing parameters rather than requiring multiple different algorithms
3Productivity
If the maneuver confidence threshold is lowered to accept higher structural risk, then the mission completion capability is improved, but the reliability deteriorates
Solution Approach 1:
The system dynamically adjusts the confidence threshold based on mission-critical factors and risk tuning parameters. When mission completion is prioritized, the threshold can be lowered to accept higher risk maneuvers, while maintaining the ability to raise it when safety is prioritized
Solution Approach 2:
The patent modifies the confidence threshold parameter based on risk tuning inputs, enabling the system to operate at different risk levels. This parameter change allows the same system to both complete challenging missions and maintain high safety standards depending on operational requirements
Data Source
AI summary
In some examples, systems and methods are described for output biasing maneuvers recommendations provided by at least one machine learning maneuver-recommendation (MLM) engine executing on an aerial vehicle. In some examples, output biasing data can be received that includes at least one risk tuning parameter that can influence which of the maneuver recommendations are selected by a maneuver decision engine executing on the aerial vehicle based on a maneuver confidence threshold for implementation by the aerial vehicle. The maneuver confidence threshold can be updated based on the at least one risk tuning parameter to provide an updated maneuver confidence threshold for the output biasing of the maneuvers recommendation provided by the at least one MLM engine. Vehicle command data for implementing a given maneuver recommendation can be outputted based on an evaluation of the updated maneuver confidence threshold.