Adaptive Controller Sign Correction for RLS Estimators
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional adaptive controllers using recursive least-squares (RLS) estimators often incorrectly determine the sign of response, leading to inappropriate adjustments in resource allocation, especially when external events impact system performance, resulting in suboptimal management of black-box computing systems.
Innovation Solution
Proposed extensions to RLS include the 'Remember' method to use a good past model, the 'Modify' method to correct the model, and the 'RunAll' method to select the best model from multiple estimators, ensuring the sign of the performance model is correct and thus guiding the adaptive controller to take appropriate actions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional RLS-based estimators are used in adaptive controllers, then the controller can automatically tune itself while the system is running, but the sign of response may be incorrectly determined leading to inappropriate adjustments
Solution Approach 1:
The patent implements a feedback mechanism where the estimated model from RLS is continuously monitored for sign correctness. When the sign is determined to be incorrect (based on performance degradation or external event detection), the system provides corrective feedback to adjust the model sign, ensuring the controller responds appropriately to system changes.
Solution Approach 2:
The patent introduces an intermediary component (sign correction mechanism) that sits between the RLS estimator and the control law. This intermediary detects when the estimated model has an incorrect sign and corrects it before the control law is applied, preventing inappropriate controller responses while maintaining the automatic tuning capability.
2Productivity
If the adaptive controller adjusts actuators based on RLS model estimates, then performance goals can be achieved, but external events not accounted for in the model can cause incorrect sign determination
Solution Approach 1:
The patent makes the model dynamic by allowing the sign parameter to change based on detected external events or performance degradation. Instead of a static model assumption, the system dynamically adjusts the model sign to adapt to changing system conditions, including external events not originally accounted for in the model.
Solution Approach 2:
The patent changes the model parameters (specifically the sign of the response) based on detected external events or performance feedback. When an external event is detected or incorrect response is identified, the system modifies the sign parameter of the RLS model to restore accuracy, allowing continued performance goal achievement despite external disturbances.
3Device complexity
If non-adaptive control methods are used, then the system structure remains simple, but the controller cannot handle systems that change too much
Solution Approach 1:
The patent segments the controller into distinct functional components: the RLS estimator, the sign detection mechanism, and the control law. This segmentation allows the complex adaptive functionality to be added while maintaining a clear, modular structure that is easier to understand and implement than fully integrated adaptive controllers.
Data Source
AI summary
According to one embodiment, a method comprises receiving, by an adaptive controller, performance measurement for a computing system. The method further comprises estimating a performance model for use by the adaptive controller, and determining whether the estimated performance model has a correct sign for approaching performance desired for the computing system. When determined that the estimated performance model has an incorrect sign, the adaptive controller takes action to determine a performance model having a correct sign for approaching performance desired for the computing system.