Nonlinear Parameter Estimation Using Concurrent Filters and Neural Inference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional inferential sensing methods struggle to accurately estimate the time-varying parameters of complex, highly nonlinear systems, particularly in applications like autonomous navigation and spacecraft operations, due to underconstrained estimation problems and the lack of generalized parameter estimation methods.
Innovation Solution
A two-step process involving a filter operating in Failure Mode 2 and a multiple concurrent filter methodology, followed by an auxiliary neural network, to estimate time-varying parameters based on input-output behaviors, allowing for unlimited parameter estimation independent of the number of system outputs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional inferential sensing methods (e.g., Kalman Filter variants) are used for parameter estimation, then the system can handle linearizable state-space equations with a limited number of parameters, but the system cannot accurately estimate parameters in highly nonlinear systems with a high number of time-varying parameters causing underconstrained estimation problems
Solution Approach 1:
The patent segments the parameter estimation problem by dividing the system parameters into two categories: identified parameters (with a limited number equal to or less than the number of independent measured outputs) and unidentified parameters (the remaining parameters). This segmentation allows the system to handle highly nonlinear systems with many parameters by focusing estimation efforts on the identified subset while treating unidentified parameters as constants or using specialized ad hoc methods, thereby resolving the underconstrained estimation problem without sacrificing accuracy.
2Reliability
If specialized ad hoc methods are used for underconstrained estimation applications, then the system can provide limited scope adaptive control solutions, but the system lacks a broadly applicable inferential sensing method for complex nonlinear systems
Solution Approach 1:
The patent creates a universal parameter estimation framework that can handle both constrained and underconstrained estimation problems across different system types. By establishing a general methodology that identifies which parameters can be estimated from available measurements and which require specialized handling, the system achieves broad applicability to complex nonlinear systems while maintaining reliable adaptive control capability through the identified parameter estimates.
3Adaptability or versatility
If the number of time-varying parameters exceeds the number of independent measured outputs, then the system can model complex nonlinear behavior, but the estimation problem becomes underconstrained and traditionally unsolvable
Solution Approach 1:
The patent segments parameters into identified and unidentified categories based on the relationship between the number of parameters and measured outputs. This allows the system to model complex nonlinear behavior with many parameters while maintaining estimation feasibility by focusing on estimating only the identified parameters that can be determined from available measurements, thereby resolving the underconstrained problem.
Solution Approach 2:
The patent uses computational models (e.g., computational fluid dynamics) to create virtual representations of the physical system. These computational models serve as copies that can be used for simulation and analysis without requiring direct measurement of all parameters, allowing the system to handle cases where the number of parameters exceeds the number of measurements by relying on the computational model's predictions for unidentified parameters.
Data Source
AI summary
The embodiments are directed to an inferential sensing system, methods and computer program product of an estimator for estimating parameters of complex nonlinear time-varying systems from scarce system output measurements. The estimator comprises a two-step process to accurately estimate the time-varying parameters of the time-varying system based on the input and output sample of the time-varying system. First, multiple filters in the high frequency processing loop, operating independently and concurrently process the input and output samples of the time-varying system to generate a hypersurface comprising time series objects. Each filter is restricted to adapt only a subset of the modeled time-varying parameters. The hypersurface comprising the time series objects is aggregated over several iterations of the high frequency processing loop. Second, the hypersurface is passed through a neural network in the low frequency processing loop to infer estimates of the time-varying system parameters.