Systems and methods for estimating dynamic system states and parameters

JP2026517580AActive Publication Date: 2026-06-02ストラトス パーセプションエルエルシー
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
ストラトス パーセプションエルエルシー
Filing Date
2024-04-02
Publication Date
2026-06-02

AI Technical Summary

Technical Problem

Conventional inferential sensing methods struggle with estimating time-varying parameters in highly nonlinear systems, leading to under-constrained estimation conditions and limiting the feasibility of adaptive control in complex systems such as autonomous navigation and artificial intelligence, especially in scenarios with a large number of time-varying parameters.

Method used

A two-step process using multiple filters and a neural network to estimate time-varying parameters, where each filter adapts a subset of parameters independently, generating hypersurfaces that are aggregated to infer accurate parameter estimates, allowing for unlimited parameter estimation regardless of the number of system outputs.

Benefits of technology

Enables comprehensive fault detection and management, autonomous control, and supervision of artificial intelligence by providing real-time, accurate estimates of time-varying parameters, enhancing system adaptability and safety in complex environments.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2026517580000001_ABST
    Figure 2026517580000001_ABST
Patent Text Reader

Abstract

The embodiments relate to an inference sensing system, method, and computer program product of an estimator for estimating parameters of a complex nonlinear time-varying system from missing system output measurements. The estimator includes a two-step process for accurately estimating the time-varying parameters of a time-varying system based on input and output samples of the time-varying system. First, multiple filters in a high-frequency processing loop operate to process input and output samples of the time-varying system independently and concurrently to generate a hypersurface containing time-series objects. Each filter is restricted to adapt only a subset of the modeled time-varying parameters. The hypersurface containing the time-series objects is aggregated through multiple iterations of the high-frequency processing loop. Second, the hypersurface passes through a neural network in a low-frequency processing loop to infer estimates of the time-varying system parameters.
Need to check novelty before this filing date? Find Prior Art