Future State Estimation via Weighted Basis Functions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for predicting future states of operation targets in continuous states are inefficient, as they require longer calculation times and do not effectively estimate states in infinite time ahead in a continuous state space.
Innovation Solution
A future state estimation apparatus that uses a storage device to store a state transition model expressed as a linear combination of weighted basis functions and an arithmetic device to calculate state transition probabilities through product-sum calculations, enabling rapid estimation of future states in continuous state spaces.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If iterative calculation method is used to predict future state, then prediction accuracy is improved, but calculation time increases
Solution Approach 1:
The patent replaces the iterative mechanical calculation process with a neural network-based computational system. The neural network is trained offline using iterative methods, but during online prediction, it provides rapid results without iteration, substituting the time-consuming iterative calculation with a pre-trained intelligent model that delivers both accuracy and speed.
2Productivity
If prediction time horizon is extended to infinite time, then predictive performance is improved, but calculation complexity increases
Solution Approach 1:
The patent performs preliminary training of the neural network offline before actual use. During this preliminary phase, the network learns from historical data and is prepared to handle infinite time horizon predictions. When deployed, the pre-trained network can rapidly predict future states without complex real-time calculations, as the heavy computational work was done in advance.
Data Source
AI summary
A storage device stores a state transition model in which a first state transition probability (state transition probability function τ), which indicates the probability that a prediction subject will transition from a first state s to a second state s′ after a first time (Δt) has elapsed, is expressed by a linear combination of weighted basis functions (model storage unit). A computation device calculates a second state transition probability (attenuation-type state transition probability function D), which indicates the probability that the prediction subject will transition from the first state s to the second state s′ after a second time (Δt×∞) has elapsed, by means of a product-sum computation of a weight matrix ∇ representing a matrix the elements of which are the respective weights of the weighted basis functions (future state prediction computation unit).