Future State Estimation Using Attenuation Transition Matrices
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for predicting future states in model predictive control, such as those used in automobiles and industrial plants, face challenges in calculating optimal control laws for distant future states due to the requirement of longer prediction times, which are constrained by computer performance and control periods.
Innovation Solution
A future state estimation method and device that calculates states of a control object and its surrounding environment in an infinite time ahead using a probability density distribution, allowing for rapid estimation independent of the time to the predicted future state, employing an attenuation type state transition matrix calculated from a state transition probability matrix.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If iterative calculation methods are used to predict future states, then prediction accuracy is improved, but calculation time increases with the time horizon
Solution Approach 1:
The patent pre-calculates and stores transition probability matrices for various time horizons before actual prediction is needed. This preliminary computation allows the system to retrieve pre-computed results during operation, eliminating the need for iterative calculations at prediction time and thus resolving the contradiction between accuracy and calculation time.
Solution Approach 2:
The patent computes transition probability matrices for a range of time horizons beyond what might be immediately needed, storing these partial results for future use. By computing more than the minimum required (excessive action), the system ensures that any future prediction query can be answered quickly with pre-computed data, avoiding iterative calculations when time is critical.
2Duration of action of moving object
If prediction time horizon is extended to infinite future, then long-term optimization is improved, but computational complexity increases
Solution Approach 1:
The patent transforms the problem from predicting specific future states to computing transition probability matrices that capture the statistical behavior over infinite time horizons. By changing the parameter representation from detailed state trajectories to aggregated probability distributions, the system can handle infinite time horizons without proportional increases in computational complexity.
Solution Approach 2:
The patent uses the transition probability matrix as a compact representation (copy) of the system's long-term behavior patterns. Instead of simulating or calculating actual infinite-time trajectories, the system creates a probabilistic model that copies the essential characteristics of long-term system evolution, enabling efficient computation and storage.
3Manufacturing precision
If detailed future state predictions are made, then control optimization is improved, but data storage requirements increase
Solution Approach 1:
The patent extracts only the essential information needed for control optimization from the full future state predictions. By computing and storing only the transition probability matrices (which contain the necessary statistical information for optimization) rather than complete detailed state trajectories, the system reduces data storage requirements while maintaining control optimization capability.
Data Source
Figure 1~2
Figure 3~4
Figure 5~7
AI summary
An objective of the present invention is to provide a future state estimation method and future state estimation device with which, given that the space in which the estimation is carried out is finite, it is possible to rapidly estimate the states of a controlled object and the peripheral environment thereof in the form of probability density distribution for infinite future. Provided is a future state estimation device characterized by comprising: a model storage part for saving a model for simulating a subject of simulation and the peripheral environment of the subject of simulation; a future state forecast result storage part for storing information obtained by estimating future states of the subject of simulation and the peripheral environment of the subject of simulation within a finite space in the form of probability density distribution, for either an infinite time or a given time step in the future; and a future state forecast computation part for carrying out calculation, which is equivalent to a series, using the model for simulating the future states of the subject of simulation and the peripheral environment of the subject of simulation in the form of probability density distribution.