RBF Gradient Calculation Hardware for Real-Time Control Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data-based models, such as Gaussian process models, require high computing capacity to achieve sufficient accuracy and struggle to calculate complex relationships within real-time requirements for control unit applications, particularly in systems like internal combustion engines.
Innovation Solution
A hard-wired model calculation unit is designed to efficiently calculate radial basis function (RBF) models using a processor core with predefined processing algorithms, reducing processing load and enabling real-time calculations by utilizing a hardware circuit that calculates gradients and model values, allowing for a lower number of nodes without compromising accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If data-based models with a large number of nodes are used to achieve sufficient modeling accuracy, then modeling precision is improved, but computing capacity requirements increase and real-time calculation becomes difficult
Solution Approach 1:
The patent segments the gradient calculation into two separate calculation paths: one for calculating the RBF model value and another for calculating the gradient. This segmentation allows the system to reuse intermediate calculation results (specifically the exponential function results) across both paths, thereby reducing redundant computations and lowering the overall computing capacity requirements while maintaining modeling accuracy with a reduced number of nodes.
2Measurement precision
If data-based models with a large number of nodes are used to achieve sufficient modeling accuracy, then modeling precision is improved, but calculation time increases and real-time requirements cannot be met
Solution Approach 1:
The patent performs preliminary calculations of intermediate values (such as the squared differences and exponential function results) that are common to both the model value calculation and gradient calculation. These preliminary results are stored and reused, eliminating redundant computations and significantly reducing the total calculation time required, thereby enabling real-time control applications.
3Measurement precision
If gradient calculation is performed separately from model value calculation, then accuracy is maintained, but processing load and calculation time increase
Solution Approach 1:
The patent merges the gradient calculation and model value calculation into a unified computational framework where intermediate results are shared. Specifically, the exponential function results computed for the model value are reused in the gradient calculation, and the squared difference terms are computed once and used in both calculations. This merging maintains mathematical accuracy while dramatically improving processing efficiency by eliminating redundant operations.
Data Source
AI summary
A model calculation unit for calculating a gradient with respect to a certain input variable of input variables of a predefined input variable vector for an RBF model with the aid of a hard-wired processor core designed as hardware for calculating a fixedly predefined processing algorithm in coupled functional blocks, the processor core being designed to calculate the gradient with respect to the certain input variable for an RBF model as a function of one or multiple input variable(s) of the input variable vector of an input dimension, of a number of nodes, of length scales predefined for each node and each input dimension, and of parameters of the RBF function predefined for each node.