Hardware Model Calculation Unit for Stable Gaussian Process Computation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing model calculation units in control units face numerical instability during calculations with unfavorable configuration data, limiting the stable calculation of data-based function models, particularly Gaussian process models.
Innovation Solution
A hardware-based model calculation unit is designed with a processor core, including a multiplication unit, addition unit, exponential function unit, configuration register, and logic circuit, which implements a MAC unit for efficient calculations, enabling stable and fast computation of Gaussian process models through input and output standardization and transformation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a model calculation unit is designed on the hardware side for efficient calculation of exponential functions, then the computation rate for Gaussian process models is improved, but numerical instability occurs during calculation with unfavorable configuration data
Solution Approach 1:
The patent implements dynamic scaling factors that are adjusted based on the input data characteristics. The logic circuit dynamically selects appropriate scaling factors from configuration data to maintain numerical stability while preserving calculation speed. This dynamic adaptation allows the system to handle unfavorable configuration data without sacrificing the hardware-accelerated computation rate.
Solution Approach 2:
The patent changes the parameters of the calculation by introducing configurable scaling factors that modify the input data before exponential function calculation. These parameter changes transform the input data into a range that ensures numerical stability during hardware-based exponential calculation, while the logic circuit manages the transformation and restoration processes efficiently.
2Productivity
If configuration data are read into the model calculation unit prior to calculation, then the calculation efficiency is improved, but the system becomes more complex in managing configuration data and parameters
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing scaling factors in the configuration data before the actual Gaussian process calculation. The logic circuit uses these pre-prepared scaling factors during calculation, which simplifies the runtime operation while maintaining calculation efficiency. The configuration data is prepared in advance with all necessary parameters for stable calculation.
Data Source
AI summary
A model calculation unit for calculating a data-based function model in a control unit is provided, the model calculation unit having a processor core which includes: a multiplication unit for carrying out a multiplication on the hardware side; an addition unit for carrying out an addition on the hardware side; an exponential function unit for calculating an exponential function on the hardware side; a memory in the form of a configuration register for storing hyperparameters and node data of the data-based function model to be calculated; and a logic circuit for controlling, on the hardware side, the calculation sequence in the multiplication unit, the addition unit, the exponential function unit and the memory in order to ascertain the data-based function model.