Likelihood Table Reconstruction Across Multiple Modulation Schemes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing likelihood generation devices are limited to specific modulation methods due to differing characteristics in likelihood distributions, making it difficult to handle various modulation methods with a single circuit.
Innovation Solution
A likelihood generation device that includes a similarity detection unit to identify modulation method-specific likelihood similarities, a likelihood table reference unit to store and retrieve relevant data, and a similarity processing unit to perform calculations, allowing for the generation of likelihoods with minimal table capacity without limiting modulation methods.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a deep neural network with many layers is used to improve classification accuracy, then recognition precision is improved, but processing time increases
Solution Approach 1:
The patent applies preliminary action by performing dimensionality reduction on the input image data before feeding it to the deep neural network. The conversion unit reduces the dimensionality of the input data to a lower dimension while preserving essential features, which prepares the data in advance for more efficient processing by the recognition model, thereby reducing processing time without significantly compromising recognition precision.
2Measurement precision
If high-resolution input data is used to improve recognition accuracy, then recognition precision is improved, but processing load increases
Solution Approach 1:
The patent applies local quality by selectively processing different regions of the input data at different resolutions. The conversion unit identifies and processes salient regions or features at higher resolution while reducing or skipping processing in less important regions. This allows the system to maintain recognition precision by preserving critical local information while reducing the overall processing load by not uniformly processing all data at high resolution.
Data Source
Figure 1A
Figure 1B
Figure 2A
AI summary
Provided is a likelihood generation device including: a similarity detection unit (101) configured to receive input of a modulation method selection signal (10) and a received value (11) to detect a likelihood similarity between information bits included in the received value (11) based on the modulation method selection signal (10), and output a likelihood selection signal (13) for specifying likelihood data to be searched for and a calculation selection signal (14) for specifying calculation to be applied to the likelihood data; a likelihood table reference unit (102) configured to register, as likelihood data, only a small region having a likelihood that is different from a likelihood of another small region in a likelihood distribution indicating a likelihood of the information bit, and extract the likelihood data based on the likelihood selection signal (13) from a likelihood table; and a similarity processing unit (103) configured to perform the calculation specified by the calculation selection signal (14) for the extracted likelihood data, to thereby acquire the entire likelihood distribution.