Odd-Even Bit Weight Equalization for Balanced DWA Component Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The data weighted averaging (DWA) algorithm struggles to achieve equal probability of selecting each component during the conversion process due to manufacturing process mismatches, leading to signal interference and poor component performance.
Innovation Solution
An odd and even bit weight equalization method and system that adjusts bit values through a weight equalization circuit, performing multiple iterations to set even and odd bits to a first weight value, ensuring each bit value has an equal probability of being adjusted to that value.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the data weighted averaging (DWA) algorithm is used to overcome manufacturing process mismatch, then component performance can be improved, but the probability of selecting each component cannot be controlled to be approximately equal
Solution Approach 1:
The patent segments the bit weighting process into separate even bit and odd bit handling procedures. By dividing the weight adjustment into distinct segments (even bits and odd bits) with independent control logic, the system can more precisely control the selection probability of each component, ensuring approximately equal probabilities while maintaining improved component performance through weighted averaging.
2Manufacturing precision
If bit values are adjusted through multiple iterations to achieve equal probability, then selection probability equality is improved, but the complexity of the weight equalization circuit increases
Solution Approach 1:
The patent implements periodic action through iterative weight adjustment cycles that alternate between processing even bits and odd bits. This periodic approach systematically achieves equal selection probabilities through repeated refinement while managing circuit complexity by using structured, repeating patterns of adjustment rather than complex one-time calculations.
Solution Approach 2:
The patent applies preliminary action by pre-determining weight values through systematic iteration before final conversion. The weight equalization circuit pre-calculates and stores optimal weight values that ensure equal selection probability, eliminating the need for complex real-time calculations during the actual conversion process and thereby reducing operational complexity.
3Manufacturing precision
If even and odd bits are processed separately with iterative adjustment, then selection probability equality is achieved, but the number of processing steps increases
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing optimal weight values through iterative adjustment before the actual conversion process. This pre-computation phase, though requiring multiple steps, eliminates the need for complex real-time calculations during conversion, thereby reducing the time loss during critical operational phases.
Solution Approach 2:
The patent uses periodic action through structured alternating processing of even and odd bits in systematic cycles. This organized periodic approach, while increasing the number of processing steps, optimizes the sequence to achieve equal selection probabilities efficiently, preventing unnecessary redundant operations and minimizing overall processing time through systematic refinement.
Data Source
AI summary
An odd and even bit weight equalization method and system are provided. The method includes processes of: determining whether or not bit values are equal to a weight value; if not, setting even and odd initial bits, and if yes, setting a next even bit of an even bit to which an even bit position is last shifted previously as the even initial bit, and setting a next odd bit of an odd bit to which an odd bit position is last shifted previously; shifting the even bit position from the even initial bit to other even bits and shifting the odd bit position from the odd initial bit to other odd bits; and adjusting the even and odd initial bits, and the even and odd bits to which the even and odd bit positions are shifted, to be equal to the weight value.