Viterbi Traceback Parallelization for Faster Decoding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing Viterbi decoders face challenges in achieving high-speed processing for high data rate decoding due to the inability to perform parallel processing techniques, leading to processing time limitations independent of hardware resources.

Innovation Solution

A Viterbi decoding processor that executes at least two backtracking steps in parallel by using both possible inputs for two processing blocks, allowing the second step to utilize the output bit of the first step, and employing a multiplexer to select between results, thereby enabling faster decoding.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of time

If sequential backtracking processing is used in Viterbi decoder, then hardware complexity is reduced, but processing time increases

Engineering Contradiction:
Improveprocessing timeVSAvoidhardware complexity
Core Design Contradiction:
Loss of timeVSDevice complexity

Solution Approach 1:

The backtracking process is divided into multiple independent steps (first backtracking step, second backtracking step, etc.), where each step processes different survivor path metrics separately. This segmentation allows parallel execution of multiple steps without increasing overall hardware complexity, as each step uses dedicated processing blocks that can operate simultaneously.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent transitions from sequential time-based processing to parallel spatial processing by executing multiple backtracking steps simultaneously using separate processing blocks. This dimensional change from temporal to spatial processing enables faster decoding without proportionally increasing hardware resources.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Productivity

If parallel processing is implemented in Viterbi decoder, then processing speed increases, but hardware resources increase

Engineering Contradiction:
Improvedecoding speedVSAvoidhardware resources
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

Processing blocks are segmented into specialized units (first processing block, second processing block, multiplexer) that handle specific backtracking steps. This segmentation enables parallel processing of multiple survivor path metrics simultaneously, increasing decoding speed while keeping each individual block relatively simple and resource-efficient.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary organization of survivor path metrics into groups corresponding to different backtracking steps before processing. This preliminary arrangement allows the parallel processing blocks to immediately access and process the required metrics without additional overhead, maximizing decoding speed while minimizing hardware resource requirements.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3996285B1Parallel backtracking in viterbi decoder
Publication Date: 2026.02.18 NXP BV
  • EP3996285B1 patent drawingFigure 1~2
  • EP3996285B1 patent drawingFigure 3
  • EP3996285B1 patent drawingFigure 4

AI summary

A Viterbi traceback processing method, system, and apparatus are provided wherein a first Viterbi traceback processing operation (MUX 514) is performed on a first survivor path metric (TMV1) by selecting, in response to a back track state (INDEX 0), a first output data bit (Ti1) for the first survivor path metric, wherein a plurality of Viterbi traceback processing operations (MUX 512, 513) are performed on respective portions of an additional survivor path metric (TMV2A, TMV2B) by selecting, in response to a shifted back track state (INDEX 1), candidate data bits (Tn1, Tn2) for the additional survivor path metric, wherein a multiplexer (MUX 518) controlled by the first output data bit selects between the candidate data bits to generate an additional output data bit (Ti2) for the additional survivor path metric such that the Viterbi traceback processing operations are performed in parallel to produce the output data bits.