Three-Stage M-CPFSK Demodulation Reducing Computational Complexity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for demodulating continuous phase frequency shift keying (CPFSK) and continuous phase modulation (CPM) signals face challenges with high computational complexity, large hardware requirements, and reduced noise immunity, especially when dealing with arbitrary modulation indices, long impulse responses, Doppler fluctuations, and fading.
Innovation Solution
A three-stage demodulation method that iterates over symbol sequences within specific block lengths, using sampling and phase information to reduce the number of symbol sequences iterated over, and employing neural networks to achieve noise immunity comparable to fully coherent Maximum Likelihood Sequence Detection (MLSD) demodulators, while being implementable in low-cost hardware.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If fully coherent Maximum Likelihood Sequence Detection (MLSD) demodulation is used, then noise immunity is maximized, but computational complexity and hardware requirements become prohibitively high
Solution Approach 1:
The patent segments the demodulation process into three distinct stages: (1) initial symbol detection using a simplified metric, (2) Viterbi algorithm processing with reduced state space, and (3) final sequence optimization. This segmentation allows the system to achieve near-MLSD performance while avoiding the prohibitive computational complexity of exhaustive MLSD by breaking down the problem into manageable parts that can be processed sequentially with decreasing complexity at each stage.
Solution Approach 2:
The patent applies partial action by implementing a three-stage demodulation process that performs sufficient processing to achieve near-MLSD noise immunity without completing the full exhaustive MLSD algorithm. The first stage performs partial detection on all possible symbols, the second stage applies Viterbi to a reduced set of candidate sequences, and the third stage performs final optimization, thereby achieving adequate performance with reduced computational effort.
2Productivity
If the number of symbol sequences iterated over is reduced for lower computational complexity, then processing speed increases, but noise immunity deteriorates
Solution Approach 1:
The patent applies preliminary action by performing initial symbol detection in the first stage using a simplified metric that identifies the most likely symbol values before applying the Viterbi algorithm. This preliminary detection narrows down the search space for subsequent stages, allowing the system to focus computational resources on the most probable symbol sequences and thereby maintain noise immunity while improving processing speed.
Solution Approach 2:
The patent implements dynamics by adaptively adjusting the number of symbol sequences processed at each stage based on the confidence metrics from previous stages. The system dynamically reduces the search space in later stages based on the results of earlier stages, allowing flexible optimization between processing speed and noise immunity depending on signal conditions and computational resources available.
3Reliability
If three-stage demodulation with iterative symbol sequence processing is used, then noise immunity approaches fully coherent MLSD, but device complexity increases compared to simple non-coherent methods
Solution Approach 1:
The patent segments the complex demodulation task into three stages with decreasing computational requirements, where each stage builds on the results of the previous stage. This segmentation allows the use of simpler hardware components at each stage rather than requiring full MLSD capability throughout, thereby achieving near-MLSD performance with more manageable hardware complexity.
Solution Approach 2:
The patent introduces intermediary structures such as symbol likelihood metrics and reduced-state Viterbi tables that mediate between the received signal and the final decoded sequence. These intermediaries allow the system to process information in a staged manner, reducing the immediate computational burden while preserving the essential information needed for high noise immunity, thus bridging the gap between simple and complex demodulation approaches.
Data Source
AI summary
Method of demodulation of M-CPFSK signal, includes receiving the M-CPFSK radio signal; moving it to zero frequency; sampling at no less than double a frequency of symbols; storing the samples with their amplitude and phase for at least L4 symbols; demodulating the sampled signal in three stages, wherein each stage includes iterating over symbol values within a block of symbols, of length is L1, L2 and then L3; in the first stage, N1 symbol sequences out of all possible symbol sequences are iterated over, at the second stage, N2 symbol sequences out of all possible symbol sequences are iterated over, and at the third stage, N3 symbol sequences out of all possible symbol sequences are iterated over, to obtain final symbol values; symbol values obtained at previous stage is used in a next stage to reduce a number of symbol sequences; and determining encoded bits based on final symbol values.