Sequence Detection Reducing Cross-Correlation Complexity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current sequence detection methods in wireless communications systems, such as those used in NB-IoT, face high complexity due to the need to calculate cross-correlation values between received RACH sequences and all possible time domain cyclic shift sequences, leading to increased computational load and inefficiency.
Innovation Solution
The method involves differentiating received sequences based on multiple granularities to reduce the number of candidate frequency domain root sequences to be correlated, thereby reducing the computational complexity by filtering out less relevant sequences and determining the correct sequence through cross-correlation analysis with a subset of candidate sequences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the base station calculates cross correlation values between the received RACH sequence and all possible time domain cyclic shift sequences of all RACH sequences, then the detection accuracy is improved, but the calculation complexity increases significantly
Solution Approach 1:
The patent segments the detection process into two stages: first performing frequency domain correlation to identify candidate root sequences, then performing time domain correlation only on these candidates. This segmentation reduces the total number of correlation operations from examining all possible sequences to examining only promising candidates, thereby reducing computational complexity while maintaining detection accuracy.
Solution Approach 2:
The patent performs preliminary frequency domain correlation analysis before the final time domain detection. By pre-identifying candidate root sequences through frequency domain processing, the system prepares a reduced set of candidates that are more likely to match the received signal, avoiding unnecessary full-scale time domain correlation with all possible sequences.
2Reliability
If all candidate frequency domain root sequences are processed for cross-correlation, then the detection reliability is improved, but the processing time increases
Solution Approach 1:
The patent divides the sequence of operations into frequency domain processing followed by time domain processing. The frequency domain stage acts as a filter that quickly eliminates unlikely candidates, allowing the time domain stage to focus computational resources on a smaller set of promising candidates, thus reducing total processing time while maintaining reliability.
Solution Approach 2:
The frequency domain correlation serves as a preliminary filtering action that identifies candidate root sequences before the more computationally intensive time domain correlation. This preliminary step reduces the number of sequences that require full time domain processing, thereby reducing processing time without compromising detection reliability.
3Reliability
If the base station performs time domain cyclic shifting on all RACH sequences to account for propagation delay, then the detection robustness is improved, but the computational load increases
Solution Approach 1:
The patent segments the correlation operations into frequency domain and time domain components, performing them in sequence rather than exhaustively combining all possibilities. The frequency domain correlation handles root sequence identification, while the time domain correlation handles cyclic shift detection only for candidate sequences, dividing the computational burden into manageable stages.
Solution Approach 2:
The frequency domain correlation performs a preliminary identification of candidate root sequences before time domain cyclic shift analysis. This preliminary action narrows down the search space, allowing the time domain processing to focus only on promising candidates rather than all possible sequences, thereby reducing computational load while maintaining robustness.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
This application relates to a sequence detection method and a device. In embodiments of this application, K candidate frequency domain root sequences may be first filtered based on a differentiation result of received first sequence and differentiation results of candidate frequency domain root sequences, and a candidate frequency domain root sequence to which the first sequence actually corresponds only needs to be determined based on the first sequence and the K candidate frequency domain root sequences. For example, there are U candidate frequency domain root sequences. In this case, a current calculation amount is calculation of U * Cs cross correlation values. In embodiments of this application, a calculation amount is calculation of only L * U + K * Cs, where Cs represents a quantity of sampled time domain cyclic shift values, and L represents a quantity of differentiation granularities. If K is less than U, the calculation amount required in embodiments of this application is far less than the calculation amount required for current sequence detection. Therefore, the technical solution in embodiments of this application can greatly reduce detection complexity of a signal receive end, and improve detection efficiency.