Layered LDPC Coding for Lower Delay Error Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current communication technologies face challenges with the complexity and time variation of communication channels, leading to inefficiencies in error control methods like ARQ and FEC, particularly with TURBO codes, which require high system costs, complex decoding, and are not suitable for real-time services.
Innovation Solution
The implementation of a layered LDPC code system that constructs and decodes data using multiple layers of check matrices, reducing system costs and decoding delay, and improving performance for high-speed data services by using quasi-cyclic LDPC codes and bit-filling or progressive edge-growth algorithms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If TURBO code is used for error correction, then error correcting capability is improved, but system cost increases due to need for CRC check bits
Solution Approach 1:
The patent applies self-service by enabling the LDPC code to perform its own error detection function through its inherent structure and decoding characteristics, eliminating the need for separate CRC check bits. The code automatically detects and corrects errors during the decoding process without requiring additional external checking mechanisms.
2Reliability
If TURBO code decoding is performed, then error correction is achieved, but decoding delay increases due to complex iteration process
Solution Approach 1:
The patent applies segmentation by dividing the LDPC code into multiple layers with different code rates. This layered structure allows the receiver to attempt decoding with lower-layer codes first (which have higher code rates and require less processing), and only proceed to higher layers if initial decoding fails. This segmented approach significantly reduces average decoding delay compared to uniform complex decoding.
Solution Approach 2:
The patent applies partial action by implementing incremental redundancy transmission where only the necessary portion of the coded data is transmitted. The receiver attempts decoding with the received data and only requests additional redundancy if decoding fails, rather than transmitting all possible redundancy information upfront. This reduces the amount of processing required and decreases decoding delay.
3Reliability
If TURBO code is used, then error correction capability is improved, but decoding complexity increases leading to huge caching requirements
Solution Approach 1:
The patent segments the LDPC code into multiple layers with progressively lower code rates. Each layer can be decoded independently with its own smaller buffer requirements. The receiver only needs to cache data for the current decoding attempt and can discard data from successfully decoded layers, significantly reducing total caching requirements compared to storing all data for a single complex decoding process.
4Productivity
If processing workload is concentrated on mobile station, then high-speed downlink transmission is achieved, but mobile station complexity and power consumption increase
Solution Approach 1:
The patent applies segmentation by creating a layered code structure where the base layer can be decoded with simpler algorithms requiring less computational power. This allows mobile stations with limited processing capability to successfully decode at least the essential information, while more powerful receivers can attempt to decode additional layers for enhanced performance.
Solution Approach 2:
The patent applies local quality by optimizing different layers of the code for different receiver capabilities. The lower layers are designed with properties suitable for mobile stations (simpler decoding, lower complexity), while upper layers provide enhanced performance for fixed or more capable receivers. This local optimization allows the same transmission to serve multiple types of receivers effectively.
Data Source
AI summary
The present invention discloses a method, an apparatus and a system for low-density parity-check (LDPC) coding and decoding. The coding method includes the following steps: constructing each layer of a check matrix of a layered LDPC code used as an error correcting code; when data is initially sent by a data transmitting terminal, performing first-layer-coding of the data to be sent by using the first layer of a check matrix of the LDPC code, sending the first-layer-coded data; when data for (n−1)th retransmission is sent by the data transmitting terminal, performing nth-layer-coding of the data by using the nth layer of a check matrix of the LDPC code, sending the nth-layer-coded data, wherein n is an integer no less than 2. It is possible to reduce the system overhead, decrease the decoding delay, and improve the decoding performance by using the technical solution of the present invention which is also adapted to high-speed data services.