Sequential Decoder Early Termination for Lower ML Search Complexity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing decoding methods in wireless communication systems face challenges with high computational complexity, especially when dealing with increasing constellation sizes and the number of antennas, leading to suboptimal performance due to limited hardware resources, particularly in deep fading channels.
Innovation Solution
A decoder system that includes a symbol estimation unit and a termination alarm monitoring unit, which reduces the weight metric of each node in the decoding tree based on a metric parameter associated with a termination alarm, allowing for anticipated termination of the decoding process while maintaining optimal performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If ML decoding techniques with exhaustive search are used, then decoding performance is optimized, but computational complexity increases significantly
Solution Approach 1:
The patent segments the exhaustive search space of ML decoding into a structured decoding tree with multiple levels and nodes. Each node represents a partial solution, allowing the decoder to explore the search space systematically rather than evaluating all possibilities simultaneously, thereby reducing computational complexity while maintaining optimal performance
Solution Approach 2:
The patent implements preliminary actions by computing metric values for each node in the decoding tree before making decoding decisions. The metric computation is performed in advance for all nodes at each level, allowing the decoder to anticipate termination conditions and prepare for early termination without sacrificing performance, thus managing computational complexity
2Productivity
If the constellation size or number of antennas increases, then communication capacity improves, but decoder complexity increases
Solution Approach 1:
The patent segments the high-dimensional decoding problem into a tree structure where each level corresponds to a dimension. This segmentation allows the decoder to handle increased constellation size and number of antennas by processing dimensions sequentially through tree levels rather than simultaneously, managing complexity while maintaining capacity
Solution Approach 2:
The patent transforms the high-dimensional decoding problem into a multi-level tree structure, adding a temporal dimension to the search process. Nodes at different levels represent different dimensions of the signal space, allowing the decoder to manage high-dimensional inputs (increased antennas and constellation size) by exploring them dimension-by-dimension
3Device complexity
If decoding is terminated early to reduce complexity, then computational resources are preserved, but decoding performance deteriorates
Solution Approach 1:
The patent performs preliminary metric computations for all nodes in advance, storing these values before termination is triggered. When early termination occurs, the decoder can immediately select the best path based on pre-computed metrics without needing to complete the full decoding process, thus preserving both performance and computational resources
Solution Approach 2:
The patent implements feedback through termination alarms that monitor decoding progress and trigger early termination when predefined conditions are met. The feedback mechanism allows the system to adaptively stop decoding when sufficient performance is achieved, preventing unnecessary computational expenditure while maintaining reliability
4Speed
If stack decoding strategy is used to improve search efficiency, then decoding speed increases, but storage requirements increase
Solution Approach 1:
The patent segments the storage requirements into node-specific metric values stored at each tree level. Rather than storing complete path information for all possible decodings, only essential metric data is stored at each node, reducing overall storage requirements while maintaining the speed benefits of the stack decoding strategy
Data Source
AI summary
There is provided a decoder for decoding a data signal received through a transmission channel in a communication system, the decoder (310) comprising a symbol estimation unit (311) configured to determine estimated symbols representative of the transmitted symbols carried by the received signal, the estimated symbols being determined from nodes of a decoding tree based on a weight metric associated with each of the node. The decoder further comprises a termination alarm monitoring unit (312) for monitoring a termination alarm depending on the current decoding computation complexity, the termination alarm being associated with a metric parameter, the symbol estimation unit being configured to reduce the weight metric of each node of the decoding tree by a quantity corresponding to a function of the metric parameter associated with the termination alarm, in response to the triggering of the termination alarm.