Precoding Vector Computation for Interference Suppression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current precoding technologies in large-scale antenna systems face high computation complexity, leading to increased energy consumption and delay, especially in 5G wireless communication systems, where Zero-Forcing (ZF) precoding methods consume significant energy and cause system delays due to complex matrix inversion operations.
Innovation Solution
A precoding method that selects a user set for partial interference suppression based on distances or channel coherence between users and the transmitting end device, computing a precoding vector for each scheduled user using a matrix formed by channel estimations of the user set, thereby reducing computation complexity while maintaining interference suppression.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If ZF precoding technology is used to eliminate interference between users, then interference suppression performance is improved, but computation complexity increases significantly
Solution Approach 1:
The patent segments the user set into multiple subsets and performs precoding computation separately for each subset. Instead of computing precoding vectors for all K users simultaneously using full matrix inversion, the system divides users into groups (e.g., based on channel coherence or distance criteria) and computes precoding vectors for each group independently. This segmentation reduces the computational complexity from O(M²K) to O(M² × (K/n)), where n is the number of segments, thereby resolving the contradiction between interference suppression performance and computation complexity.
2Reliability
If ZF precoding technology is used to suppress interference, then user interference is eliminated, but energy consumption increases to 80% of total energy consumption
Solution Approach 1:
The patent reduces energy consumption by segmenting the user set and performing precoding computations on smaller subsets. Since energy consumption is directly proportional to computation complexity, dividing the full user set K into n smaller subsets reduces the total energy required for precoding from E_full to approximately E_full/n, thereby resolving the contradiction between interference suppression performance and energy consumption.
3Reliability
If matrix inversion is performed for each coherent bandwidth in ZF precoding, then interference suppression is achieved, but system delay increases to 30% of coherence time
Solution Approach 1:
The patent reduces system delay by segmenting the user set and performing precoding computations on smaller subsets. Since the time consumption for matrix inversion is proportional to the square of the number of users, dividing the full user set K into n smaller subsets reduces the total computation time from T_full to approximately T_full/n. This segmentation approach maintains interference suppression performance while reducing the delay from 30% of coherence time to a more acceptable level, thereby resolving the contradiction between reliability and time loss.
4Device complexity
If MF precoding technology is used, then computation complexity is reduced, but interference between users cannot be eliminated
Solution Approach 1:
The patent combines the advantages of both MF and ZF approaches through segmentation. For each user subset, the system computes precoding vectors that provide interference suppression similar to ZF, but only for the users within that subset. This segmented approach achieves partial interference suppression with computational complexity similar to MF precoding, thereby resolving the contradiction between computation complexity and interference suppression performance by finding a middle ground between the two extreme approaches.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
The present invention discloses a precoding method and device. In the present invention, acquiring, by a precoding device, all scheduled users on a resource block, and selecting, within a range of all the scheduled users on the resource block, a user set used for partial interference suppression for each scheduled user on the resource block; for each scheduled user on the resource block, computing a precoding vector for a current scheduled user according to a matrix formed by channel estimations of all users in the user set used for the partial interference suppression corresponding to the current scheduled user, and precoding data of the scheduled users according to the computed precoding vector. Precoding computation complexity is reduced by using the present invention.