Massive MIMO Downlink Channel Estimation Using Compressed Sensing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional downlink channel estimation methods in massive MIMO systems face inefficiencies due to assumptions of channel reciprocity and the need for extensive training sequences, leading to high overhead and reduced resource allocation for data transmission.
Innovation Solution
A method using compressed sensing (CS) for downlink channel estimation, where a base station determines and transmits a dictionary to user equipment (UE) for each UE, allowing for accurate channel estimation based on feedback information, thereby reducing pilot overhead and improving channel state information (CSI) accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional downlink channel estimation methods are used with extensive training sequences, then channel estimation can be performed, but overhead increases and resource allocation for data transmission decreases
Solution Approach 1:
The patent extracts the essential channel characteristics by representing the channel as a sparse vector in a transformed domain using a dictionary matrix. Instead of transmitting extensive training sequences, only the sparse channel coefficients and dictionary information are exchanged, significantly reducing overhead while maintaining estimation accuracy.
Solution Approach 2:
The patent transforms the channel estimation problem from the time domain to a sparse domain using a dictionary matrix, changing the representation parameters. This transformation allows the channel to be described by fewer non-zero coefficients, reducing the training sequence requirements while preserving estimation precision.
2Productivity
If channel reciprocity is assumed in TDD mode, then downlink channel estimation can be performed based on uplink pilots, but channel state information accuracy deteriorates due to non-ideal hardware and calibration errors
Solution Approach 1:
The patent implements a feedback mechanism where the user equipment estimates the downlink channel using received training sequences, quantizes the channel state information, and feeds it back to the base station. This direct downlink estimation approach avoids the inaccuracies of assuming channel reciprocity while maintaining estimation efficiency.
3Reliability
If downlink CSI is fed back from user equipment for precoding design, then precoding can be optimized, but feedback overhead increases
Solution Approach 1:
The patent extracts only the essential channel information by representing the downlink channel as a sparse vector with non-zero coefficients in a transformed domain. Instead of feeding back complete channel matrices, only the sparse coefficients and their positions are transmitted, significantly reducing feedback overhead while maintaining precoding performance.
4Quantity of substance
If compressed sensing with dictionary-based representation is used, then training length is reduced, but system complexity increases due to dictionary construction and selection
Solution Approach 1:
The patent performs preliminary action by pre-defining a set of candidate dictionary matrices at both the base station and user equipment before channel estimation. These dictionaries are constructed based on statistical channel properties and are stored locally, eliminating the need for real-time dictionary construction and reducing processing complexity during actual operation.
Data Source
AI summary
Method and apparatus for downlink (DL) channel estimation in massive MIMO are provided. A base station (BS) support massive MIMO may select a dictionary for a user equipment (UE) in its cell and transmit information for constructing the dictionary to the UE. The UE constructs the dictionary based on the information received and performs compressed sensing (CS)-based DL channel estimation for the BS. The UE also sends feedback information to the BS, which include information that is useful for determination of the dictionary by the BS.