CSI Feedback Scheme for FD-MIMO Using Partial Precoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current LTE systems face challenges in providing a high-performance, scalable, and flexible channel state information (CSI) feedback framework for Frequency Division, Multiple Input, Multiple Output (FD-MIMO) communications, especially in Frequency Division Duplex (FDD) scenarios, where the previous PMI-based feedback approach is inadequate and excessive in terms of feedback requirements.
Innovation Solution
The proposed solution involves a CSI feedback scheme that uses a finite set of basis functions/vectors/precoders to reduce the number of coefficients that need to be quantized and reported, allowing for partial precoding of CSI-RS, which enables efficient resource utilization and improved CSI-RS coverage, particularly in scenarios with closely spaced large 2D antenna arrays and low mobility.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If a finite set of basis functions/vectors/precoders is used to reduce quantization coefficients, then overhead in CSI feedback is reduced, but measurement precision of channel state information may be degraded
Solution Approach 1:
The channel state information feedback is segmented into two parts: long-term channel statistics (spatial correlation, angular spread) that change slowly and are fed back less frequently, and short-term channel variations that are fed back more frequently. This segmentation allows reduced feedback overhead for the dominant long-term components while maintaining accuracy through periodic updates.
Solution Approach 2:
The system performs preliminary measurement and feedback of long-term channel statistics (spatial correlation matrices, angular spread parameters) before actual data transmission. These pre-acquired channel characteristics are stored at both eNodeB and UE, enabling efficient short-term CSI feedback without repeatedly transmitting full channel matrices.
Solution Approach 3:
The system changes the parameter representation from full channel matrices to compact parameter sets including spatial correlation matrices, angular spread parameters, and dominant eigenvectors. This parameter transformation reduces feedback dimensionality while preserving essential channel characteristics for accurate CSI computation.
2Quantity of substance
If partial precoding of CSI-RS is implemented, then CSI-RS resources are saved, but device complexity for processing partial precoding increases
Solution Approach 1:
The system dynamically adjusts the precoding configuration for CSI-RS based on channel conditions and resource availability. The eNodeB can flexibly select between full precoding and partial precoding modes, and adjust the number of CSI-RS ports and precoding matrix indicators (PMI) based on current system state, optimizing resource usage while managing complexity.
Solution Approach 2:
The patent introduces an intermediate precoding stage where a first precoding matrix is applied to CSI-RS ports, followed by a second precoding matrix for final transmission. This intermediate precoding structure allows the system to reduce CSI-RS resources by selecting fewer ports for the first precoding while maintaining overall signal quality through the combination of both precoding stages.
3Productivity
If a 2D antenna array is used for FD-MIMO, then data rate and coverage are improved, but feedback requirements become excessive
Solution Approach 1:
The system transitions from one-dimensional antenna arrays to two-dimensional antenna arrays, adding the elevation dimension to the traditional horizontal dimension. This 2D configuration enables spatial multiplexing in both dimensions, increasing data rate capacity. Simultaneously, the patent reduces feedback overhead by exploiting the structured nature of 2D arrays through separable precoding and reduced-rank CSI feedback mechanisms.
Solution Approach 2:
The 2D antenna array is segmented into separate horizontal and vertical sub-arrays, each with its own precoding matrix. This segmentation allows independent optimization of each dimension and reduces the overall feedback requirement compared to treating the 2D array as a single large matrix, since the precoding can be factorized into smaller component matrices.
Data Source
Figure 1a~1c
Figure 2
Figure 3~4
AI summary
The present disclosure relates to a pre-5th-Generation (5G) or 5G communication system to be provided for supporting higher data rates Beyond 4th-Generation (4G) communication system such as Long Term Evolution (LTE). User equipment (UE) is configured to receive two types of channel state information (CSI) reference signals (CSI-RS), the first containing more antenna ports and measured less often than the second. A first CSI report is based on first type CSI-RS and a second CSI report on the second type CSI-RS. The second type CSI-RS is at least partially precoded in response to the first CSI report, is precoded for the specific UE, and is transmitted more often than the first type CSI-RS, while the first type CSI-RS is not precoded. The first CSI report contains a first precoding matrix indicator (PMI) parameter and the second CSI report contains only a second PMI parameter of a two-PMI codebook, where the first PMI parameter is a long-term and wideband PMI. The first PMI is derived from measuring the first type CSI-RS in a plurality of subframes and resource blocks.