Multi-Antenna Channel Estimation via Temporal Replicas
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional multi-antenna systems face challenges in accurately estimating transmission channels due to increased noise levels and reduced spectral efficiency, particularly in high-speed communication systems, where differential modulation is not effective and pilot symbols are necessary for channel estimation.
Innovation Solution
The method involves determining a basic learning sequence with replicas that are temporally shifted, allowing for the calculation of a matrix A constructed from learning sequences and the Fourier matrix, enabling the estimation of impulse responses by multiplying demodulated pilot symbols with the pseudo inverse of this matrix, thereby improving channel estimation accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If pilot symbols are inserted in the time-frequency frame for channel estimation, then channel estimation accuracy is improved, but spectral efficiency is reduced
Solution Approach 1:
The patent segments the pilot symbols into multiple learning sequences, each transmitted by different transmitting antennas. Each learning sequence is further divided into multiple replicas that are temporally shifted relative to each other. This segmentation allows the receiver to obtain multiple channel estimates from different replicas and average them, improving estimation accuracy while using a controlled number of pilot symbols, thus managing the trade-off with spectral efficiency.
Solution Approach 2:
The patent introduces a temporal dimension by creating multiple replicas of learning sequences with different time shifts. Instead of only using pilot symbols in the frequency domain, the system exploits the time domain by transmitting replicas at different time positions. This additional temporal dimension provides more information for channel estimation without proportionally increasing the number of pilot symbols in the frequency domain, thereby improving accuracy while maintaining spectral efficiency.
2Productivity
If multiple transmitting antennas are used to increase capacity, then system capacity is improved, but the complexity of channel estimation increases
Solution Approach 1:
The patent segments the channel estimation task by assigning different learning sequences to different transmitting antennas. Each antenna transmits a learning sequence with multiple temporally shifted replicas. The receiver processes each learning sequence independently to obtain channel estimates for corresponding channels. This segmentation divides the complex MIMO channel estimation problem into multiple simpler single-antenna estimation tasks, reducing overall computational complexity while maintaining the ability to estimate all channels in the multi-antenna system.
Solution Approach 2:
The patent uses copying by creating multiple replicas of the same basic learning sequence pattern, where each replica is a temporally shifted version of the others. These copies are transmitted by different antennas or at different times. The receiver uses these copies to obtain multiple observations of the channel impulse response and averages them to improve estimation accuracy. This copying approach provides redundant information that simplifies the estimation process and reduces complexity compared to using entirely different pilot patterns for each antenna.
3Ease of manufacture
If differential modulation is used in multi-antenna systems, then implementation simplicity is improved, but noise level increases and performance degrades
Solution Approach 1:
The patent uses copying by creating multiple replicas of the same learning sequence pattern with temporal shifts. These copies provide redundant channel state information that enables coherent detection. By having multiple copies of the same pilot pattern transmitted by different antennas or at different times, the system can average the channel estimates and improve signal reliability without using differential modulation, thus avoiding the 3dB performance degradation while maintaining implementation simplicity through coherent detection with channel estimation.
Data Source
Figure 1~4
Figure 5~6
Figure 7~8
AI summary
The invention relates to a method and a transmission device adapted for a system comprising Nt transmission antennas TXi and Nr reception antennas RXj. the outputted symbols are inserted in a time frequency frame containing a training sequence. Each training sequence is built from replicas of a base training sequence. At least one of said training sequence includes at least two replicas one of which at least is out of phase.