ML-Based Measurement Compression for Wireless Reference Signals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current wireless communication systems face challenges in efficiently compressing and transmitting measurements for multiple reference signals (RSs), leading to increased signaling resources and latency due to the high bit requirements for reporting numerous RSRP measurements.
Innovation Solution
The implementation of machine learning (ML) models for encoding and decoding measurements, allowing for the generation of codewords that represent a quantized version of RSRP measurements, thereby reducing the bit requirement and enabling efficient compression and decompression of measurements for multiple RSs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional methods are used to report RSRP measurements for multiple reference signals, then measurement precision is maintained, but signaling resources and bit requirements increase significantly
Solution Approach 1:
The patent changes the parameter representation from raw RSRP values to quantized codewords generated by ML models. This transforms the measurement data into a compressed format that requires significantly fewer bits while maintaining measurement precision through the trained ML model's reconstruction capability.
Solution Approach 2:
The patent creates a compressed representation (codeword) that is a simplified copy of the original measurement data. The ML model learns to map between the original RSRP measurements and the compressed codewords, allowing the system to transmit only the essential information while preserving measurement accuracy.
2Measurement precision
If detailed RSRP measurements for multiple reference signals are transmitted, then measurement accuracy is improved, but transmission latency increases
Solution Approach 1:
The patent extracts only the essential information from the full RSRP measurements through the ML model's quantization process. By removing redundant information and retaining only the critical measurement characteristics, the system reduces the data size for transmission while maintaining accuracy, thereby reducing transmission latency.
Solution Approach 2:
The ML model performs preliminary quantization and compression of measurement data before transmission. This pre-processing step transforms the raw measurements into optimized codewords that require fewer bits and can be transmitted more quickly, reducing overall transmission latency while preserving measurement accuracy.
3Quantity of substance
If machine learning models are used to compress measurements, then signaling resources are reduced, but system complexity increases
Solution Approach 1:
The patent replaces traditional mechanical compression methods (such as fixed-rate quantization or simple encoding schemes) with machine learning-based compression. The ML models automatically learn optimal quantization strategies from training data, providing adaptive compression that reduces signaling resources more effectively than conventional methods.
Solution Approach 2:
The ML models perform self-optimization through training on measurement data, automatically adapting to different signal conditions and compression requirements. This self-learning capability allows the system to achieve efficient compression without requiring complex manual configuration or reconfiguration, managing system complexity through automated adaptation.
Data Source
AI summary
Various aspects of the present disclosure generally relate to wireless communication. In some aspects, a user equipment (UE) may receive multiple reference signals (RSs). The UE may measure the multiple RSs to generate measurements for the multiple RSs. The UE may encode the measurements using a machine learning model to generate codewords representing a quantized version of the measurements. The UE may transmit the codewords. Numerous other aspects are described.