Wireless Communication With AI-Based SL-CSI Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional communication systems face inefficiencies due to the occupation of wireless resources by bit data in SL-CSI reports, leading to a shortage of resources available for user data.
Innovation Solution
Implementing AI-based encoding and decoding of channel state information using learned encoding and decoding models in wireless communication systems, where a first terminal device encodes and a second terminal device decodes the channel state information, reducing the data volume of SL-CSI reports.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional encoding methods are used for SL-CSI reports, then the channel state information can be transmitted accurately, but the data volume occupies excessive wireless resources
Solution Approach 1:
The patent transforms the channel state information from conventional bit representation to a compressed parameter representation using AI encoding models. The encoding model learns the statistical characteristics of channel state data and represents it with fewer parameters, thereby reducing the data volume while maintaining the essential information for accurate channel state reconstruction at the receiver side.
Solution Approach 2:
The patent extracts the essential features of channel state information using AI-based encoding models, separating the critical channel characteristics from redundant data. The encoding model identifies and extracts only the most important parameters needed for channel state representation, transmitting these extracted features instead of the complete original data set.
2Reliability
If more wireless resources are allocated for SL-CSI reports, then the channel state information can be transmitted with higher reliability, but the resources available for user data decrease
Solution Approach 1:
The patent changes the parameter representation of channel state information from detailed bit-level data to compressed feature parameters. This parameter transformation maintains the reliability of channel state information transmission by preserving the essential channel characteristics in a more efficient format, thereby reducing the resource allocation needed for SL-CSI reports and increasing resources available for user data transmission.
Solution Approach 2:
The patent segments the channel state information transmission into two parts: compressed feature parameters transmitted over the air interface, and model reconstruction algorithms executed locally at the receiver. This segmentation reduces the amount of data that needs to be transmitted reliably over the wireless channel, thereby reducing the resources dedicated to SL-CSI reports while maintaining transmission reliability.
3Quantity of substance
If AI encoding models are implemented, then the data volume of SL-CSI reports is reduced, but the complexity of encoding and decoding processes increases
Solution Approach 1:
The patent performs preliminary training of AI encoding and decoding models during an offline phase using historical channel state data. The models are pre-trained to learn the statistical characteristics and optimal compression representations of channel state information. This preliminary action allows the models to be deployed with minimal real-time processing complexity, as the heavy learning work has already been completed during the training phase.
Solution Approach 2:
The patent uses pre-trained AI models that can be copied and deployed across multiple terminal devices. Once a model is trained and validated, it can be replicated and distributed to multiple devices, reducing the complexity burden on individual devices. The models serve as reusable components that can be instantiated multiple times without retraining, thereby managing complexity at the system level rather than at each device level.
Data Source
AI summary
A communication system includes a first UE and a second UE communicating with the first UE by SL communication, the first UE transmits, to the second UE, encoded channel state information, which is channel state information encoded by using an encoding model learned for the channel state information, which is information for reporting the channel state of the SL communication, and the second UE decodes the encoded channel state information by using a decoding model learned for channel state information.