AI-Assisted Data Block Transmission for XR Communication Load Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Wireless communication systems face challenges in supporting complex Computer Graphics (CG) and Extended Reality (XR) services due to the high volume of data required, necessitating enhancements and improved performance evaluation.
Innovation Solution
Implementing artificial intelligence (AI) prediction models to selectively transmit data blocks, generating predicted and differential data blocks to reduce data transmission while maintaining communication quality, and utilizing distributed training methods for enhanced model efficiency and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If AI prediction models are used to predict and transmit only differential data blocks, then data transmission volume is reduced, but communication quality may deteriorate
Solution Approach 1:
The patent segments data transmission into two parts: predicted data blocks generated by AI models and differential data blocks containing only changes. This segmentation allows the system to transmit minimal data while maintaining quality, as the receiver can reconstruct full data blocks by combining predictions with differentials.
Solution Approach 2:
The patent implements feedback mechanisms where the receiver provides information about prediction accuracy and communication quality to the transmitter. This feedback loop enables dynamic adjustment of the prediction model parameters and transmission strategy, ensuring communication quality is maintained while minimizing data transmission volume.
2Measurement precision
If prediction models are trained using traditional methods, then training accuracy is achieved, but training efficiency is low
Solution Approach 1:
The patent applies preliminary action by pre-training prediction models using historical data and pre-computing prediction parameters. This allows the model to make accurate predictions during actual communication without requiring real-time training, thereby improving training efficiency while maintaining accuracy.
Solution Approach 2:
The patent uses partial action by training models selectively on relevant data patterns rather than attempting to learn everything. The model focuses on capturing essential characteristics of data streams, achieving sufficient accuracy with reduced training data and computational resources, thus improving training efficiency.
3Measurement precision
If data transmission and model training are performed sequentially, then training accuracy is improved, but time consumption increases
Solution Approach 1:
The patent merges data transmission and model training operations by performing training in the background during transmission time. The receiver can train updated prediction models while receiving data streams, and the transmitter can send both data and training information simultaneously. This merging eliminates sequential execution and reduces overall time consumption while maintaining training accuracy.
Solution Approach 2:
The patent ensures continuity of useful action by overlapping transmission and training operations. Instead of stopping transmission to complete training, the system continuously performs both functions simultaneously, ensuring that data transmission never halts while training progresses in parallel, thus reducing time loss.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Disclosed are communication methods for artificial intelligence assistance, and related devices. An exemplary communication method comprises: receiving a data stream including at least a first data block and a second data block; sending the first data block to a user equipment (UE); based on the first data block, using a prediction model to generate a second predicted data block; based on a difference between the second data block and the second predicted data block, generating a second differential data block; and sending the second differential data block to the UE.