5G NR Data Transmission with Adaptive Modes for Training Sets
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data transmission mechanisms in 5G NR systems fail to meet the high-rate and high-error-tolerant requirements of neural network training sets, leading to bottlenecks in data scheduling and inefficient transmission of large data sets.
Innovation Solution
A data transmission method utilizing first verification, retransmission, and RLC modes, activated by MAC CE, DCI, or UCI, to ensure high-rate and high-error-tolerant data transmission, including non-verification, non-detection, and non-feedback mechanisms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional data transmission mechanisms are used, then data transmission can be performed with verification and retransmission, but the transmission rate is limited and error tolerance is low, causing long transmission time for large data sets
Solution Approach 1:
The patent dynamically adjusts the verification mechanism based on data type. For first type data (neural network training sets), the system disables verification to maximize transmission rate, while maintaining verification for other data types. This dynamic adaptation resolves the contradiction by making the system flexible enough to prioritize speed for error-tolerant applications while preserving reliability for critical data.
Solution Approach 2:
The patent changes the verification parameter from 'enabled' to 'disabled' specifically for first type data transmission. By modifying this key parameter based on data type classification, the system achieves high-rate transmission for training sets while maintaining appropriate error handling for other data, thus resolving the contradiction between transmission rate and reliability.
2Reliability
If verification and retransmission mechanisms are enabled, then data transmission reliability is improved, but transmission overhead increases and transmission speed decreases
Solution Approach 1:
The patent applies different quality levels of verification to different data types. For first type data (training sets), verification is completely disabled, achieving maximum transmission speed. For other data types, standard verification is maintained. This local differentiation resolves the contradiction by applying appropriate reliability measures only where needed, minimizing overall transmission time.
Solution Approach 2:
The patent segments data into different types (first type and other types) and applies different transmission mechanisms to each segment. This segmentation allows the system to optimize transmission parameters for each category, disabling verification for error-tolerant training data while maintaining it for critical data, thus reducing overall transmission time without compromising essential reliability.
3Adaptability or versatility
If traditional scheduling mechanisms are used, then data transmission can be managed, but bottlenecks occur for large data set transmission and error tolerance is insufficient
Solution Approach 1:
The patent introduces a new parameter - data type classification - that changes how scheduling decisions are made. By categorizing data as first type (training sets) or other types, the system adapts its error tolerance and verification parameters accordingly. This parameter change enables high error tolerance for training data without requiring complex scheduling mechanisms, as the classification itself drives the transmission strategy.
Solution Approach 2:
The patent makes the scheduling mechanism dynamic by adjusting verification and error handling behavior based on real-time data type identification. For first type data, the system dynamically disables verification and adopts a send-and-forget approach, achieving high error tolerance. For other data, traditional scheduling is maintained. This dynamic adaptation resolves the contradiction by simplifying scheduling for specific data types while maintaining versatility across different data categories.
Data Source
AI summary
Embodiments of the present application provide a data transmission method, an apparatus, a device and a storage medium, where in the method, a new transmission manner is provided for the transmission of high-rate and high-error-tolerant first type of data, and the transmission manner includes at least one of a first verification mode, a first retransmission mode, and a first RLC mode, by using the new transmission mode to transmit the first type data, thus a transmission rate is ensured.