5G NR Data Transmission Modes for High-Rate Error-Tolerant Datasets
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 datasets.
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 for neural network training sets, then data can be transmitted with standard verification and retransmission protocols, but transmission rate is limited and error tolerance is insufficient for high-rate services
Solution Approach 1:
The patent implements dynamic switching between different RLC modes (Acknowledged Mode, Unacknowledged Mode, Transparent Mode) based on service requirements. For high-rate services like neural network training, the system can dynamically select UM or TM modes to achieve higher transmission rates, while for reliability-critical services, AM mode provides robust error correction. This dynamic adaptation resolves the contradiction between transmission rate and reliability.
Solution Approach 2:
The patent changes key transmission parameters including verification mode (CRC checking), retransmission behavior, and error handling mechanisms based on service type. For high-rate services, the system modifies parameters to reduce verification overhead and enable faster transmission, accepting higher error rates. This parameter adaptation allows the system to optimize for either speed or reliability depending on requirements.
2Productivity
If standard verification modes are used for high-rate data transmission, then data integrity can be ensured, but transmission overhead increases and transmission rate decreases
Solution Approach 1:
The patent applies different verification qualities to different data types and service scenarios. For high-rate services like neural network training, the system uses simplified verification (UM mode with optional CRC) or no verification (TM mode), accepting lower integrity for faster transmission. For critical services, full AM mode verification is applied. This localized quality adjustment resolves the contradiction between verification complexity and transmission rate.
3Reliability
If retransmission mechanisms are enabled for all data transmissions, then data reliability improves, but transmission latency increases and throughput decreases
Solution Approach 1:
The patent dynamically controls retransmission behavior based on service requirements. For high-rate services, the system disables or minimizes retransmission mechanisms (using UM or TM modes), accepting potential data loss for faster transmission. For reliability-critical services, AM mode provides robust retransmission. This dynamic control resolves the contradiction between retransmission reliability and transmission latency.
4Productivity
If traditional scheduling mechanisms are used for neural network training data, then resource allocation follows standard protocols, but data scheduling efficiency is insufficient for large-volume high-rate transmission
Solution Approach 1:
The patent creates a universal RLC framework that can handle both traditional services and new high-rate services like neural network training through a single multi-functional architecture. The RLC layer can adaptively switch between AM, UM, and TM modes, and can handle various data types (eMBB, URLLC, mMTC) with a unified mechanism. This universality improves scheduling efficiency for large-volume data transmission while maintaining compatibility with existing services.
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.