Proactive Uplink Grant Scheduling via ML Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In grant-based wireless communication systems, the request-grant process for uplink data transmission leads to significant network access latency, negatively impacting user experience due to back-and-forth scheduling requests and inefficient resource utilization.
Innovation Solution
Implementing a machine learning (ML) engine to predict uplink data requirements and feed proactive grants to the layer 2 scheduler, using an on-chip ML engine to analyze downlink data patterns and provide predictive insights for proactive uplink grant scheduling, thereby reducing latency and improving throughput.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a request-grant process is used for uplink data transmission in grant-based systems, then resource allocation control is improved, but network access latency increases significantly
Solution Approach 1:
The system performs preliminary actions by proactively scheduling uplink grants before the UE actually needs to transmit data. The gNodeB predicts future uplink data requirements and schedules grants in advance, eliminating the need for reactive request-grant cycles and significantly reducing access latency while maintaining reliable resource allocation control.
2Reliability
If a request-grant process is used for uplink data transmission, then scheduling control is improved, but user experience deteriorates due to back-and-forth proceedings
Solution Approach 1:
The system schedules uplink grants proactively in advance of actual data transmission needs, eliminating the frustrating back-and-forth request-grant cycle. Users experience faster data transmission without compromising scheduling control, as the network predicts and prepares grants before the user initiates transmission.
Solution Approach 2:
The system uses feedback mechanisms to continuously monitor UE data patterns and adjust proactive grant scheduling accordingly. This feedback loop ensures that scheduling control remains accurate while improving user experience by anticipating actual transmission needs based on historical data patterns.
3Device complexity
If reactive scheduling is used, then scheduling simplicity is maintained, but resource utilization efficiency decreases
Solution Approach 1:
The system transitions from reactive to proactive scheduling by predicting future uplink data requirements and scheduling grants in advance. This preliminary action improves resource utilization efficiency by preparing grants before data arrives, while the complexity increase is managed through machine learning models that automate the prediction process.
Solution Approach 2:
The system employs machine learning models that enable the network to self-predict and self-schedule uplink grants autonomously. This self-service capability improves resource utilization efficiency by eliminating reactive scheduling delays, while the ML models handle the complexity of prediction and optimization automatically.
Data Source
AI summary
In a grant based system, a user equipment (UE) sends data in an uplink in a request-grant process. The UE first sending a scheduling request, a gNodeB processing the request and scheduling a grant sometime in future, then UE then either sending data if the grant is sufficient or requesting for another grant with more capacity to accommodate data sending. Such a proceeding could cause serious latency in network access. Described in the present patent disclosure are embodiments to reduce the access time by giving proactive grants through inspecting downlink (DL) data sent to the UE or uplink data being transmitted from the UE. The uplink data may be predictive since it maybe in lieu of requirement for sending a TCP acknowledgement for the DL TCP data scheduled earlier. For voice calls, a ML system for system may be deployed to predict when proactive UL grants may be given.