Learning-Based UL Scheduler Using Unified Status Reporting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional UL scheduling related reporting procedures in wireless communications systems, such as BSR, DSR, and PHR, operate independently and do not provide simultaneous or complete information for AI/ML-based scheduler training, leading to inaccurate predictions and decreased performance.
Innovation Solution
Introduce a new type of UL status information comprising buffer and delay status, including remaining delay of data units and UL power information, which is collected and reported by the UE to the gNB for training an AI/ML-based scheduler.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional UL scheduling reporting procedures (BSR, DSR, PHR) are used independently, then device complexity is reduced and ease of operation is maintained, but measurement precision of UL status information deteriorates and reliability of AI/ML-based scheduler training decreases
Solution Approach 1:
The patent combines multiple independent reporting procedures (BSR, DSR, PHR) into a unified UL status information reporting mechanism. The network entity configures a single reporting procedure that collects buffer status, delay status, and power headroom information together, providing comprehensive UL status for AI/ML-based scheduler training while maintaining manageable device complexity through standardized configuration.
2Reliability
If comprehensive UL status information is collected for AI/ML-based scheduler training, then reliability of scheduling decisions improves, but loss of information in conventional independent procedures is reduced
Solution Approach 1:
The unified UL status information reporting mechanism serves multiple functions simultaneously: it provides buffer status information for resource allocation, delay status information for QoS management, and power headroom information for power control. This multi-functional approach ensures comprehensive information availability for AI/ML-based scheduler training without requiring separate independent procedures.
3Productivity
If independent reporting procedures are used, then ease of operation is maintained and device complexity is low, but productivity of scheduler training decreases due to inaccurate predictions
Solution Approach 1:
The network entity performs preliminary configuration of the unified UL status information reporting mechanism before AI/ML-based scheduler training begins. By pre-configuring the reporting parameters, periodicity, and information types, the system ensures that comprehensive UL status information is readily available for efficient scheduler training without requiring complex real-time adjustments during operation.
Data Source
AI summary
Various aspects of the present disclosure relate to techniques for a learning-based scheduler. A user equipment is configured to receive a configuration for capturing uplink status information, the uplink status information comprising buffer status information, delay status information, or a combination thereof. The user equipment is configured to capture the uplink status information according to the configuration. The user equipment is configured to report the uplink status information to a learning-based scheduler.