ML-Based Access Class Barring for Wireless Network Congestion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current access class barring (ACB) systems in wireless networks face challenges in dynamically adjusting parameters to manage congestion effectively, particularly during events and emergencies, due to the difficulty in determining optimal thresholds and values for Barring Factor and Barring Time, leading to inefficient capacity utilization and lack of real-time feedback on parameter effectiveness.
Innovation Solution
A machine learning computing system is integrated into the base station to predict barring parameters based on time, traffic, and location data, allowing for dynamic adjustment of Barring Factor and Barring Time in System Information Blocks (SIBs) sent to user equipment, enabling better capacity utilization and reduced congestion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If static ACB parameters are used in traditional systems, then device complexity is reduced, but network capacity utilization deteriorates during varying load conditions
Solution Approach 1:
The patent implements dynamic ACB parameters by training a machine learning model (e.g., neural network, random forest, or gradient boosting) that continuously learns from network load patterns and predicts optimal Barring Factor and Barring Time values in real-time, allowing the system to adapt to varying traffic conditions rather than using fixed static parameters
Solution Approach 2:
The system employs self-service through automated machine learning models that independently analyze network conditions, predict congestion patterns, and adjust ACB parameters without manual intervention, enabling the network to self-optimize capacity utilization while managing complexity through algorithmic autonomy
2Adaptability or versatility
If manual threshold adjustment is used for ACB parameters, then ease of operation is improved, but adaptability to dynamic conditions deteriorates
Solution Approach 1:
The patent replaces manual mechanical adjustment of ACB parameters with automated machine learning systems that process network data, identify congestion patterns, and automatically tune Barring Factor and Barring Time values, substituting human operational complexity with algorithmic adaptability to dynamic network conditions
Solution Approach 2:
The system implements feedback mechanisms where the machine learning model continuously monitors network load, ACB effectiveness, and congestion patterns, using this feedback to iteratively refine parameter predictions and improve adaptability to changing conditions while eliminating manual tuning requirements
3Productivity
If dynamic ACB parameters are implemented, then network capacity utilization is improved, but loss of time for parameter determination worsens
Solution Approach 1:
The patent applies preliminary action by pre-training machine learning models offline using historical network data to learn congestion patterns and parameter relationships before deployment, so that during real-time operation the models can rapidly predict optimal ACB parameters without extensive computation, reducing parameter determination time while maintaining high capacity utilization
Data Source
AI summary
Systems and methods for ML based ACB are provided herein. In an example, a system includes BBU(s), RU(s) communicatively coupled to the BBU(s), and antenna(s) communicatively coupled to the RU(s). Each respective RU of the RU(s) is communicatively coupled to a respective subset of the antenna(s). The BBU(s), the RU(s), and the antenna(s) are configured to implement a base station for wirelessly communicating with UEs in a cell. The system includes a machine learning computing system configured to: receive time data, traffic data, and location data; and determine predicted barring parameter(s) for the base station based on the time data, the traffic data, and the location data. The system is configured to: adjust barring factor(s) and/or barring time(s) in an information message based on the predicted barring parameter(s) for the base station; and send the information message with the adjusted barring factor(s) and/or barring time(s) to UEs in the cell.