Radio Resource Allocation Using Machine-Trained Functions

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current radio communication methods face inefficiencies in resource allocation and synchronization between disjoint radio nodes using the same channel, leading to increased control signaling overhead and reduced spectrum efficiency.

Innovation Solution

A method involving the collection and analysis of radio traffic data sets to train machine-trainable functions, allowing for the determination of utilization schemes and allocation of radio resources without a central granting procedure, enabling synchronized resource allocation and reduced overhead.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If a central granting procedure is used for resource allocation, then resource allocation control is improved, but control signaling overhead increases

Engineering Contradiction:
Improveresource allocation controlVSAvoidcontrol signaling overhead
Core Design Contradiction:
ReliabilityVSLoss of information

Solution Approach 1:

The patent enables radio nodes to autonomously determine resource utilization schemes using machine-trained functions without requiring central granting procedures. Each node independently analyzes traffic data sets and determines its own resource allocation, eliminating the need for extensive control signaling between central controller and nodes.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent introduces machine-trained functions as intermediaries that process traffic data and generate resource utilization schemes. These trained functions act as mediators between raw traffic data and resource allocation decisions, enabling distributed nodes to make informed decisions without direct central control signaling.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Productivity

If disjoint radio nodes use the same radio channel independently, then spectrum efficiency is improved, but synchronization between nodes deteriorates

Engineering Contradiction:
Improvespectrum efficiencyVSAvoidsynchronization between nodes
Core Design Contradiction:
ProductivityVSStability of the object's composition

Solution Approach 1:

The patent trains machine models at multiple radio nodes using identical or equivalent traffic data sets and training configurations, ensuring that all nodes operate at the same decision-making level. This equipotential approach ensures synchronized resource utilization schemes across disjoint nodes while maintaining independent spectrum usage.

Inventive Principle:
Principle #12Equipotentiality

Solution Approach 2:

The patent uses rectification indicators to detect and correct deviations in machine model parameters between nodes. When synchronization drift is detected, the system adjusts parameters such as training configuration, data collection settings, or model architecture to restore synchronization while maintaining efficient spectrum usage.

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If machine learning models are trained with high accuracy requirements, then resource utilization scheme determination is improved, but training time and complexity increase

Engineering Contradiction:
Improveresource utilization determination accuracyVSAvoidtraining time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent implements iterative training where machine models are trained incrementally using available traffic data sets. Rather than requiring complete and perfect training data before deployment, the system uses partial training results and continuously refines models as more data becomes available, reducing initial training time while maintaining adequate accuracy.

Inventive Principle:
Principle #16Partial or excessive action

Solution Approach 2:

The patent employs feedback mechanisms where rectification indicators monitor the synchronization and performance of trained models across nodes. This feedback enables the system to identify when retraining or parameter adjustment is needed, optimizing the balance between model accuracy and training time by only retraining when necessary.

Inventive Principle:
Principle #23Feedback

Data Source

PatentUS20240049268A1Methods and apparatuses for radio communication
Publication Date: 2024.02.08 ROBERT BOSCH GMBH
  • US20240049268A1 patent drawing
  • US20240049268A1 patent drawing
  • US20240049268A1 patent drawing

AI summary

A method for operating a first apparatus. The method includes: collecting at least one first radio traffic data set that is associated with a utilization of at least one radio channel; training a machine-trainable function based on the collected at least one first radio traffic data set; collecting at least one second radio traffic data set, which is associated with the utilization of the at least one radio channel, wherein the second traffic data set is different from the first data set; determining, using the machine-trained function, at least one utilization scheme associated with the at least one radio channel based on the at least one second radio traffic data set; allocating at least one radio resource on the at least one radio channel according to the determined utilization scheme; and transmitting data via the at least one allocated radio resource of the at least one radio channel.