Wireless Device Prediction Information Provisioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current AI and ML solutions for energy efficiency in wireless communication systems lack flexibility and accuracy in predicting radio signal measurements, leading to suboptimal energy savings and potential QoS issues due to unawareness of UE-specific capabilities and conditions.
Innovation Solution
A method that provisions wireless devices with prediction information using denoising autoencoders and candidate noising patterns, tailored to UE capability information, allowing the device to assess and decide on radio signal measurement predictions, thereby improving energy efficiency and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by moving object
If AI models are deployed at the UE to predict radio signal measurements, then UE power consumption is reduced, but prediction accuracy may deteriorate due to limited UE computational resources and lack of network-side processing capabilities
Solution Approach 1:
The patent segments the AI prediction system into two parts: a training phase performed at the network side (base station) where complex model training occurs with full data access, and an inference phase executed at the UE with pre-trained lightweight models. This segmentation allows each component to operate in its optimal environment, resolving the contradiction between energy savings and accuracy.
Solution Approach 2:
The network performs preliminary actions by training the AI models offline using extensive measurement data and computational resources, then delivers the trained models to the UE. This preliminary training ensures that the UE receives optimized models that achieve high prediction accuracy without requiring the UE to perform computationally intensive training operations, thus saving power while maintaining accuracy.
2Ease of operation
If generic AI models are used for prediction, then model deployment is simplified, but prediction accuracy deteriorates due to lack of UE-specific optimization
Solution Approach 1:
The patent changes the parameter of model personalization by incorporating UE-specific parameters (device type, antenna configuration, mobility characteristics) into the model training process. The network trains separate models or adapts existing models based on these parameters, ensuring each UE receives an optimized model that achieves high accuracy while maintaining a standardized deployment process through automated model selection and delivery.
3Measurement precision
If comprehensive UE capability information is collected, then prediction accuracy is improved through UE-specific optimization, but device complexity and signaling overhead increase
Solution Approach 1:
The patent makes the UE capability information collection mechanism universal by reusing existing 3GPP standardized capability reporting frameworks that UEs already employ for other purposes (e.g., network configuration, resource allocation). By leveraging these existing multi-functional reporting mechanisms, the system obtains necessary UE-specific information without adding dedicated signaling procedures, thus avoiding increased complexity and overhead.
Data Source
AI summary
The present disclosure relates to a computer-implemented method (100), performed by a first network node, for provisioning a wireless device with prediction information for allowing the wireless device to predict a radio signal measurement between the wireless device and a base station, wherein the prediction information comprises a denoising autoencoder and at least one candidate noising pattern. The method comprises: receiving (110), from the wireless device, wireless device capability information: obtaining (120), based on the wireless device capability information, the denoising autoencoder and/or the at least one candidate noising pattern for predicting a radio signal measurement; and transmitting (130) an indication of the prediction information to the wireless device. The present disclosure also relates to a first network node, a wireless device and a computer program.