Wireless Device Energy Estimation Using ML and Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing wireless devices face challenges in accurately estimating the energy required for data communication, particularly in IoT devices, as traditional power analyzers are bulky and expensive, and machine learning models struggle to adapt to varying environments without pre-training for all possible scenarios.
Innovation Solution
A wireless device equipped with memory circuitry, processor circuitry, and a wireless interface that obtains cellular and network parameters to determine energy cost parameters, allowing for continuous adaptation and refinement of energy estimates without the need for dedicated power analyzers, using machine learning models like feed forward neural networks and support vector machines.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If power analysers are used to measure energy required for data communication, then measurement precision is improved, but device complexity and cost increase due to bulky hardware requirements
Solution Approach 1:
The patent replaces physical power analyser hardware with a machine learning-based software model that runs on the wireless device itself. The energy parameter is calculated using ML algorithms that process cellular parameters (signal strength, noise, interference) and network parameters (data size, transmission time, retransmissions) to estimate energy consumption without requiring external measurement equipment.
Solution Approach 2:
The wireless device performs its own energy measurement function by using its built-in processor to execute machine learning models. The device self-determines energy parameters needed for transmission decisions without relying on separate measurement instruments, thereby eliminating the need for bulky power analysers while maintaining measurement capability.
2Adaptability or versatility
If machine learning models are pre-trained for all possible environments, then adaptability is improved, but device complexity and training requirements increase significantly
Solution Approach 1:
The patent implements dynamic adaptation where the machine learning model is initially trained on general data and then continuously fine-tuned using real-world operational data from the specific environment. The model adapts its parameters over time based on actual transmission outcomes, allowing it to adjust to local conditions without requiring exhaustive pre-training for all possible scenarios.
Solution Approach 2:
The patent performs preliminary training with general data to establish a baseline model, then uses ongoing operational data to progressively refine the model for specific environments. This staged approach allows the system to start functioning immediately with reasonable accuracy while continuously improving adaptability to local conditions without requiring complete pre-training.
3Speed
If energy parameters are determined using only cellular parameters before transmission, then decision speed is improved, but measurement precision deteriorates due to lack of actual transmission feedback
Solution Approach 1:
The patent implements a feedback mechanism where the energy parameter determined before transmission is compared with the actual energy consumption measured after transmission. The difference between estimated and actual values is used to update and refine the machine learning model, improving future energy estimates. This closed-loop feedback system maintains fast decision-making while progressively enhancing accuracy.
Data Source
AI summary
A wireless device includes a memory circuit, processor circuitry, and a wireless interface, and is configured to obtain a cellular parameter indicative of cellular channel quality and determine, based on the cellular parameter, a first energy cost parameter indicative of energy required for performing a communication of data. The wireless device transmits the data. The wireless device obtains a network parameter indicative of a network condition in which the transmission of the data occurred. The wireless device determines, based on the cellular parameter and the network parameter, a second energy cost parameter. The wireless device determines, based on the second energy cost parameter, an update parameter. The wireless device, upon the update parameter meeting a criterion, update the first transmission cost parameter based on the second transmission cost parameter and the cellular parameter.