IoT Device State Change Prediction for Power Conservation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Battery-operated IoT devices face high power consumption due to Wi-Fi connectivity, which was designed to optimize bandwidth and range rather than power efficiency, necessitating a method to predict state changes to conserve power by activating or maintaining a sleep state.
Innovation Solution
A method using a trained machine learning model, specifically a temporal convolutional network, to predict state changes in IoT devices by forming input vectors from current state values and temporal features, enabling the optimization of wake-up and sleep intervals to reduce power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If Wi-Fi connectivity is added to battery-operated IoT devices to enable network communication, then network connectivity and data transmission capability are improved, but power consumption increases significantly
Solution Approach 1:
The system dynamically adjusts the operational state of the Wi-Fi transceiver based on predicted device states. The transceiver alternates between active and sleep modes according to real-time predictions, optimizing the balance between network connectivity availability and power consumption. This dynamic adaptation allows the device to maintain network functionality when needed while conserving energy during periods of predicted inactivity.
Solution Approach 2:
The system performs preliminary prediction of device state changes before actually transitioning between states. By using machine learning models to predict future device states based on historical data and current context, the system proactively activates or deactivates the Wi-Fi transceiver in advance, preventing unnecessary power consumption while ensuring connectivity is available when state changes are anticipated.
2Loss of energy
If the device stays in the lowest power mode as much as possible to conserve battery, then power consumption is reduced, but the ability to respond to network events and maintain connectivity is worsened
Solution Approach 1:
The system implements a feedback loop where machine learning models continuously predict device states based on historical data, current sensor readings, and contextual information. These predictions feed back into the Wi-Fi transceiver control decisions, creating a closed-loop system that adapts to actual device behavior patterns. This feedback mechanism ensures the device maintains reliability by being connected when predictions indicate state changes are likely, while conserving power when changes are unlikely.
3Speed
If the device frequently activates to check for state changes to maintain responsiveness, then connectivity responsiveness is improved, but power consumption increases
Solution Approach 1:
Instead of continuous monitoring or frequent periodic checks, the system uses event-driven periodic activation based on ML predictions. The Wi-Fi transceiver activates only during predicted state change events rather than following a fixed periodic schedule or running continuously. This approach maintains response speed by being available when needed while dramatically reducing power consumption by staying dormant during predicted stable states.
Data Source
AI summary
A method (700) for predicting, for each device included in a set of devices, whether the device will change state at a particular future point in time. The method includes, for a first device within the set of devices, obtaining a first state value indicating the current state of the first device. The method also includes, for a second device within the set of devices, obtaining a second state value indicating the current state of the second device. The method also includes forming an input vector, the input vector comprising the first state value, the second state value, and a temporal feature (e.g., a set of one or more time values indicating the current time). The method also includes inputting the input vector into a trained machine learning (ML) model. The method also includes, after inputting the input vector into the trained ML model, obtaining a probability vector into the trained ML model, the probability vector comprising, for each device included in the set of devices, a state change prediction value indicating a likelihood that the device will change state at the particular future point in time.