Scheduling Sensor Duty Cycling Based on Event Detection Using Bi-Directional Long Short-Term Memory and Reinforcement...

Area of Science:

• Computer Science
• Artificial Intelligence
• Internet of Things

Background:

• Smart homes facilitate human activity detection using Deep Learning and sensor data.
• Detecting interdependent human activities and managing sensor energy depletion are key challenges.

Purpose of the Study:

• To propose an Energy and Event Aware-Sensor Duty Cycling scheme.
• To address challenges in human activity detection and sensor energy management in smart homes.

Main Methods:

• Utilized Bi-Directional Long-Short Term Memory (Bi-LSTM) for predicting expected events.
• Employed Jaccard Similarity Index for localizing unexpected events with Monitor and Hibernate Sensors.
• Optimized event tracking using Q-Learning algorithm.

Main Results:

• Achieved a significant improvement in activity detection accuracy from 94.12% to 96.12%.
• Demonstrated substantial reduction in sensor energy consumption through effective Monitor Sensor rotation.
• Extended the overall smart home network lifetime.

Conclusions:

• The proposed scheme effectively balances human activity detection accuracy and sensor energy efficiency.
• The integration of Bi-LSTM, Jaccard Similarity, and Q-Learning offers a robust solution for smart home monitoring.
• This approach contributes to more sustainable and reliable smart home systems.