AS2ICRL: Energy Management Model using Inverse Constrained Reinforcement Learning based Adaptive Sleep Scheduling in Wireless Sensor Networks

Sensor nodes have very limited battery life energy management becomes an important issue in WSN. Most of the older energy-saving techniques find it hard to achieve a good trade-off between conserving energy, maintaining network connectivity, and maximizing efficiency in the delivery of data. We propose adaptive sleep schedule-based inverse constrained reinforcement learning (AS2ICRL), which optimizes the sleep scheduling of WSNs intelligently through inverse constrained reinforcement learning (ICRL). The model dynamically adjusts node activity according to the environment, varying network conditions, and traffic; thus, the methodology has attained 28.7% energy savings and 35.4% improvement in network lifetime compared with earlier proposals. AS2ICRL guarantees communication that is efficient and trustworthy by achieving 94.2% classification accuracy, F1-score of 92.1% with a robustness index gain of 15.3%, while maximizing energy efficiency. Experimental evaluations establish the potential of AS2ICRL to significantly increase network stability, adaptability, and overall performance, which favors its deployment in energy-aware IoT and WSN applications.