Publication

Abstract : Energy consumption is a major challenge in wireless sensor networks, despite improvements in hardware and protocols. Limited battery capacity and energy-intensive transmissions, especially in unpredictable environments, drive the need for continued research into better energy-saving methods. Areas such as human habitats, industrial settings, and forests deploy Internet of Things applications, where obstacles impact communication quality and energy consumption. Current energy-saving strategies in low-power networks select parents based on metrics such as residual energy, hop count, and transmission count, but these methods overlook environmental and interference factors. To reduce the network’s energy consumption, the present work introduces an innovative method for environment-aware power prediction that integrates classification techniques. This approach is designed to enhance routing protocols for low-power and lossy networks by updating their objective functions and enabling dynamic power-level selection during routing. The proposed model is implemented in the Cooja simulator. The performance is compared with standard objective functions. The simulation results show a remarkable 6.92% increase in packet delivery rate (PDR), an 18.11% reduction in delay, and a 35.84% decrease in power consumption in obstacle-prone simulation environments.