EDL-IoT-Net Method and Experimental Design for Energy, Trust, and Congestion Aware Routing in IoT Wireless Sensor Networks

Energy depletion, unreliable relay behavior, and queue growth interact during multi-hop forwarding in IoT based Wireless Sensor Networks. The preceding research stage established five gaps that prevent current routing methods from treating these conditions within one decision process. We define EDL-IoT-Net as a routing method that combines a Deep Q-Network with a Temporal Convolutional Network and an Echo State Network. Six normalized state variables describe residual energy, energy decay, composite trust, congestion, delay, and packet loss. Causal convolutions encode recent traffic changes, the reservoir state records nonlinear dynamics, and the DQN selects an eligible next-hop node from the fused representation. A separate congestion head permits class-wise predictive testing without replacing the routing objective. The experimental protocol fixes a 60-node network, an 8,000-sample state dataset, an 80/20 stratified partition, five-fold cross-validation, conventional and learning-based comparators, and component ablation. The method links each knowledge gap to a measurable variable, a model operation, and a defined test while reserving achieved performance for the subsequent experimental stage.