A novel spatio-topological embeddings for efficient & model-free redundant node placement in 6G IoT networks

The development of 6G has accelerated the usage of IoT for data collection. With continuous usage and ubiquitous connectivity, the batteries of the nodes soon deplete and create a coverage hole in the network, imposing connectivity challenges in the 6G IoT. The problem has been dealt with a novel set of spatial features with topological embeddings extraction using the graphical convolutional network (GCN). Further, the deep deterministic policy gradient in the continuous action space trains the agent for the optimal placement of redundant nodes. The complete methodology with spatio-topological features has seen an improvement of up to 13.1% in energy residual and 19.4% in the uniform load distribution compared to the state-of-the-art methods with stable network connectivity. In addition, the analysis was performed under various environmental conditions with varying hole density and sensor density. The proposed scheme has also shown improvement under adverse coverage conditions.