The goal of this project is to design and evaluate efficient heterogeneous resource management by adaptive power control, throughput enhancement and interference management for D2D communication. The originality is to exploit machine learning (ML) techniques to improve the existing state of the art works. Further, by exploiting unmanned aerial vehicle (UAV) for weak signal detection, and devices accurate position evaluation is an important objective. The deployed setup with UAV assisted connectivity is one of the novel contribution of this project. In addition, context aware and reliable D2D multi-hop routing and network connections to ensure high end-to-end throughout and low end-to-end energy consumption and delay is another core objective.