In national security and public safety (NSPS) scenarios, the concept of device-to-device (D2D) clustering allows user equipment (UEs) to dynamically form clusters and thereby allows for local communication with partial or no cellular network assistance. We propose and evaluate a clustering approach to solve this problem in this thesis report. One of the key components of clustering is the selection of so called cluster head (CH) nodes that are responsible for the formation of clusters and act as a synchronization and radio resource management information source. In this thesis work we propose a weighted CH selection algorithm that takes into account UE capability, mobility and other information and aims at balancing between energy efficiency, discovery rate and cluster formation time. Numerical results show that the clustering approach consumes more energy but it can achieve a much higher discovery rate and communication rate for the system. Simulation results indicate that the weighted clustering approach is a viable alternative in NSPS situations.