We consider a robotic network composed of mobile robots capable of communicating with each other. We study the problem of collectively tracking a number of moving targets while maintaining a certain level of connectivity among the robots, by moving them into appropriate positions. The distances of the robots to each other and to the targets are used to define a communication and target tracking graph, respectively. We formulate the combined global objective as a Semi-Definite Program (SDP) and propose a non-iterative distributed solution consisting of localized SDP's which use information only from nearby neighboring robots. Numerical simulations illustrate the performance of the algorithm with respect to the centralized solution.