Abstract This paper presents a novel range-free localization algorithm that has been originally designed to help in the characterization of human behavior by turning connectivity traces of mobile nodes into localization information. It is based on an error function that uses connectivity between nodes and information about their maximum velocity and that is solved by iterative minimization using unconstrained optimization. The resultant trajectories are evaluated with respect to the localization solution space (LSS), a multi-dimensional space consisting of all solutions that satisfy completely the conditions of the problem. The algorithm is evaluated through extensive simulations in scenarios with different levels of connectivity and under regular and irregular conditions in the communications. The complementarity with other localization algorithms is presented by using their results as initialization stage for this algorithm. Finally, it is applied to a real database that provides information about Bluetooth-based connectivity and motion of people in an office scenario, rendering satisfactory results and hence, validating the practicability of the proposed algorithm as a framework to obtain localization traces.