Autotuning and system parameters monitoring are crucial aspects of modern motion control algorithms. At-theedge controllers need to detect system changes and perform autotuning autonomously without requiring excessive computational burdens. Two algorithms to estimate the main dynamics of mechanical systems using integral figures of merit are proposed. The algorithms implementations are analyzed in offline and online modes. They have been tested in simulation, using an elastic model as a testbed, and in a hardware-in-the-loop PLCcontrolled system. The results show the effectiveness of the methods compared to a recursive least square method.