Extended State Observer Based Parameter Identification of the Hovercraft System

In this paper, a problem of online identification of the dynamic parameters of the hovercraft system is considered. A novel approach is presented, which employs an extended state observer and takes advantage of the total disturbance estimate as a basis of gradient identification method. This enables simultaneous estimation of both the state of the plant and the values of the system parameters. Although the identification algorithm is synthesized under the assumption that the velocities of the vehicle are measurable, the experimental results show that these can also be estimated and only availability of the positional measurements are required in practical scenarios. The results of simulations and experiments are presented as a proof-of-concept of the extended state observer based identification.