-This paper proposes autonomous hybrid system (AHS) with estimator based inverse dynamics controller along with extended kalman filter and artificial neural network based state estimators ensuring best performance and robustness by minimum ISE in controlling non-measurable state variables of autonomous hybrid systems. With the help of experimental setup of benchmark model of AHS (hybrid three-tank system), the detailed performance comparison of these proposed methods was made both qualitatively and quantitatively with that of existing method (AHS with unscented kalman filter and inverse dynamics controller (UKFC)) under different operating conditions such as servo and regulatory operations, process-model parameter uncertainties, initial condition mismatch, and different types of faults in the system, subsequently the results are reported.