Torben Knudsen’s research within control engineering has focused on stochastic systems in the area of system identification, Kalman filtering, sensor fusion, statistics and stochastic processes. In the area of classical time series analysis he has been working with transfer function models as e.g. Box Jenkins models where the focus has been on modeling, parameter estimation and forecasting. For linear state space Markov models in discrete time he has contributed to the subspace methods for estimating the complete model including the stochastic part. For these type of models also some work on using the unscented method for stochastic MPC has been done. For continuous time Markov models, also called stochastic differential equations (SDE) he has been working with the unscented method for estimating states and parameters even for models with state dependent diffusion, jump diffusion and multi models. Within application areas the focus has been on modeling and control of wind energy but the above methods has also been used for very different problems as modeling and control of type 2 diabetes and forecasting of the spread in the resent Covid19 epidemic.