How good is the support of modeling hybrid systems with MTK in the summer of 2026?
It used to be incomplete/ buggy for a long time. Did this change?
I am looking at the control of electrical machines. The controller can be outside of MTK, but it would be nice if not only the model, but also the observer could be modeled in MTK. See, for example: IM_AWES_bench.jl/src/estimators/rotor_flux_observer_discrete.jl at main · c-nicomar/IM_AWES_bench.jl · GitHub
What do you mean by hybrid system modelling? Just having both continuous-time dynamics (modelled by ODEs) and periodic sampling in one model (aka sampled-data systems)? I noticed that this is what some people regard as a hybrid system. Or do you need aperiodic, perhaps event-driven sampling? Or do you mean hybrid automata defined by modes, flows, guards, invariants and resets (in the spirit of Stateflow charts)? Or yet something else?
Just discrete and continuous time, like in Simulink. A discrete-time observer and Kalman filter, and a continuous-time machine model.