Adding a tuner that tunes only the costs and/or the optimizer of a controller will help increase the rate at which controllers can be iterated over. This is particularly useful with the current task framework as testing multiple configurations of tasks requires many iterations but it is not always necessary to re-optimize the system model that is being used until a properly behaving task definition is designed.
Adding a tuner that tunes only the costs and/or the optimizer of a controller will help increase the rate at which controllers can be iterated over. This is particularly useful with the current task framework as testing multiple configurations of tasks requires many iterations but it is not always necessary to re-optimize the system model that is being used until a properly behaving task definition is designed.