The use of multirotor UAVs is becoming widespread in various fields, such as aerial photography, surveillance, and payload transportation. In this work, a robust control system is designed to enable pose tracking of an unknown slung load, carried by a team of multirotors. A nonlinear geometric control design approach is adopted, to explicitly take into account the nonlinear and coupled dynamics of the system. An online optimization is developed to ensure reliable and efficient transport of the payload. Finally, the control system is augmented with an online parameter estimator, to reduce high levels of uncertainty in the payload parameters (mass, moments of inertia, and attachment points). It is shown that the suggested approach provides excellent payload pose tracking performance, even for cases of high initial uncertainty in the payload parameters.