Artificial Pancreas (AP) systems consist on a subcutaneous insulin pump connected to a continuous glucose monitor (CGM) through a control algorithm that calculates the appropriate insulin dose according to CGM readings. Though great research has been carried in this topic, there are still many limitations that make full closed-loop glycemic control challenging. These limitations include large delays in CGM readings and in the insulin action, intra and inter-patient variability, and the fact that there is no negative action to counteract insulin, among others. Here, several steps towards full closed-loop control are taken. First, hypoglycemia is mitigated using constant and time-varying IOB constraints in hybrid-loop and closed-loop controllers with meal announcement. Then, a switching signal generator is proposed in order to enable automatic switching between closed-loop controllers in a switched control scheme, thus eliminating the need for meal announcement. Lastly, a dual-hormone controller is presented in order to improve the performance achieved by a fully closed-loop, single-hormone control algorithm. Also, in the context of the COVID 19 pandemic, a remote monitoring platform is employed in order to monitor patients in the intensive care unit, thus reducing healthcare profesional exposure. Then, a closed-loop algorithm is tested in the first ambulatory artificial pancreas clinical trials in Argentina.