Applying Deep Reinforcement Learning to Cable Driven Parallel Robots for Balancing Unstable Loads: A Ball Case Study

The current pandemic has highlighted the need for rapid construction of structures to treat patients and ensure manufacturing health care products such as vaccines. In order achieve this, transportation materials from staging area deposition is needed. future, this could be achieved through automated sites that make use robots. Toward in paper a cable driven parallel manipulator (CDPM) designed built balance highly unstable load, ball plate system. system consists eight cables attached end effector can extended or retracted actuate movement plate. hardware was utilizing modern processes. A camera using image recognition identify pose on used inform development control consisting reinforcement-learning trained neural network controller outputs desired platform response. nested PID each motor realize For controller, three different model were compared assess impact varying complexity. It seen less complex resulted slower response flexible more output high frequency oscillation actuation signal resulting an unresponsive concluded showed promise future with potential improve state art.