Hospitals teem with busy healthcare professionals in seemingly ceaseless motion. Several companies and institutions focus on robotic assistants to alleviate demands on clinical staff members. Last month we wrote about engineers from the University of Toronto who are developing Pepper, a four-foot tall robot designed help nursing staff gather information from emergency department patients and provide companionship with emotional modeling that reflects how patients feel. Earlier this year we covered Vecna’s VGo telepresence robots that move from room to room to check patient status while nurses seated at a central station use the robot’s cameras and microphones to see and hear patients while working on their charts at the station.
Austin, Texas-based Diligent Robotics, created Moxi which is a social intelligent healthcare robot. Moxi supports clinical staff teams by performing non-patient-facing logistical tasks such as delivering lab samples, gathering and delivering supplies for new patient admissions, and transporting heavy linen bags during room cleaning. This leaves nurses with more time for direct patient care and tasks requiring advanced skills and training. Moxi has a mobile base and utilizes autonomous AI mapping to travel about the facility. It can grasp, carry, and place materials appropriately using its flexible robotic arm and precise gripper. One of Moxi’s key features is social intelligence: the ability to move its head and LED eyes in gestures that communicate what it is doing or about to do so that observers will feel comfortable letting the robot perform its tasks.
Diligent Robotics has paid attention to the potential concerns of patients, visitors, and even clinical staff, any of whom might have misgivings about an obviously very strong robot operating with apparent independence. Even though many people smiled at Star Wars’ competent if impatient R2D2 and polylingual but often confused C3PO, real-life interaction with robots is still outside the norm for most. Layering Moxi’s task-focused autonomy with humanoid facial clues softens the interaction.
