Healthcare resources are being stretched further and further, leaving holes in care that require novel approaches to address. Tasks that were performed by people in the past are now being handled by machines, and perhaps the most novel machines of all are the robots. We’ve seen robots to dispense medication, greet hospital waiting room patients, and even comfort pediatric patients.

Computers have long been used for training and educational applications, as they have endless patience and can record every action by the student. Researchers at the University of Southern California have taken this one step further to create a robot designed to help children on the autism spectrum learn more effectively. Because of their condition, many autistic children lag behind in a standard classroom environment.

The system presents a math game on a touchscreen panel while a friendly green robot sits and watches. The robot interacts with the child and encourages engagement through conversation, facial expressions, and movement. The systems were placed in the homes of 17 children for a month, and they recorded video and audio of the children’s interactions during game play. Further analysis of the data found that the robots were able to correctly identify whether or not the student was engaged with the activity with a 90% accuracy rate.

This means that such a system will be able to monitor the child’s engagement during training sessions. If the child is not focused on the activity at hand, the robot will try to re-engage the child. The result will be a patient tutor who will be able to patiently monitor the child’s progress on a given individualized lesson, and nudge focus back to the task if the child’s attention drifts. The result is that the child may learn essential skills more effectively than in traditional classroom settings.