Advertisement

Sensitive types

Reading Time:2 minutes
Why you can trust SCMP
0

Robots are becoming increasingly nimble devices, with the capability to assemble cars, clean floors and dance. Alas, like the fictional Terminator, they lack sensitivity.

But that is changing. For example, speech-recognition software made by Israeli firm Nice Systems is programmed to distinguish when a customer is getting angry with an agent so a call-centre system can automatically redirect the call to a manager. Scientists in the United States have gone a step further by developing a robot that can decode emotion. Nilanjan Sarkar, an associate professor of mechanical engineering at Vanderbilt University in Tennessee, has invented a method that allows a robot to help in the treatment of autistic children. In initial experiments, Sarkar used a robot to accurately read their emotional state based on physical responses, such as heart rate and temperature.

The robot worked with children suffering from autistic spectrum disorders (ASD), which are characterised by the impairment of social and interactive skills.

Sarkar's original plan was to use his method, which monitored a person's emotional state from key physiological measurements, to iron out kinks in dialogue between robots and humans. But when his nephew was diagnosed with ASD, Sarkar decided to use the technique to help children with the condition.

It helps that children love robots. To make his emotion-sensing robot even more appealing, Sarkar designed it to play games.

In his study, Sarkar hooked six teenagers with ASD up to physiological sensors and let them play games, such as Pong, with his robot. The amassed data on the moods of the participants pinpointed states such as engagement and anxiety with what was claimed to be more than 80 per cent accuracy.

With robots correctly reading their emotional states, those with ASD are expected to be better able to monitor and control their feelings.

Advertisement
Select Voice
Choose your listening speed
Get through articles 2-3x faster
1.1x
220 WPM
Slow
Normal
Fast
1.1x