Scientists from MIT have presented a new algorithm that allows the robot to dress people without interfering with them. The device can be used as an assistant for people with limited mobility.
The researchers explained that robots have great potential to help people with reduced mobility. But this is a complex task that requires agility, user safety and speed from the devices. Now scientists at MIT have developed an algorithm that ensures this balance.
MIT added that it is relatively easy for a person to help someone else dress, because we instinctively know where and how to hold a garment, how a person can bend an arm, how fabric will react to movement, and so on. However, robots need to be taught all of this from scratch.
The MIT team, led by Shen Li, has developed an algorithm that redefines the safety of robots by allowing movement to prevent collisions with it. This allows the robot to come into contact with a person without harm to health in order to complete its task, provided that the impact on the person is negligible.
The system for dressing a person worked even if at that moment he was doing other things – for example, checking the phone. This is achieved by combining several models for different situations, rather than using one model as before. “Our multifaceted approach integrates set theory, human security constraints, predicting human movements, and communication management for safe human-robot interaction,” said Zakori Erickson of Carnegie Mellon University.
The research is still in its early stages of testing, but the scientists’ ideas can be used for more than just dressing people. They can potentially be applied in various fields of assisted robotics and other types of assistance for people with limited mobility.
