Jeonbuk researchers let pilots steer drones by gesture and sense them by vibration

Researchers at Jeonbuk National University in South Korea have built a wearable interface that lets a person control a drone with hand gestures and sense where it is through vibrations, without watching a screen. The work was published in the journal IEEE Robotics and Automation Letters.

A sensor on the back of the hand reads 12 gestures as commands, and a belt with 36 vibration motors turns the drone's position into patterns that show direction, relative height and distance. The gesture model was 97.58% accurate. In 1,200 indoor trials, 12 participants identified direction, height and distance together correctly 82.75% of the time, taking 3.1 seconds on average.

With a real drone, one novice participant reached the target in eight of nine trials using touch alone, taking 45.1 seconds on average, against nine of nine and 31.4 seconds with visual feedback. The researchers say the results show the idea can work, but that distance cues and response time need to improve and larger outdoor studies are needed.

Professor Kee-Ho Yu, who led the team, said the aim was to link what the user intends with clear information about where the drone is.