Scientists have taught drones to obey muscle movements
Compact drones have firmly established themselves in the modern world, becoming not only a gaming gadget, but also a means of applied purpose in various fields, including the military. Often, the drone is controlled using remote controls or special applications, but scientists were able to implement another method. Now flying devices have learned to respond to hand movements and subordinate their actions to her muscle contractions. The engineers managed to develop a special system of bracelet and electrodes, which together form a kind of controller. It reads muscle movements and then translates them into drone commands. As a result, if, for example, the operator raises his hand up, the controlled drone will do the same - the process of interaction between the controller and the device is based on this principle. The construction works as follows. The bracelet is placed on the lower arm, and several pairs of electrodes are attached to different muscle areas. The bracelet records the movements of the muscles in the area of the forearm, and the electrodes - those that are located above. As a result, the system can measure the activity of virtually all parts of the arm. Specially designed gesture recognition algorithms allow the drone to "understand" where the hand is being deflected and follow in that direction. While the drone on the remote control responds to keystrokes, the new system of bracelet and electrodes subordinates its actions to gestures. For example, to launch the drone forward, it is enough to clench a fist, and to turn the device up / down or right / left - just turn the brush in the appropriate directions. Using controls, the drone can be rotated along the vertical axis by turning the arm as needed.
During testing, the researchers tested about 1,500 gestures, 82% of which the drone was able to accurately recognize. The authors of the project call the main advantage of their development the absence of the need for special settings of the bracelet and electrodes for each individual operator. That is, anyone can put the entire system on their hand, while the drone will accurately identify all movements. This is not the first time that researchers have attempted to create control of a drone using gestures. A number of large companies producing quadcopters have similar projects, but their control systems are based on the availability of cameras. Therefore, in terms of uniqueness, the new development, although not a novelty, has some personal features that distinguish it favorably from previous projects. So, most of the gesture control systems for flying devices are based on cameras. Therefore, at long distances or at the onset of darkness, it will be impossible to use such systems. Other types of brutal control are based on different devices, for example, gyroscopes or accelerometers, but in certain cases (several movements in a row) they can show a strong error. In comparison, the new wristband system with electrodes has a minimal error, while the drone under its control showed performance even in the dark. The Topic of Article: Scientists have taught drones to obey muscle movements. |
