When was the last time you single handedly fixed a ceiling panel? Probably a long time ago owing to the dreadful experience you previously had between balancing screws and panel to failing altogether and hurting yourself.
This is the time when you wish to borrow extra pair of arms from Dr. Octopus. Why wish when they are soon to become a reality?
A “Supernumerary Robot Limbs (SRL)” project is being carried out at MIT’s d’Arbeloff Laboratory for Information Systems and Technology. Under this project an extra pair of arms can be attached directly over your spine for conducting a two-man job. These robotic limbs are light weight mounted on a frame attached to a backpack. The structure supports your spinal cord for carrying extra weight without any strain.
How to control these robotic arms when your own hands are busy too? You cannot use a remote control, joystick or game pad in this situation. Hence, the researchers programmed the arms to decide for themselves when and where to help. Special sensors are added to the human and the robotic wrists. These sensors determine when the human is on task and when to tell robotic arms to help. These robotic limbs learn by demonstration i.e. another human has to first manually practice the robotic arm to perform a certain task.
The MIT lab demonstrated the robotic limbs working through two experiments. The first one was to simply hold a ceiling panel in place while a human fastens the panel. The robotic arms relax when they feel less pressure is required for the panel. The second experiment supported a construction worker trying to drill holes in a skyscraper hundred meters above the ground. In this case, the robotic limbs are attached to the waist to grab the beams and act as a balancer for the worker.
This technology can help senior citizens, construction workers, assembly operators or anyone requiring extra help in day to day tasks.
MIT team will be adding further improvements to the Supernumerary Robot Limbs in terms of weight reduction, increase strength and designing better control algorithms.