“Our long-term goal is to make robots much more versatile than they are today,” Stelian Coros, an adjunct professor at Carnegie Mellon’s Robotics Institute, told Digital Trends. “One way to make progress towards this goal is to give them the ability to change how they use their bodies based on the environment they find themselves in and the tasks they are trying to complete.”
This represents a departure from current approaches, where robot limbs are designed with very specific tasks in mind — such as legs which are intended only for walking, or arms which are designed solely to manipulate other objects. While using limbs for multiple purposes is rare in terms of robotics, however, it is widespread in the natural world. Mother gorillas, for example, will often walk on three limbs as they carry their babies around.
Adding this level of versatility to robots means improvements need to be made both in hardware and software. “We mainly focused on the software aspects for this work, namely on the challenge of developing new control solutions that allow robots to quickly figure out how to change gaits as they dynamically switch the roles of their limbs,” Coros continued.
In the real world, the idea of a robot that is able to vary its movement and functionality will allow researchers to build entirely new kinds of smart robots, including miniaturizing them by doing away with excess limbs — since the remaining ones can double up by performing multiple actions. “Ultimately, the idea is to develop a system that you can feed high-level tasks to — something like ‘go over there and pick up that thing’ — and the robot will choose how best to make that happen, no matter how many legs it starts (or finishes) with,” the researchers note in a paper describing their work.
For now, however, there’s still plenty more to be done. “This is a very early research prototype, so commercialization is not on our horizon at the moment,” Coros said. “We would first like to see these robots become increasingly autonomous, dexterous and agile, and this goal poses important challenges in terms of how robots are designed, sensing and actuation technologies, controls and [artificial intelligence].”
- Why teaching robots to play hide-and-seek could be the key to next-gen A.I.
- Man uses brain-controlled prosthetic robot arms to eat a Twinkie
- The best robots at CES 2021
- Chess. Jeopardy. Go. Why do we use games as a benchmark for A.I.?
- The future of augmented reality is earbuds, not eyeglasses