iCub open source platform designed robotics evolution

Robotics research in the last 10 years comes from a standard open source platform for research on artificial intelligence (AI). The iCub humanoid robot is built in Italy and is available in laboratories across Europe, U.S., South Korea, Singapore and Japan.

More than 100 researchers contributed to developing iCub skills and researchers at IIT-Istituto Italiano di Tecnologia said in Science Robotics the importance of the research platform as a way to challenge the development of robotics technology.

Tinuku iCub open source platform designed robotics evolution

iCub is an unfinished evolution from the beginning to the present where hardware and software tailor research needs and highlight the benefits for robotics. The researchers reuse and improve the results of each other toward the goal of creating a smarter machine.

"We emphasize what we have learned so far by working to build an open source robot community where researchers need a standard platform for humanoid robotics," said Giorgio Metta, IIT-Istituto Italiano's deputy scientific director and iCub facility coordinator.

"Using software created by others, replicating experiments, and benchmarking is an important element of standard platform robotics research is an enabler. The evolution of hardware and software is very important to keep both," Metta said.

iCub was developed in 2004 as part of a European project co-ordinated by IIT and researchers have released iCub 1.0, iCub 2.0 and iCub 3.0. The platform integrates the latest robotics research results covering the entire spectrum of AI, machine learning, human-robot interaction, and language acquisition.

This humanoid is the size of a five-year-old child and is able to crawl, sit, walk balanced, physically interact with the environment and recognize objects. The design uses a full-body electronic skin system that is sensitive to give a sense of touch.

"iCub 3.0 version is being created, but we are planning another revision that comes from requests to use robots in a clinical setting, specifically to design HRI experiments and training for children diagnosed with autism spectrum disorder (ASD)," Metta said.