Brain Operating System
Robots are not yet a part of everyday life because they are difficult to program, unable to adapt to changing environments, and are costly to customize at the point of use. Brain Corporation developed BrainOS (Brain Operating System) to address these issues. It allows robots to perform specialized tasks and adapt to new environments through training, reducing years of costly programming to mere days of teaching. BrainOS intends to enable any robotics company with a business idea to quickly launch and iterate its products, leading to an explosion of consumer applications.
Brain Corporation will provide the Brain Operating System running on a single-board computing platform and reference designs, so that building a robotics product will be as easy as creating a mobile app.
Home Service Robots
Science fiction writers of the last century believed that we would be taken care of by multi-tasking humanoid robots. In contrast, BrainOS and high-performance, low-cost mobile processors will help create teams of small, affordable robots, each trained to do a single task.
These task robots could be trained to navigate in cluttered family environments, avoid obstacles and use simple manipulators to pick up household objects for cleaning, sorting, or moving. They may also be capable of taking out the garbage, yard maintenance, or cleaning up after pets. Such robots may be pre-trained by manufacturers but also include some end-use adaptation defined by the consumer.
Recent developments in home automation and the Internet of Things have provided for sensors and actuators that can be embedded inside the home, turning it into a sophisticated “robot”. Currently, programming such a system is difficult for consumers – far more difficult than programming an early VCR. BrainOS replaces this programming with training.
Installed on a smart hub, BrainOS could allow users to customize their homes by simply showing examples of what they want. For instance, an elderly person may teach remote-controlled lights to turn on or off in response to a hand gesture or an everyday activity such as climbing into bed. A TV could learn to pause when somebody enters the room or mute when you want to take a phone call. Automated cabinets and drawers could be taught to open and close just by presenting examples of desired behaviors. Instead of us adapting to our houses, BrainOS allows the house to adapt to us.
Teaching a BrainOS-controlled robot is fun; it is akin to watching a pet learn new tricks. This trained behavior process may be embedded into a new generation of autonomous toys, turning them into robotic pets who learn and adapt.
Visual perception, an important part of BrainOS, enables robots to learn to sense and avoid obstacles and autonomously track objects – a desirable feature for the growing fleet of consumer UAVs (unmanned aerial vehicles).
BrainOS will enable training of single-task agricultural robots just by presenting them with examples of desired behaviors – harvesting, weeding, precision spraying, or hoeing. Being small and affordable, such robots might cover an entire strawberry or tomato farm, performing their tasks tirelessly and efficiently.
With 3D-printers becoming mainstream, hobbyists are able to download, print, and assemble robots such as Brain Corporation’s own self-balancing eyeRover™. Developers using BrainOS can create and share their designs and trained behaviors with others, potentially leading to a surge in 3D-printed custom service robots ready to download for any chore.
For more information visit Brain Operating System Beta FAQ.