Sarcos Robotics announced today that its subsidiary, Sarcos Defense, has been awarded a contract for a remote-controlled upper body variant of it’s Guardian XO Full Body Exoskeleton. The Office of Naval Research has awarded the contract for the Guardian DX. The “mobile arms” would be attachable to a base such as wheeled or tracked vehicle. They would be installed at the end of lifts or bucket trucks and be used to address maintenance and logistic needs at high or difficult to reach places.
This, in itself, is not a surprising award for Sarcos Robotics. Teleoperation has been one of the oldest applications for exoskeleton technology. Prototype exoskeletons have been used to collect movement data from their users in order to control robotic manipulators for decades; just see the section on remote teleoperation on the CyberneticZoo. Sarcos Robotics has years of experience developing its own human amplifier type systems such as the Guardian GT.
“Similar to our teleoperated Guardian® GT robot for heavy, dexterous work, but designed at human scale, the Guardian DX robot can be teleoperated to perform intricate tasks that require human-like dexterity,” said Ben Wolff, chairman, and CEO, Sarcos Robotics. “Examples of such tasks include the use of portable sensors for non-destructive structural testing and inspections, the use of portable power tools for grinding, cutting and welding at height, as well as lifting and manipulating heavy components weighing up to 200 lbs. Because the Guardian DX robot is kinematically equivalent to the upper body of humans, operators are able to manage the Guardian DX robot intuitively at typical human speeds by relying on their reflexes, instincts, and judgment to perform complex tasks in unstructured, often hazardous environments that historically have only been able to be completed by people directly.”
Artificial Intelligence (AI)
Additionally, last week Sarcos Defense was awarded another contract, this time by the U.S. Air Force, to Develop “Smart” Dexterous Robotic Systems with advanced artificial intelligence. “The artificial intelligence (AI) system that would enable robotic platforms, like the Guardian DX robot, to learn how to perform tasks with human-like movement through positive reinforcement and imitation machine learning (ML) technologies, known as Cybernetic Training for Autonomous Robots (CYTAR™).”
This can have very interesting implications for the usability and integration of powered exoskeletons. Powered exoskeletons, by design will never replace the user. A powered exoskeleton’s purpose is to combine human ingenuity, instincts, and reflexes with the endurance, strength, and tenacity of a robotic system. However, for a powered exoskeleton, the onboard motors and controllers have to guess what the user is attempting to do. In most cases, such as standing or walking, that can be relatively simple. However, a sudden event like a slip or a jerk to avoid an obstacle can result in complicated body movements by the user. This is further compounded by the complexity of the Guardian XO, which accounts for the forces and the inertia of the exoskeleton. AI could be instrumental not to replace the user, but to make powered exoskeletons easier and more intuitive to operate.
The second likely advantage of merging AI with exoskeleton technology is that it will increase its integration. Exoskeletons do not operate in a vacuum, and that is even more true for a military variant. The device needs to communicate and synchronize with a myriad of peripherals. For example, a soldier kit could have one single power source that supplies energy to a powered exoskeleton, communication solution, and night vision goggles. AI could help to automatically redistribute power as needed depending on the situation without distracting the exo user.
For now, however, Sarcos will apply its AI work on the Guardian DX:
“This is a unique opportunity to leverage a robotic system that is kinematically equivalent to the human body to lay the foundation for teaching robots how to move and accomplish tasks in the real world, the same way humans do,” said Denis Garagić, chief scientist, advanced systems and AI, Sarcos Robotics. “This success-based teaching will speed up the learning process of AI-enhanced perception, reasoning, and decision-making techniques currently being implemented across Sarcos’ platforms. Implementation of such an AI-based system will enable autonomous situational awareness, which can radically reduce the cognitive load on the operator while dramatically increasing precision as it augments human performance.”
“We are excited to collaborate with Sarcos to leverage the company’s unique robotic systems to develop a novel approach to success-based AI systems” said Dr. Alok Das, Senior Scientist at AFRL’s Center for Rapid Innovation. “We will be working with end-users throughout the Air Force to refine the CYTAR platform for Air Force-specific applications.”
Sarcos.com Press Releases link