Since the DARPA Robotics Challenge Finals in June of 2015, roboticists at the Italian Institute of Technology (IIT) have been working to improve the capabilities of their custom made humanoid disaster robot, WALK-MAN. WALK-MAN is part of a much larger European Commission-funded project, which involves a variety of research institutes and universities all contributing to the development of different aspects of the robot, from simulation to perception to locomotion to manipulation.
After a solid five years of work, the WALK-MAN project is now at its final validation phase, and it’s gotten one last major upgrade to help it prepare to be helpful in the disasters we’re certain to have in the future.
For background on WALK-MAN, make sure and check out this in-depth article that we posted in 2015, just before the DRC. The version that IIT is announcing today has a number of hardware improvements, starting with a redesigned frame made of aluminum, magnesium alloys, and titanium. These lightweight materials shaved 31 kilograms off of WALK-MAN, bringing its down from 133 kg to just 102 kg.
With less mass to haul around, especially upper body mass, the robot is now faster, more dynamic, and better able to balance itself, all very important things for moving around the sorts of disaster areas it’s intended for. The upper body actuators are new as well, and even with a smaller battery, improved efficiency means that WALK-MAN can operate for about 2 hours on a 1 kWh battery.
The original WALK-MAN was perhaps not the slimmest of robots, but the new version is more compact in both width and depth, making it easier to go through doorways. IIT has collaborated with the University of Pisa on biomorphic 19-DoF hands that can robustly grasp a wide variety of objects, and the arms are strong enough to carry 10 kg around for more than 10 minutes.
To demonstrate the robot’s new capabilities, WALK-MAN took part in a validation scenario representing an industrial plant damaged by an earthquake, where gas leaks and fire are present. Not somewhere you’d want to send a human into, which is the whole point of robots like these.
For more on what’s new with WALK-MAN, we spoke with Nikos Tsagarakis, a senior researcher at IIT and WALK-MAN project lead.
IEEE Spectrum: What is new about the new version of WALK-MAN?
Nikos Tsagarakis: The new WALK-MAN has an upgraded upper body that is lighter and more powerful than the first release of the robot. The new upper body allowed us to reduce the overall weight of the robot 101 kilograms. Despite this reduction, the arms have 40 percent higher strength with a payload capacity of 10 kg each, compared to the 7 kg of the first version. The onboard computational power has been increased and robot software has been substantially enhanced, enabling the robot to use robust real-time control modules with higher complexity. This allows it to perform faster actions and tasks than the original system.
What specific problems were you trying to address by making these updates—and what are some of the lessons learned from building the first version of the robot for the DRC in 2015?
The main issues were related to the dimensions and weight of the robot, as well as with the physical performance of the arms. Reducing the weight of the robot had also benefits for locomotion as well. The reduced dimensions of the upper body provide higher flexibility to operate within human environments, like passing through doors. The new version can lift heavier loads for longer periods of time, walk faster and more stably, and it consumes less power and executes more complex coordinated motions and actions.
How much of what we see in the video is autonomous, and how much is teleoperated?
The robot is under the command of the human operator for about 80 percent of its actions, with some local autonomous functions related to whole body motion actions and stabilization.
What is the status of the project now, and what will happen to WALK-MAN after the project ends?
The WALK-MAN project will have its final meeting and review very soon. During the last phase of the project the validation of the platform was performed through a series of task scenario, which the robot had to perform under the control of the human operator. It has been an extremely successful project and was recently selected by the European Commission to be featured as one of their success stories. We will keep working on the WALK-MAN platform beyond the project with a focus on improving its locomotion and manipulation capabilities.
What is the biggest challenge that you are working on with WALK-MAN right now?
The main challenges are to increase gradually the autonomy of the robot control, and in this way achieve faster execution performance. Achieving faster locomotion as well as demonstrating more robust physical interaction performance are two other specific challenges we want to tackle. In particular, we hope to demonstrate a richer repertoire of “loco-manipulation” skills that incorporate a certain level of autonomy to cope with uncertainties on the manipulation loads and locomotion disturbances.
[ WALK-MAN Project ]
Evan Ackerman is a senior editor at IEEE Spectrum. Since 2007, he has written over 6,000 articles on robotics and technology. He has a degree in Martian geology and is excellent at playing bagpipes.
Erico Guizzo is the digital product manager at IEEE Spectrum. He oversees the operation, integration, and new feature development for all digital properties and platforms, including the Spectrum website, newsletters, CMS, editorial workflow systems, and analytics and AI tools. He’s the cofounder of the IEEE Robots Guide, an award-winning interactive site about robotics. An IEEE Member, he is an electrical engineer by training and has a master’s degree in science writing from MIT.