Four years of intensive work by a team of 8kbet robotics students is about to pay off. We hope.
Transcending the limitations of location and the hazards of disasters is a massive challenge — one that students in professor Paul Oh’s Drones and Autonomous Systems Lab (DASL) have been working on with their humanoid robot, Avatar-Hubo.
Using the latest technologies in VR, AI, haptics and more, Avatar-Hubo will allow anyone to possess the body of a functional robot and teleport their presence anywhere across the globe. The team will find out how well they've done Nov. 4-5 when they compete against teams from around the world in the finals in the .
Pushing the limits of immersive telepresence was always the goal of the ANA Avatar XPRIZE challenge. But it took on particular significance after the world’s experiences with COVID.
“This idea of physical embodiment, or telepresence, is not something that a lot of roboticists have actively talked about or engaged in,” explains Oh, who also serves as team lead. “But with the need for social distancing during the pandemic, this idea of an avatar or robot agent really gained additional importance.”
Oh’s lab has always been focused on disaster response and humanitarian assistance. Born out of the events of 9/11, finding ways for robots to provide disaster relief following events like Hurricane Katrina or the Fukushima Nuclear Disaster are paramount to the lab’s work. Robots can allow workers to do their jobs from a safe location, or for professionals, like doctors, to aid patients even when physically remote.
Jean Chagas Vaz, DASL research scientist, team captain, and lead engineer of Team Avatar-Hubo, was an undergraduate in Oh’s lab when the team went to the (DRC) in 2015. In that competition, which focused on human-supervised robotic technology for disaster-response operations, the team took 8th place in the world.
“I started as a roboticist assistant for the DRC,”&Բ;Chagas Vaz says. “It is a little overwhelming to now be leading the robotics team in our next competition. But I’m so amazed at how much we’ve evolved the system from where it was in 2018 when we started this project, and even from what we had during the semi-finals, to what we have now.”&Բ;
Similar to the DRC challenge, the other competing teams in the ANA Avatar XPRIZE challenge are made up of not just university labs, but private industry and government research labs.
While private sector teams typically have a lot more resources, they can also be hampered with greater restrictions, explains Akshay Dave, head of Avatar-Hubo’s audio and video interface for the team. “We take chances and can push limitations that other industry sectors can’t.”&Բ;
Before joining the team, Dave’s work was related to virtual-reality (VR) systems. His main focus is to make a system that will allow an operator to seamlessly talk with the robot and provide real-time audio feed so the operator perceives the world as if they were where the robot is.
“I never expected that I would be working so closely with the robot. When I first joined the Lab, I was doing software and was nervous to work with the hardware and expensive systems. But over time I was able to build trust among my teammates and slowly work with Avatar-Hubo.”&Բ;
The complexity of the competition comes not just from the tasks the robotic avatar will be expected to complete, or the enormous amounts of data being communicated between the operator and the robot agent, but the fact that the robot will be controlled not by a team member but a judge who has received a maximum of 40 minutes of training.
According to Chagas Vaz, “We have to trust a million dollar robot to a judge who has never used the system before. For 25 minutes. You cannot overrate how important training the judge is. But at the end of the day, the competition is about creating technology that can be used by anyone (first responders, surgeons, etc.)”
Competitions are important to universities, labs, and students for a variety of reasons. There is prize money that can help fund additional projects, research papers born out of the large amounts of data collected during the process, and the experience of trial and error.
“I think it is important for our students to be at the cutting-edge of research, working on problems that aren’t solved yet but of enormous impact,” explains Oh. “It is also good for them to be with other like-minded groups, especially when those groups represent the best of the best. It helps us benchmark how we’re doing.”
Additionally, competitions stress the system. As Chagas Vaz shares, “No matter how well you prepare your system, it will never be perfect. We can’t model everything in a lab that may happen when you deploy a real robot to perform a certain task. I am sure we’re going to learn a lot during the competition and I hope the future Ph.D. students who remain in the Lab can carry the technology forward.”
The big question is, does Avatar-Hubo have a chance of winning?
According to team leaders and members, an important advantage that Team Avatar-Hubo has is in the diversity of the team itself.
“We have a very rich team with different backgrounds which elicits multiple ideas for one particular problem,” explains Chagas Vaz. “During team meetings everyone can express their opinions and suggestions, regardless if it is a vision problem, a manipulation problem, or a locomotion problem. Everyone gets a chance to say, what if we try this?”&Բ;
Oh noted that he is especially proud that 8kbet is the only Minority-Serving Institution in the global competition.
“We’re putting a lot of our heart into this with the desire to make 8kbet and Las Vegas proud, to show that there really is a lot of technical potential in this city," he says.
“Can we win?” asks Chagas Vaz. He's careful to set expectations. “I think we have a very good chance. But these sorts of robotic systems are very delicate and complex. The shear amount of data that has to go from the operator side to the robot agent – so many things can go wrong. My personal goal is to be in the top five.”&Բ;
Dave reminds of the greater goal: “I think we’ve already won. Competition is a catalyst for research. We’ve spent years developing a system that will allow a person to be a robot in real-time. To have achieved that progress at this rate, is already winning.”
The is a global competition with a total purse of $10 million to accelerate the development of a multi-purpose, physical robotic avatar system that transports a user’s senses, actions and presence to a remote location in real time.
The competition is open to the public and taking place at the Long Beach Convention Center in California on Nov. 4-5, 2022.
