Logo: University of Southern California

Improved Robot Communication - While the Building's on Fire

USC Viterbi researchers are developing a wireless network for robots to autonomously communicate and work together, amid potential disaster areas
By: Sam Corey
November 04, 2015 —
A team of robots scan the premises of a burning building, giving firefighters knowledge of the environment before they enter. (Illustration: Peter Bollinger)

Imagine a person trapped on the top floor of a burning building, with a handful of firefighters scrambling to rescue them and extinguish the fire.

Developing a rescue and evacuation plan requires quick thinking, precise preparation and anticipation of any potential obstacle, such as a floor or a wall collapsing from intense heat.

Firefighters risk their lives by entering burning buildings that on average reach 1,000-1,400 degrees Fahrenheit. They must observe their surroundings and quickly map out the best rescue route amid a chaotic, crumbling structure.

But what if a team of robots can survey the building instead?

If a floor or a wall collapses, firefighters don’t have any way of knowing this happened, and if they are in the building while it occurs, they risk significant injury or death.

Rather than having a group of firefighters enter the building blindly, a group of robots can work together to deliver an invaluable snapshot of the burning building so firefighters will know what obstacles they face before they ever step inside.

This information could make it easier for firefighters to rescue people, reduce their risks and allow them to douse the flames more quickly.

Two groups of USC Viterbi professors and doctoral students are working to make this a reality.

USC Viterbi Ph.D. students Jason Tran and Pradipta Ghosh, with the guidance of Bhaskar Krishnamachari, associate professor in the Ming Hsieh Department of Electrical Engineering, are focusing on the movements and networking of a group of autonomous robots.

Meanwhile, USC Viterbi Ph.D. student Shangxing Wang is working with Krishnamachari and Nora Ayanian, assistant professor in the Department of Computer Science, to develop algorithms that will be used by Tran and Ghosh to evaluate with mobile robots that are able to communicate with each other wirelessly.

Tran and Ghosh are currently developing a test bed to see if their mobile robots can use radio communication to help each other maneuver around a room full of obstacles to get from one side to another.

If successful, a team of robots could accurately pinpoint and communicate the details of the room with the least human input possible, which would mean less people in the control room and more people in the field.

Although robotics companies, such as Google-owned Boston Dynamics, already have prototypes of humanoid robots that can navigate through different environmental terrains, they still depend on humans controlling each individual robots. Additionally, these robots cannot communicate with each other without human direction.

“The vision is the robots would notify firefighters where to go, and where not to go,” Tran said. “These robots could detect a survivor’s location or determine if the temperature and atmospheric conditions of a specific room may be too much for a human to handle. Robots are typically equipped with various sensors to gather these kinds of information.”

“The end goal is for robots to be able to go into a burning building and assess the situation,” Ghosh said. “Based on their feedback, the firefighters can plan their rescue mission. Hopefully in the future, rescue robots can also be added that will do the firefighting in very dangerous situations.”

Tran and Ghosh are still developing these mobile robots. The doctoral students hope the mobile robots, along with the wireless networks needed for them to communicate, will become functional by next year.

“The biggest challenge is interlinking them all together,” Ghosh said. “For everything to work together, the robots have to work as a whole, not as individuals. Keeping the communication and networking between the robots intact all the time will also be difficult, since the communication in such dangerous situations is very flaky and intermittent.”

Meanwhile, Wang is concentrating on developing algorithms to withstand realistic wireless communications problems. These include any potential hindrance to wireless connections caused by obstacles within different environments.

For example, if a team of robots is mapping out a burning building, the collapse of a ceiling could block the signal required for robots to communicate. If Wang’s algorithms can adapt to different settings and sporadically added hurdles, then robots will be able to communicate with high efficiency, regardless of the circumstance.

Ghosh and Tran’s robots will use Wang’s algorithms to improve communication between their mobile robots.

“I want to see how we can design algorithms to make sure robots can stay connected and communicate with each other at a high data rate when they are performing a task in such an unknown obstructive environment,” Wang said.

If Wang’s algorithms can withstand any hurdle nature throws at these robotics networks, her research will have many useful applications including: creating detailed maps of environments; area surveillance; delivering food and supplies; rescuing people in disasters such as Hurricane Katrina; hostage rescue and establishing wireless networks at massive events, such as a World Cup game.

In simplest terms, these teams of robots will be able to help people accomplish tasks more efficiently.

The research by Wang, Krishnamachari and Ayanian has already earned international recognition.

On Sept. 15, their paper, titled “The Optimism Principle: A Unified Framework for Optimal Robotic Network Deployment in An Unknown Obstructed Environment,” was presented at the IEEE/RSJ International Conference on Intelligent Robots and Systems in Hamburg, Germany.

“It’s great to see that people are interested in our research after all our hard work,” Ayanian said.