Animal behavior inspires robotic research

Tiara Winata

Researchers at Tech are making strides in the field of robotic intelligence through projects that look to both human and animal behavior as inspiration. Dr. Ronald Arkin, Director of Tech’s Mobile Robot Laboratory and Regents’ Professor in the College of Computing, and his team are investigating these behaviors and working to implement them in robotic systems.

Dr. Arkin and his students and colleagues are working to discover how they can transfer animal instincts to robots, enhancing their capabilities.

“We’ve spent a good amount of time looking at human behavior, but we always look at animals as well. You’d be surprised how smart they are. I think in many ways our intelligence when viewed by ourselves is a little overrated. Animals have to be extremely smart to survive in their environments… they have to be successful predators, and most of them are.” Arkin said.

“Animals have to be extremely smart to survive in their environments… they have to be successful predators, and most of them are.”

Funded by the Office of Naval Research, Dr. Arkin is currently working on a basic research challenge that also involves navigation techniques. He is collaborating with another faculty member, Frank Dellart, an Associate Professor whose work is in the areas of robotics and computer vision.

Dr. Arkin and Dr. Dellart are investigating ways of implementing a concept called optical flow, which they compare to driving down the highway and seeing everything streaming past. As this happens, people can recover depth with a single eye, thereby creating a ‘depth map’ of the world that relates the distance of objects from a particular viewpoint.

The research is looking to change the way robots visually perceive their environments in a process known as mental rotation, which would make use of the depth map to create a frame of reference robots can follow directionally.

“Humans and some higher-level primates have this capability to perform mental rotations, which is similar to IQ tests where you identify objects with different shapes. As we import those functions into robots to perform navigational tasks, our hunch is that they can be used for giving advice, the same way that humans use maps,” Arkin said.

There may be practical implications associated with the project, but Dr. Arkin emphasizes that the process itself has to be fully understood before they can see if it does fit a certain role in robotic intelligence.

“We’re not trying to understand in this case how to make the navigation process necessarily better, but how to use really solid techniques to help robots get from point A to point B and to understand why we as humans have this capability as well,” Arkin said.

Again inspired by biology, Dr. Arkin is leading a team on various projects for the U.S. Department of Defense investigating the potential for robots to learn how to deceive other robots and people as well, especially on the battlefield. Dr. Arkin emphasized that deception, which is at this point a relatively understudied area of robotics initially requires trust. They studied this first in the situational context of knowing when and how to deceive.

To this end, Dr. Arkin’s team has been studying other examples of deception, which is rampant in nature between predators and prey.

One of the examples that Dr. Arkin cited is the behavior of the eastern gray squirrel. Squirrels hide their nuts from competitors and occasionally patrol their nut caches, but when another squirrel or a predator appears, they change their patrolling pattern and start visiting empty caches to throw them off the trail. When Arkin and his team tested similar behavior in their system, it took longer for a competitor to find its objective when the robot implemented a deceptive strategy.

Another ongoing project considers handicap principle, which deals with the dishonesty of wasting resources to show strength. Along these lines, Dr. Arkin’s team is beginning to study antelope, which exhibit a behavior called “stotting,” where, rather than running away from a predator, they jump up and down, indicating to the predator that they are fit enough escape. This sometimes convinces the predator to abandon the hunt.

Where deception comes into play, Dr. Arkin hopes to learn how to implement the technique of feigning strength where strength does not exist in the software for their robots.

“When you start opening that Pandora’s Box, what should be done with this new capability?”

“Suppose there are some unfit antelope that jump up and down, knowing that they couldn’t outrun the lion, but hoping that the lion won’t call their bluff. The real question is, like in poker, knowing when to gamble,” Arkin said.

“Sun Tzu said that all warfare is based on deception. The results of this test were unsurprising but very interesting in terms of what we can do with this capability. If you want robots to be on patrol and to guard your resources, from ammunition to first aid supplies or even people hiding, you don’t want the robot leading the enemy directly to them,” Arkin said.

Dr. Arkin does emphasize a moral obligation that we have to reflect on the ethics of robotics studies It is his hope that people will examine as a community and a society what guidelines and restrictions should be made, especially in the military.

“When you start opening that Pandora’s Box, what should be done with this new capability? I believe that there is a potential for non-combatant casualties to be lessened by these intelligent robots, but we do have to be very careful about how they’re used and not just release them into the battlefield without appropriate concern. Robots could do things that put soldiers at risk, and there is the possibility of human augmentation with cyborg soldiers as well that could change the structure of society,” Arkin said.

Dr. Arkin teaches a course CS 4002 on Robotics and Society, examining the effects of advanced technology on our modern world. He encourages broad discussion on the implications of recent advances in robotic intelligence.

“We, historically, have not been as concerned with the technology that we are creating as we should have been, and we need to change that. I’m glad I have the opportunity to raise these questions in my class with students as well, and to gain a better understanding myself through their opinions,” Arkin said.