She wants to make robots behave better

Researcher Hannah Pelikan at Linköping University, Sweden, believes that we will see increased conflicts between humans and robots in the future. There may be several reasons for this, one being that the robots are quite clumsy in social contexts. In her research, she films everyday encounters between humans and machines to see what happens.

The little robot is rolling along the streets of Northampton. It looks like a white box on wheels and is loaded with groceries to be delivered to an address in the city. Two scientists are following it, using their cameras to capture what is happening.

A helpful lady presses the button for the robot to get a green light at the zebra crossing. A man pulls hard on its antenna. Some children try to stand in its way. But these are isolated events. There is something else that interests the scientists more.

“We saw so many instances where people had to adapt to the robot. This works sometimes, but over and over again you see the same problem,” says Hannah Pelikan, who researches robot-human encounters.

She is a postdoc at the Department of Culture and Society at Linköping University. She carried out this particular study together with two British colleagues: Stuart Reeves at the University of Nottingham and Marina Cantarutti at the University of York. They got the idea after they saw videos about robot misadventures that people had posted online. In March this year, the finished scientific article won an award at a major international conference in the United States.

Some years prior to that, in her bachelor thesis, Hannah Pelikan had used video to document the encounter between humans and machines. Her thesis was converted into a scientific article that she got to present at an international conference. From the reactions, she realised that she had actually come up with something new by filming the encounter between robots and humans and then conducting careful analyses for research purposes.

“I’m not the only one doing this, but there was enough room to make a career out of it,” she says.

Today we are used to robotic lawn mowers and vacuum cleaners, but Hannah Pelikan does not think it will stop there. The next step is most likely robots used for the transport of goods and people. The delivery robots in Northampton are just one example.

Robo-taxis have been tested in several cities in the United States for years. This has not been free of conflict. In San Francisco, emergency services have complained that the cars have blocked the road in emergencies, and taxi drivers and other citizens have protested. The companies behind the self-driving cars, on the other hand, claim that they are significantly safer than those driven by humans.

“You might ask if we should even have these robots. But I think there will be more and more mobile autonomous systems in our surroundings. The question then is how to make this work without creating too much social conflict,” says Hannah Pelikan.

Conflicts of different kinds occur. In Northampton, researchers study the small annoyances that can occur in everyday life. When people meet each other in the street, it’s like a semi-conscious dance. We show consideration and anticipate each other’s movements. Even unexpected situations can usually be resolved smoothly. For the delivery robot, this is difficult. It stops suddenly and makes pedestrians stumble, it can become paralysed in the face of the unexpected and it annoys people who have the street as their workplace.

So how do you make the robot a little better at dancing the way we do? Well, starting from observations in real life instead of just ideas on the drawing board is a good method, according to Hannah Pelikan.

“I hope that our research can show those who work at robotics companies that they must take the time to go out one afternoon and follow their product in everyday life on the street. Then they would see that it often works a little differently from how it was intended.”

A lot is about timing. During her doctoral studies, Hannah Pelikan spent many hours onboard Linköping University’s self-driving buses. One thing she tested was different sound signals to make the interaction between vehicles and cyclists flow better.
A signal at the right time makes traffic flow smoothly, a signal at the wrong moment can startle a passer-by. What is the right timing can only be determined by on-site observations.

Hannah Pelikan thinks that increased robotisation requires a deeper social debate about the problems that can arise, the purpose of the technology and what the human role will be. She also sees a risk of increased conflicts when jobs are threatened or robots are present in environments where people have not chosen to have them. In such cases, making them show a little more consideration is a small step on the way.

“But I’m not afraid that robots will take over the world anyway, because we’re not there at all. In some ways, that’s reassuring too.”