Can Pepper the robot be a good playmate?
Robots can actually be good playmates, but only if they behave in ways that make sense to people.
Norwegian University of Science and Technology
image:
Pepper and master's student Suraj De exchange information.
view moreCredit: Photo: Dafina Marku, NTNU
Researchers at the Norwegian University of Science and Technology (NTNU) have investigated what it is like to play a physical game with or against a robot that both looks and behaves like a person.
They conducted a controlled laboratory experiment with Pepper, a social robot designed to interact with humans.
“Our observations show that robots can actually be good playmates, but only if they behave in ways that make sense to people,” said Yavuz Inal, associate professor at NTNU’s Department of Design in Gjøvik.
So, humanoid robots can make good playmates, but their design must take into account gameplay modes, pace, role and order of play. If the robot suddenly acts like an overzealous seven-year-old who simply has to win, while also moving a bit stiffly and not quite understanding the rhythm of the game, we quickly get annoyed. We expect robots that are more natural, responsive and flexible than the current technology is capable of delivering.
When robots resemble us, we expect more
Humanoid robots such as Pepper are designed to resemble humans, in both their appearance and behaviour. They have heads, eyes, hands and facial expressions that make it easier for us to relate to them, but which can lead us to automatically expect them to behave somewhat like humans as well.
Previous research has shown that these types of characteristics increase engagement, whether the context is health, education, or pure entertainment.
Human versus machine in trash can basketball
In the study, the participants played a physical version of trash can basketball with Pepper. The human participants and Pepper the robot threw scrunched up balls of paper at a trash can from carefully chosen positions that made the game suitably challenging.
The researchers tested two gameplay modes: either humans and the robot were on the same team, or they played against each other. In addition, the order of play varied; sometimes the human participants started first, sometimes they let the robot start.
This enabled the researchers to investigate how gameplay mode and order of play affected engagement, motivation, emotional responses and enjoyment of physical activity.
“The study shows that even small adjustments in pace and order of play can be crucial in determining whether playing against a robot is perceived as fun or frustrating,” explained Inal.
The participants in the study enjoyed the game most when it was started in cooperative mode. However, there were also situations where the participants enjoyed playing against the robot. Many of the participants found competitive gameplay more exciting and motivating, especially when they were able to start the game themselves and thus felt more in control.
Beating Pepper produced a distinct sense of mastery, and some individuals admitted that it was especially satisfying when the robot missed the trash can. For these participants, competitive mode provided a clear goal and a sense of being challenged in a fun way.
Annoying when Pepper is too slow or too eager
It quickly became clear that the participants tended to get annoyed with the robot for many of the same reasons that we get annoyed with family members who are a little too competitive when playing Ludo or Monopoly.
Frustration increased especially when Pepper started the game in competition mode. The robot’s stiff movements, long pauses before each throw and slightly overzealous attempts to ‘win’ led people to expect more and become frustrated when Pepper failed to come across as either dynamic or particularly human-like. Some participants described the experience as like playing against an “overworked printer with arms”.
“When robots are invited to take on the role of playmates, we expect them to behave like proper players. If they don’t, we quickly become impatient,” asserted Inal.
The results offer a glimpse into a future where robots not only assist us at work and in the healthcare system, but also participate in everyday activities such as play, exercise and games.
Reference:
Yavuz Inal, Deepti Mishra, Suraj De: ‘The effect of cooperative and competitive human-robot interaction on player experience’ https://www.sciencedirect.com/science/article/pii/S1875952126000273
Androgynous and 120 centimeters tall, Pepper is designed to be unintimidating in person.
Credit
Photo: Suraj De, NTNU
Journal
Entertainment Computing
Method of Research
Experimental study
Subject of Research
Not applicable
Article Title
The effect of cooperative and competitive human-robot interaction on player experience
WVU researcher develops soft and ‘squishy’ robotic hand for harvesting delicate fruits like strawberries
West Virginia University
image:
In the WVU Benjamin M. Statler College of Engineering and Mineral Resources Robiotics Lab run by researcher Anand Mishra, a soft robotic gripper developed for fruit picking holds a strawberry.
view moreCredit: WVU Photo/Jennifer Shephard
A new robotic gripper designed to pick fruit could mean supermarket shoppers enjoy better quality produce as well as new options like pawpaws, a West Virginia University researcher said.
Anand Mishra, assistant professor in the Department of Mechanical, Materials and Aerospace Engineering at the WVU Benjamin M. Statler College of Engineering and Mineral Resources, was part of a team of roboticists behind the development of a soft gripper capable of gauging the size, curvature, color and ripeness of fruits like strawberries and avocados.
He said the gripper’s soft “fingers,” made of silicone and polyurethane, handle delicate fruits gently, while multiple sensors enable precise decisions about which fruits are ripe for harvest.
“Our gripper’s quick, accurate inspections and harvesting can reduce spoilage of fruits and lower supply chain costs,” Mishra said. “Fruit inspections are critical for harvesting decisions that have traditionally been made by human workers. However, using human workers for harvesting involves challenges such as labor shortages, health concerns and inaccuracies in picking.
“Recently, robots have emerged as a promising technology to reduce spoilage and improve distribution efficiency. But most current robotic systems are designed for indoor greenhouse applications, even though over 99.9% of crops are cultivated in outdoor environments. And rigid robotic systems can bruise ripe fruit with their bulky, hard grippers, while struggling with sensor reliability.”
In contrast, he explained, his team’s soft robotic gripper can inspect the fruit with both tactile and visual sensors that identify the perfect time for picking, and it can harvest a fruit like a strawberry simply by twisting the stem, no cutting required.
In a paper in Nature Communications, Mishra and his coauthors detail their findings from research originally conducted in the Organic Robotics Lab of Rob Shepherd at Cornell University, and now continuing in Mishra’s WVU lab.
The five-fingered gripper they developed looks and acts like a hand — but it also resembles a starfish. Mishra said those associations with natural organisms aren’t accidental.
“My research group at WVU is called the Robiotics Lab because we focus on robots that mimic biology. We design squishy, squeezy, rubbery robots inspired by animals. These robots move differently than traditional hard robots. Imagine a robot that moves like an octopus, for example. Because our robots are soft, they absorb vibrations and mechanical forces differently than traditional robots. And their tactile sensing is different in the way they touch objects and discern pressure and shape.”
Those qualities are important for fruit farmers because ripeness is a major challenge in farm-to-market distribution. While some crops, like citrus fruits, can be commercially harvested over a period of weeks, others like strawberries and raspberries have a window of only a day or two. Making it harder, those fruits with shorter ripening periods are usually soft with thin skin, meaning they bruise and spoil easily. Strawberry farmers, for instance, can see post-harvest losses of up to 25%.
But because many fruits, such as avocados, don’t reliably look different after ripening, visual assessments aren’t enough. Robots, like humans, often need to squeeze a fruit to know how ripe it is. In addition to driving significant financial losses for farmers and retailers, the difficult question of when to harvest means that some fruits, like the delicate, quick-to-spoil pawpaw, can’t be successfully distributed beyond their local areas.
The soft robotic gripper Mishra and his colleagues developed can mitigate those challenges. Within each of its fingers, stretchable optical fibers serve as tactile and curvature sensors, while the palm houses a miniaturized camera and distance sensor. After testing the gripper on supermarket strawberries and then a living strawberry plant, the researchers showed the gripper can open and close in under two seconds, lift weights of up to a kilogram — more than 16 times its own weight — and achieve nearly 100% accuracy in shape prediction. It can sense not only stiffness and bending, but also slipping, so it can detect when its grasp on a fruit is unstable.
While the gripper’s multifunctional soft sensing system is focused on reducing food waste and financial losses in agricultural supply chains, Mishra pointed to uses beyond fruit harvesting, and even beyond farming.
“This system offers applications in space exploration, health care, food handling and underwater manipulation,” he said. “In biomedical robotics, for instance, the integration of curvature and tactile sensing could enhance wearable and rehabilitation devices. There are many possibilities inherent in the technology’s capacity for multi-object grasping, manipulation and environment interaction. Like the animals they’re based on, soft robots can adapt to new situations.”
WVU researcher Anand Mishra’s robotic gripper is capable of gauging the ripeness of the fruit through sensors.
Credit
WVU Photo/Jennifer Shephard
Journal
Nature Communications
Article Title
Sensor fusion of touch & vision in soft manipulators for fruit picking
No comments:
Post a Comment