How Do We Get to Rosey the Robot?


To kick off the third day of RIA’s Robotics Week, a panel of robotics executives discussed how the industry is poising itself for the implementation of complex household robots. What advancements does the industry need to achieve close human-robot interaction? What will that robot look like? Will it be treated like the beloved “Jetsons” character?

Robots of the future won’t be designed to do one task, especially if they are designed for home use. How to use robot technology in more than one capacity starts with the mobility of the robot in environments not designed for robotic maneuvering.

“To our mind, one of the first things to unlock is the mobility challenge,” said Michael Perry, vice president of Business Development at Boston Dynamics. “Before Rosey the Robot can happen, it needs to deal with getting upstairs and dealing with the piles of clothes in the house.”

Perry said that his company has started applying that mobility challenge in “semi-unstructured environments” in which a robot is exposed to the chaos and changing environment that humans face daily.

A household is an environment where advanced robotics and learning capacities have to come into play, but the same technology used in industrial robots won’t always cut it.

“Too often you see robots that were designed from an industrial mindset being put in a home,” said Aaron Edsinger, CEO and co-founder of Hello Robot. “They weigh 300 lb and they can’t even reach to the back of a countertop.”

Edsinger noted aspect ratio, length and movement of an arm and the base are all important considerations for household robots. The robot also needs to understand how to get around obstacles, which is where advancing learning sensors will help. Bottom line: The robot has to get to where it needs to be and needs to understand what it needs to do.

While the industry innovates these advanced sensors, robots will have to rely on human perception and instincts to fill the gap. And humans have to be ready for that. In a non-household setting, robots can be used in the form of exoskeletons for the amplification of human actions.

“Our whole focus is leveraging the human intelligence and human sensor system to really partner with the intelligence and sensor system and computing power in our robots,” noted Ben Wolff, chairman, CEO and director of Sarcos Robotics. “When you’re talking about unstructured environments and complex tasks, we’ve got to rely on human intelligence, reflexes, instincts and judgement to make a lot of these decisions.”

In relation to collaboration, robots are relatively dependent on those who work near them. Esben Ostergaard, CEO of REInvest Robotics, discussed the different types of collaboration and how misrepresented collaboration can lead to confusion. “It’s not robots working hand-in-hand with humans in my perspective,” he said. “The idea with these collaborative robots was that we gave the ownership of the machine to the people working near it.”

This, according to Ostergaard, has allowed small and medium-sized companies to take more ownership over their automation processes.

Regardless of environment—be it in the suburbs, on a city street or on a factory floor—the human psyche and perception needs to be accounted for a successful close-contact robot implementation. Physical attributes of robots, which came up in an earlier session, appeared once again. What will the future robot look like, and what psychological factors should robot manufacturers consider?

Edsinger, whose company makes direct consumer robots, contended that if a robot in a household performed any sort of useful task, an emotional bond would be forged regardless of humanoid appearance or not. “We deliberately made our robot very utilitarian in appearance,” he said. “We don’t see that social aspect is necessary in the very utilitarian way people want these devices.”

Perry agreed with Edsinger, echoing that, for now, a humanoid appearance is not necessary for a connection. He also cited research on human-robot interaction and said there will be things to learn over time. “The type of value that we can add is primarily in remote or dangerous environments or dull, dirty and dangerous work,” Perry added.

The panelists agreed that there are major advancements coming in the world of robotics in the next few years. The areas of greatest advancement, according to the panelists will be in wearable devices, machine learning, and mobile and remote manipulation.

“We need to be aware of the reasons why we’re building the robots like we’re building them. We’re not trying to build copies of ourselves,” said Ostwegaard. “It’s about empowering people. It’s about giving people the ability to do more with their time and energy.”



Source link