November 02, 2017      

SANTA CLARA, Calif. — At RoboBusiness 2017 here earlier this fall, the keynote speakers provided their personal insights into where the automation industry is headed and what still needs to be done. Melonee Wise, CEO of Fetch Robotics Inc., shares her thoughts about robotics and AI for logistics with Managing Editor Steve Crowe in the brief video below.

Among her other experience and accolades, Wise was manager of robot development at Willow Garage. She was also the closing speaker at this year’s RoboBusiness conference and expo. San Jose, Calif.-based Fetch makes mobile robots for warehouse and other environments.

Wise gives a preview of her address, in which she tried “to help people understand where we’ve come from in terms of technology, machine learning, and automation; where we’re at right now; and really where I think we need to be going and some of the hurdles and obstacles as we head in that direction.”

As for her own company, “Fetch is driving toward lights-out operations in warehouses,” Wise says. “We’re working on the movement of goods, [piece] picking, machine tending, and kitting.”

In addition to industrial automation, Crowe notes that Fetch has been working recently with research institutions on robotics and AI for logistics and other applications.

“Our mobile manipulation platform is being used at innovation centers of Toyota, Honda, Panasonic, and then they’re at universities like the University of Michigan, the University of San Diego,” Wise says. “People are doing a lot of different things with them. At the University of Washington, they’re doing a kind of ‘program the robot for the home’ journey.”

“That program is really focused on enabling research at all sorts of different levels,” she says. “Like OpenAI has been using it for their AI gym.”

Fetch’s platform is robust because it was designed for commercial applications. “We’ll be targeting it for more specific applications in the next two years [such as] machine tending, cell operations,” Wise says.

The company is focusing on broadening its market because there’s “a big need; it’s a good cross-section of our customer base,” she says.

Learning from customer feedback

Among the conference participants was a Fetch Robotics end user, says Crowe. Cindy Traver, senior director of operations at RK Logistics Group Inc., spoke about “Creating a Winning Team — Moving From Discrete Investments to Enterprise Scale” during the Chief Robotics Officer (CRO) Network Summit preceding RoboBusiness 2017.

Traver was also in a session on “How We Increased Our Revenue With Robots,” in which she discussed RK Logistics’ use of Fetch robots and what it overcame to obtain returns on investment.

“RK was one of our earliest customers. They’ve had a live deployment for over a year,” says Wise. “They were early adopters and had a lot of tolerance for the learning curve.”

“We learned about making robots easy to use,” Wise recalls. “Moving to the cloud came out of working with them. They didn’t have IT staff or support on premises during deployment. That was a big change. It pays off, helps when they have questions.”

Because RK Logistics changes its warehouse setup every few weeks, Wise says, “we started to realize that our 3PL [third-party logistics] customers aren’t doing static process operations. We needed to build a tool set to support that.”

Crowe mentions a Forbes article describing Fetch’s use of Skymind’s tools as an example of AI for logistics.

“We’re looking at AI, machine learning, or deep learning and the many ways it can help improve capabilities to do things quickly and be relevant to big data sets,” Wise replies.

For instance, “there’s a gray area around safety,” she continues. “There’s a method for training against data sets for certain things in the environment, knowing where all the robots are at all times, but it’s harder if they’re on and off Wi-Fi.”

“An algorithm that allows for the ability to detect robots, helps us do other more complex behaviors in the future,” Wise says. “Using AI or machine learning, we’re creating a lexicon for things in the environment, the ability to understand the environment.”

AI for logistics can also lead to “nascent behaviors,” so robots can do certain things if they encounter a robot or a person, she says.

More on Mobile Robots, AI for Logistics:

Follow your passion

Crowe congratulates Wise on receiving many tech industry honors. “It’s not something I strive for,” she says. “What’s important to me and I care about … I have a personal goal of wanting to move the robotics needle just a little bit.”

“It’s such a big field, if I can play even a small part in it, it’s a win, but there’s a lot of work to do,” Wise explains. “People take for granted ROS and PR2, but Willow Garage was a $100 million effort.”

Crowe asks Wise if she has advice for young women looking to get into robotics.

“If you’re passionate about something, you should do it,” she says. “Many people may try to shape your opinion of yourself and what you should be doing in terms of your career, in terms of life. The most important thing is to recognize what’s important to you and not be dissuaded from it.”

“It’s important for parents, mentors to build that confidence — that’s true for men and women,” Wise says.