Robots require Manual, Human Learning to Ease Human Lives

Toyota Research Institute (TRI), on September 30, disclosed a few projects, including a ceiling-mounted robot that may help us one day with household chores. This system is one example of how TRI envisions the future of robotics and artificial intelligence (AI). The company is focusing on robotics and AI technologies for "amplifying, rather than replacing, human beings," says TRI CEO Gill Pratt.

In other words, Toyota is keen to develop robots, not for convenience or to do human jobs. Instead, it aims to allow people to continue to live and work independently with the growing age.

The 20-minute-video by Toyota depicts various its vision of robotics 15 to 20 years from now. It also displays multiple scenarios, "where the application of robotic capabilities is enabling members of an aging society to live their lives independently despite the challenges that getting older brings." The video explains TRI's perspective on how robotics will integrate with humans in our daily lives over the next two decades.

For TRI, Both reliability and cost are product development tasks. "The hardware that we use in the laboratory for doing experiments, we don't worry about cost there; instead, we focus on the product," shares Mr. Pratt. However, "what research we do, we have to think about if the research is possible or successful for it to end up in a product that has a reasonable cost."

It is because "if a customer can't afford what the company comes up with, perhaps it has some educational value, but it will not make a difference in their quality of life in the real world," elaborates Mr. Pratt. "So, we think about cost very much from the beginning."

TRI's vision is also the same for reliability. The company works hard to make its control techniques robust to broaden variations in the ecosystem. For instance, Russ Tedrake is manipulating dishes in a sink and a dishwasher, both in physical testing and in simulation; TRI is doing thousands and now millions of different experiments to ensure that it can handle the edge cases and it works over a wide range of conditions.

TRI is putting a massive amount of effort into bringing robotics out of the age of performing demonstrations. There is a history of robotics where things were not reliable for some time. Therefore, it would catch the robot succeeding just once and then display that video to the world, and people would get the misimpression that it worked all the time.

TRI is also exploring how to use light so that people can be ambiently aware of one another across distances. It experimented with a prototype called 'glowing orb' to keep people connected. Although this prototype didn't work out, it explored different modalities for keeping people in touch.

Robotics Assistance in Grocery Shopping

TRI's researchers worked on another prototype and discovered that grocery shopping is an integral part of life, and for many older adults, getting groceries always delivered is not necessarily the right answer. "Getting up and getting out of the house keeps you physically active, and many people prefer to continue doing it themselves," says Steffi Paepcke, senior UX Designer at Toyota Research Institute. "But it can be challenging, especially if you're purchasing heavy items that you need to transport."

"We had a prototype that assisted with grocery shopping, but when we focused on Japan, we found out that the inside of a Japanese home needs to stay inside, and the outside needs to stay outside," cites Steffi. "A robot that traverses between both domains is probably not the right fit for a Japanese audience, and those were some valuable lessons for us."

Robotics to Ease Humans Lifestyle

TRI is exploring things such as the gantry robot as a long-term vision to make it human friendly to make changes to households. "We don't want to give people a robot and assume that they're not going to change anything about their lifestyle," states Max Bajracharya, Vice President of Robotics at Toyota Research Institute. "We have evidence from people who use automated vacuum cleaners that people will adapt to the tools you give them, and they'll change their lifestyle."

"So we want to think about what is that trade between changing the environment, and providing people robotic assistance and tools," he adds.

TRI's present gantry system is a prototype and requires significant infrastructure. In the prototype phase, researchers are trying to understand whether the gantry system is worth to bypass navigation challenges and coming up with its pros and cons. Before launching, they are reviewing if it is the right approach to solve problems.

TRI aims to amplify people, and achieving this will require robots to be in a loop with people in some form. Using people in a slow loop with robots, like teaching them or helping them when they make mistakes, gives a robot a crucial advantage over one that has to do everything with 100% accuracy. In other words, in unstructured human environments, robots will encounter corner cases and require learning to adapt. People will likely play a pivotal role in helping robots learn.