Home Household chores Toyota works on robots for complex situations – Like housework

Toyota works on robots for complex situations – Like housework


Toyota Research Institute

Robots have come a long way but still face incredible challenges when it comes to tasks and environments that seem fairly mundane to humans. This is what makes the video below from the Toyota Research Institute (TRI), which shows robots solving complex tasks in unstructured home environments so fascinating.

“Our goal is to create robotic capabilities that amplify, not replace, human capabilities,” said Max Bajracharya, vice president of robotics at TRI. “Training robots to understand how to operate in home environments poses particular challenges due to the diversity and complexity of our homes where small tasks can add up to big challenges. “

The new video, which was released to coincide with National Selfie Day, is a bit silly, but the strides are significant on the ground. The roboticists at TRI demonstrate here that they have trained robots to understand and operate in situations that completely confuse most other automation systems, especially when it comes to recognizing and responding to transparent surfaces and reflective, a major obstacle for machine vision. As a statement from the TRI explains, since most robots are programmed to react to objects and geometry in front of them regardless of the context of the situation, they are easily fooled by a glass table, a shiny toaster. or a transparent cup.

To date, this has kept robots largely confined to strict job designations most often performed in predictable environments like factories and warehouses. Bringing robots into the real world – what is happening most dramatically right now in the autonomous vehicle world – is much more complicated, requiring these complex and potentially dangerous systems to constantly take into account and confront the unexpected, this which carries enormous risks of failure.

“To overcome this, TRI roboticists have developed a new training method to perceive the 3D geometry of the scene while detecting objects and surfaces,” Bajracharya continued. “This combination allows researchers to use large amounts of synthetic data to train the system.” Using synthetic data also reduces the need for time-consuming, costly, or impractical data collection and labeling.

The research is part of TRI’s mission to develop technologies for active vehicle safety and automated driving, robotics and other human amplification technologies. Veteran roboticist Dr Gill Pratt leads the TRI.