Robots need to be nicer around humans

Nvidia is working on technology to make humanoid robots safer in workplaces, as the company pushes further into robotics.

Its Halos software, based on systems used in self-driving cars, is designed to help robots better understand what is happening around them. 

The goal is to enable humanoids to make quick, safer decisions when working near people, handing over items, or helping with physical tasks.

That is important because many robots today still rely on basic safety systems. If they detect a person nearby, they often slow down or stop completely. 

This reduces risk, but also makes them less useful in busy workplaces.

Nvidia is pairing Halos with its IGX Thor hardware, giving robots the computing power to analyse their surroundings and respond in real time. 

The system can also connect to external sensors, such as warehouse cameras, so machines can spot risks before they happen.

The company is also creating a safety lab where robot makers can test their machines before seeking regulatory approval.

The push comes as robotics becomes a bigger focus for AI companies.

Barclays estimates humanoid robots could generate $200 billion in revenue by 2035.

Nvidia is already working with companies like Agility Robotics, whose Digit robot has been used by customers including Toyota facilities in Canada.

Agility’s CTO Pras Velagapudi said humanoid safety is more complex than self-driving car safety. 

Cars mainly need to avoid contact, while humanoid robots may need to touch, lift or move objects near people. 

They also need enough strength to be useful, which makes safety harder to get right.

For now, the industry is starting with warehouses and logistics, where environments are more controlled. 

Retail, healthcare, construction and homes could come later, but those settings will be much harder to manage.

They’ll be nice to you before they replace you! - MV