Three things I read this week on Physical AI, with a short note on each and a link to the source. The notes are mine. The reporting is theirs, so go and read it.
A warehouse grid that doubled
THG Fulfil has expanded the AutoStore system at its Omega warehouse in Warrington, UK, from the original 120,000 bins to 240,000. The grid now runs 184 robots, 84 more than at launch, with 64 ports where items come out. Storage capacity goes from 4.2 million units to 8.4 million. Robotics & Automation News
My note: A cube of bins with robots on top is a very tidy world. Fixed grid, known positions, nothing wanders in. That is why it grows by addition: more bins, more robots, same rules. The harder physics starts at the port, where one item has to come out of the bin.
Robots on building sites, a conference report
Robotics & Automation News reports from the first Groundworks Summit, a construction robotics conference held on 11 November at UCL Here East in London. Robotics companies and contractors discussed what it takes to bring automation into everyday building work. One speaker put the sector’s productivity growth at 1 per cent, and the founder of Monumental said its bricklaying robot can place 800 bricks a day. Robotics & Automation News
My note: A building site is the opposite of a factory. The floor is mud, the map changes every day, and the product is built once. A machine that works there needs perception, control and mechanics that agree with each other. That is also most of my answer to what a Physical AI engineer does.
Cleaning robots that take directions from the ceiling cameras
Richtech Robotics is showing its Intelligent Cleaning System at ISSA Show North America in Las Vegas, 17 to 19 November. The system connects a building’s existing indoor CCTV cameras to the company’s Dust-E cleaning robots: the cameras identify areas that need cleaning, and a robot is sent there. Robotics & Automation News
My note: The eyes do not have to ride on the body. A fixed camera sees the whole floor. The robot sees the metre in front of it. Put the two together and the robot stops searching and starts going. Where the computing sits then matters for how quickly it reacts, which is the point of edge AI.
More: all Physical AI news notes · what Physical AI is.
