Engineering notes
How it's built and how it's tested.
AI
What 538 positions and 1 millisecond taught us about on-robot AI
An overnight simulation run: a cloud model labels, a small model learns, and answers come back about 3,000 times faster. What it proved, and what not.
Safety
No AI in the control loop: how a MoLogica robot stays safe
Six safety rules built into the architecture, from a whitelist that refused 'chase the intruder' to a watchdog that stops the robot in 0.94 s.
Method
How we test: the proof gate
One script, the full scenario, end to end. How the proof gate works, what passed in a simulated depot, and why it ran on two robot bodies.
Hardware
CPU, GPU and where each earns its keep on a robot
Physics on fast CPU cores, images and models on the GPU, and why the trained inspection model ran in about 1 ms on a GPU against 15 ms on a CPU.
Guide
Edge AI
On-robot AI answers in milliseconds, keeps working with no network, keeps footage on site and costs nothing per call. How it gets there, and CPU versus GPU explained.
Guide
Hardware-agnostic robotics
Robot bodies are becoming commodities. The value is in the software that plans, supervises, inspects and learns, running on whichever body does the job best.
Ready to see robots patrolling your site?
Before any hardware is bought, we can build a simulated copy of your site and show the fleet patrolling it at night while you configure the checks in plain language.