The situation
The wearable device stays in constant communication with the safety system. When that communication drops, the system does the safe thing and stops robotic operations.
Before this program, getting operations running again after a dropout required a manual reset. That kept robots stopped for much longer than the dropout itself.
What I did
I owned the program on the product side and worked closely with the firmware and wireless engineering teams who built the recovery mechanism. When communication drops now, the system pauses operations for a short time to let the connection restore. Once it's back, robotic operations resume on their own, with no human intervention.
We shipped it across the whole fleet of 15,000 devices.
What happened
Short communication losses no longer turn into long stoppages waiting for a manual reset. Automated recovery unlocked about $8M in value.
A related project
Around the same time, I reviewed the spare parts kits we send to new buildings against expected failure rates, and helped build fault analysis tooling so field teams can replace parts before they fail. That cut launch spare and consumable costs by about 70%, or $2M over three years.
Technical details of the devices, the wireless system and the recovery logic are left out on purpose. I'm happy to go deeper on the approach in conversation.