Planning Uptime for a Hospital Delivery Robot Fleet

What uptime does a hospital delivery fleet need?

A hospital delivery fleet needs enough ready robots, at the hours it is busiest, to clear its real queue with margin to spare. No public standard sets that number for hospital delivery, and we did not find a verified uptime or downtime rate for Moxi or any other hospital delivery fleet. Planning starts with workload, and the best public workload figure comes from Children's Hospital Los Angeles.

According to PYMNTS, citing an August news release from the hospital, its Moxi robots had completed more than 40,000 deliveries, representing more than 16,000 hours of work. Those are hospital-reported totals. They describe demand the fleet absorbed, not how often the robots were ready to take it.

What do the CHLA figures actually show?

Divide 16,000 hours by 40,000 deliveries and the result is roughly 24 minutes per delivery. That is our arithmetic, and it lines up with an older CHLA blog post in which pharmacy chief Carol Taketomo said staff recoup 20 to 30 minutes per delivery. A missed run hands that work straight back to a nurse or pharmacy tech.

The same older post reported more than 2,500 deliveries in just over four months, with 25 to 30 deliveries per day. Two robots ran around the clock and used only about half their capacity. That half capacity detail matters more than the headline. A fleet sized to its average day has headroom for a bad afternoon, while a fleet sized to its average with no spare has none.

CHLA has since grown its fleet from two robots to three. According to Diligent's rollout material, utilization rose more than 10 percent in the second quarter alone. Those are company-issued figures, so read them as direction rather than audited results.

How should you set an uptime target?

Start with availability: the hours a unit was ready to take work, divided by the hours it was scheduled to run. Then set the target against the demand curve, not a single number lifted from a brochure. Vendors can quote a battery figure, but the queue only experiences availability.

Work the target in four steps. Choose the operating window, since a fleet on a round the clock schedule carries a different burden than one on a day shift. Count requests by the hour, not by the day. Work out how many units must be ready at the busiest hour to clear the queue inside your response standard. Then set the availability target so the fleet holds that count with a margin, and test it for a month before you trust it.

Battery claims feed the plan without setting it. Diligent describes Moxi 2.0 as running up to 9 hours on a single charge, with 30 percent faster charging, enabling up to 18 hours of daily operation. Charging time comes out of the schedule, so a unit that charges at noon is not available at noon, whatever the spec sheet says.

What belongs in a service log?

A log that only records that a robot went down will not explain why, and it will not help you set the next target. Capture the fields below for every fault, and review them weekly with the people who staff the queue.

• Timestamp and unit identifier
• Fault type and the code the robot reports
• Location and route segment at the moment of the fault
• Battery level at dispatch and at return to service
• Network state: onboard only, wireless link, or cellular fallback
• Recovery method: remote fix, onsite reset, or a staff member moving the unit by hand
• Time from fault to return to queue
• Deliveries affected, and whether each was delayed, queued again, or cancelled

What does a backup route look like when one unit is down?

When a unit drops out, the first decision is which deliveries move, and that decision should be written down before any fault happens. Pharmacy and time sensitive specimen runs go first. Routine supply moves wait, and the charge nurse or unit clerk hears in advance that the wait is expected.

Next, rebalance the queue across the units still running. If you run three robots and one fails, the remaining two must absorb the peak, which is why the peak count from the earlier step decides whether a single failure is a delay or a gap. A fleet of two running near full load has no such cushion.

A backup route also means physical redundancy. Map at least one alternate elevator and corridor for each main path, and confirm the robot can take it without staff help. Set an escalation rule as well: if a unit stays out past the threshold you chose, a person with authority to pull staff or add capacity gets the call. Diligent's Moxi 2.0 material says the robot improved autonomy and recovery and resolves more edge cases independently. That should reduce how often the call comes in, but it does not remove the need for one.

How do network and charging shape the plan?

Autonomy that runs onboard keeps a robot moving when the wireless link drops, which is why network state belongs in the log. Diligent says Moxi 2.0 runs its safety and autonomy functions onboard, can complete tasks without a wireless connection, and falls back to cellular where hospital networks have dead zones. Check that claim on your own floors, especially older wings and corridors with thick construction, before you count on it.

Where does Service Robot Co. fit in?

Service Robot Co. is a full service commercial robot integrator for US businesses, and we are OEM neutral. We pick the right robots across manufacturers, then finance, deploy, integrate, train, and service every unit through a nationwide US engineer network, so one vendor covers the whole lifecycle. We lease robots only, on a lease running three years, with one fixed payment quoted after a site assessment.

That site assessment is where the demand curve above gets measured: peak hours, route lengths, elevator waits, and network coverage. A target built from the actual floor plan needs far less guesswork than one copied from a vendor slide.