The metric will be "cost of time spent cleaning" / "cost of sending out a dirty car." If this ratio is > 1 the "recommended cleaning schedule" will be the lowest priority item in the entire fleet.
The cost of cleaning goes way down when you do it in bulk at a service center rather than individual lyft/uber drivers trying to do it. A standardized car also helps.
If you can incorporate the cost of large horizontal demand spikes into the off hours and you can find cheap enough labor to fill it, perhaps.
The time spent travelling out of service, in cleaning, and back into service are all lost opportunities. Hopefully you can clean a very large number of cars in a very short period of time.
The cars may be standard. The messes, obviously, will not be.
The have multiple charging centers strategically placed throughout the city. These seem to currently only have security guards there. The logistics of having cleaning staff there and trying to match their schedule to expected charging times is probably not very difficult but also not very reliable either.
The win they do have, that I did not consider is, they have cameras _in_ the car. So visible cleanliness is something they can manually check before and after the rid and schedule for service if required; however, it currently seems that this requires the vehicle to go to the larger centralized maintenance facility, which I guessing takes quit a bit more time than the auxiliary charge only lots.
Not trying to be super pessimistic, but mixing distributed autonomous operations with centralized manual service, especially in an urban environment, seems fraught with novel challenges.
The cars will all charge overnight. Clean them there. Getting to and from the charging stations / cleaning depots is roughly free. It’s just the electricity which is like 5 cents a mile roughly. Utilization at night is really low anyways so not much opportunity cost.
The overall point being way more efficient to clean than an Uber.
Once you have a large enough fleet of roughly the same shape, it starts to make sense to build some automated cleaning thing. I'm picturing like, a long tube with vacuum ports in it that can just be shoved in by a robot arm to vacuum each seat. Unless there's a wet mess, that'll basically get it 90% of the way there.
I think we'll see such self-driving taxi interiors optimized for staying clean and then ease of cleaning. At the limit, think like a stainless steel kitchen that can just be sprayed down. Or these sort of self-cleaning public bathrooms: https://www.youtube.com/shorts/z81KtV9w5fo?feature=share
