Industries / Transport and vehicles / Automotive / Car rental
Cleaning and turnaround
Cleaning and turnaround is the work between one rental and the next: checking the returned car, refueling or recharging it, washing and vacuuming it, and parking it ready for the next renter.
4
Milestones achieved, from 2012 to 2023.
2
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
2012
Wash water used again
A quick turnaround area (QTA) is the prep yard where returned rental cars are fueled, washed and vacuumed. The design and construction standards for Seattle-Tacoma International Airport's rental car facility, dated October 2012, say its QTA car washes supply "a minimum of 90% recycled water" through a reclaim system. Vacuum, compressed-air and washer-fluid lines sit at the fuel islands, and wash runoff is pretreated before it reaches the sewer.
Credit
Port of Seattle
2018
Washing cars on rainwater
On 6 March 2018 the Port of Portland opened a $67 million QTA at Portland International Airport, three times the size of the one the rental companies built in 1989, which had five wash spaces and 24 fueling spaces. At peak shifts 220 workers can clean and fuel 430 cars an hour, at 72 fueling positions and 16 wash bays. An on-site well, a wash-water reclaim system and about 2.83 million gallons of rainwater a year, collected from the roof deck, were designed to let the washes run without city water, by the Port's account.
Credit
Port of Portland; Mackenzie (architects); Stantec (fuel and wash systems); Hoffman Construction
2020
Every car disinfected
In 2020 the large rental companies published COVID-19 cleaning programs, which the travel management company BCD Travel compared in July 2020. Enterprise Holdings' Complete Clean Pledge, announced in May 2020 for its Enterprise, National and Alamo brands, added disinfecting, "with particular attention to more than 20 high-touch points," to washing and vacuuming. Hertz's Gold Standard Clean set a 15-step routine that ended with a seal on the car. Avis's Safety Pledge used disinfectants from Reckitt, the maker of Lysol, and Europcar sealed the driver's door with a sticker and put sanitized keys in sealed envelopes.
LessonSurfaces proved a minor route for the virus. A 2021 risk model by A. K. Pitol and T. R. Julian put the risk of infection from one touch of a contaminated surface below one in a million at the infection rates then observed, and found that hand disinfection cut risk however common the disease was, while the benefit of surface disinfection depended on it.
Credit
Enterprise Holdings; Hertz; Avis Budget Group; Europcar
2023
A charging depot for rentals
From February 2023 Zeem Solutions charged Hertz's returned electric cars at a depot about a mile from Los Angeles International Airport: 70 to 100 cars a day, dropped off in waves of 10 to 20, with an average charge of about 40 minutes. Hertz used mostly DC fast chargers because it needed a total turnaround of about four hours, and the depot draws on a 10-megawatt power feed, by Auto Rental News's account after a site visit in August 2023.
LessonRecharging electric rental cars between rentals can need its own site, staff and a power supply that Auto Rental News compared to that of a town of more than 20,000 people.
Credit
Zeem Solutions (Paul Gioupis); Hertz
Watch for
Robot arms clean the cabin
Aseon Labs says its servicing pods use cameras and two industrial robot arms on floor tracks to plug in a charging cable, remove rubbish, store lost belongings in a locker, wipe surfaces and vacuum, in a cycle of about 30 minutes. When the arm meets a mess it cannot handle, it stands down and sends pictures to the depot. The pods are built for robotaxis, and Aseon says its demonstration program will establish what share of cars they can clean alone. Separately, Not a Tesla App reported in July 2026 that state filings list a "Cleaning Robot" at Tesla's Austin robotaxi hub.
Achieved whena fleet operator or an outside tester publishes the share of cabins cleaned with no person involved and how the next customers rated them.
Claimed by
Aseon Labs, October 2026
Chargers above the prep lanes
Portland International Airport has installed 10 dual-port ChargePoint fast-charging dispensers hung above its QTA. They can charge up to 20 rental cars at once while the cars are cleaned, inspected and prepared, and the Port says mounting them overhead cut the conduit and wiring needed and keeps them clear of moving cars.
Achieved whenthe Port or a rental company publishes how many charging sessions the system delivers a day and how often a charger is out of service.
Claimed by
ChargePoint and the Port of Portland, August 2026
Open
01
A clean-or-dirty cabin test
A public set of at least 1,000 photos of empty car cabins, taken the way a return agent or a camera pod would take them, labeled for rubbish, stains, pet hair, ash and belongings left behind. Score AI vision models on it: the share of problems found, and false alarms per 100 clean cars.
The briefCredit · unclaimed02
Open model of rental charging
Open-source software that takes an airport's rental returns by hour, the battery level cars come back with and the chargers on site, and reports how many chargers a QTA needs, the peak power in megawatts and how long renters would wait for a charged car. Check it against the published Dallas/Fort Worth result below, then run it for ten airports.
The briefCredit · unclaimed
Open 01 · the brief
A clean-or-dirty cabin test
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A public set of at least 1,000 photos of empty car cabins, taken the way a return agent or a camera pod would take them, labeled for rubbish, stains, pet hair, ash and belongings left behind. Score AI vision models on it: the share of problems found, and false alarms per 100 clean cars.
Why it is still open
Aseon Labs says its pods inspect each cabin with cameras before cleaning, but it publishes no accuracy figures and says its demonstration program will establish what it can handle. The nearest public set, from Nilusha Jayawickrama and colleagues at Aalto University (2023), holds 1,379 images from two research cars, taken by one fixed wide-angle camera above the rear seats and labeled for trash and valuables left behind; their model reached 89% accuracy across the item classes. It does not label stains, pet hair or ash, and it holds no photos taken by hand across many car models. TICaM, from the German research center DFKI (2021), labels people's actions and objects while a car is being driven, not an empty cabin.
roboticsandautomationnews.comjournalofbigdata.springeropen.comarxiv.org
Built with what exists
Phone cameras; volunteers' cars with staged rubbish, washable test soils and planted items; the open-source labeling tool Label Studio; object detectors such as Ultralytics YOLO and general AI models that read images.
First step
Photograph 50 cabins of different models, each with and without five planted items (a phone, a bottle, a wrapper, a toy and sunglasses), and label them.
Done looks like
1,000 labeled photos from at least 30 car models published under an open license, with stains, pet hair and ash labeled as well as belongings, and at least three models scored on the share of problems found and false alarms per 100 clean cabins.
Open 02 · the brief
Open model of rental charging
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
Open-source software that takes an airport's rental returns by hour, the battery level cars come back with and the chargers on site, and reports how many chargers a QTA needs, the peak power in megawatts and how long renters would wait for a charged car. Check it against the published Dallas/Fort Worth result below, then run it for ten airports.
Why it is still open
The National Renewable Energy Laboratory (NREL, since renamed the National Laboratory of the Rockies) modeled a fully electric rental fleet at Dallas/Fort Worth International Airport, returned at 20% charge and charged to 80% on 350-kilowatt chargers, and found it would need 264 DC fast chargers and draw 29.7 MW at peak. Its September 2025 fact sheet on the EVI-Rental tool gives no public code, and a GitHub search in November 2026 found no repository for it; the laboratory's open fleet simulator, HIVE, is built for ride-hail fleets, not rental returns.
Built with what exists
Python with the SimPy simulation library; airport passenger statistics from the US Bureau of Transportation Statistics; published charging curves for common electric models; an AI coding assistant.
First step
Code the Dallas/Fort Worth scenario from the fact sheet's assumptions and compare the charger count with 264.
Done looks like
Code published under an open license that reproduces NREL's Dallas/Fort Worth charger count within 10% from the fact sheet's stated assumptions, and results for ten airports giving chargers, peak megawatts and renter wait times.