Home / Columns / Using algorithms to imagine a future world for Uber’s air taxis
Claudia d'Ambrosio
π Science and technology π Mobility π Energy

Using algorithms to imagine a future world for Uber’s air taxis

Claudia D'Ambrosio, CNRS researcher at the Computer Science Laboratory of the École Polytechnique (LIX*)

Uber’s fleets of elec­tri­cal air­craft are set to take to the skies in Dal­las, Los Ange­les and Mel­bourne with­in the next three years. Dubbed UberAir and based on eVTOL (elec­tron­ic ver­ti­cal take-off and land­ing) vehi­cles, the ser­vice promis­es to see us into the next decade with urban ‘on-demand avi­a­tion’ solu­tions being released by 2023. Instead of tak­ing a long-dis­tance taxi, Uber will be offer­ing pas­sen­gers trans­port via bite­size zero-emis­sions, elec­tron­ic air­craft. Even though the vehi­cles them­selves are set to cost 20 times as much as nor­mal cars, Uber says that the eVTOLs will be cheap­er for com­muters over the long-term. What’s more, they are set to pro­vide an effec­tive way for com­muters to share journeys. 

Whilst there may be rea­son to believe that the idea is a far-fetched pipe dream, in real­i­ty the tech­nol­o­gy is just around the cor­ner. Some of the chal­lenges still to over­come include find­ing bat­ter­ies that are both pow­er­ful and light­weight enough, pilot train­ing and obtain­ing the nec­es­sary cer­ti­fi­ca­tion and autho­ri­sa­tions. But these are tech­no­log­i­cal and regle­men­tary bar­ri­ers Uber hope to iron out soon enough. A big ques­tion, though, is how to antic­i­pate the infra­struc­ture require­ments and devel­op urban air traf­fic man­age­ment tools to sup­port such a sys­tem. For that, Uber teamed up with Insti­tut Poly­tech­nique de Paris in 2019 to cre­ate an inter­na­tion­al aca­d­e­m­ic research chair on “Inte­grat­ed urban mobil­i­ty”, which I hold. Our aim is to antic­i­pate the needs of the urban infra­struc­tures around eVTOL, help­ing Uber to bet­ter under­stand how the sys­tem can fit into urban landscapes. 

Tomorrow’s needs imagined

In down­town Los Ange­les, for exam­ple, there are already at least 40 pre-exist­ing heli­pads. How­ev­er, most are pri­vate­ly con­trolled. Could they offer a solu­tion? If so, what are the pow­er require­ments? In my research we design algo­rithms that could help Uber scout out the dif­fer­ent options, so that the com­pa­ny can make con­cert­ed, strate­gic deci­sions. This is what oper­a­tions research algo­rithms do: help peo­ple make ‘opti­mal decisions’. 

To pro­vide an anal­o­gy, think about the exam­ple of Google Maps; the soft­ware uses an algo­rithm to cal­cu­late the fastest route between two points whilst tak­ing into account real-time infor­ma­tion like traf­fic or road­works. The deci­sion prob­lem the pro­gram solves is referred to as ‘short­est path’. Put sim­ply, it breaks the jour­ney down into small­er seg­ments and cal­cu­lates the quick­est route per seg­ment by ask­ing “which is fastest?”. In my work as the research chair, I design sim­i­lar algo­rithms based on the dif­fer­ent pos­si­ble sce­nar­ios that will be account­ed for. A fleet of eVTOL air­craft will bring with it a whole new set of con­straints that we have nev­er seen before. The world is like­ly to look dif­fer­ent in 2023! 

Future choic­es anticipated

Such algo­rithms pro­vide a way of study­ing effects of deci­sions made in this future world. What fac­tors should be tak­en into account? What com­po­nents will need to be installed? What will the ener­gy require­ments be for eVTOLs are elec­tric and how will infra­struc­tures need to be adapt­ed to sup­ply it? 

In the ear­ly stages, eVTOLs will be pilot dri­ven and mon­i­tored via con­trol tow­ers. But even­tu­al­ly the goal is to make them dri­ver­less. In my research, I seek to under­stand how the air­borne vehi­cles will be able to respond to unex­pect­ed events. If an emer­gency vehi­cle needs to take flight, the tra­jec­to­ry of the eVTOLs will have to be altered rapid­ly as a consequence. 

We also need to be sup­ple enough to change in response to shift­ing needs. The pan­dem­ic is just one exam­ple of a chal­lenge the project will face. If we con­sid­er ground trans­porta­tion, for exam­ple, cur­rent­ly Uber­Pool – the ser­vice which allows pas­sen­gers to share vehi­cles – is not an avail­able option. With the cur­rent risk of infec­tion, trav­ellers can­not share taxis. But then again, this will prob­a­bly change by 2023. What may stay with us though is the change in habits. In Paris, bike users went up by around 60 % after the con­fine­ment. We can fac­tor this into our sim­u­la­tions. Changes in lifestyles are an inter­est­ing avenue to explore.

Contributors

Claudia d'Ambrosio
Claudia D'Ambrosio
CNRS researcher at the Computer Science Laboratory of the École Polytechnique (LIX*)

Claudia D'Ambrosio works on theoretical and practical problems in operations research. At the Computer Science Laboratory of the École Polytechnique (*LIX: a joint research unit of the CNRS, École Polytechnique - Institut Polytechnique de Paris), she studies mathematical and algorithmic tools for decision making. She is a research director at the CNRS, responsible for the chair with Uber "Integrated urban mobility" and a lecturer at the École polytechnique.