Helping automated vehicles communicate with vulnerable road users

By Heather Mon­tague

As auto­mat­ed vehi­cles become more preva­lent, there are con­cerns about the lack of com­mu­ni­ca­tion between the vehi­cles and oth­er vul­ner­a­ble road users (VRUs) like pedes­tri­ans and cyclists. Dr. Pavlo Bazilin­skyy, Assis­tant Pro­fes­sor at TU Eind­hoven, research­es how auto­mat­ed vehi­cles inter­act with these vul­ner­a­ble road users in order to improve safe­ty and poten­tial­ly save lives.

An advo­cate of open sci­ence, Bazilin­skyy used a grant from the 4TU.ResearchData FAIR Data Fund (Spring 2021) to refine his dataset from a cou­pled sim­u­la­tor capa­ble of run­ning immer­sive sim­u­la­tions with dozens of peo­ple, and record­ing of a real traf­fic sit­u­a­tion from a portable sen­sor.

Human factors

With a human behind the wheel, it is com­mon for dri­vers to com­mu­ni­cate with pedes­tri­ans and oth­er road users via eye con­tact or hand ges­tures. “In the future there will be no dri­ver in the car and as a result the par­a­digm will shift,” said Bazilin­skyy. “If you’re in a self-dri­ving vehi­cle, maybe you’re read­ing a book, work­ing, look­ing down or even sleep­ing. A pedes­tri­an might not know if they should cross the street.”

His research focuss­es on how to com­mu­ni­cate with peo­ple out­side of the car by means of arti­fi­cial things like dis­plays or sounds. One method of doing so is exter­nal human-machine inter­faces (eHMIs) which are cur­rent­ly being devel­oped to com­mu­ni­cate inten­tion or pro­vide advice to pedes­tri­ans on when it is safe to cross. For exam­ple, if an autonomous car approach­es an inter­sec­tion it could show by means of a light what it intends to do, explained Bazilin­skyy.

Virtual simulation

To test design con­cepts like this, some of Bazilinkskyy’s research involves the use of a cou­pled vir­tu­al-real­i­ty sim­u­la­tor. Although it’s still a work in progress he said they are devel­op­ing a tool that can be used for research with mul­ti­ple humans in the same envi­ron­ment. “In my domain, exper­i­ments are most­ly done with one per­son or one car and the rest is sim­u­lat­ed,” he said. “But in real life if you imag­ine your­self at an inter­sec­tion in Man­hat­tan there could be 200 peo­ple around you, maybe even 2000.”

“In my domain, exper­i­ments are most­ly done with one per­son or one car and the rest is sim­u­lat­ed, but in real life if you imag­ine your­self at an inter­sec­tion in Man­hat­tan there could be 200 peo­ple around you, maybe even 2000.”

-Pavlo Bazilin­skyy, Assis­tant Pro­fes­sor, TU Eind­hoven

For Bazilin­skyy, it’s impor­tant to use a large and rep­re­sen­ta­tive sam­ple pop­u­la­tion in order to pro­duce robust, reli­able and repro­ducible research. To achieve this, he uses crowd­sourc­ing which enables respon­dents from all over the world to par­tic­i­pate in exper­i­ments.

Clean and shiny data

A long-time advo­cate of open sci­ence, Bazilin­skyy says that mak­ing data FAIR is intu­itive for him. In fact, he esti­mates that the data from his project was already approx­i­mate­ly 90% FAIR. But with a lot of code and data already in an open domain, he applied for the grant to try and make it a bit more organ­ised. “Through this fund­ing, which I’m real­ly grate­ful to have received, I’m more focussed on pre­sentabil­i­ty of my data,” he said. “Now I have made it a bit nicer, clean­er and shinier.”

Part of Bazilinskyy’s work also involved mak­ing a game tai­lored for aca­d­e­m­ic research, which is con­fig­ured to save data in an effi­cient, robust and syn­chro­nised way. So, he used his fund­ing to hire a free­lance pro­gram­mer to help devel­op the data a bit fur­ther and to struc­ture his GitHub repos­i­to­ry. “Before FAIR data I focussed more on the num­bers inside of my repos­i­to­ries and the actu­al data, I made sure that the data is easy to access and open,” he said. “But now I’m try­ing to step it up a bit and make it more pre­sentable.”

Keep it simple

Using the FAIR data fund, Bazilin­skyy also worked on a host pan­el for his sim­u­la­tor. That basi­cal­ly means that if some­one down­loads data through the GitHub link then it will be much eas­i­er to con­fig­ure and will also make it a bit more open. “I’m try­ing to make the sim­u­la­tor in such a way that you don’t need to be a sci­en­tist or a pro­gram­mer to use it,” he said. “The pub­lic can play around with it, even a kid could run the sim­u­la­tor. In fact, some­times I try to tai­lor my research towards edu­ca­tion in schools.”

FAIR from the start

For Bazilin­skyy, mak­ing his research open and FAIR has always been a pri­or­i­ty. “I nev­er even thought about any oth­er way to approach things,” he said. “We live in the 21^st cen­tu­ry and it’s quite easy to make your things open if you spend a bit of time. And if you think about it from the first day of a project, it can be done with­out much work at all.”

“I think sci­ence needs to be open by default because we do sci­ence for the ben­e­fit of the whole soci­ety, not just for our own aca­d­e­m­ic career.”

- Pavo­lo Bazilin­skyy, Assis­tant Pro­fes­sor, TU Eind­hoven

There’s also the mat­ter of cred­i­bil­i­ty for Bazilin­skyy. “95% of the time when I look at work by some­one else, if I don’t see data and code my lev­el of trust in that work decreas­es. I think sci­ence needs to be open by default because we do sci­ence for the ben­e­fit of the whole soci­ety, not just for our own aca­d­e­m­ic career.” He also not­ed that hav­ing done a PhD in human fac­tors, he knows that humans are not per­fect. “If I see data, I trust that way more than humans these days.”

Future goals

So far, Bazilin­skyy says they have pub­lished one paper that result­ed from the cou­ple sim­u­la­tor. “It was a mas­sive project I did with my stu­dents, but it shows the pow­er of this tool,” he said. “My goal is by the end of this year to make a major new release of the tool and that will be the main deliv­er­able.” Dur­ing his PhD and post­doc at TU Delft, Bazilin­skyy also vol­un­teered as a data cham­pi­on, a group of researchers at the fore­front of inno­va­tion in top­ics such as research data, code man­age­ment and open sci­ence. And he hopes to con­tin­ue this work by help­ing start up a data cham­pi­ons pro­gramme at TU Eind­hoven.