The UTC Chair “Smart and Trans­par­ent Sur­faces for the Auto­mo­bile of the Future (SITAF)”, launched in Novem­ber 2019, is the first such chair to be fund­ed by the UTC Inno­va­tion Foun­da­tion. It is an indus­tri­al-inten­sive chair devot­ed to ”smart” glass sur­faces con­duct­ed in close rela­tion­ship with the com­pa­ny Saint-Gob­ain Seku­rit based in Compiègne.

The SITAF chair deals with var­i­ous issues relat­ed to the devel­op­ment of the auto­mo­tive cock­pit of the future and more specif­i­cal­ly with the role of glaz­ing in it. “When we talk about the car of the future, we imag­ine an autonomous vehi­cle with inno­v­a­tive glaz­ing. The tech­no­log­i­cal inno­va­tions imple­ment­ed with­in this Chair are pro­duced by Saint-Gob­ain. Through these inno­va­tions, new ques­tions arise and will lead us, as research sci­en­tists and engi­neers, to fram­ing new prob­lems and propos­ing new tools to solve them. It is our role to trans­late the prob­lems posed by indus­try into sci­en­tif­ic ques­tions”, explains Del­phine Brancherie, a lec­tur­er-cum-research sci­en­tist at UTC’s Rober­val lab­o­ra­to­ry, who has been in charge of sci­en­tif­ic coor­di­na­tion since 2020. The research work was car­ried out through two the­ses by PhD stu­dents super­vised by lec­tur­er-cum-research sci­en­tists from the Rober­val and Heudi­asyc laboratories. 

In the Rober­val mechan­i­cal engi­neer­ing lab­o­ra­to­ry, an acoustic the­sis is being con­duct­ed by Alexan­dre Berthet, aged 27; it relates to the devel­op­ment of super-ele­ments for the dynam­ic sim­u­la­tion of vis­coelas­tic mul­ti­lay­er struc­tures. “The aim is to respond to an impor­tant prob­lem raised by indus­try, name­ly con­fi­den­tial­i­ty. The sub­sys­tem we are inter­est­ed in here is the wind­screen, made up of an assem­bly of glass lay­ers and an inter­me­di­ate poly­mer lay­er, the vis­coelas­tic behav­iour of which must remain con­fi­den­tial in order to be able to exchange numer­i­cal mod­els with car man­u­fac­tur­ers”, he explains. “With­in the frame­work of the Chair, the devel­op­ment of super-ele­ments for the wind­screen was the sub­ject of my the­sis work, which began in Novem­ber 2019 and was defend­ed last Decem­ber, the main objec­tive being to enable the exchange of glaz­ing mod­els while ensur­ing the non-dis­clo­sure of con­fi­den­tial information.”

Deciphering encoded data

Antoine Mon­tesinos’ intern­ship dealt with the opti­mi­sa­tion of a method aimed at reduc­ing the numer­i­cal mod­el of a car wind­screen. The post-doc of Christophe Lan­glois, 29, is enti­tled “Exper­i­men­tal iden­ti­fi­ca­tion of the dynam­ic behav­iour of a car wind­screen and reverse engi­neer­ing process from a reduced mod­el”. It was con­duct­ed from Feb­ru­ary 2022 to Feb­ru­ary 2023. “My work con­sist­ed of detect­ing what is hid­den, in find­ing the mechan­i­cal prop­er­ties of the encrypt­ed wind­screen mod­el. This is called reverse engi­neer­ing. The pur­pose here is to eval­u­ate the pos­si­bil­i­ty of iden­ti­fy­ing the mechan­i­cal char­ac­ter­is­tics of con­fi­den­tial mate­ri­als from the sup­ply of the wind­screen mod­el alone.

For the indus­tri­al­ist, con­tribut­ing to the gen­er­a­tion of a state of the art on reverse engi­neer­ing meth­ods applied to lam­i­nat­ed struc­tures con­tain­ing a vis­coelas­tic mate­r­i­al, but also of exper­i­men­tal results from the vibra­to­ry analy­sis of a lam­i­nat­ed struc­ture and com­par­ing them with numer­i­cal results is very inter­est­ing. “We are very demand­ing on the per­for­mance of our glaz­ing. They must be the best to ensure the safe­ty of the occu­pants and pro­vide them with opti­mum com­fort in the car. To achieve this, we want to co-devel­op inno­v­a­tive solu­tions with car man­u­fac­tur­ers while retain­ing con­trol of the mate­ri­als and tech­nolo­gies used in the glaz­ing. With­in the frame­work of the chair with the UTC, we have there­fore cho­sen to devel­op an encryp­tion method that will enable us to make the essen­tial prop­er­ties of our glaz­ing avail­able to car man­u­fac­tur­ers. The research work on deci­pher­ing data aims at ver­i­fy­ing the robust­ness of this encryp­tion method and thus of our tech­no­log­i­cal mas­tery. The chair with the UTC is very impor­tant for us beyond the sci­en­tif­ic work, as it puts us in touch with a pool of tal­ent and a whole ecosys­tem of skills. This pro­vides a great oppor­tu­ni­ty for Saint-Gob­ain,” says Jérôme Gilles, deputy direc­tor of Saint-Gob­ain Recherche Com­piègne. The UTC Rober­val lab­o­ra­to­ry had already forged links with the Saint-Gob­ain Group on oth­er projects involv­ing glass and metal.

“The research work is bear­ing fruit. We know the Saint-Gob­ain com­pa­ny well. For doc­tor­al stu­dents, a the­sis is also train­ing, for exam­ple in numer­i­cal mod­el­ling,” con­cludes Jean-Daniel Cha­zot, head of the acoustics and vibra­tion for engi­neers in the mechan­i­cal engi­neer­ing depart­ment of the Rober­val laboratory.

In the beginning, there is the UTC Foundation for Innovation

The UTC Foun­da­tion was the cat­a­lyst that allowed to bring togeth­er indus­tri­al­ists and UTC aca­d­e­m­ic skills in a part­ner­ship approach favourable to build projects and explore tomor­row’s tech­no­log­i­cal solu­tions along­side the actors of inno­va­tion. The Foun­da­tion, cre­at­ed in 2018, finances and sup­ports this Chair. “The Chair on intel­li­gent glass sur­faces was the first pro­gramme sup­port­ed by the Foun­da­tion. It is a strong mark­er of a priv­i­leged and last­ing part­ner­ship with the Saint-Gob­ain group, a his­tor­i­cal play­er in the region for its glass activ­i­ties. It is also a world actor in inno­va­tion, essen­tial both in research to advance on the tech­no­log­i­cal stakes of tomor­row, and in train­ing: many stu­dent-engi­neers of the UTC car­ry out their intern­ship there, and many grad­u­ates join the group”, under­lines Sylvie Lemon­nier-Morel, Gen­er­al Sec­re­tary of the UTC Foun­da­tion for Inno­va­tion. The voca­tion of the Foun­da­tion is indeed to raise funds, to pro­mote by its actions the devel­op­ment of the UTC and to sup­port inno­v­a­tive projects which are mean­ing­ful for the society.

An “augmented reality” thesis at Heudiasyc

In March 2023, Bap­tiste Wojtkows­ki, 26 years old, will sub­mit his the­sis enti­tled “Val­i­da­tion of mul­ti­modal
for the mul­ti­sen­so­ry cock­pit in an immer­sive vir­tu­al envi­ron­ment in the case of dri­ver inat­ten­tion.” He was super­vised by Indi­ra Thou­venin, a lec­tur­erre­search sci­en­tist from the UTC-Heudi­asyc lab­o­ra­to­ry and Veron­i­ca Teichrieb from the Voxar Labs in Brazil, but also by Pauline Mer­veilleux with the Saint- Gob­ain Group. “Indus­tri­al­ists have realised that ful­ly autonomous vehi­cles will not be pos­si­ble in the near future. These vehi­cles still need to be taught how to han­dle com­plex sit­u­a­tions such as a lack of road mark­ings, faulty vehi­cle sen­sors, how to accel­er­ate and slow down and, above all, how to inform the dri­ver of the rea­son for tak­ing back con­trol, in short, to ensure that the dri­ver under­stands,” sum­maris­es the com­put­er engi­neer­ing stu­dent. It was main­ly these last two aspects that inter­est­ed us. I have real­ly enjoyed work­ing on this sub­ject for the past three years in a lab­o­ra­to­ry with a high qual­i­ty sci­en­tif­ic team. The Region co-financed this the­sis, via Fed­er funds, as did Alexan­dre Berthet’s acoustics-based thesis.

The Heudiasyc laboratory and the multisensory cockpit

Three ques­tions to Indi­ra Thou­venin, a teacher-researcher from the UTC-Heudi­asyc lab­o­ra­to­ry who co-super­vised with Veron­i­ca Teichrieb from the Voxar Labs in Brazil, a the­sis enti­tled “Val­i­da­tion of mul­ti­modal feed­back for the mul­ti-sen­so­ry cock­pit in an immer­sive vir­tu­al envi­ron­ment in the case of dri­ver inattention”.

What is the research con­text that led to this thesis?

It is known today that the autonomous vehi­cle presents too great a risk, and that the pres­ence of the dri­ver is required in cer­tain com­plex or ambigu­ous sit­u­a­tions. This is known as the high­ly auto­mat­ed vehi­cle (HAV), which allows the dri­ver to take back con­trol of the vehi­cle. Regain­ing con­trol is a com­plex cog­ni­tive task involv­ing a trans­fer of knowl­edge from the automat back to the dri­ver. We there­fore pro­pose aids with aug­ment­ed real­i­ty dis­played on the wind­screen. These aids must adapt to the sit­u­a­tion but also to the dri­ver’s atten­tion (or inat­ten­tion), oth­er­wise the dri­ver will no longer look at them or will be dis­turbed by these per­ma­nent dis­plays. We there­fore need to describe the dri­ver’s state, with sen­sors but also a mod­el that takes into account his state, his actions and his gaze in order to gen­er­ate feed­back at the right time and in the right place on the wind­screen. Our semi-autonomous dri­ving sim­u­la­tor allows us to test sce­nar­ios, to dis­play adap­tive feed­back in a vir­tu­al envi­ron­ment and to check that this feed­back helps to regain con­trol, on a motor­way for example.

So what is your sci­en­tif­ic con­tri­bu­tion to the indus­tri­al real­i­ty of the Saint-Gob­ain Group?

Our sci­en­tif­ic con­tri­bu­tion is to pro­pose a mod­el that will be use­ful for intel­li­gent glaz­ing because the sur­face of the wind­screen becomes an aug­ment­ed real­i­ty inter­face. Many projects already exist on this theme and are indus­tri­alised. But what we are pro­vid­ing are descrip­tors of the dri­ver’s state, for the motor­way dri­ving sit­u­a­tion. We are also pro­vid­ing proof that the mod­el is use­ful, because we are test­ing it in a sim­u­la­tor and we have results on the inter­est of our approach. UTC-Heudi­asyc is a joint research unit — between the CNRS and the UTC. This project is part of the “Auton­o­my of mobile robots in inter­ac­tion with humans” theme. The objec­tive is to con­trol and nav­i­gate an autonomous sys­tem in an open and com­plex envi­ron­ment and to study how the robot can gain auton­o­my in the pres­ence of the human operator.

How is the the­sis “Val­i­da­tion of mul­ti­modal feed­back for the mul­ti­sen­so­ry cock­pit in an immer­sive vir­tu­al envi­ron­ment in the case of dri­ver inat­ten­tion” rel­e­vant to the auto­mo­bile of the future?

The VHA becomes a mul­ti-sen­so­ry cock­pit, a sort of cab­in equipped with sen­sors, visu­al and audio dis­plays, or even allow­ing the dri­ver to feel the touch of the sys­tem. This can be com­pared to the cab­in of an air­craft in which the autopi­lot is in con­trol. But the pilot has to stay alert and inter­vene quick­ly in case of a prob­lem. Humans can­not quick­ly check all the sen­sors, and need so-called adap­tive inter­faces, i.e., they adapt to their state at a giv­en moment. These are mul­ti­modal inter­faces because they are also in visu­al mode, sound mode and hap­tic mode, which is the sense of touch.