Actus Monde

Cycle AI: Utilizing artificial intelligence to ensure safer cycling in cities

ecf.com (en anglais) : comment l'Intelligence artificielle peut mesurer la sécurité des routes pour les cyclistes

Cycle AI wins first place at the first ever Euro­pean Cyclists’ Fed­er­a­tion Hackathon, orga­nized in asso­ci­a­tion with the Euro­pean Insti­tute for Inno­va­tion and Technology’s (EIT) Urban Mobil­i­ty depart­ment. Their pro­posed solu­tion will encour­age more cycling in our cities by using arti­fi­cial intel­li­gence (AI) tech­nol­o­gy to assess road safe­ty lev­els for cyclists.

What is it?
Ensur­ing cyclists’ safe­ty is one of the key obsta­cles for encour­ag­ing more cycling in our cities. At the inau­gur­al ECF Hackathon on Novem­ber 13th, 2020, Cycle AI pre­sent­ed what could be a major break­through for solv­ing this prob­lem. Their imple­men­ta­tion of state-of-the-art tech­nol­o­gy to ensure safer cycling impressed the jury pan­el, who unan­i­mous­ly award­ed their solu­tion first place in recog­ni­tion of its ben­e­fits for both infra­struc­ture devel­op­ers and infra­struc­ture users.

The Cycle AI team is made up of Luís Rita, Por­tu­gal; Joan Calvera, Spain; and Miguel Peliteiro, Por­tu­gal. Using crowd­sourced data from their web­site, their pro­posed AI mod­el will esti­mate a safe­ty score for roads by ana­lyz­ing images to detect poten­tial haz­ards for cyclists. The mod­el will then use this data to auto­mat­i­cal­ly process images based on Google Street View, widen­ing the scope of analy­sis to a nation­al scale. (The call for data is still open and active­ly look­ing for con­tri­bu­tions. If you would like to con­tribute any known dan­ger­ous hotspots for cycling, you can sub­mit your infor­ma­tion here).

Read more on ecf.com

Tags
En voir plus

Thierry Roch

Président de l'au5v, Association des Usagers du Vélo, des Voies Vertes et Véloroutes des Vallées de l'Oise.

Vous aimerez également...

Laisser un commentaire

Votre adresse de messagerie ne sera pas publiée. Les champs obligatoires sont indiqués avec *

Close
Close