Vicomtech-IK4 participates in AUTOPILOT, the biggest pilot Project on Autonomous Driving and Internet of Things co-funded by the European Commission

The AUTOPILOT project was presented in a public event held in Versailles last February with representatives of the European Commission and other European personalities.

AUTOPILOT (AUTOmated driving Progressed by the Internet Of Things) is a pilot project that will study and test IoT solutions for autonomous driving in five locations: Versailles (FR), Helmond (NL), Tampere (FI), Daejeon (KR) and Vigo (ES). Every possible scene for autonomous driving will thus be covered: Urban Driving, Highway Driving, Automated Valet Parking, Platooning, etc.

The consortium is formed by 43 participants that include universities, research centres, companies automotive sector and their suppliers, and ICT companies.

Vicomtech-IK4 participates in this project by supplying technology for the development of new perception sensors that will support vehicle operation in the aforementioned scenarios.

With this project Vicomtech-IK4 takes a step forward in their commitment to research and development of new sensors for the automation of driving, based on advanced perception (artificial vision) systems and artificial intelligence for vehicle scene understanding.

Vicomtech-IK4 is currently leading three key projects in this area within the H2020 framework, where they collaborate with the main actors in driving automation, such as Honda, Valeo, Intel, IBM, NVIDIA, TomTom, TASS International, etc.

Those three projects are aligned to tackle different challenges related to autonomous driving in open environments; the projects are: InLane, Cloud-LSVA and VI-DAS.

Autonomous driving in open scenarios, where traditional vehicles are, still requires overcoming certain obstacles, amongst which the most important one is the complete understanding of the scene around the vehicle. This research line is currently a priority, and both vehicle manufacturers and suppliers are fostering it.

Up to date, research has been mainly carried out on specific functionalities that help reach a bit more autonomy on different driving aspects, such as staying within the lane, keeping distance with the car in front of us, stop&go functionalities in traffic jams, etc.

In this sense, artificial intelligence is very important since the aim is to achieve full understanding of the scene. To get this, it is essential to design an advanced system able to foresee circumstances in the scene, and to take appropriate decisions by using computer vision and deep learning technologies. To deal with this complex approach, it is necessary to implement the next generation sensors, to develop high definition maps, to analyse every possible situation in the scene, to design new perception functionalities, sensors for driver status and attention monitoring, as well as algorithms for control and prediction of the evolution of the scene (what other vehicles around us are going to do).