The CD6 presents a car fitted with AI and optical sensors that gathers thousands of data points to improve autonomous driving Nov 05, 2024

Download

Researchers Santiago Royo (in the foreground), Gerard de Mas (in the background, left) and Adrià Subirana (in the background, right)

Autonomous driving is the future of urban mobility, especially in an increasingly ageing society. In the medium term, people will travel in vehicles capable of making real-time decisions without human support in all types of situations. Initial tests have already been conducted in cities like San Francisco, Phoenix and Singapore. Autonomous driving will help address environmental issues by optimising fuel consumption and traffic management, reducing accidents and benefiting groups such as the elderly and people with disabilities.

Professor Santiago Royo, the director of the Centre for Sensors, Instruments and Systems Development (CD6) of the Universitat Politècnica de Catalunya - BarcelonaTech (UPC) and the principal investigator of USEFUL, is convinced that "achieving this transformation requires large volumes of accurate data from sensors operating in various modes, enabling autonomous systems to interpret their surroundings and act safely, even in complex environments. Thus, we need to train advanced artificial intelligence (AI) algorithms. If we want them to be reliable in any circumstance, we must supply them with as diverse a range of data as possible, to generate the most realistic and general information possible that will allow AI to learn and make sensible decisions."

Artificial intelligence, sensors and algorithms
In this context, the Strategic Research Agency of the Spanish Ministry of Science, Innovation and Universities, through its Ecological and Digital Transition programme, has provided the UPC's CD6 with 273,000 euros in funding to launch the USEFUL project. The project is also supported by companies such as Beamagine (a UPC spin-off), Anello Photonics and QNAP, and the Terrassa School of Industrial, Aerospace and Audiovisual Engineering (ESEIAAT) and the Terrassa School of Optics and Optometry (FOOT).

In this project, the CD6 has provided a low-emission vehicle with an integrated and computerised optical sensor system, featuring a high-resolution solid-state LIDAR (a type of laser sensor with a photodetector that measures distances) supplied by the Beamagine spin-off; a high-precision geolocation and odometry system that operates without GPS (measuring data from moving sensors) provided by Anello Photonics; and a high-capacity data storage system donated by QNAP.

The system that the vehicle will use to navigate the streets of Terrassa over a 6-month period incorporates a processing unit approved for autonomous driving systems. This unit manages data acquisition from all sensors and executes artificial intelligence neural networks for specific real-time applications such as obstacle perception and 3D mapping of the environment.

Through this project, the CD6 will obtain millions of images from multiple sensors, some of which are groundbreaking for autonomous vehicles, and will provide data suitable for designing reliable autonomous driving systems.

The data will be anonymised (not identifying pedestrians and vehicles) and will be added to a public dataset allowing other researchers to develop related algorithms. Santiago Royo explains: "We want the dataset to be as diverse as possible in terms of the scenarios captured, such as rural, urban and motorway, including all kinds of vehicles, especially the more vulnerable ones like bicycles and scooters, and also capturing data under real weather conditions and in realistic settings where conditions may not always be favourable for every sensor. In fact, these data will be used to develop advanced applications enabling road object detection, environmental mapping, predicting the movement of road agents and creating digital twins, among other applications, with higher reliability."

In addition, USEFUL will allow for testing different AI training methodologies to determine the most effective way to achieve safe perception systems. In the future, the vehicle will make it possible to test new types of sensors for autonomous driving, validate new sensor configurations and develop detailed mapping applications for all kinds of settings.

30 years of technology transfer
The UPC's CD6 is a research group specialised in optical and photonic engineering, located on the Terrassa Campus and part of the TECNIO Network. The CD6 focuses on transferring technological innovation results to industry. Currently, it includes more than 40 researchers working on projects in fields such as health, industry 4.0, urban mobility, precision agriculture, smart cities, space and sustainability. Throughout its nearly 30-year history, the CD6 has filed 76 patents and incubated 15 spin-offs. It is also a member of 10 industrial clusters and associations.

Download

Researchers Gerard de Mas and Adrià Subirana inside the car

Download

Adrià Subirana working on the vehicle's interior equipment

Download

The car incorporates various types of optical sensors

Download

The vehicle will collect real-time data over six months