Business story #2 Ambient intelligence: Valeo uses visual recognition to improvepassengers comfort
What will be the future of cars? Autonomous certainly. But how is it going to impact the passenger experience?
The race between traditional car manufacturers and high-tech giants to put the first autonomous vehicle on the road has created a strong impetus to re-imagine life on board. As this revolution will leave room to a collective car model instead of traditional ownership, comfort criteria will take over performances in passengers’ enjoyment.
Among these companies, the automotive manufacturer Valeo stands out by taking up the challenge of an individualized experience within a collective space. It focuses on Ambient Intelligence, and develops a series of technologies that will allow them to better understand users in order to anticipate their needs and improve their experiences.
With sensors not only outside but also inside the vehicle, the intelligent car of the future will be able to interact with its passengers and provide personalized features.
In January, the company presented its concept-car at the latest edition of CES (Consumer Electronic Show) in Las Vegas. Under the name “Smart Cocoon”, Valeo displayed several innovations within a car model. Among these innovations is an individualized thermal control system. By monitoring passengers’ information such as gender, heartbeat or morphology, the car
automatically adjusts the temperature according to each passenger’s specific needs. Beyond comfort, this innovative thermal system also pushes forward electric car performances as it provides significant energy savings.
To achieve this, Valeo’s Thermal System Team combined infra-red cameras installed inside the car cockpit together with a deep-learning system in order to analyse passengers’ gender and clothing layers. Thanks to deepomatic’s platform, they were able to develop a custom image recognition system to make the in-car cameras capable of automatically extracting this information.
Afterwards, the challenge lied in deploying the resulting artificial intelligence system into the car.
Both from a connectivity and security point of view, it is crucial that the system runs locally. As deep-learning technology requires a lot of computing power, deepomatic’s engineers adapted the image recognition models so that they could run directly in the car without access to the cloud and on restricted power hardware.
The intelligent approach to driving using image recognition for a better travel experience is only at its beginning. There is a lot more to explore when it comes to automatically understanding passengers interactions with the vehicle through motion or face recognition for example, or to contribute to a more secure ride by detecting signs of fatigue or inattention.
Redefining standards of comfort and safety, these solutions open the door to new traveling experiences.
Thanks for sharing!
To learn more, check out our website. Don’t forget to tweet @deepomatic and give the team your ideas and views in the comments below.
Eléonore Tezenas, Product Marketing Manager at Deepomatic