The Automotive market is growing rapidly and evolving from “dumb” autos to smart electronic cars. Analysts estimate that by 2020, 120 million cars will be sold worldwide, and that China will be the major automotive market, responsible for nearly one third of annual new vehicle sales worldwide (34.7 million). North America will be the second largest sales region, with nearly 20 percent of annual global sales volume (20.7 million) and Western Europe will retain third place until 2020 (16.3 million).
Cars Getting Smarter
Autonomous vehicle, a new market segment also known as driver-less car, is pushed by Google and leading car manufacturers, requires heavy processing horsepower and multiple sensors. Eventually, much of these technologies will leak to the rest of the automotive market.
ADAS (Advanced Driver Assistance Systems) and Infotainment systems are becoming more and more sophisticated and play an important role in differentiating between brands and car models. Car manufacturers are integrating more sensors and computing power into their cars to support customer’s desires and safety regulations.
Taking Human Machine Interface (HMI) to the next level
For the time being, until driver-less cars will become common (which could take 10 years or even more), cars will be driven by people with the assistant of ADAS systems installed in the car. Taking human machine interface to the next level will allow vehicles to closely monitor the driver’s responsiveness and state-of-mind, while taking some decision on behalf of the driver, helping him to arrive safely to his destination. This is exactly why we at CEVA initiated the FANCI project.
FANCI (Face and body Analysis Natural Computer Interaction) is a collaborative project funded by the European Commission, targeting the development of state of the art NUI (Natural User Interface) capabilities. Automotive applications are a primary target for such capabilities.
FANCI partners include CEVA as the provider of the imaging and vision platform and computer vision libraries for the human-machine technologies, Harman Becker Automotive Systems as the automotive industry expert, KeyLemon as the provider of face detection and voice software, nViso as the provider of emotion detection software, SoftKinetic as the provider of 3D vision technologies and gesture software, Tobii Technology as the provider of eye tracking software, and the University of Siena as the developer of the application layer and the user API.
A Creative Automotive Application-Suite Demonstrator
The project will focus on intelligent user sensing methods, including face analysis, eye tracking, emotion detection, driver authentication, hand gesture and head gesture. FANCI is integrating all of these capabilities demonstrated by building a creative automotive application-suite demonstrator, including an application programming interface (API) for software application developers.
The FANCI automotive platform will demonstrate several applications including: Driver Identification, Smart Navigation Assistant, Smart Music Player, Smart Safety Assistant, Intuitive Multimodal adjustments, Personal POI (points of interest), Assistant Smart Dashboard, Usage based Insurance, Enhanced Speed Limits Detection, Face Authentication etc.
Here is a scenario that demonstrates the above-mentioned applications:
Kate is driving on the highway and she has missed to look at a street sign which lowers the allowed speed from 65 mph to 55 mph. The system cross verifies that she has also not slowed down to 55 mph. In this case, the system alerts Kate with a visual and/or auditory message of the sign that she has likely missed to notice. Many current infotainment systems already detect and recognize road signs (e.g., speed limit). Leveraging eye gaze tracking, the system can determine whether the driver has looked at them or not, and appropriately alert the driver.
The FANCI project is centred around the integration of these multi-modal human-machine applications running on a single imaging and vision platform from CEVA, realizing significant cost and power savings for the implementation of these functions in embedded products.
Ceva