Sensors in the Headlamps for Better Automated Driving Functions

Our development is modeled on nature. Dragonflies, for example, do not miss anything. Thousands of tiny eyes register every movement in a flash. Furthermore, the insects also have 360° all-round vision. It is therefore no great surprise that IAV and the ZKW Group have called their joint project “Dragonfly”, because the focus is on providing autonomous vehicles with sensor systems that are also capable of registering all relevant signals from their surroundings.

But this demands a large number of individual sensors. And the number is sure to increase even further, at the latest with the introduction of highly automated driving functions on SAE level 4 and 5, including above all cameras as well as radar and lidar sensors. Some companies, particularly in the USA, accommodate the complex technology in roof superstructures. The advantage is that the superstructures are easily transferred to other vehicle platforms. On the other hand, they spoil the look of the vehicle, are detrimental to its aerodynamic drag and restrict the use of parking garages.

IAV and ZKW therefore advocate a different solution. Both companies want to fit the sensors and corresponding evaluation algorithms in the “smart corners” of the vehicle. This refers to the front and rear headlamps. “There are a number of advantages here. The sensors are relatively well protected from damage and also from dirt, for example with cleaning systems, the lotus effect and anti-fog coating. Thermal management is already present, and there is also scope to accommodate the interfaces to the vehicle”, says Carsten Simon, who is in charge of IAV’s Perception & Concepts team.

ZKW’s contribution to the Dragonfly project consists in supplying the headlamps with integrated sensors. IAV gets the test vehicle ready and supplies the algorithms for HMI and for autonomous driving. The company already has a wealth of experience in this respect gained, for example, from the Symbioz project for Renault (highway chauffeur) and the HEAT project (driving on SAE level 4 in an urban area). Altogether, IAV test vehicles have already completed several hundred thousand kilometers in autonomous driving mode and provided valuable insights.

The engineers focus on various scenarios for their tests: firstly, as a “highway chauffeur” that can see ahead and back and is capable of changing lane automatically. Secondly as an assistant for crossroads or for parking garage exits with the ability to see round corners. And thirdly, as a solution to fulfill the
future ENCAP 2024 requirements including the ability to detect an approaching cyclist when maneuvering the vehicle out of a parking space.

Since mid-2019, a test vehicle has been providing proof that the Dragonfly concept already works well in practice. IAV and ZKW have fitted the front headlamps of a production vehicle with four cameras: two to see to the front and two to register adjacent traffic at the sides. “In a crossroads situation, for example, the cameras in the corner give a clear view with good visibility of vehicles crossing from the right or left, far superior to what a human driver can see, particularly at complex crossroads”, reports Simon.

And that is not all. The new concept also scores better than the standard setup today (a camera in the middle of the windscreen, together with radar and lidar sensors as well as front and surroundings cameras in the radiator) by offering better protection from parking rowdies and other environmental factors. “It is therefore quite likely that in future, numerous sensors for vehicles on SAE level 4 and 5 will in fact be fitted in the front and rear headlamps. There is even a smart module for driver assistance SAE level 2 consisting of a complete package of sensor systems and functions for integration in headlamps”, says Simon.

At the moment, the test vehicle only has cameras at the front headlamps. But that won’t be all. The next step will add two lidar sensors for better depth information and even more detailed environment recognition thanks to sensor fusion. The plan is then for two cameras in the rear headlamps to provide information about the traffic behind the vehicle. With the fourth upgrade, the test vehicle will be equipped with a whole range of sensors. The already integrated systems will be joined by two additional cameras at the front and two lidar and radar sensors at the rear. By this point at the very latest, the test vehicle will truly deserve the project name.

The article was published in automotion 01/2020, the automotive engineering magazine of IAV. Here you can order the automotion free of charge.