The key factor in autonomous driving: prediction

Given the complexity of mobility in large cities and in search of better traffic management, autonomous driving is one of the future key elements in improving traffic. In order to provide an optimal solution to the current situation, and in order to improve the safety of urban traffic, some manufacturers such as Bosch have focused on an element that, a priori, may not be as important as others: streetlights.

What does MEC-View consist of?

The German manufacturer’s project is called MEC-View and its operating architecture is based on streetlights equipped with video and lidar sensors, which is a type of radar, and which, combined with today’s advanced mobile technology, makes it possible to offer vehicles relevant information in real time, such as detecting any type of obstacle in an agile and effective way.

The partners that have worked on this project, which has been developing for more than 3 years, have been Bosch, leader of the consortium, Mercedes-Benz, Nokia, Osram, TomTom, IT Designers and the universities of Duisburg-Essen and Ulm. This last city, Ulm, has been the testing ground for this technology, the experience gained from which will allow the improvement and development of automated driving and mobile technology to continue.

Through cameras installed on streetlights, up to 6 meters high, a 360-degree view is obtained, allowing all the information collected to be transferred to the automated vehicles that circulate through the streets and highways of this city. This type of camera is required at a great height to be able to achieve a complete vision that, at ground level, may not include all the incidents or situations generated, such as a vehicle coming out of a ford. The purpose of these cameras is to extend the range of vision that the sensors of a vehicle cannot reach.

From the information generated by the sensors themselves and that obtained by the sensors installed in the lampposts, a complete vision of the real-time situation of the road is obtained. All this will help achieve excellence in autonomous driving and, consequently, improve driving for all drivers.

More projects underway

Other companies, such as Uber, have developed technologies with similar characteristics, as is the case of the Multinet system. This offers, like the MEC-View project, the detection of obstacles, even predicting what decision pedestrians, cyclists or other vehicles that cross their autonomous vehicles will make.

Using artificial intelligence, this system is able to predict, at the moment, what is the most possible trajectory of the object, as well as develop a plan for it. The great difference of this system, compared to other similar ones or to MEC-View itself, is that MultiNet is based on the uncertainty of the behavior and movement of objects. It is capable of human reactions and, by collecting data before use, in real time, define a prediction of a behavior and refine it to obtain the possible potential trajectories. Thanks to this system, Uber has confirmed that it exceeds, in precision, other options on the market, its system being between 9% and 13% more precise.

Improving safety on the road for drivers and pedestrians is, once again, a priority for leading technology companies, which seek to minimize risks and ensure that we can circulate with greater peace of mind. It is clear that autonomous driving remains one of the great future goals of the automotive industry. How many years will it take for this phenomenon to become generalized?

Related Posts