Drones support validation of automated driving functions
For the development and validation of automated driving functions, fka has created a database with the help of which relevant traffic scenarios can be made available to customers from the automotive industry. For this purpose, data from various sources such as accident databases, simulations, data from vehicles equipped with sensors – and now also drones – are first harmonized and then further processed with a uniform process chain.
Today’s traffic data acquisition by vehicles equipped with sensors has some disadvantages. The vehicle can only ever record the concrete situation in the immediate vicinity. As a result, thousands of kilometres are required to record a relevant number of different traffic situations.
In addition, research vehicles equipped with conspicuous sensors influence the behaviour of other road users and thus the quality of the data. Drones avoid these typical limitations of established traffic data acquisition methods and can record thousands of kilometers of vehicle data in a very short time.
The trajectory of each vehicle, including vehicle type, size and maneuver, is automatically extracted from the data collected. By using modern algorithms, the positioning error is typically less than ten centimeters. In addition to the safety validation of highly automated vehicles, the data sets are also used for many other tasks such as the analysis of traffic patterns or the parameterization of driver models.
The data collected and processed by fka offer naturalistic vehicle trajectories of different traffic scenarios. Currently, drone data sets of motorways, urban intersections and roundabouts are available, recorded in Germany and the USA. Further countries and traffic scenarios are in progress. The data now comprise more than 110,000 data sets.
The functional reliability of automated driving functions can be analyzed in suitable test environments thanks to the knowledge of the situation. By displaying relevant driving scenarios in highly dynamic driving simulators or on test tracks such as fka’s Aldenhoven Testing Center, it is possible to combine real driving tests with subsequent validation by simulation.
Under the name levelXdata, fka bundles its expertise for the validation of all stages of automated driving – from the acquisition of traffic data using drones or vehicles to the extraction of the trajectories of all road users and the use of the results as scenarios for the validation of automated driving functions.
More information: https://www.fka.de/en/
Related articles
Drones learn autonomic navigation from cars and bicycles
Applying V2X technology to drones
Virtual testing of navigation and location of autonomous vehicles
Concept vehicle brings steer-by-wire to sports car study
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
