Dynamic Ground Truth

Project Description

The safety, security and effectiveness of modern ADAS (advanced driver assistant systems) are highly dependent on the systems’ ability to sense their physical operating environment, characterize it in a virtual ‘scenario’, infer information from that scenario and respond accordingly. ADAS and automated driving functions (ADF) demand that cars are able to perceive objects such as streets, houses, trees, vehicles, pedestrians or animals, formulate a composite operating scenario, infer future states of that scenario and maneuver appropriately, i.e. apply brakes. 

In order to be effective and ensure security and safety of users, any system like this needs to be thoroughly tested. Testing of ADAS/ADF functionalities is very different from testing of classical control functionality, e.g. an engine control system. The vehicle becomes the actuating variable and the objects in the environment react on the input of the vehicle. 

The aim of DGT (Dynamic Ground Truth) is to raise the development process of ADAS and ADF to a mature level equivalent to today’s processes for classical powertrain development by reliably capturing environment data (Ground Truth) using a high resolution sensor system, the so-called reference measurement system. The design and development of such a sensor system (reference measurement system) and the implementation of a functional prototype are targets of this research project. 

The dedicated components of this system can be compared to the object-model of the ADAS control unit in real-time by running it on the development laptop of an engineer and observing the behavior of the automated vehicle. All data is recorded and can therefore be used for post-processing as well. This also allows the generation of scenarios for environment simulation tools which enable repetition of the real-world scenario in a simulated environment. 

The results of these projects are of utmost importance for the partners to develop ADAS and ADF functions in a reliable and reproducible way for an industrial market field currently considered a mega-trend in the automotive industry. This project is absolutely mandatory to maintain and even further extends the market leading positions of AVL and TTTech as suppliers of on-road measurement equipment.

Contribution

In the context of the project “Digital Ground Truth” TTTech will continue the ongoing research and development of novel and innovative technologies for autonomous driving platforms. The time-critical data aggregation of various onboard data streams from different sensors and control units is crucial for subsequent processes like real-time control or data analytics. The data recording and aggregation are essential for testing and development of ADAS/ADF functionalities. By providing the vehicle network data synchronous with externally collected environmental data, a data fusion of both data streams can provide an overall picture of the benchmarking situation. In the frameworks of the project, TTTech will develop prototype software components for the next generation of automotive data logging systems, enabling reliable data aggregation for higher data volumes of next generation ADAS platforms.

Funding

bmvit and ffg

Digital Ground Truth (DGT) receives funding from the Austrian Ministry for Transport, Innovation and Technology under the funding program "Mobilität der Zukunft", grant agreement no 860820.

Digital Ground Truth (DGT)

Digital Ground Truth (DGT)

Duration: 01.08.2017 – 31.07.2020 (3 years)

Funding: BMVIT

Coordinator: AVL List GmbH

Keywords: autonomous driving, advanced driver assistance systems, ADAS, automated driving, control unit platform, big data aggregation and analysis

Contact: Gerald Fritz
gerald.fritz@tttech.com