Are connected automated vehicles particularly predictive? To investigate this, the vehicles have to prove to ZF that they can communicate reliably with their environment. A specially developed communication simulator from dSPACE makes it possible to perform a prediction test.
A highly automated vehicle is waiting at a red traffic light. The traffic light turns green, but the vehicle does not drive off. There is a good reason for this: An emergency response vehicle is respon-ding to an incident and is entering the intersection. But how could the automated vehicle foresee the situation? A technology called vehicle-to-everything (V2X) communication enables precisely this “peek around the corner” and thus predictive driving.
More Safety Through V2X Communication
The increasing connectivity of vehicles opens up new possibilities for road safety. With the aid of vehicle-to-everything (V2X) technology, vehicles can communicate with each other as well as with the surrounding infrastructure. This enables anticipatory perception of the traffic situation, which significantly improves the safety of all road users. Oliver Maschmann of ZF explains: “ZF is pursuing the goal of further developing and validating advanced driver assistance systems (ADAS) and functions for automated driving (AD) with V2X technologies.” Automated vehicles equipped with these technologies are referred to as connected autonomous vehicles (CAV). They combine the advantages of connectivity with functions for autonomous driving to make road traffic safer and more efficient.
Challenges in the Validation of V2X-Based Systems
The development of CAV requires a comprehensive testing and validation strategy. One particular challenge is the realistic creation of test environments for various traffic scenarios.
- Static aspects: Virtual worlds must emulate the real world as accurately as possible. An exact digital twin of the real environment must be created from real environmental data. Map data from georeferenced road networks provides support for this.
- Dynamic aspects: The real traffic situation must be recorded by means of test drives, reproduced for testing, and supplemented by synthetic scenarios. Safety-critical driving situations in particular, such as near collisions, are often inadequately represented in real data and must be simulated in a targeted manner.
- High precision: To support the tracking accuracy of ADAS/AD, all data must be georeferenced, i.e., recorded together with the global navigation satellite system (GNSS) data. Geodata with an accuracy of a few centimeters is also required for the simulation of synthetic scenarios to ensure the validity of the tests. This is where the GNSS simulator is used, which is fed directly from the real-time simulation.
A Virtual Intersection as a Basis
“A particular challenge for automated driving systems is safe and anticipatory behavior at intersections,” reports Fabian Mürdter of ZF. To enable these investigations, a complex intersection was simulated in detail in the existing road network. This includes the accurate modeling of lanes and cycle paths, including precisely positioned traffic signs and traffic lights. This was achieved by capturing real data (sensor data such as camera, radar, and lidar data) and transferring it to the 3D world using high-precision map data (see Digitalization of urban traffic areas, dSPACE Magazine 2024).
Generation of Dynamic Test Scenarios
Heiko Simpfendörfer of ZF explains the procedure: “The real traffic, including all available communication data, was recorded through test drives: raw data from the perception sensors, GNSS information, and V2X communication data.” To test corner cases such as suddenly appearing pedestrians or unpredictable driving maneuvers by other vehicles, synthetic scenarios were generated from the real test data. The dSPACE scenario generation workflow was used for this. The process can meanwhile be highly automated using the Traffic Virtualizer tool. “The goal is to make the scenarios so flexible that all the necessary test cases can be generated in the shortest possible time by varying the parameters,” adds Heiko Simpfendörfer. These are, for example, parameters such as speed, acceleration, vehicle distance, the number of road users and traffic light switching phases, etc. All parameters can be varied independently of each other so that an infinite number of scenarios are available for the simulation. “With triggers set in the simulation, traffic light phases and entry times at intersections could be varied to enable constantly changing traffic flows,” says Sören Stamm about a trick for more depth of variation. At the same time, the run-time-independent reproducibility of the various scenarios was maintained throughout the entire transport network. The simulations were created and carried out using the simulation tool suite ASM (Automotive Simulation Models) and the sensor-realistic simulation software AURELION.
V2X-Specific Test Definitions
In addition to the intersection topology and traffic light signals, the V2X messages can also be used to transmit data on other road users and infrastructural features such as roadworks or accident sites. Incorporating such details into the simulation ensures that the automated/autonomous system is prepared for a variety of real-life traffic situations,” reports Fabian Mürdter. A special V2X user interface is available in the ASM parameterization software ModelDesk for the preparation of simulation scenarios with V2X information.
Validation Under Real-Time Conditions
“In order to be able to take all safety-critical aspects into account, it is necessary to test the real electronic control unit under real-time conditions,” says Oliver Maschmann. For this overall system test, all electronic control units and relevant components must be integrated or simulated in a hardware-in-the-loop (HIL) simulator. To be able to carry out realistic, georeferenced V2X simulations in real time, a SCALEXIO HIL system for ADAS/AD validation was expanded to include a GNSS simulator and a V2X simulator that communicate with the ADAS electronic control unit under test via a radio link. Software tools from dSPACE are used to control the V2X and GNSS simulators directly from the traffic simulation with the corresponding data: “The ASM simulation supplies the precise V2X and GNSS reference information at run time to the V2X and GNSS simulators, which send this data to the ADAS/AD electronic control unit via a radio interface,” reports Sören Stamm from ZF. The integration and configuration of V2X message types such as MAP (lane information) and SPaT (traffic light phases) with tools from dSPACE (V2X Interface for WaveBee) enable detailed and realistic traffic flow simulations. Oliver Maschmann on the properties of the simulator: “The precise synchronization of the signals and data with the HIL simulator makes it possible to plausibly stimulate the ADAS/AD electronic control unit and simulate and test the vehicle’s responsiveness at the overall system level.” The test evaluation ensures that all systems are coordinated and work correctly.
Sensor Simulation Options
“The sensor simulation can be carried out at different levels of detail, for example, to include the processing stages of individual sensors completely or not at all in the simulation,” explains Heiko Simpfendörfer. The processing of the sensor data therefore ranges from object lists generated with ground truth sensor models, which follow a geometric approach to calculate a nearest point based on object-specific bounding box information of the objects, to the simulation of raw sensor data with highly realistic physical sensor models that take into account all of the contours and material properties of the individual objects. The ground truth approach is realized with ASM, while the simulation of the raw sensor data is carried out with AURELION. The raw sensor data is made available to the electronic control unit via the Environment Sensor Interface Unit, among other things.
Advantages and Benefits for ZF
By combining real measurement data with synthetic test scenarios, a realistic environment simulation was created that significantly improves the validation of CAV. The most important advantages are:
- Greater safety: By integrating V2X data, automated vehicles can “peek around the corner” and thus react to potential dangers earlier.
- Efficient test processes: The ability to test scenarios reproducibly and in real time reduces development times and costs.
- Seamless integration: The combination of dSPACE HIL systems and V2X simulations enables an end-to-end test strategy from the early development phase to final validation.
- Expanded test options: Real recorded situations can be varied by adding additional participants, e.g., ambulances, thus expanding the scenario space to be tested.
- Wide range of sensor simulation options: The dSPACE environment enables a variety of sensor simulation options to increase simulation depth and expand the setup. This is all done while simultaneously providing the corresponding ground truth data.
Conclusion – The 7th Sense in Simulation
The collaboration between ZF and dSPACE makes a decisive contribution to increasing the safety of automated vehicles in urban traffic. The practical tests and high flexibility of the simulation tools provide ZF with valuable insights for the further development of technologies for connected and autonomous vehicles. This makes it possible to test the “7th sense” of these vehicles in the simulation.
Courtesy of ZF
dSPACE MAGAZINE, PUBLISHED JUNE 2025
The article was created in close cooperation with the following persons:
Oliver Maschmann
Oliver Maschmann is responsible for HiL Solutions at ZF in Friedrichshafen, Germany.
Fabian Mürdter
Fabian Mürdter is a development engineer at ZF in Friedrichshafen, Germany.
Heiko Simpfendörfer
Heiko Simpfendörfer is a development engineer at ZF in Friedrichshafen, Germany.
Sören Stamm
Sören Stamm is a development engineer at ZF in Friedrichshafen, Germany.
At a Glance
Task
- Development of dynamic V2X scenarios for the validation of connected vehicles in intersection areas
Challenge
- Creation of flexible logical scenarios that are highly variable
- Enriching the scenarios with V2X information
Solution
- Converting real traffic data into logical scenarios with dSPACE’s scenario generation workflow
- Preparation of V2X information with special tools from dSPACE and a V2X simulator
- Time-correlated simulation of sensor data, V2X data, and GNSS data with the dSPACE HIL simulator SCALEXIO
Advantages
- Flexible V2X real-time scenarios are reproducible and help to reduce development times and costs.
- The simulator makes it possible to validate the reliable function of V2X-capable ADAS/AD electronic control units.
