The Task
The steadily growing number of functions for automated driving requires a powerful test environment that can handle the exponentially increasing effort for the tests in early development phases. To achieve this with utmost efficiency, a simulation-based evaluation of vehicle software for functions for highly automated driving and self-driving cars is a necessity.
The Challenge
Comprehensive tests of functions for automated driving require:
- The automated simulation of different traffic scenarios on a wide range of road networks and with different environmental conditions
- A large number of tests to cover scenario variations and randomized testing
- The simulation of millions of test kilometers
- Consistent management of test parameters and a result analysis
- The reuse of tests in all development phases
The Solution
Multiple instances of ASM Traffic are simulated in parallel on a VEOS cluster.
The seamless dSPACE tool chain opens up powerful and efficient possibilities for realistic simulations of autonomous driving. These include:
- The graphical definition of road networks and traffic scenarios in ModelDesk
- The simulation of many different traffic scenarios comprising a multitude of road users with ASM Traffic
- Testing multiple traffic scenarios by means of PC clusters with a high number of VEOS® instances for parallel execution, which reduces the time required for testing to a minimum
- Data management and traceability via dSPACE SYNECT®
- The seamless reuse of tests for MIL, SIL and HIL simulation on platforms like PC, VEOS, and SCALEXIO