Autonomous driving starts in the virtual world
Autonomous driving represents a future in which mobility is safer, more convenient, and more accessible. Reliably detecting its surroundings, making split-second decisions and taking people to their destination comfortably, the interaction between sensors, software and control devices within the vehicle must be perfect. Virtual vehicle development is key to this vision: It allows realistic simulations and comprehensive tests of complex systems, so they can be shown to be safe long before the first vehicle hits the road.
Continuous test
and validation platform
- Test your autonomous driving stack end-to-end or in single components
- Test early algorithms (MIL) and production-intent software (SIL)
- Test single ECUs or integrated ECU networks including mechanical hardware (HIL and VIL)
- Generate any imaginable scenario utilizing sophisticated editing tools, scripting interfaces and industry standards
- Physical Sensor Models to generate synthetic sensor raw data
- Real-time capable simulation of full AV sensor suites
- Detailed and modular vehicle model to achieve very high correlation to physical vehicle behavior
- Collect and annotate data from the real world
- Replay and inject recorded data with hardware-in-the-loop
- Connect proving ground and the virtual world with vehicle-in-the-loop
- Automation and integration of testing solutions into your CI/CD pipeline
- High scalability and efficiency of cloud-based simulations
- Management solutions for vehicle models, scenarios and test fleets for full traceability
Use cases
End-to-end development processes for automated driving: Simulation, ODD and standardization
The approval of automated vehicles has been the topic of intense debate in recent years – particularly with regard to the validation of specific operational design domains (ODD). The potential for savings in the commercial vehicle sector is immense. This is where insights from the passenger car sector, such as from open tests, AI-generated scenarios and simulation-based validation come in. What is needed are transferable standards and credible simulations. End-to-end integration into development processes enables continuous and efficient validation.
Simulation-based position data analysis in the development of stable platooning functions
Platooning describes several trucks driving closely behind each other automatically to reduce fuel consumption. Error-free platooning depends on the individual vehicles’ position data and the stability of the convoy. This article illustrates how platooning algorithms can be efficiently developed and validated using simulation.
Risk-free validation of autonomous driving functions using virtual prototypes and optimized decision logic
Autonomous driving functions must show a reliable and correct response in any given situation. Virtual prototypes form the basis for testing the sensor behavior and optimizing the decision logic. This means that even complex systems can be validated comprehensively – without risk, reproducibly and away from real road traffic.
Systematically validating complex SAE Level 3+ driving functions
Scenarios for SAE Level 3+ driving functions systematically cover complex traffic situations, such as bottlenecks due to construction work or lane changes in heavy traffic. Generating scenarios on the basis of functional requirements enables the reproducible and targeted validation of automated systems in a reliable, scalable and standard-compliant manner.
Validating autonomous driving functions with systematic scenario coverage and automated assessment
The systematic implementation of a wide range of scenarios that simulate all relevant driving conditions and situations ensures the validation of autonomous driving functions. Automated assessments are used for the comprehensive analysis of system reactions, enabling the reliable evaluation of functionality under realistic operating conditions.
Systematic evaluation of safety-critical scenarios to boost the reliability of autonomous driving functions
Autonomous driving functions require extensive long-term testing with a focus on rare and/or safety-critical events. Automated filing and the structured classification of results enable a systematic evaluation that identifies any weak points at an early stage, minimizes risks and improves long-term system reliability.
