AUTO_INFRA: Validation lab for autonomous driving vehicles
With support from
Challenge
The validation of autonomous driving functions is time-consuming and expensive. This is due to the complexity of the software and hardware architectures and the integration of these systems. Virtual simulators are unable to take all different operating conditions and functional safety requirements into account. In addition, real-world testing can only be done on a system-level. Subsystems errors can therefore not be analysed and solved in isolation, which leads to a lengthy, costly trial-and-error process.
Project goals
The goal of this project is to create a state-of-the-art validation lab for autonomous systems. New hardware-in-the-loop infrastructure to guide the validation process for automated tasks, integrating a staged approach from virtual simulations to real-life testing. To close the gap between testing with a simulated vehicle in a virtual environment, and testing with a real vehicle in a real-world, restricted environment.
- A real-time HIL simulation environment to simulate vehicles, road surfaces, lighting, weather conditions, etc.
- Perception sensors that are simulated can be gradually replaced with real sensors or high fidelity models
- Replay of recorded road data with existing environmental conditions, and insert new actors (e.g. crossing pedestrian) in a way that closed loop tests can be performed
- Inject object data on the sensor interfaces to enable the possibility to inject typical sensor errors
Economic value
With this project Flanders Make grabs the opportunity to offer test- and validation infrastructure to the industry in Flanders. This infrastructure will not only facilitate the economic valorisation of the knowledge and knowhow of autonomous mobile systems in the region, but will also speed up the time-to-market. The infrastructure allows us to stimulate companies via validation and demonstration to invest in product and process innovation with autonomous systems.