VaViM - Validation of virtual models used for simulation of autonomous vehicle systems
In order to achieve scalability and robustness, verification and validation (V&V) of self-driving vehicle systems need to be largely performed virtually. This requires validated models on sensors, vehicles and environments where the systems are to operate. The aim of this pre-study is to identify state-of-the-art (SoA) and high-priority research questions within methods for validating models required for virtual V&V of self-driving vehicle systems within a fenced area, and to formulate a joint project application with identified key actors for further studies.
Expected results and effects
The pre-study is expected to result in a SoA report and in a joint project application with identified key actors for further studies. The long-term goal is to obtain objective measures of the reliability of virtual V&V of self-driving vehicle systems. With a working method where validated models and real tests are used side by side, an efficient and predictable continuous V&V of the product´s quality and safety during its entire life cycle is enabled; pre-study phase, development phase; commissioning; maintenance & system optimizations.
Planned approach and implementation
1. SoA; Examination of the latest research and application in the validation of simulation models in the field of autonomy and relevant adjacent areas with similar problems. Analysis of other actors´ methodology and results in similar areas.
2. Research questions; connection to new partners and formulation of a larger joint project application for further studies.
3. Feasibility study; Modeling effort with model of vehicle, surrounding physical environment and autonomy model. Implementation of simpler physical comparative tests, results and conclusions.
Final presentation
Project period
February 2022 - August 2022
Contact
Stefan Thorn, Volvo Autonomous Solutions
Partners
Volvo Autonomous Solutions, Blekinge Tekniska Högskola
Vinnova number
2021-05178
