The Autonomous and BASELABS hosted a virtual Chapter Event on Safety & Sensor Data Fusion in order to extend the Global Reference Solutions’ scope towards challenges in the field of environmental sensing and data fusion.

The topic plays a crucial role in the “sensing part” of the general “Sense->Plan ->Act” pipeline implemented in self-driving vehicles. Flaws in the sensor fusion can jeopardize the safety of the overall system responsible for the self-driving functionality. Therefore, sensor fusion solutions must be developed with the highest safety level in mind.

The Chapter Event agenda was structured as follows:

13:00 – 15:30 FOCUS I: Data fusion for SAE Level 2-4

  • Fail-operational algorithmic architectures for L4: e.g., algorithmic diversity
  • Overview of current standards and initiatives, e.g. ISO23150
  • Gaps in current standards (interfaces, architectures)
  • Safe and scalable architectures for data fusion
  • Sensor setups and configurations
  • Discussion on whether we should standardize an architecture for L2 data fusion?
  • Fail-operational sensor setups/architectures for L4

15:30 – 16:00 BREAK

16:00-18:00 FOCUS II: Validation of data fusion systems

  • Overview of current standards and initiatives, e.g. OSI, OpenDrive, OpenScenario
  • Gaps in current standards
  • Best practices and case studies
  • Simulation as part of the validation chain

Speakers

The Chapter Event on Safety & Sensor Data Fusion brings together a diverse array of technical experts: