Unmanned Systems Technology 028 | ecoSUB Robotics AUVs I ECUs focus I Space vehicles insight I AMZ Driverless gotthard I InterDrone 2019 report I ATI WAM 167-BB I Video systems focus I Aerdron HL4 Herculift

12 October/November 2019 | Unmanned Systems Technology A start-up in Israel has developed a common and standard Scenario Description Language for verifying and validating systems in autonomous vehicles (writes Nick Flaherty). The Measurable Scenario Description Language (M-SDL) developed by Foretellix, provides a common, human- readable, high-level language to simplify the capture, reuse and sharing of scenarios. It can easily specify any mix of scenarios and operating conditions to identify previously unknown hazardous edge-cases. This allows developers to monitor and measure the coverage of the autonomous functionality that is critical to proving the safety of an autonomous vehicle design, independent of tests and testing platforms. It is similar to high-level description languages such as VHDL and Verilog, which are used to design and verify complex silicon chips. Foretellix has also made M-SDL open so that software development tool vendors, suppliers and developers can use it. It is working with the Association for Standardization of Automation and Simpler system safety tests Software testing Measuring Systems (ASAM) standards committee on the next generation of the OpenScenario standard. This is necessary because developers of autonomous systems cannot be sure the tests on the systems are actually orchestrating desired scenarios or evaluating test coverage as intended. Safety methods and metrics are currently based on the number of miles driven in simulation and road testing, as well as the number of disengagements from autonomous operation to a human operator. The coverage of these tests though is insufficient, non-scalable, and not easily shared or reused. They also don’t provide mechanisms for identifying previously unknown hazardous edge- case scenarios or aggregate coverage metrics across all virtual and physical testing platforms. “The ability to achieve measurable safety of autonomous vehicles is still being limited by a lack of standards, methods and metrics that inhibit reuse and sharing, are insufficient or don’t scale,” said Ziv Binyamini, CEO of Foretellix. “We are actively supporting ASAM in its efforts to create an open language standard.” Companies using M-SDL include AVL List, Volvo, Unity Technologies, Horiba Mira, TUV SUD, Automotive Artificial Intelligence, Metamoto and Vector Zero, as well as researchers at the Trustworthy Systems Lab of Bristol University, in the UK, and the Advanced Mobility Institute of Florida Polytechnic University. M-SDL can easily specify any mix of scenarios and operating conditions to identify previously unknown hazardous cases Schematic of the human-readable M-SDL ecosystem