Issue 39 Unmanned Systems Technology August/September 2021 Maritime Robotics Mariner l Simulation tools focus l MRS MR-10 and MR-20 l UAVs insight l HFE International GenPod l Exotec Skypod l Autopilots focus l Aquaai Mazu

35 makers are increasingly supplying these detailed simulation models to system developers, Tier 1 subsystem providers and car OEMs early in the design cycle. That is essential in allowing system developers to test out the performance of the hardware and software design long before it is built. Standards But simulation is also moving up in its importance as the result of a new standard, and even into the operation of autonomous systems, both embedded in the system and also in the cloud. The new standard is ISO 21448, also known as the ‘safety of the intended functionality’, or SOTIF. Alongside the ISO 26262 standard, which is based around ensuring faults are not designed into a component, ISO 21448 applies to functionality that requires proper situational awareness in order to be safe without a fault occurring. SOTIF is based on a statistical process, so it needs millions or billions of data points to demonstrate that. This is not possible to do in the real world, even though self-driving car makers have fleets of vehicles travelling millions of kilometres for testing. Simulation allows many more miles to be travelled with synthetic data that accurately emulates the real-world environments but under the control of the developers. SOTIF is driving new methodologies linked to a statistical demonstration that an autonomous vehicle is safer than a human driver. This is a fundamental change in the industry; it is not just a Lidar sensor maker demonstrating that the sensor is suitable in a safety-critical ISO 26262 design, but that the sensor is safe under certain conditions. That is why the component supplier has to deliver the model of the sensor as soon as possible for the OEM to simulate the entire vehicle with all the other components. This can be done with the physical sensor – the hardware-in-the-loop (HIL) approach – but having the simulation model brings significant advantages. It allows engineers to focus on specific risky situations at the edge of the statistical array or even at the corners of multiple influences where sensors could get confused or give inconsistent data. These edge and corner cases represent situations that might happen only once in a million kilometres in the real world, but simulation enables these scenarios to be tested millions of times every day. For example, one car maker has modelled a traffic jam scenario with 13 parameters. The combinations Simulation tools | Focus Having a simulation model allows engineers to focus on specific risky situations, where sensors could get confused or give inconsistent data Unmanned Systems Technology | August/September Accurately simulating the physics of rain is a key challenge for simulation tools (Courtesy of Ansys)