Mike is the Managing Director of Claytex and a technical expert in modelling and simulation using Modelica and FMI. Mike studied Automotive Engineering at Loughborough University and then worked at Ford and Rover on powertrain simulation. After starting Claytex as a consultancy, the company has grown and also become a specialist partner for Dassault Systems, training provider and simulation tool developer. Claytex now works with Formula 1 and NASCAR teams as well as Automotive OEM’s to deliver models and tools covering many different applications helping to create next generation of products.
One of the objectives for automotive OEMs over the past decade has been to strive to reduce the number of prototypes and the amount of physical testing required. Yet as the system complexity has increased, the number of different types of tests that need to be carried out has risen substantially. Having been asked to develop a complete autonomous vehicle, as part of our partnership in the Innovate UK project StreetWise, we have developed Radar, LiDAR, ultrasound, and camera sensor models so that we can connect the complete AI control system and run that inside the virtual environment. The objective of the project was to define the scenarios for autonomous systems that need to be tested, how to manage and characterize these and, finally, how to automate the virtual testing in intelligent ways to measure the system performance. Developing simulators for the testing of autonomous vehicles and advanced driver-assistance systems (ADAS) is fundamentally dependent on sensors and the way in which those sensors perceive the world. Moreover, in modelling a sensor to input into a complex simulation environment there can be no substitute for real world data. In the early stages of an autonomous vehicle or ADAS project, a developer may choose to deploy a generic sensor model that, in simulation, functions as though in perfect conditions. The next level is a device specific model: a model of a particular LiDAR, for example, which in simulation will produce the same output data as the actual device within a project. As a project develops a developer will want to have sensor models that replicate all the noise effects that we see in the real world. The focus moving forward is on how to continue to improve the sensor models, how to obtain better simulations for cameras to support machine vision applications and how to achieve better physics-based models of Radar and LiDAR.