Ansys AVxcelerate Sensors
Test and Validate Sensors for Self-Driving Cars

To test sensors in scenarios, add several cars to create complex situations, such as following a car and monitoring the path of a crossing car simultaneously. Each vehicle in a scenario can be either static or automatic, enabling evaluation at a specific point of interest or along a predefined trajectory. Sensor simulation follows ego-vehicle dynamic motion.

Sensor data output, such as lightning strike hit points, material properties, etc., and vehicle parameter input or output, such as position, orientation, speed, and steering wheel, are all available. Thanks to the consistent combination of data from multiple sensors, the simulation enables you to validate the model of a smart sensor’s behavior or its fusion algorithms. Deterministic and real-time modes are both supported.

Software allows you to simulate the actual camera model in edge case driving situations. It simulates all components of a camera, such as the lens system, imager and pre-processor. For automotive front-facing cameras, the windshield can also be considered in simulation. Consider the optical and spectral properties of the environment in visible range, along with the optical properties of the lens system and optoelectronic properties of the imager. With the addition of plugins, the simulation can manage dynamic adaption. Camera simulation creates raw images, which are used to test and validate perception algorithms either as models-in-the-loop, software-in-the-loop or hardware-in-the-loop.

You will benefit from powerful ray-tracing capabilities to recreate sensor behavior and be able to easily retrieve sensor results through a dedicated interface. IR emitter world model IR properties (is this “The IR emitter will model IR properties?” and receiver electronics are considered in the simulation which can output from raw signals – waveforms – to point clouds. This solution provides a unique way to collect virtual sensor information during real-time drives and use the information to develop autopilot code.

VRXPERIENCE’s GPU Radar feature provides the capability to perform full physics-based radar scenario simulations in real time at frame rates greater than 30 frames per second. The simulations consider multi-bounce reflections and transmissions from dielectric surfaces. Multichannel and MIMO radars can be simulated using the linear scalability of GPU Radar. With the addition of GPU radar, VRXPERIENCE now provides the ability to perform ADAS and Autonomy scenario simulation with full-physics models of all key sensors - cameras, radars and lidars. The data collected out of the radar model is used to efficiently stimulate the algorithm of radar ECU digital signal processing to quickly improve the accuracy and robustness of automotive radars in edge cases. Ansys VRXPERIENCE comes with a library of objects with dielectric properties defined.

Model-and Software-in-the-Loop (MiL,Sil): Perform massive scenario variation by leveraging cutting-edge testing — on-premise and in the cloud. Assess perception performance by varying parameters across countless driving scenarios.