Probabilistic Sensor Models for ADAS/AD Simulations

The Task

Algorithms of autonomous vehicles are tested by means of simulation in complex traffic scenarios to validate them at an early stage. This requires sensor models with which the simulated vehicles can detect their environment and react to it.

The Challenge

To feed the ECU software of autonomous vehicles with sensor signals in simulations, they must be processed exactly as output by a real sensor. If the sensor behavior itself does not have to be taken into account, it is sufficient to calculate classified objects and their respective motion data.

The Solution

ASM Traffi c contains generic models for different sensor types, such as radar and lidar. The sensor simulation provides ideal data (ground-truth information), which can optionally be superimposed with probabilities of events (probabilistic effects). For example, superimposition is used to simulate a typical measurement noise of radar. The simulation returns a lists of classified objects (vehicles, pedestrians, cyclists, traffic signs etc.) as well as their coordinates and motion data (distance, relative speed, relative acceleration, relative azimuth and elevation angle). These models are suitable for SIL simulation with virtual ECUs (V-ECUs) as well as for HIL simulation in hard real-time with real ECUs.