Physics Based Sensors - AVSimulation

Physics Based Sensors

Physics Based Sensors

Our Physics Based Sensors Pack is specially designed for engineers who need to design or validate sensor perception algorithms. It offers a complete solution for advanced and realistic modeling facilitating validation by simulation.

DISCOVER Physics Based Sensors

The Pack consists of several tools and two types of sensor models:

Level 3, referred to as "Full Physics," aims for maximum realism without compromising on physics accuracy. This level enables precise reconstruction of raw signals to support perception algorithms. The Radar sensor is the first available sensor at this level.
Level 2, known as "Real Time Physics," applies physical principles optimized for efficient calculations. It also features generic processing algorithms to produce intermediate outputs. Two sensors are modeled at this level: Radar and Lidar.

Benefits

Optimized performance and easy handling

Improved flexibility

Features

Optimized modeling of a real time physics based radar and lidar sensor (L2)

Developed in collaboration with OKTAL-SE

More details

it relies on the proven OKTAL-SE EM solver capabilities:

EM interactions (masking, shadowing, coupling of targets)
Multiple paths phenomena
- Scattering computation using Physical Optics (PO)
- Multiple reflections computation using Geometrical Optics (GO)
Clutter model based on textures (for details)

Characteristics

At Level 2 the radar model provides:

Real time computation
Optimized Ray-tracing
Included generic processing which enables to provide a Radar point cloud in real-time that exhibits all typical EM phenomenon like coupling between targets or with the environment, multiple reflections, ghost target, etc.

Parameters

Radar:
- Frequency
- Aperture
- Resolution
- Emission and reception antennas: type (Generic, Gaussian, Diagram), polarization
Computation:
- Ray tracing sampling grid
- Subdivision criteria
- Intensity threshold
Display options
Debug option

Outputs

Network message:
- Radar plot number
- For each: azimuth, distance, speed, intensity
Plot display in debug mode

High level modeling of a full physics radar sensor (L3)

Developed in collaboration with OKTAL-SE

More details

it relies on the proven OKTAL-SE EM solver capabilities:

EM interactions (masking, shadowing, coupling of targets)
Multiple paths phenomena
- Scattering computation using Physical Optics (PO)
- Multiple reflections computation using Geometrical Optics (GO)
Clutter model based on textures (for details)

Characteristics

At Level 3 the radar model provides:

A rich and complete package that includes additional tools to build, customize and analyze EM simulation.
RAW data to enable real signal processing integration
Range-doppler maps

Parameters

Radar:
- Frequency
- Aperture
- Resolution
- Emission and reception antennas: type (Generic, Gaussian, Diagram), polarization
Computation:
- Ray tracing sampling grid
- Subdivision criteria
- Range doppler map parameters: dynamic range, domain and distance sampling, Doppler and velocity sampling
Debug option

Outputs

Raw data: EM Fields (amplitude, phase) in range gates
Range doppler map:
- Display
- Direct access in memory
- Video and images

Tools for preparing the simulation and analyzing the results

Helps to enrich the description of the scene and its objects with materials describing their physical properties

Uses Cases

Operational safety

Operational safety is at the heart of discussion surrounding autonomous vehicles. The Physics Based Sensors Pack allows you to check the influence of the scenario and the environment on perception. A precise modeling of the propagation allows the user to identify the weak points of the perception algorithms in complex situations, and to analyze and improve them in order to help anticipate risks and design a safe and efficient vehicle.

Want to discover how SCANeR can change the way you test and validate ?

Ask for a free trial now