Highly automated or autonomous driving relies on sensor technologies such as lidar, radar, camera, ultrasonic or GPS to interpret the vehicle surrounding and locate itself in the environment. The trend towards shorter development cycles in the automotive industry requires hardware benchmarking and vehicle packaging as well as model, software and hardware in the loop early in the development cycle.
The Tronis® sensor package provides all the means for vehicle and sensor simulation early in the development cycle. It gives you real-time access to raw data being generated close to the analog digital interface of each sensor and thus allows evaluating and prototyping of post processing and control algorithms.
The Tronis.Camera module acts as a virtual camera attached to the vehicle. You can precisely adjust the position of the monocular or stereo multiple camera and parameterize them according to your needs, e.g. in terms of camera resolution, view ankle etc. The virtual camera generates raw data of image flows and makes it available to the TWT.Control module in real-time. Likewise, you will be able to evaluate the performance of the camera in various scenarios as e.g. in night drives, adverse weather conditions, glazing sunlight or further critical visual situations. An additional post processing also extends the camera model to provide desired lens effects, e.g. vignetting or distortions.
In contrast to a stereo camera or a laser scanner, a lidar sensor does not perceive the environment from one snapshot, but instead scans the topology of the environment point by point. The movement of the vehicle must be updated at high frequencies, whereby the update frequency of the classic rendering pipeline of the gaming engine is not high enough. In order to guarantee high-frequency data acquisition, the efficient rayshooting technology is used.
Tronis.Radar gives a virtual representation of a flow of raw data images. It is based on 3D-raytracing working in real-time. In the same way as the camera module, you can precisely adjust the position and parameterize the virtual sensor according to your hardware configuration. You can adjust path lengths, signal amplitudes and frequency shifts. As is a real radar, ghosting effects resulting from multiple reflection are taken into account, too. Tronis.Radar provides a C++ interface for accessing the data.
Complex environments require easy tools for editing: Import or create your own 3D environment and exploit the full capabilities of the Unreal engine. You are free to use a pre-built environment to adjust it to your needs, e.g. to simulate a highway drive, an inner city cruise or a parking garage drive thru.
In addition, you can extend your environments with typical assets such as road and traffic signs, buildings and foliage. And most importantly, you can keep it simple by using a graphical interface adjusted to the needs of an engineer.
With Tronis® even complex effects, e.g. shadows of wind-blown foliage, reflections of a wet road or rain, are considered. When looking at a road sign, in contrast to the first-generation simulation environments, not only textures are visible, but also the fine reflection honeycombs, which can increase the visibility of the sign depending on the light conditions.
The use of large 3D worlds also enables virtual trips within mountain, forest, city and lake areas.
Testing highly automated driving centers around representative, repeatable and meaningful scenarios. Once you have configured your environment, you will be able to extend it with dynamic objects which come along with their own behavior or the possibility to define such a behavior. Other vehicles will approach, pedestrians will cross the road and traffic lights will give way.
- Driver assistance systems
- Localization and mapping
- Training of neural networks for e.g. object recognition
- Radarsimulation and sensors
- Automatic creation of 3D maps
You have your sensors configured, your environment built and your scenarios defined? Get ready for running your own individual simulation! Use our interfaces to the Unreal engine to test and prototype your own control or signal processing algorithms implemented on Matlab/Simulink. Use signal data from different sensor sources to test your virtual vehicle in different environmental conditions or track corner cases in which your vehicle behavior deviates!
Customization: Tronis® provides all the means for building your own environment for prototyping highly automated driving systems. However, much is the same way any given environment, scenario and sensor will help you to build your own simulation, customization may be required to match the environment perfectly with your needs. We may build your environment, we have visualized your scenario and we may build your sensor model and a lot more. Tronis® combines a prototyping environment with given assets with the possibility of project-driven customization.