BeamNG Software Ecosystem

Our Services

BeamNG GmbH supports its customers in their product development for driver’s training and advanced driver assistance system (ADAS) through tailored features and by providing freely available support libraries. Above diagram highlights features of the core simulation BeamNG.tech and their respective area of application. In addition it also shows the dependencies of our freely available libraries that target ADAS development.

BeamNG Software Ecosystem

Our Services

BeamNG.tech is a versatile simulation platform suitable for the development of state of the art advanced driver-assistance systems (ADAS). We provide a highly customizable simulation environment that supports development in terms of data acquisition and X-in-the-loop development. With our custom build physics engine and detailed vehicle models, we are taking data acquisition to the next level by combining a high degree of physical realism with elaborate sensor models and photo-realistic imagery.

In support of SOTIF’s validation and verification approach, we rely on the scenario-based development paradigm. This allows the user to quickly and systematically collect data and evaluate systems in clear-cut conditions, either programmatically or by using our Flowgraph editor that does not require any previous programming experience. To facilitate interoperability with other frameworks we also provide the freely available python interface BeamNGpy.

Be it by selecting from our large variety of vehicle models and levels, or by integrating your own models into the simulation, BeamNG helps you bring your product on the way to completion.

Sensors

Camera

We extended the classic camera sensor to include a collection of additional information. This allows it to extract as much information as possible from the simulation. Besides RGB images, it provides pixel-perfect information about depth, objects classes and object instances. It is also easily parameterized with regard to the classical camera settings such as field of view and resolution and can be attached to any part of the vehicle to match your technology demonstrator.

Lidar

Our sensor implementation mimics the behavior of real Lidar sensors. As a rotational Lidar, it generates a point cloud by relying on ray-tracing. Just like any of our sensors, it is highly customizable and generates a perfect scan of the environment. Soon it will also provide ground truths along the generated point clouds, accelerating your research and product development.

IMU

Our IMU sensor captures all the details of the agent’s driving dynamics. While the other sensors provide the data for intelligent decision making, the IMU sensor is suitable to take into account the comfort of any vehicle passenger. For non-emergency systems a steady driving behavior is essential for its successful deployment. Due to our simulation generating detailed vehicle dynamics, the IMU sensor provides the necessary means to develop a product taking into account the customer.

Ultrasonic Sensor

Based on the graphical information from the simulation engine we devised a custom algorithm that mimics the behavior of real ultrasonic sensors. Belonging to any modern vehicle equipment, the ultrasonic sensor is also indispensable in our sensor suite.

Damage Sensor

Unique to BeamNG.tech, the damage sensor provides comprehensive information about the state of the vehicle. With this sensor it is not only possible to assess the state of the outer vehicle parts but also the state of the individual powertrain components. Although lacking a real-world counterpart, this sensor is nevertheless an indispensable tool in assessing the quality of any autonomous driving assistant without putting any real-world hardware at risk.

BeamNG GmbH supports its customers in their product development for driver’s training and advanced driver assistance system (ADAS) through tailored features and by providing freely available support libraries. Above diagram highlights features of the core simulation BeamNG.tech and their respective area of application. In addition it also shows the dependencies of our freely available libraries that target ADAS development.

BeamNG is a precise, powerful tool that can be used to support driver’s training applications. From top to bottom, BeamNG is designed to simulate true-to-life scenarios that any driver may encounter. Whether it’s an obstacle from the realistic environment, or a detailed, accurate collision via the soft-body physics engine, it all works in tandem to effectively prepare trainees for the road.

Environments & Level Customization

It’s crucial to provide an authentic challenge when training drivers. That’s why BeamNG’s urban environments replicate real-world cities from Europe and America, allowing trainees to safely explore genuine obstacles and develop their skills. It’s this level of Intricacy and detail that makes the program so effective.

Trainees don’t need to feel limited to just the standard environments provided. Level and scenario customization tools can be taken advantage of in order to cater to the trainee. This flexibility means future drivers can work on their areas of weakness, or test themselves against sophisticated AI traffic scenarios.

Soft-body Physics

The BeamNG physics engine is using a bottom-up approach representing complex interactions in real time. It allows simulation of unique situations not covered in other real time simulators which provides more accurate driving dynamics (weight distribution, force propagation, friction, etc.) and more accurate damage model, allows simulating of faulty/broken parts

We have covered more information on the fundamentals of the BeamNG’s physics here

Drivetrain System

BeamNG keeps options open as far as the drivetrain system is concerned. Trainees can feel comfortable knowing their specific needs are met, whether that means driving manual, automatic, rear wheel drive, front wheel drive, and more.

Setup

In order to provide the most immersive training environment possible, BeamNG offers triple monitor support. This makes a devoted simulation setup much more feasible. Additionally, motion platform support is provided to heighten the realistic scenarios.

Force Feedback

Force feedback is a critical feature for drivers to control simulated vehicles. It transmits a haptic response to human drivers that helps them to control their vehicle and contributes to an authentic driving experience. Active and  effective feedback system can help the trainees with decision making while driving.

Not only visual factors are important but also other sensory modalities, such as force feedback through a steering wheel or torque, provides valuable information for the trainees. Although BeamNG.tech do not supply hardware but it is capable of transmitting authentic driving experience with our force feedback support system.

Data Collection & Analysis

Informed feedback is vital to the learning process. BeamNG comes equipped with a host of data extraction capabilities that are easy to access. This makes performance assessment much more immediate and clear.