Sensor simulations are one of the driving forces behind the dynamic development of automated transportation and autonomous solutions. In the evolving industry, RGL (Robotec GPU Lidar) stands out as a performant alternative to existing solutions.

Developed to overcome the limitations of traditional CPU-based simulations, RGL harnesses the power of GPU acceleration to deliver unprecedented efficiency, scalability, and flexibility.

The list of potential applications is long, from the automotive industry to agriculture, mining, and multi-industry drone inspection.

RGL is designed for practical, real-life problem-solving.

Here is a short guide to RGL created by our RGL team.

Introducing RGL

RGL is a cross-platform (Windows and Linux) C/C++ library developed by Robotec.ai for simulating lidars on CUDA-enabled GPUs, accelerated by RTX cores (if available).

It is freely available in a GitHub repository with plugins for O3DE, Gazebo, and Unity.

The Evolution of RGL

RGL was born out of the need to enhance the performance of lidar simulations in the automotive industry and robotics. Traditional methods were hampered by the limited processing power of CPUs, creating significant bottlenecks. Our team recognized the potential of GPUs’ massive processing capabilities early on and set out to develop a tool that could meet the complex demands of modern simulations. Over the years, RGL has evolved from basic lidar simulations to a robust library packed with advanced features like velocity distortion, Gaussian noise handling, and instance segmentation.

Unmatched efficiency

GPU acceleration is at the core of RGL‘s efficiency. Unlike CPUs with a strictly limited number of cores, GPUs boast thousands capable of processing many calculations simultaneously. This architecture allows RGL to simulate millions of lidar rays per frame, achieving real-time processing and detailed simulations unmatched by traditional solutions.

Scalability and Flexibility

RGL’s scalability is a game-changer, allowing users to simulate from a single lidar sensor to a complex network without any performance drop. This is particularly beneficial in automotive simulations, where comprehensive environmental scanning is crucial. Thanks to our C-API, RGL’s flexibility ensures seamless integration, making it a versatile and adaptable solution for industries ranging from automotive to agriculture and beyond.

Overcoming Challenges with Innovation

Developing RGL has not been without its challenges. One major hurdle was accurately modeling lidar interactions in diverse environments, particularly where traditional methods struggled to maintain accuracy and performance.

Our team, armed with deep knowledge of physics and optics, tackled this by combining theoretical research with practical experiments. This led to the development of sophisticated algorithms capable of handling these interactions in real time without compromising performance.

Transforming Autonomous Vehicle Development

In the automotive sector, RGL accelerates the development of autonomous vehicles by providing high-fidelity environmental simulations in real time.

Engineers can quickly iterate and test various scenarios, enhancing vehicle navigation, collision avoidance, and object detection systems. The ability to simulate millions of lidar rays ensures that every detail is captured, leading to more reliable and safer autonomous systems.

Expanding Horizons in Agriculture and Mining

Beyond automotive, RGL significantly benefits industries like agriculture and mining. In agriculture, it aids in simulating autonomous tractors and drones, improving precision farming techniques through accurate field mapping and obstacle identification.

In mining, RGL helps simulate autonomous haul trucks and other machinery, optimizing routes and ensuring safety in challenging environments. The high efficiency and detail provided by RGL are crucial for these applications, where real-time data processing is essential.

Getting Started with RGL

For those interested in exploring what RGL can offer, our GitHub repository provides detailed documentation, sample projects, and a comprehensive API guide. We also offer direct support for specific queries or custom integration needs.

Robotec GPU Lidar represents the future of high-efficiency lidar simulations. Whether you are in the automotive industry, agriculture, mining, or any field that relies on autonomous systems, RGL is designed to meet your needs.

Want to keep up with Robotec? Follow us on LinkedIn!

Author:

Piotr Mrozik, Software Engineer at Robotec.ai