The LGSVL Simulator is a simulator that facilitates testing and development of autonomous driving software systems. It enables developers to simulate billions of miles and arbitrary edge case scenarios to speed up algorithm development and system integration. The simulator is under active development on Github, and we hope to grow our developer community to best serve the autonomous vehicle and robotics industries in accelerating the testing and validation of autonomous systems.
The simulator is fully integrated with the popular open source platforms Apollo (from Baidu) and Autoware (Autoware Foundation). This means that a developer working with either of these platforms can simply download our latest binaries and run the simulator to test their algorithms, whether they are for localization, perception, or path planning.
By running an autonomous vehicle, or modules of the autonomous stack, in a safe and reproducible yet realistic 3D environment, developers are able to run tests and iterate much more quickly on their algorithms. Our simulator provides real-time outputs from sensors including camera, LiDAR, RADAR, GPS, and IMU. Environmental parameters can also be changed, including map, weather, traffic, and pedestrians.
The LGSVL simulator is developed by the Advanced Platform Lab at the LG Electronics America R&D Center, formerly the LG Silicon Valley Lab.