Zvision, a leading provider of solid-state MEMS LiDAR solutions, announced at CES2023 that it will work with the NVIDIA Isaac Sim advanced simulator to create powerful tools enabling the validation of industrial sensing solutions, satisfying the demands of high-performance LiDAR simulation for more application scenarios and fully adaptive testing.
NVIDIA Isaac Sim, a tool for robotics simulation applications and synthetic data generation, is used by robotics experts to train and test robots more efficiently, recreate realistic robot interactions within specified settings, and validate the performance in different scenarios that go beyond the real world. With Isaac Sim, companies have dramatically reduced the costs of development, testing, and data acquisition, and accelerated time-to-market for commercial deployments.
NVIDIA Isaac Sim features many scenarios and application environments, using its accurate physics-based rendering to reproduce multiple shapes and sizes of obstacles to form a complete functional scene. The platform also allows users to validate the performance state and technical capabilities of LiDAR against the complexity and uncertainty of a given scene. With the digital twin, in combination with a variety of surroundings and different use scenarios available in Isaac Sim, developers in the robotics field can determine which performance modes are most suitable for the selected application based on solutions for complex edge cases.
As NVIDIA's robotics simulation platform, Isaac Sim offers the benefit to develop, test and evaluate robotic systems and algorithms in a simulation environment. Developers and researchers can also use Isaac Sim’s many tools, including simulation of robot dynamics to test control algorithms, simulation of robot sensors to generate realistic cameras, depth and segmentation images, LiDAR perception, IMU, simulation of testing algorithms in different environments and condition scenarios, and providing training sets of random objects and attributes.
The MEMS LiDAR portfolio created by Zvision includes the ML-30s+ short-range LiDAR and the ML-Xs forward long-range MEMS LiDAR. The ML-30s+ short-range LiDAR is the best for perception with detection range of 25m@R=10% and an ultra-wide field of view (FoV) of 140°X70°. Robots can view surrounding objects more clearly, and the higher quality image like point cloud captured with an angular resolution of 0.44°x0.44°. The FOV of 140°, which is larger than the typical 120°, not only covers more wider area, but it can also detect objects in the corresponding edge area faster and sooner without adjusting the heading angle. In order to detect and identify objects more accurately in blind areas, the vertical FoV of 70° is set at 20° above and 50° below the horizontal. The ML-Xs, a MEMS long-range LiDAR with outstanding exceptional capability, also equips a 1550nm laser with effective range of 250 meters and an FoV of 120°x25°. The ultra-long range-sensing capability can accurately cover further locations and identify obstructions in the forward path. As a result, the features and performance of the ML series LiDAR model as defined by Isaac Sim simulation software will further advance the development of industrial sensing solutions.
As a member of NVIDIA's Jetson partner ecosystem, Zvision will continue to strengthen its collaboration and support for NVIDIA Isaac Sim to develop comprehensive robot sensing capabilities and deliver high-performance LiDAR solutions.