Gazebo gets cross-platform GPU-accelerated SONAR simulation via WGPU
Replace CUDA dependency: new WGPU plugin runs sonar on any GPU hardware.
The Gazebo community is hosting its May 2026 meeting on the 27th, featuring a key presentation by Naitik Pahwa from NSUT, Delhi, on a WGPU-based SONAR plugin. The project directly addresses a long-standing limitation in underwater simulation: the existing sonar plugin is built on CUDA, tying it exclusively to NVIDIA GPUs. This new work migrates the GPU-accelerated pipeline to wgpu, a modern cross-platform graphics API that abstracts over Vulkan, DirectX 12, and Metal. The goal is to maintain the high performance of real-time sonar rendering while drastically improving portability.
The plugin is designed to slot seamlessly into Gazebo and ROS 2, the standard stack for robotics simulation. By leveraging wgpu, developers can run the same sonar simulation on AMD, Intel, or even Apple GPUs without altering code. This matters for underwater robotics teams who need reliable virtual testing of acoustic sensors but may not have access to NVIDIA hardware. The meeting will also include updates from Open Robotics on Gazebo's roadmap. The session is virtual, recorded, and open to all, with sign-up via Luma.
- Replaces CUDA dependency with WGPU for cross-platform GPU acceleration (Vulkan, DirectX, Metal).
- Maintains high-performance real-time sonar simulation for underwater robotics in Gazebo.
- Integrates with ROS 2, removing NVIDIA hardware lock-in for sonar testing.
Why It Matters
Enables underwater robotics simulation on any GPU, democratizing access to high-fidelity sonar testing.