Vendor Agnostic SONAR Implementation

This project develops a vendor-agnostic GPU-accelerated SONAR simulation plugin for Gazebo by migrating the current CUDA-centric approach to a wgpu-based compute pipeline. Current existing physics-based underwater sonar plugin rely on CUDA for speed, which limits users to NVIDIA GPUs and complicates deployment in mixed hardware environments (AMD/Intel/Apple Silicon GPUs, and future-proofed stacks). Recent advances in cross-platform GPU APIs and the Rust wgpu ecosystem make it feasible to run high-throughput compute workloads across Vulkan / DirectX / Metal backends via a unified abstraction to utilize GPGPU capabilities. This project uses that foundation to implement sonar image synthesis on the GPU without vendor lock-in, while keeping Gazebo/ROS 2 integration clean and upstream-friendly.