TurboMap: GPU-Accelerated Local Mapping for Visual SLAM
Parsa Hosseininejad, Kimia Khabiri, Shishir Gopinath, Soudabeh Mohammadhashemi, Karthik Dantu, Steven Y. Ko
TL;DR
TurboMap addresses bottlenecks in visual SLAM local mapping by offloading GPU-friendly tasks—Search for Triangulation and Map Point Fusion—to the GPU, maintaining GPU-resident keyframe storage, and integrating a GPU-based Local Bundle Adjustment solver, while CPU-side logic handles keyframe culling. Built atop ORB-SLAM3 with CUDA, it achieves average local-mapping speedups of $1.3\times$ on EuRoC and $1.6\times$ on TUM-VI without sacrificing trajectory accuracy, across desktop and embedded platforms. The approach reduces latency variability and improves robustness under high mapping loads, highlighting the practical potential of GPU-accelerated local mapping for real-time SLAM on diverse hardware. The work is generalizable to other global-map SLAM systems and provides a public implementation for broader adoption.
Abstract
This paper presents TurboMap, a GPU-accelerated and CPU-optimized local mapping module for visual SLAM systems. We identify key performance bottlenecks in the local mapping process for visual SLAM and address them through targeted GPU and CPU optimizations. Specifically, we offload map point triangulation and fusion to the GPU, accelerate redundant keyframe culling on the CPU, and integrate a GPU-accelerated solver to speed up local bundle adjustment. Our implementation is built on top of ORB-SLAM3 and leverages CUDA for GPU programming. The experimental results show that TurboMap achieves an average speedup of 1.3x in the EuRoC dataset and 1.6x in the TUM-VI dataset in the local mapping module, on both desktop and embedded platforms, while maintaining the accuracy of the original system.