eInfochips presents
GPU-Accelerated Real-Time SLAM Optimization for Autonomous Robots
Thursday, July 10, 2025 | 8:00am PDT | 11am EDT | 17:00 CEST
Webinar Overview
Optimizing execution time for long-term and large-scale SLAM algorithms is essential for real-time deployments on edge compute platforms. Higher throughput of SLAM output provides faster map refresh rate and quicker decision-making scenarios. RTABMap is a state-of-the-art and popular SLAM algorithm used in autonomous mobile robot space. RTABMap comes with core SLAM libraries able to perform Lidar-based and Vision-based SLAM with multiple sensor configurations. In this talk, we will discuss how Lidar-based SLAM using PointCloud data can be made faster by leveraging GPU-based libraries on NVIDIA platforms. The detailed comparative results and methodology will provide good insight into how RTABMap functions execution time can be reduced for the AMR applications. This webinar will also provide idea about the effective ways in which a SLAM algorithm can be accelerated on resource-constrained devices.
In this webinar, we have discussed
- Explore specific GPU-based libraries and frameworks for Lidar SLAM acceleration
- GPU libraries integration with RTABMap, focusing on modifications and optimizations
- Provide benchmarks and performance comparisons showing the impact of GPU acceleration on SLAM execution time
Register on Webinar
Speakers
Samarth Shukla holds a B. Tech in Robotics and Automation from Parul Institute of Technology. He has close to two years of experience working on autonomous mobile robots, robotic manipulators, SLAM, and navigation using the ROS2 framework. His work also involves CUDA programming, digital twin systems, and simulation using NVIDIA Isaac Sim. In addition, he has contributed to projects involving artificial intelligence and machine learning, and has been actively using C++, Python, and GPU programming in his development work.
Vishal Bhapkar
AI Solution Engineer, eInfochips (An Arrow Company)
Vishal Bhapkar holds a B. Tech in Computer science, Pune University. He has around 4 years of experience working on AI Computer vision, perception and robotic operations like SLAM, and navigation using the ROS2 framework. His work also involves 3D computer vision based activities like Image classification, object detection or segmentation, CUDA programming, and simulation using NVIDIA Isaac Sim, Gazebo etc. In addition, he has contributed to projects involving AI/ML based pipelines with GPU Programming on edge devices with C++ and Python.
