Description
In case of embedded devices or applications, throughput plays the key role in terms of computation time, limit, memory, and executing it in real time. While working in robotics, for autonomous mobile robots, mapping and navigation are the major algorithms that direct the robot to move and behave in certain applications. For the robot to have the highest computational power, quick response and execution is necessary to avoid any collision or any kind of ambiguities in detection and mapping. Visual SLAM needs to process huge image data through the sensor and convert it into point cloud and other data that requires high performance computing hardware.
Necessary hardware configurations and algorithmic parameters changes significantly affecting the use case should be aware of and need to be taken care of.
This study has two main aims; one to identify and examine the camera performance in different hardware configurations and the second is to measure how some hardware configurations of different boards and some parameter changes affect the output rate of VSLAM algorithm.
The focus is mainly VSLAM and Jetson Orin Boards here. Let’s learn more about it.
Project Highlights
- VSLAM Algorithm
- ToF and RealSense
- Workflow of RTAB-Map with ToF/RealSense Camera
- Assessment Parameters
- Results on Jetson Orin Devices