Breaking Through The Difficulty of Robot Localization in GNSS Denial Environment

In today's era of intelligent development, mobile robots have become important tools in many industries. However, achieving high-precision positioning remains a huge challenge in environments where Global Navigation Satellite System (GNSS) signals are limited or denied. In response to this issue, researchers have proposed an improved adaptive general corner detection (AGAST) algorithm and applied it to a vision/inertial fusion positioning system, significantly improving the positioning accuracy of mobile robots.

Innovative Application of AGAST Algorithm

The AGAST algorithm optimizes the Visual Odometry (VO) algorithm through local histogram equalization and adaptive threshold detection. This improvement not only improves the quality of feature point extraction, but also enhances the stability and accuracy of visual odometry in complex environments. Specifically, local histogram equalization can effectively handle lighting changes and ensure the quality of feature point extraction under different lighting conditions. Adaptive threshold detection dynamically adjusts detection parameters based on environmental changes, further improving the robustness of feature point detection.

Factor Graph Optimization: Integrating Vision and Inertial Navigation Systems

On the basis of optimizing the visual front-end using the AGAST algorithm, researchers adopted the Factor Graph Optimization (FGO) algorithm to deeply integrate the visual odometer with the Inertial Navigation System (INS). Factor graph optimization is an optimization method based on probability graph model, which can effectively integrate information from different sensors and improve the overall performance of the positioning system. Through this fusion method, mobile robots can achieve high-precision positioning in environments with missing GNSS signals.

Experimental verification: significantly improved positioning accuracy

To verify the effectiveness of the improved algorithm, researchers conducted tests on publicly available indoor and outdoor datasets. The results showed that the improved AGAST algorithm achieved a 22.8% improvement in localization accuracy on indoor datasets compared to the mainstream VINS Mono algorithm, and a 59.7% improvement in localization accuracy on outdoor datasets. This significant improvement indicates that the AGAST algorithm has excellent performance in handling complex environments and varying lighting conditions.

Wide application prospects

The improved mobile robot vision/inertial navigation fusion positioning system not only performs well in GNSS denied environments, but also has broad application prospects. In fields such as logistics, security, and agriculture, mobile robots often need to perform tasks indoors or in complex environments. By introducing the AGAST algorithm, these robots will be able to navigate and locate more accurately, improving work efficiency and task completion.

The proposal and application of the Adaptive General Corner Detection (AGAST) algorithm provide an innovative solution to the high-precision positioning problem of mobile robots in GNSS denied environments. Through the deep integration of vision and inertial navigation systems, this technology not only improves positioning accuracy, but also lays a solid foundation for the application of mobile robots in various complex environments. In the future, with the continuous advancement and improvement of technology, the AGAST algorithm will lead the development trend of mobile robot positioning technology, creating a more intelligent and efficient new era.

https://www.mide-act.com/industries/Robotics.html

https://www.mide-act.com/geodetic-antenna-supplier/