Design and Implementation of Indoor Visual Autonomous Navigation System for Quadrotor UAV Based on Binocular Camera and IMU

Addressing the challenge of unmanned aerial vehicles (UAVs) navigating indoors without satellite positioning, this study introduces a novel method combining binocular camera and inertial measurement unit (IMU) data. This approach overcomes the structural differences between visual and inertial information, integrating them effectively for enhanced indoor autonomous navigation. The research presents a gradient-free path planning framework, pivotal for enhancing the UAV's ability to execute efficient and stable flight in intricate indoor settings. The cornerstone of this system is an advanced path planning algorithm, adept at swiftly identifying optimal routes and adapting to environmental shifts. The system's efficacy was rigorously evaluated through simulations and real-world tests, confirming its precision in positioning and path planning without reliance on satellite data. The findings demonstrate the system's capacity for precise localization and effective route determination, utilizing an integration of visual and inertial inputs. This research significantly contributes to the autonomous flight capabilities of UAVs, offering a valuable reference for future UAV autonomous navigation systems. By delving into visual-inertial fusion navigation technology, the study paves the way for expanded indoor UAV applications.