Project Abstract

Autonomous navigation of Unmanned Aerial Vehicle (UAV) plays an essential role in designing robust industrial UAVs for various applications. One interesting application is the ability to navigate indoors and under tree canopies. Such drones will be extremely valuable for security and smart agriculture applications.

Recently, diverse sensors and multi-sensor fusion algorithms have been proposed for such applications; however, most approaches are vision-based. For this project, we will be designing a novel multi-sensor fusion algorithm based on Deep Learning (DL) algorithms, which supports obstacle avoidance and situational awareness in a commercial drone frame with multiple proximity sensors to reinforce the vision-aided system.

Here, a set of multi-sensors represents a combination of an Inertial Measurement Unit, RGB-D camera, Lidar, Ultrasonic sensor, Infrared-based Night Vision Camera, and Infrared Distance Measuring sensor. This will enhance the drone’s autonomous navigation capability even in unexpected environments. We are currently working alongside researchers from the University of Southampton Malaysia, Aerodyne Bhd and MIMOS Bhd to develop and implement the required algorithm for DL DL-aided sensor fusion algorithm.