Georgia Tech Miniature Autonomous Blimp (GT-MAB)

GTSR has developed the Georgia Tech Miniature Autonomous Blimp (GT-MAB), which is a small and low-cost aerial robotics platform for controls and communication research. The GT-MAB can fly for longer durations compared to typical drones, and are much safer. Additionally, blimps share similar dynamics with underwater vehicles, which makes the GT-MAB a practical solution for prototyping and testing control algorithms for both aerial and underwater robots. The GT-MAB can carry sensors like cameras, single-point lasers, and an inertial measurement unit (IMU). Because the GT-MAB’s payload must be less than 80g, most computation is performed offboard on a base station computer. Data can be transferred between the base station and GT-MAB through various communication modes.

For my senior capstone project (Fall 2018 – Spring 2019), my team is planning on developing additional hardware and software for the GT-MAB such that the GT-MAB is capable of guiding humans to different locations in indoor environments. Often times, it is difficult for people to find the room they need to get to in an unfamiliar building. Even if a map is available, it can be difficult to interpret. The motivation for developing a robotic guide is to create a platform that can save people time by bringing them directly toward their intended destination, and potentially reduce foot traffic in cluttered indoor environments. This could be used in airports to guide people to their gate, or in sports stadiums to help people get to their seat. Future applications of this platform could include using the GT-MAB as a guide for evacuation procedures, since the GT-MAB’s height can help people easily identify and follow the GT-MAB to the nearest emergency exit.

This project is called GT-MAB Guide. For example, a user will speak into the GT-MAB’s microphone, asking a question like “Where is the cafeteria?”. The GT-MAB will acknowledge the request, and navigate toward the target destination. For this project, we will be implementing a novel deep learning approach to localization. This project is under active development, and this section will be updated once the GT-MAB Guide is complete!

search previous next tag category expand menu location phone mail time cart zoom edit close