FlytBase Blog
Search and Rescue Drone

FSU team builds a Search and Rescue Drone using FlytBase

FSU Team with Search and Rescue Drone

A team of research students from Florida State University (FSU) used FlytBase to build a Search and Rescue Drone, which won the BEST PROJECT award.

Introduction

Six drone researchers from Florida State University (FSU) came together to participate in the annual FSU Senior Day in Robotics category. The team formed in September 2016 was led by Halil Yonter, a senior computer engineering student from Turkey. They were tasked with creating a novel Unmanned Aerial Vehicle (UAV), capable of scanning disaster zones and identifying unique objects of interest.

FSU Drone Team

From left to right: Sarah Hood (Dynamic Systems Engineer), Shawn Cho (Software Engineer), Alexandra Borgesen (Computer Engineer), Cody Campbell (Hardware Engineer), Halil Yonter (Team Leader), Peter Burchell (Mechanical Engineer)  

Drones at Florida State University

The Emergency Management and Homeland Security Program is part of the College of Social Sciences and Public Policy at Florida State University. They have been utilizing drones extensively to deliver world class education and disaster recovery. In search for a better alternative to their search and rescue efforts, they sponsored the project to increase autonomy and to reduce the human effort involved in Search and Rescue Operations.

Limitations of previous FSU UAVs

UAVs used by FAMU-FSU’s Emergency Management and Homeland Security (EMHS) Program can autonomously scan an area but provide no feedback on image contents, and lack a user-friendly interface for inter-process communication. The team was tasked  with creating a novel UAV, capable of scanning disaster zones and identifying objects of interest.

To address the problem, the team needed to come up with an innovative flight control architecture, featuring a powerful onboard computer capable of live image processing for object detection. The project also required integration of algorithms for color filtering and pedestrian tracking.

“The current methods applied by EMHS fail to deliver the desired quality of flight or transmission efficiency. The inferior quality of images transmitted by the current UAV meant there was a high chance of a potential risks being overlooked.”

How did FlytBase help?

“FlytBase provided all the fundamental tools and interfaces to successfully develop the drone application” says Halil Yonter, the team leader.

“Because of the strong and reliable foundation delivered by FlytBase, all that was left to do was customize the framework for our specific needs. This was highly beneficial as it allowed us to spend more time on the particular aspects of our project.”

The team realized they needed a controlling entity to oversee the operations of subsystems and to handle data transfer. The team initially decided to use Robot Operating System (ROS), a generic OS used in a variety of robotic applications. But due to the lack of specialized features and APIs designed for UAV applications, the team turned their attention to FlytOS, a framework for drone applications, built on ROS. FlytOS provides a complete drone development environment, including, APIs for navigation, telemetry, various intelligent features (computer vision, machine learning), along with web/mobile SDKs and a simulator for testing drone applications. The team reached out to FlytBase to get access to FlytOS for NVIDIA TX1.

As FlytBase is compatible with all major companion computers available in the market, Team FSU were able to integrate FlytOS with NVIDIA TX1 and build their application on top. The primary aspects of the project were an autonomous flight system with object detection capabilities. FlytOS, along with the bundled APIs, SDKs and simulator, provided the necessary tools for building a stable and reliable autonomous drone application. The available FlytBase Vision APIs helped team to accelerate development of vision algorithms, and the NVIDIA TX1 CPU provided the necessary processing power to meet intensive computing requirements of vision tasks.

The team’s task was further simplified with the use of FlytBase’s native mission control interface, FlytConsole.

The Results

search and rescue drone

The final solution is a fully equipped UAV, capable of fully autonomous flight and object detection. The drone, named “Saurus” by the team, increases the ability to conduct reliable and efficient search and rescue missions, by eliminating manual processes in favor of increased autonomy of the UAV, both for, autonomous flight, as well as, object detection through real time image processing.

object detection

The Search and Rescue Drone was a big success, as it addressed several limitations of the previous systems. The team worked for over eight months to implement the project, and their hard work and tenacity fetched them the BEST PROJECT award in the Robotics category during FAMU-FSU Senior Design Day.  

A proud Halil acknowledged the impact of using FlytBase by saying – “Our experience with FlytBase has been great. Therefore, FlytBase would be our first choice for any kind of drone application in the future.”

FlytBase is proud to be a part of this success story, and congratulates Halil Yonter and other members of team on winning the Best Project Award!

 

Add comment