A team of researchers from Texas and Tenlessian universities to the USA developed Dronechase system based on Raspberry Pi |
---|
A team of researchers from Texas and Tenlessian universities in the USA has developed a system called Dronechase. This system, which is based on Raspberry Pi, allows for the continuous tracking of drone flights even in cases of visual contact loss. The Dronechase system utilizes a combined approach by incorporating both acoustic and visual signals, enabling it to function effectively in conditions where visibility is limited. The researchers have described the system as having a high level of mobility and affordability due to its use of Raspberry Pi. They also have plans to further enhance its functionality in order to track multiple drones simultaneously. |
The composition of the Dronechase system |
---|
Dronechase consists of six microphones, video cameras, and a low-power Raspberry Pi computer. This entire system is mounted on a movable platform, such as a car, allowing for continuous monitoring of the drone without being tied to a specific location. The researchers utilized the Yolov5 visual model during the training of Dronechase. This model automatically generated tags indicating the position of the drone in the images. Subsequently, these tags were utilized by the acoustic model to learn how to determine the drone’s location solely based on sound. |
The effectiveness of the Dronechase system |
---|
The Dronechase system has achieved impressive results in terms of its ability to localize drones. Even in situations where visual contact is lost, such as when a drone is obscured by trees or in darkness, the system is capable of localizing the drone with an accuracy of 61% solely through the audio signal. This level of accuracy significantly surpasses traditional methods of visual tracking. The researchers believe that the system’s affordability and mobility are key distinguishing features. They highlight the potential of combined acoustic and visual approaches to enhance drone safety. |