TUKAN Interceptor is an affordable, fully autonomous AI interceptor with a 3D-printed airframe. The project brings together airframe and software development with practical testing. It is a low-cost defensive fire-and-forget drone designed to neutralise airborne or ground targets without requiring FPV flying experience. The main idea is to remove the human pilot from the guidance process, making the system immune to radio jamming.
The operator selects the flying target on the LCD screen and confirms the selection (or does so remotely). The drone then takes off automatically from a box or a Hecthor UGV by ProLab engineering (or from a handheld launcher) and flies towards the target fully autonomously.
Interceptor can also be launched from a handheld launcher. The operator simply aims at the target and confirms the selection; the drone then continues towards the target fully autonomously.
The entire drone airframe is manufactured using 3D printing. In addition to a standard flight controller (FC) and electronic speed controller (ESC), it carries a Raspberry Pi and a dedicated Hailo-8 AI accelerator for fast target recognition.
Advanced targeting: The system is designed to refine the point of impact when close to the target (for example, specifically targeting the rear of a tank or the junction between an aircraft’s fuselage and wing).
The software uses a container architecture so detection models can be easily swapped to suit the weather or target type.
The current design is the result of four generations of development and extensive testing. In the fourth generation, we moved from a rounded shape to an angular one, allowing the electronics to fit inside the airframe more efficiently and in a more organised layout.
Real-world testing also has consequences: hundreds of destroyed propellers and frames, along with piles of burnt-out boards and components. Yet each one has brought us closer to a reliable interceptor.
