QUICK-SWARM: Quick Unmanned Identification and Control with Knowledge-based SWARM detection

We present a cutting-edge solution for the real-time detection and classification of drone swarms, leveraging advanced AI algorithms. Our system boasts: Ultra-Fast Processing Capabilities: Our system ensures swift analysis and response to drone swarm activities, crucial for timely intervention or management. To achieve this, we leverage cutting-edge processing technology across four advanced hardware platforms: NVIDIA Jetson boards (GPU), Tenstorrent AI accelerator, FPGA boards, and the Intel Loihi neuromorphic platform. This strategic choice optimizes detection speed, supports diverse sensor data, and maintains low power consumption, surpassing the limitations of traditional CPU-based classification. Layered Approach Based on Multiple Sensors: Employing a layered detection strategy, we integrate data from various sensors including radar, RF, and DVS camera. This approach enhances the accuracy and reliability of drone swarm identification across diverse environmental conditions and time of the day. Processing Multiple Drone Features: Our AI algorithms are adept at analyzing a wide range of drone features, encompassing size, shape, flight patterns, and communication signals. This comprehensive analysis enables thorough swarm assessment and threat identification. Adaptability to Various Hardware Choices: Designed for flexibility, our solution seamlessly integrates with a spectrum of hardware configurations, accommodating diverse operational requirements and infrastructure setups. Utilizing frameworks such as OpenVINO facilitates efficient model deployment and optimization across different hardware platforms. By combining cutting-edge technology with adaptable architecture, our Real-time AI-based Drone Swarm Detection and Classification system offers unparalleled effectiveness and scalability in safeguarding critical assets and public safety against emerging drone threats.