ReScience develops AI-powered swarm intelligence, cooperative algorithms, and autonomous mission-management solutions for fixed-wing Unmanned Aerial Vehicle (UAV) platforms. Our expertise combines embedded artificial intelligence, multi-UAV path planning, cooperative target detection, automated task handover, and real-time flight control integration.

We support the complete autonomous system lifecycle, from AI model training and algorithm prototyping to hardware-in-the-loop testing, communication network design, flight-data analysis, and field demonstrations.

Our Capabilities

AI-Powered Object Detection & Edge Computing

  • Custom computer vision and deep learning model training (Transfer Learning)
  • Edge-AI deployment and optimization (TensorRT, achieving high FPS on onboard flight computers)
  • Target search, feature extraction, and automated detection
  • Post-mission flight telemetry and target positioning data packaging

Swarm Coordination & Task Handover

  • Centralized and decentralized decision-making architectures
  • Automated leader/manager UAV coordination and approval protocols
  • Dynamic task reconfiguration and real-time swarm behavior rules
  • Cooperative positioning, target tracking handover, and automated return-to-base (RTH) management

Autonomous Path Planning & Area Scanning

  • Grid-based area scanning for homogeneous UAV swarms
  • Automated waypoint generation based on spatial boundaries
  • Route planning in obstacle-rich and unconstrained environments
  • Dynamic route regeneration centered on detected targets

Field-Demonstrated Swarm Experience

ReScience collaborated with Eskişehir Technical University and the Presidency of Defense Industries (SSB) on the AI-Powered Autonomous Swarm Intelligence project. The initiative demonstrated a multi-UAV swarm of low-cost, expendable fixed-wing platforms performing cooperative target search, positioning, and task handover in real-world flight conditions.

Key Project Achievements:

  • Field Demonstration: Successfully validated autonomous grid-based scanning, target detection, cooperative position confirmation, and task handover in competitive operational trials.
  • Rapid Target Detection: Achieved target detection and localization in under 3 minutes and 12 seconds during field tests.
  • Onboard AI Performance: Deployed optimized TensorRT models delivering ~16 FPS processing on low-power onboard flight computers.

Our Engineering Approach

We combine theory-driven algorithmic design with rigorous simulation and experimental field validation:

We work closely with platform manufacturers and system integration teams to align algorithms with hardware payloads, ground control stations, and mission constraints.

Application Areas

Our capabilities support applications including:

  • Autonomous fixed-wing UAV swarms
  • Cooperative search, detection, and target localization
  • Grid-based area coverage and border surveillance
  • Decentralized multi-agent coordination and task allocation
  • Low-cost, expendable (attritable) aerial sensing systems
  • Edge AI vision applications for airborne platforms

R&D Collaboration

ReScience collaborates with defense industry partners, government organizations, technology companies, and academic institutions in AI model optimization, swarm algorithm development, flight computer integration, field flight testing, and technical consultancy.