Roadmap

Developments

Six phases from simulation foundation to full multi-drone deployment in realistic field conditions. Each phase builds directly on the last.

Completed
In Progress
Upcoming
01
Foundation Completed

Establishing the simulation environment and core infrastructure. Docker-based development workflow, ROS 2 workspace structure, Gazebo world setup, and baseline X500 drone model integration with PX4 autopilot. Verified basic takeoff, hover, and landing in simulation.

  • ROS 2 Humble
  • Gazebo Simulation
  • PX4 Integration
  • Docker Environment
  • X500 Model
  • MAVROS Bridge
02
Perception Completed

Onboard camera pipeline feeding a RF-DETR detection model for top-down human identification. ROS 2 perception nodes publishing detections with confidence scores.

  • YOLO26/RF-DETR
  • RGB Camera
  • Detection Node
  • Confidence Scoring
  • Roboflow Dataset
03
Flight Systems In Progress

Full autonomous flight capability in simulation. Waypoint navigation, velocity control, geofencing, and mission abort logic. Controller parameter tuning for stability under payload variation. OFFBOARD mode control via ROS 2 topics.

  • Waypoint Nav
  • Offboard Control
  • Velocity Commands
  • Geofencing
  • Mission Logic
  • Controller Tuning
04
Autonomy Upcoming

Closing the loop between perception and flight. The drone will autonomously plan search patterns, respond to detections, update a shared map, and execute target confirmation approaches. Includes GPS-denied localisation using visual odometry and onboard IMU fusion.

  • Search Patterns
  • Visual Odometry
  • Map Building
  • Reactive Planning
  • GPS-Denied Nav
  • Target Approach
05
Swarm Upcoming

Extending single-agent capability to a coordinated multi-drone swarm. Decentralised task allocation ensures coverage without a single point of failure. Inter-agent communication handles shared map updates, deconfliction, and dynamic reassignment when nodes drop out or locate a target.

  • Multi-Agent Sim
  • Task Allocation
  • Coverage Planning
  • Deconfliction
  • Shared Map
  • Node Dropout Handling
06
Realism Upcoming

Hardening the system against real-world conditions. High-fidelity forest environments with wind simulation, canopy occlusion, dynamic lighting, and adverse weather modelling. Hardware-in-the-loop trials. Transition from simulation to physical drone deployment.

  • Forest3D Worlds
  • Wind Simulation
  • Lighting Variation
  • HITL Testing
  • Hardware Trials
  • Field Validation