project 01 · Robotics / Optimization / Simulation
Secure DropAutonomous Drone Routing
A research platform for planning and simulating multi-drone parcel delivery missions.
- Python
- Gazebo
- Qt
- Google OR-Tools
- A*
- Fleet allocation
- Constraint handling
- LiDAR
- ROS / PX4 tooling
// problem
Delivering parcels with a fleet of drones means deciding which aircraft carries what, in which order, and when to recharge. All while respecting battery limits, delivery windows, terrain, wind, and obstacles.
// solution
A Python research platform combining routing optimization with Gazebo simulation, so mission plans can be generated, executed in simulation, and repaired when sensed obstacles block a route.
200+
regression tests in the validation set till now
5+
aircrafts tested in bounded Gazebo sensing demonstrations
01
Mission planning
The planner formulates delivery as a constrained vehicle routing problem: parcels are allocated across a selected fleet, and each aircraft receives an ordered route that respects payload, battery state, and delivery time windows.
Charging stations are modelled as schedulable resources, so routes can include recharge stops when a mission would otherwise exceed an aircraft's energy budget.
02
Simulation & sensing
Plans are executed in Gazebo, where simulated LiDAR sensing detects static obstacles. When an obstacle blocks a planned leg, A* graph search repairs the route locally.
A Qt-based interface gives researchers control over scenarios, fleet configuration, and replanning during simulation runs.
03
Validation
The project emphasizes reproducible testing: the research documentation records 200+ regression tests in the validation set till date, alongside successful bounded multiple aircraft Gazebo sensing demonstrations.
All results are simulation results. Real-world flight testing will begin in the next phase of research.
next project
Vora Smart AC Controller