- UAV Units: Multiple small-propeller quadcopter/hexacopter units
- IMU Sensors: Inertial Measurement Units for pose estimation
- Flight Controllers: Individual controllers for each UAV unit
- Communication Hardware: Wireless network modules for inter-UAV communication
- Power Systems: Distributed battery systems across UAV layers
- Gazebo: 3D robot simulation environment
- Custom Control Algorithms: Hierarchical cascading control implementation
- Matrix Operations: Linear algebra for control allocation
- Network Protocol: Grouped communication system for data management
- Cascading Control Method: Hierarchical top-down control architecture
- Mixer Algorithm: Thrust-torque to propeller speed conversion
- Thrust-Torque Converter: Layer-to-layer command translation
- Control Allocation Matrix: Dynamic propeller speed distribution
- Flight Control Software: Custom implementation for multi-layer coordination
- Simulation Framework: Gazebo-based testing environment
- Mathematical Modeling: Propeller aerodynamics and thrust calculations
The structural design and control code for the distributed UAV system cannot be publicly released at this time. This README provides a comprehensive description of the system architecture, methodology, and results.
For detailed technical discussions or collaboration inquiries, please contact via email.
This project proposes an innovative multi-UAV platform architecture that leverages multiple small-propeller UAVs in a layered configuration to achieve superior payload capacity, operational safety, and flight redundancy compared to conventional single large-propeller UAVs.
Design Goals:
- Provide sufficient thrust for takeoff with payload
- Enable controlled movement with attached payload
Relationship between Propeller Size and Maneuverability:
Thrust (T) = CT × ρ × ω² × rp⁴
Torque (Q) = CM × ρ × ω² × rp⁵
Where:
- CT = Thrust coefficient
- CM = Torque coefficient
- ρ = Air density
- ω = Propeller speed (angular velocity)
- rp = Propeller radius
Under the same thrust and torque requirements:
- Enhanced Safety: Smaller propellers generate less thrust per unit, reducing operational hazards and improving safety during human interaction
- Improved Stability: Multi-unit structure provides inherent redundancy and stability through distributed load
- Better Maneuverability: Smaller propellers offer superior control precision
- Scalability: Easier to add or remove units based on mission requirements
- Propose Innovation: Design a scalable and reconfigurable UAV platform based on small-propeller UAVs
- Develop Control Method: Create cascading control method for multi-UAV coordination
- Ensure Stability: Guarantee flight stability across all mission profiles
- Demonstrate Scalability: Validate payload scalability through experimental testing
- Evaluate Redundancy: Test flight redundancy with unbalanced configurations
Key Features:
- Hierarchical Structure: Multi-layer organization with cascading control
- Top Layer Unit: Estimates pose data from IMU sensors and serves as control center
- Top-Down Communication: Calculates and transmits desired thrust-torque commands to lower layers
- Multi-Airframe Support: Compatible with various UAV configurations (Quad, Hexa, Hybrid)
Control Hierarchy:
Top Layer (Control Center)
↓ Thrust-Torque Commands
Middle Layer(s) (Execution Units)
↓ Thrust-Torque Commands
Bottom Layer (Execution Units)
Network Grouping System:
- Reduces data workload and stabilizes network communication
- Prevents network congestion in multi-unit systems
- Only UAV units in the same network group are allowed to exchange messages
- Hierarchical message routing for efficient data flow
Purpose: Transmit vehicle data across all UAV units in the hierarchy
Procedure:
- Top layer receives high-level control commands (desired position, velocity, attitude)
- Convert to actuator control data (desired thrust and torques)
- Process through control allocation/mixer
- Perform matrix multiplication for command distribution
- Output propeller speeds for each unit in the layer
- Pass down thrust-torque requirements to next layer
Mathematical Framework:
Top Layer: [Thrust, Roll, Pitch, Yaw]ᵀ
↓ Matrix Multiplication
Middle/Low Layers: [Thrust₁, Thrust₂, ..., Thrustₙ]ᵀ
Purpose: Calculate final desired thrust and torque for each UAV unit
Key Concept:
- From top layer viewpoint, consider middle and low UAV units as virtual propellers
- Each lower-layer UAV unit acts as a controllable thrust generator
Method:
- Middle and low UAV units should not produce pitch and roll torques
- Focus on vertical thrust generation and yaw control
- Minimize lateral disturbances to maintain formation stability
Control Strategy:
Lower Layer Units:
- Primary: Generate vertical thrust (Fz)
- Secondary: Maintain yaw alignment
- Constraint: Minimize pitch (τₓ ≈ 0) and roll (τᵧ ≈ 0) torques
Function: Converts desired thrust and torque inputs into individual propeller angular speeds
Input:
- Desired total thrust (F_total)
- Desired roll torque (τ_roll)
- Desired pitch torque (τ_pitch)
- Desired yaw torque (τ_yaw)
Output:
- Individual propeller angular speeds: [ω₁, ω₂, ω₃, ω₄, ..., ωₙ]
Mathematical Operation:
[ω₁] [M₁₁ M₁₂ M₁₃ M₁₄] [F_total]
[ω₂] = [M₂₁ M₂₂ M₂₃ M₂₄] × [τ_roll]
[ω₃] [M₃₁ M₃₂ M₃₃ M₃₄] [τ_pitch]
[ω₄] [M₄₁ M₄₂ M₄₃ M₄₄] [τ_yaw]
Where M is the control allocation matrix specific to the UAV configuration.
Configurations Tested:
- 2-Layer Quad: Two layers of quadcopter units
- 2-Layer Hexa: Two layers of hexacopter units
- 3-Layer Hybrid: Three layers with mixed configurations
Simulation Objectives:
- Validate stability across different configurations
- Demonstrate scalability by adding/removing layers
- Test reconfigurability with different airframe types
- Evaluate UAV performance in controlled environment
- Measure dynamic response characteristics
Simulation Environment:
- Physics engine: Gazebo with realistic aerodynamics
- Sensor simulation: IMU, GPS, altitude sensors
- Wind disturbance modeling for robustness testing
Physical Prototypes Built:
- 2-Layer Quad Platform: 8 UAV units total
- 2-Layer Hexa Platform: 12 UAV units total
- 3-Layer Hybrid Platform: Mixed configuration
Testing Objectives:
- Verify cascading control method on physical hardware
- Conduct indoor flight tests in controlled environment
- Evaluate dynamic response: Yaw, Pitch, Roll angles
- Measure tracking accuracy and stability
- Calculate and compare error between simulation and actual implementation
- Validate control allocation under real-world conditions
Test Environment:
- Indoor flight arena with motion capture system
- Real-time pose estimation and tracking
- Safety protocols for multi-UAV operation
Test Configuration: 3-Layer Hybrid platform
Findings:
- Single UAV: Maximum payload capacity ~1.0 kg
- 2-Layer Quad UAV: Additional payload capacity >4.1 kg (>300% improvement)
- 3-Layer Hybrid: Further scalability demonstrated
Conclusion:
- Multi-UAV platform successfully achieves payload scalability
- Linear relationship between number of layers/units and payload capacity
- System maintains stability even with maximum payload
Scalability Analysis:
Single UAV: 1.0 kg max payload
2-Layer Quad: >4.1 kg max payload (+310%)
3-Layer Hybrid: (Expandable further)
Test Configuration: 3-Layer Hybrid with unbalanced configuration
Objective:
- Evaluate flight redundancy with asymmetric unit distribution
- Compare performance with full model configuration
- Test fault tolerance and recovery
Results:
- Unbalanced configuration successfully followed the given mission
- No loss of stability despite asymmetric layout
- Control algorithm automatically compensated for imbalance
- Mission completion rate: 100% even with non-uniform configuration
Key Finding: The proposed cascading control method demonstrates high redundancy and can adapt to:
- Unit failures during flight
- Asymmetric configurations
- Partial system degradation
- Dynamic reconfiguration
- Principle: Multiple small propellers instead of large single propellers
- Benefit: Lower generated thrust per unit = much safer operation
- Application: Safe for operation near humans and in confined spaces
- Risk Reduction: Reduced kinetic energy per rotating component
- Verification: Both simulation and actual implementation validated
- Flexibility: Different airframes (Quad, Hexa, Hybrid) are compatible
- Adaptability: Can switch configurations based on mission requirements
- Modularity: Easy to modify platform structure without redesigning entire system
- Evidence: Payload experiments show clear scalability
- Method: Increase payload capability by deploying more layers or UAV units
- Performance: >300% payload improvement with 2-layer configuration
- Future: Theoretically unlimited scalability with proper control
- Proof: Unbalanced model testing demonstrated high redundancy
- Capability: System remains operational with partial failures
- Compatibility: Control method works with unbalanced configurations
- Reliability: Fault-tolerant design ensures mission completion
- Structure: Hierarchical cascading control with top-down command flow
- Allocation: Matrix-based control for propeller speed distribution
- Estimation: IMU-based pose estimation at top layer
- Frequency: Real-time control loop with high-frequency updates
2-Layer Quad:
- 4 UAV units per layer
- 8 total units
- Optimal for compact operations
2-Layer Hexa:
- 6 UAV units per layer
- 12 total units
- Enhanced stability and redundancy
3-Layer Hybrid:
- Mixed configuration with variable unit count per layer
- Maximum payload capacity
- Highest redundancy level
- Architecture: Grouped network with hierarchical routing
- Protocol: Isolated message exchange within network groups
- Overhead: Reduced communication burden through intelligent grouping
- Reliability: Redundant communication paths
- Payload Capacity: >4.1 kg (2-Layer Quad)
- Stability: Maintained across all tested configurations
- Redundancy: 100% mission completion with unbalanced configuration
- Control Frequency: Real-time responsive control
The multi-UAV scalable platform is suitable for:
- Heavy Load Transportation: Construction materials, equipment delivery
- Logistics: Warehouse automation, inventory transport
- Emergency Supply Delivery: Disaster relief, medical supplies
- Crop Spraying: Heavy pesticide tank transport
- Monitoring: Large sensor payload deployment
- Harvesting Assistance: Equipment and produce transport
- Search and Rescue: Equipment delivery to inaccessible areas
- Firefighting: Fire suppressant delivery
- Medical Emergency: Defibrillator and medical supply transport
- Infrastructure: Bridge and building inspection with heavy cameras
- Power Lines: Inspection equipment transport
- Oil & Gas: Pipeline monitoring with sensor arrays
- Multi-Agent Systems: Testbed for distributed control algorithms
- Swarm Intelligence: Coordination and formation control research
- Aerial Manipulation: Future development platform
- Payload Testing: Comprehensive maximum capacity evaluation
- Energy Analysis: Detailed power consumption profiling
- Optimization: Battery placement and weight distribution
- Endurance: Flight time maximization strategies
- Real-Time Switching: Seamless transition between configurations
- 2-Layer Quad mode ↔ 2-Layer Hexa mode ↔ 3-Layer Hybrid mode
- Adaptive Matrix: Automatic control allocation matrix adjustment
- Hot Swapping: Enable configuration changes during flight
- Fault Detection: Automatic reconfiguration upon unit failure
- ToF Sensors: Time-of-Flight sensors for precise landing
- Vision Systems: Camera integration for autonomous navigation
- Multi-Sensor Fusion: Combine IMU, GPS, vision, and ranging sensors
- Environmental Awareness: Enhanced obstacle detection and avoidance
- Consumption Measurement: Detailed energy profiling across configurations
- Efficiency Analysis: Power-to-payload ratio optimization
- Battery Management: Intelligent charge distribution system
- Flight Planning: Energy-optimal trajectory generation
- Outdoor Flight Testing: Real-world environmental validation
- Long-Duration Missions: Extended flight time evaluation
- Payload Mechanisms: Advanced attachment and release systems
- Weather Resistance: Testing under adverse conditions
- Autonomous Operations: GPS waypoint navigation and mission planning
Propeller Thrust-Torque Model:
T = CT × ρ × ω² × rp⁴
Q = CM × ρ × ω² × rp⁵
Power = Q × ω
Control Allocation:
[F] [A₁₁ A₁₂ ... A₁ₙ] [ω₁²]
[τx] = [A₂₁ A₂₂ ... A₂ₙ] × [ω₂²]
[τy] [A₃₁ A₃₂ ... A₃ₙ] [...]
[τz] [A₄₁ A₄₂ ... A₄ₙ] [ωₙ²]
Cascading Control Transfer:
Layer i: [Fi, τi_roll, τi_pitch, τi_yaw]ᵀ
↓
Layer i+1: [F₁, F₂, ..., Fₘ]ᵀ (thrust distribution)
Small Propeller Advantages:
- Safer operation: Lower kinetic energy per propeller
- Better redundancy: More units = higher fault tolerance
- Improved maneuverability: Faster response time
- Easier scaling: Modular addition/removal of units
Trade-offs:
- Increased complexity in control coordination
- More communication overhead
- Higher initial hardware cost
- Requires sophisticated control algorithms
| Feature | Single Large UAV | Multi-UAV Platform |
|---|---|---|
| Payload Capacity | ~1 kg | >4.1 kg (2-Layer) |
| Safety | High risk (large propeller) | Low risk (small propellers) |
| Redundancy | Low (single point of failure) | High (distributed) |
| Scalability | Fixed capacity | Modular expansion |
| Reconfigurability | None | Multiple configurations |
| Complexity | Low | High |
| Cost | Lower initial cost | Higher initial cost |
| Maintenance | Single unit | Multiple units |
This research successfully demonstrates a novel multi-UAV platform that addresses the fundamental limitations of conventional single large-propeller UAVs. The hierarchical cascading control method enables:
- ✅ >300% payload improvement over single UAV configurations
- ✅ 100% mission success rate with unbalanced configurations
- ✅ Multiple airframe compatibility (Quad, Hexa, Hybrid)
- ✅ Zero stability loss during redundancy testing
- Cascading Control Architecture: Novel hierarchical control method for multi-layer UAV coordination
- Virtual Propeller Concept: Treating entire UAV units as controllable thrust generators
- Network Grouping: Efficient communication management for distributed systems
- Dynamic Scalability: Proven ability to scale payload capacity linearly with system size
The platform represents a significant advancement in heavy-load UAV transportation, demonstrating that distributed small-propeller systems can outperform conventional large-propeller designs in:
- Payload capacity
- Operational safety
- System redundancy
- Configuration flexibility
Both simulation and physical implementation validate the feasibility of deploying such systems in real-world applications, with clear pathways for future enhancement and commercial deployment.
-
Structural Design Files:
- CAD models and mechanical design specifications
- Hardware configuration details
- Assembly instructions and blueprints
-
Control System Code:
- Cascading control algorithm implementation
- Mixer and thrust-torque converter code
- Control allocation matrix calculations
- Network communication protocols
- Flight controller firmware
-
Simulation Setup:
- Gazebo model configurations
- Custom physics parameters
- Test scenario scripts
For academic collaboration, detailed technical discussions, or research inquiries:
📧 Contact via email for:
- Control algorithm discussions
- Collaboration proposals
- Technical implementation details
- Research partnership opportunities
- Access to published papers and supplementary materials
When contacting, please include:
- Your institutional affiliation
- Research interests and background
- Specific technical questions or collaboration goals
- Intended use of information (academic/commercial)
- Research paper on multi-UAV cascading control (contact for details)
- Conference presentations and proceedings
- Propeller aerodynamics theory
- Multi-agent control systems literature
- UAV payload optimization studies
- Gazebo simulation framework documentation
- Flight control system design principles
- Distributed UAV systems
- Swarm robotics coordination
- Hierarchical control architectures
- Heavy-lift drone technologies
Special thanks to all team members, advisors, and laboratory staff who contributed to the development, testing, and validation of this multi-UAV platform.
- ✅ Concept Development: Complete
- ✅ Simulation Testing: Complete
- ✅ Physical Prototyping: Complete (2-Layer Quad/Hexa)
- ✅ Performance Validation: Complete
- 🔄 3-Layer Hybrid Testing: In Progress
- 🔄 Energy Optimization: In Progress
- 📋 Dynamic Control Allocation: Planned
- 📋 Outdoor Flight Testing: Planned