🚁 Lyapunov-based deep neural network - Intermittent Feedback

Real-time simulation with Jacobian-based neural network adaptation and intermittent feedback loss

Controller Type

🎯 Composite Adaptive
📍 Tracking Only
⚙️ PID Control
Simulation Stopped

Initial Conditions

Position (m):
x: 0.10, y: -0.10, z: 0.05
Orientation (rad):
φ: 0.02, θ: -0.02, ψ: 0.01
Velocity (m/s):
ẋ: 0.00, ẏ: 0.00, ż: 0.00
Angular Rate (rad/s):
φ̇: 0.00, θ̇: 0.00, ψ̇: 0.00

Neural Network

32
3

Control Gains

8.0
15.0
25.0
12.0
30.0

Adaptation Gains

5.0e-5
15.0
30.0

Intermittent Feedback

Environment

0.5
0.25
50
5.0
0.000
Position Error (m)
0.000
Orientation Error (rad)
0.000
Control Effort
0.000
Adaptation Rate
0.000
Prediction Error
ACTIVE
Feedback Status
🚁 3D Quadrotor Simulation Environment
📊 Position Error
🎮 Control Effort

🎯 Select Initial Condition Scenario

Hovering Start

Near-zero initial conditions, perfect for smooth trajectory tracking

Large Displacement

Start far from origin with random orientation - tests adaptation

Moving Start

Non-zero initial velocities and angular rates

Inverted Start

Large initial roll/pitch angles - challenging recovery

High Angular Rate

Fast initial rotation - tests angular control

Random Conditions

Completely randomized initial state