Model Predictive Control Algorithm for Enhanced Stability and Maneuverability in Quadrotor Drones Operating Autonomously

Abstract

The invention provides a novel Model Predictive Control Algorithm that significantly enhances the stability and maneuverability of autonomous quadrotor drones, addressing key limitations of existing control methods. Its real-time adaptability and robustness make it suitable for diverse applications, setting a new standard in UAV control systems.

Specification

Description:Field of Invention:
This invention relates to autonomous unmanned aerial vehicles (UAVs), particularly quadrotor drones. It involves the design and implementation of an advanced Model Predictive Control (MPC) algorithm to enhance flight stability and maneuverability in dynamic environments.
Background of the Invention
The demand for autonomous quadrotor drones is increasing across industries such as delivery services, surveillance, agriculture, and disaster management. Key challenges include achieving high stability during flight and maneuvering in complex, dynamic environments.
Existing controllers like Proportional-Integral-Derivative (PID) controllers often fail to handle non-linear dynamics and multi-constraint operations effectively. Advanced control methods such as MPC have been explored, but issues related to computational overhead and real-time adaptability remain.
This invention introduces a novel MPC algorithm optimized for real-time implementation, overcoming the computational and stability limitations of existing solutions.
________________________________________
Summary of the Invention
The invention provides a Model Predictive Control Algorithm tailored for quadrotor drones, enabling precise trajectory tracking, obstacle avoidance, and enhanced stability under dynamic conditions.
Key features include:
1. Dynamic System Modeling: The invention models the quadrotor as a non-linear system, considering six degrees of freedom (6-DOF) and real-time sensor feedback.
2. Robust Predictive Framework: Incorporates an adaptive horizon-based predictive model to anticipate drone dynamics and external disturbances.
3. Real-Time Optimization: Utilizes a fast optimization algorithm to solve constrained optimization problems, ensuring feasibility and stability.
4. Energy Efficiency: Balances control efforts to minimize energy consumption without compromising performance.
5. Multi-Constraint Handling: Integrates physical constraints, such as actuator limits, collision avoidance, and regulatory flight boundaries.
________________________________________
Detailed Description of the Invention
1. Quadrotor System Model
The quadrotor dynamics are represented using six states:
• Translational Position (x, y, z)
• Rotational Angles (roll, pitch, yaw)
Equations of motion are derived from Newton-Euler principles:

2. Model Predictive Control Design
The MPC algorithm optimizes the control inputs (uuu) over a finite time horizon (TTT) to minimize the cost function:

The algorithm is designed to predict future states and adjust control inputs dynamically.
3. Obstacle Avoidance
A collision-free trajectory is ensured using a soft constraint-based approach. The constraints are incorporated into the cost function:

4. Real-Time Computation
To address computational constraints, the invention employs:
• Parallel Computing Frameworks: Divides computations across multi-core processors.
• Approximation Techniques: Uses first-order methods like gradient descent to approximate solutions.
5. Robustness to Disturbances
The algorithm accounts for external forces such as wind disturbances by integrating disturbance observers and adapting prediction models.
, Claims:Claims
1. System Model Integration: A non-linear predictive control system specifically tailored for quadrotor drones operating in real-time, considering translational and rotational dynamics.
2. Predictive Optimization Algorithm: A multi-objective optimization approach to ensure trajectory tracking and stability under environmental disturbances.
3. Constraint Handling: Dynamic inclusion of actuator limitations, obstacle avoidance, and regulatory boundaries.
4. Energy Efficiency Mechanism: Adaptive control effort minimization for prolonged battery life.
5. Real-Time Implementation: Parallel processing framework ensuring real-time computations for high-speed operations.

Documents