Design, Modeling and Control of a Novel Morphing Quadrotor

Abstract

In this paper, the design, modeling and control of a novel morphing quadrotor are presented. The morphing quadrotor can fly stably and accurately in the air while simultaneously undergoing shape transformation, regardless of the asymmetry of the model. The four arms can rotate around hinges on the main body of the quadrotor to form various topological models. The arms are not in the same plane, so they can overlap with each other. In the extreme case, the width of the morphing quadrotor can be reduced to the diameter of a single rotor to allow the quadrotor to fly through narrow gaps more easily. Reinforcement learning (RL) with an extended-state approach is introduced in this paper to optimize the attitude control law and enable automatic adaptation to model changes. A deterministic policy gradient (DPG) algorithm based on an actor-critic structure with four neural networks in a model-free approach is used to train the controller. Finally, a linear programming method named fast simplex algorithm is presented to solve the control allocation problem of morphing quadrotors in real time with affordable computational cost in this paper. The controller has been tested on our real morphing quadrotor platform and achieves excellent flight performance.

Deep reinforcement learning achieves multifunctional morphing airfoil control

“…29,30 RL supports the complexity of a morphing aircraft by producing a controller that learns to use morphing control surfaces and operates in several configurations, as well as by determining the best configuration for a given flight situation. [31][32][33] Although these shape changes are achieved using traditional methods, such as servos and motors, some have implemented RL in smart material based morphing simulations. 34,35 One case presented by Goecks et al implemented deep deterministic policy gradient (DDPG) with a shape memory alloy (SMA) actuated airfoil.…”

Modeling and Control of Reconfigurable Quadrotors Based on Model Reference Adaptive Control

“…Author in reference [13] presents the design and control of a variable-pitch quadrotors to overcome the maneuverability limitations of the traditional quadrotors, meanwhile give the detailed analysis of the potential benefits of this kind of quadrotors and experimental tests results. A novel morphing quadrotors which can change its topological modes by rotating hinges was proposed in reference [14], and reinforcement learning is used to optimize the attitude control laws, the experimental test on a real morphing quadrotor platform was performed and results validate the excellent performance of the proposed control laws. A reconfigurable aerial robots chain was presented in reference [15], the chain can cross narrow sections, morph its shape and therefore has excellent extensibility.…”

“…Inspired by the work of [14], herein we give the calculation formula of the center of mass and inertia of the whole system below. Where UAV m is the total mass of the UAV system, i m is the mass of the ith rotor and connected motor, armi m is the mass of the ith arm of the UAV system, i r is the displacement vector of the ith rotor in the normal initialized body frame, armi r is the displacement vector of the ith arm in the normal initialized body frame.…”

“…Where UAV m is the total mass of the UAV system, i m is the mass of the ith rotor and connected motor, armi m is the mass of the ith arm of the UAV system, i r is the displacement vector of the ith rotor in the normal initialized body frame, armi r is the displacement vector of the ith arm in the normal initialized body frame. when the real center of mass of the UAV configuration has been calculated, the inertia moment of the UAV configuration based on the new center of mass can be calculated using parallel axis theorem as described in the reference [14].…”

A Generalized Multivariable Adaptive Super-Twisting Control and Observation for Amphibious Robot

“…It just needs some trials and flight data for training the system to control quadrotor robot. The fuzzy control, neural network control and reinforcement learning control are considered as typical learning-based control techniques [11]- [16]. However, the control technology based on learning needs enough training data, and cannot always ensure the stability of system.…”