Abstract:Unmanned helicopter has broad application prospects, both in civil and military field. Helicopter is a strong coupled with many phenomena, inherently unstable, high-order, time-varying nonlinear complex system. It is a great challenge to investigate the system identification of helicopter, particularly when the non-stationary flight regimes are considered. In this paper, we address the system identification as dynamic regression. Inspired by the feature expression ability of deep learning, we use Deep convolutional neural networks (CNN) to represent dynamics model of helicopter. The parameters of the network are directly learned from the real flight data collected from helicopter. We provide model initialization method and optimization details for training. Since it captures the hidden states in aerobatic maneuvers without a-priori, the proposed identifier manifests strong robustness and high accuracy, even for untrained aerobatic maneuvers. The effectiveness of the proposed method is verified by various experiments on the real-world flight data from Stanford Autonomous Helicopter Project. Specifically, the Deep CNN identifier improves 71.60% overall in RMS acceleration prediction over the lasted Deep Rectified Linear Unit (ReLU) Network Model.

Modeling Hover Dynamics of Small-scale Unmanned Helicopter Based on Least Square Support Vector Machine

Modeling Angular Dynamics for Model-Scale Helicopter Using Support Vector Machines

Small unmanned helicopter modeling method based on a hybrid kernel function PSO-LSSVM

Grey-box modeling of a small-scale helicopter using physical knowledge and Bayesian Techniques

Grey-box Modeling of Small-Scale Unmanned Helicopter Based on Bayesian Technique

Using Svm To Model And Control Nonlinear Dynamical Systems

A Modeling Method of Miniature Unmanned Helicopter Based on Parameter Identification

Open loop system identification for a quadrotor helicopter system.

Frequency Domain Identification of a Small-Scale Helicopter with Least Square Method

Frequency-domain Identification of a Small-Scale Unmanned Helicopter with Harmony Search Algorithm

Statistical modeling angular dynamics for model-scale helicopter.

Modeling And Lqr Control Of Small Unmanned Helicopter

Vertical Modeling and Autonomous Landing of a Mini Unmanned Helicopter

Control System Design and Flight Testing for a Miniature Unmanned Helicopter

Time-Domain Model Identification of A Small-Scale Unmanned Helicopter

A Hybrid Control Method for an Unmanned Scale Helicopter Attitude

Black-Box System Identification for Low-Cost Quadrotor Attitude at Hovering

Support Vector Machine Based Modeling of Nonlinear Systems with Hysteresis

Modeling the Yaw Dynamics of an Unmanned Helicopter Through Modes Partition Method

SVM based nonparametric model identification and dynamic model control

Deep Cnn Identifier For Dynamic Modelling Of Unmanned Helicopter