Flight Control System Design for Autonomous Aerial Surveys of Volcanoes

TL;DR

This paper presents a novel control system for a quadrotor operating in the extreme, temperature-varying environment near volcanoes, incorporating path planning, eigenvalue selection, and sensor fusion for robust autonomous flight.

Contribution

It introduces a new method for eigenvalue selection in systems with moving references and develops an integrated control approach for volcanic environment conditions.

Findings

Successful simulation of height stabilization in MATLAB

Effective noise reduction using Kalman filter

Validated control strategy for extreme environmental conditions

Abstract

The controller for a quadrotor working in severe environment is developed in this study. Here, the severe environment indicates the temperature-varying air near the volcano. The controller overcomes the intensively changing temperature above the crater of the volcano which biases the nominal dynamics (25 Celcius). The target Volcano is picked as Satsuma-iojima located in a tiny insular South to the mainland of Japan. The temperature distribution is contributed from previous research by Geological Survey of Japan. To guarantee that the control signal is under the input saturation, a path planning method is developed. Picking the eigenvalue for a system with a moving reference is a novel topic; the method to develop a controller with specific requirement is created for the first time. This method might be referred as a standard way for designers/engineers in developing controller with a moving target/reference in further study. In controlling part, a state feedback controller is designed to stabilize the height of the quadrotor. The eigenvalue of the feedback controller is picked based on the method developed in Chapter Path Planning. And a PID controller is designed to control the attitude. The result of these are verified in a simulator written in MATLAB. At last, a Kalman filter is applied in height control to combine the measurement noise from IMU and laser scanner and the system noise caused by the changing temperature. Another reason for developing a Kalman filter is that IMU readout provides with acceleration. While the velocity is not achieved by sensor directly. The result of height control with a Kalman filter is verified in MATLAB Simulink.