Control System Using Time‐State Control Form for Balloon Robot

Abstract

In this study, we considered an air ship robot with a body composed of a helium‐filled balloon. The robot can move in the air utilizing pectoral fin motion. The body of the robot has high cushioning properties, and it does not use a propeller as a propulsion mechanism. From this perspective, the robot can be considered safe. It can perform a stable floating flight in a room in which people and the robot are close to each other. As potential applications of the robot, we consider the monitoring of satellites and establishing advertising platforms. In this study, we construct an automatic control system for such practical applications. The robot cannot maneuver sideways; therefore, it can be considered a robot with nonholonomic constrains, based on Brockett's study. Various control models have been devised concerning motion with nonholonomic constraints; however, a control system using time‐state control is deemed efficient owing to excellent flexibility and expandability. Therefore, we seek to implement a control system using a time‐state control form. The outline of the proposed control method is as follows. First, we converted the motion model into a time‐state control form using a time state and a state control unit. Second, we constructed a feedback control system that stabilized the state control unit. Outputs of the propulsion velocity and the turning angular velocity were obtained from the control system. We measured the propulsion force and then, on their basis, derived the output patterns. After selecting the motion from the output patterns, we constructed a control system through the simulation and investigated the characteristics of each gain. Based on these results, we considered that gain k1 adjusted the speed of reaching the target axis, gain k2 adjusted the amount of turning, and gain k3 regulated the output corresponding to the propulsion velocity. We conducted an experiment to test the established control system and confirmed that the robot was able to orbit. Based on the experimental results, we conclude that the robot can be controlled according to the user's intention by setting gains k1 and k2 sufficiently small when moving along the target axis and by setting gains k1 and k2 large when moving toward the target point. © 2021 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC.