Sang-Deok Lee

Experimental verification of stability region of balancing a single-wheel robot: An inverted stick model approach

Gyroscopically manipulated lateral motion of a single-wheel based robot system has a parametric oscillation problem. In this paper, we analyze the stability regions for balancing the single-wheel robot system and then propose how to determine the offset range of its suppression controller. Since the roll motion is coupled with the pitch motion, the parametric oscillation problem should be analyzed. The dynamics model is parameterized as a Mathieus equation from an inverted stick model. The stability region of a pitch oscillation parameter using the phase trajectory is analyzed from simulation. Simultaneously, its suppression control performance by determining the offset range is examined. Experimental studies are conducted to verify the analysis for the stable balancing control and to confirm the stability region based on the determined offset value.

Model-Based Rolling Motion Control of an One-wheeled Robot Considering the Pitching Motion of a Gyroscopic Effect

In general, a yawing motion concept is used for the lateral control of one wheel robot where the gimbal system is located horizontally. In this paper, another concept of the vertically located gimbal system is presented for the same purpose. Although the vertical concept undergoes an instability more easily than the horizontal one, the pitching motion of the gyroscopic effect is considered. Firstly, the trade-off relation between two balancing concepts are investigated by comparing the gyroscopic mechanism. Secondly, the dynamic model for the problem of the proposed concept is derived using the oscillatory inverted stick model. Thirdly, the stability of the model is analyzed using the phase trajectory method. Finally, the control performance of the system by a vibration controller is simulated.

Analysis of Relationship between Body and Gimbal Motion Through Experiment of a Single-wheel Robot Based on an Inverse Gyroscopic Effect

Control Moment Gyro (CMG) has been used as an indirect actuator of a single-wheel robot system GYROBO, developed at Chungnam National University. The flip motion of the gimbal system produces the gyroscopic motion onto the body system while the body motion also produces the gyroscopic motion onto the gimbal system inversely. In this paper, the intuitive equation of the inverse gyroscopic effect is derived as the direct relation between the rate of the body system and the rate of the gimbal system. Experiments on the inverse gyroscopic effect under the chaotically generated disturbance are conducted. Experimental data are approximated by a linear equation using the least square method.

Experimental studies on Q filter design of a disturbance observer for a one-wheel robot

In this paper, experimental studies on designing Q filters in the disturbance observer (DOB) based motion control for balancing a one-wheel robot are presented. The robot is simply modeled as a second order system and the corresponding Q filters are designed to form a DOB control structure. Two Q-filters such as Q20 and Q31 are investigated and analyzed in a view point of disturbance rejection performance and sensor noise immunity. The performances of two Q filters are evaluated by realizing a DOB in the digital system for controlling the balance of a single-wheel mobile robot.

RLS model identification-based robust control for gimbal axis of control moment gyroscope

Control moment gyroscope(CMG) is an indirect actuator used for the attitude control tasks where actuators cannot be directly applied. Control of the gimbal position becomes important to maximize the induced torque. As a torque amplifier, the gimbal axis of CMG should be controlled to produce fine and regulated torque. Since the torque performance is dependent upon the return performance of the gimbal axis, both the magnitude and direction of gyroscopic torque should be controlled to achieve agility and regulation. In this paper, a robust current controller is designed with an angle restriction property by the current compensation designed through the model identification by real-time recursive least square(RLS) algorithm. Firstly, the configuration and mechanism of designed CMG is presented and the angle restriction problem is stated. Secondly, RLS algorithm is designed for online identification, implemented, and validated. Thirdly, the proposed current control scheme of gimbal axis is verified by experiments.

Experimental Studies of the Q Filter Design for a Disturbance Observer for the Balancing Control of a One-Wheel Robot

This paper presents the design of Q filters for the disturbance observer(DOB) to balance a one-wheel robot. The one-wheel robot is an under-actuated system so that its control is not easy. To improve the balancing control performance in the DOB configuration, different Q filters are designed and tested. Three different Q filters are tested:Q, Q, and Q and their performances are compared in terms of the disturbance rejection ability and sensor noise immunity to find the best Q-filter. The analysis shows that Q can be realized in the DOB structure. Experimental studies were conducted to verify the performance of Q and Q filters in the DOB structure and the one-wheel robot was successfully controlled to maintain the balance.

Design of a Fuzzy Compensator for Balancing Control of a One-wheel Robot

For the balancing control of a one-wheel mobile robot, CMG (Control Moment Gyro) can be used as a gyroscopic actuator. Balancing control has to be done in the roll angle direction by an induced gyroscopic motion. Since the dedicated CMG cannot produce the rolling motion of the body directly, the yawing motion with the help of the frictional reaction can be used. The dynamic uncertainties including the chattering of the control input, disturbances, and vibration during the flipping control of the high rotating flywheel, however, cause ill effect on the balancing performance and even lead to the instability of the system. Fuzzy compensation is introduced as an auxiliary control method to prevent the robot from the failure due to leaning aside of the flywheel. Simulation studies are conducted to see the feasibility of the proposed control method. In addition, experimental studies are conducted for the verification of the proposed control.

An Experimental Study on Realtime Estimation of a Nominal Model for a Disturbance Observer: Recursive Least Squares Approach

In this paper, a novel RLS-based DOB (Recursive Least Squares Disturbance Observer) scheme is proposed to improve the performance of DOB for nominal model identification. A nominal model can be generally assumed to be a second order system in the form of a proper transfer function of an ARMA (Autoregressive Moving Average) model. The RLS algorithm for the model identification is proposed in association with DOB. Experimental studies of the balancing control of a one-wheel robot are conducted to demonstrate the feasibility of the proposed method. The performances between the conventional DOB scheme and the proposed scheme are compared.

Experimental verification of physical relation between a gimbal system and a body system: GYROBO

CMGs(Control Moment Gyros) provide an access to control the bodys attitude with gimbals minimum effort. This merit comes from the fact that CMG is based on the gyroscopic effect induced by the motion of a gimbal system with a constantly spinning flywheel. CMG has been used in GYROBO, a single-wheel mobile robot developed at Chungnam National University. The rate of the gimbal motion combined with the angular momentum generated by the rotating flywheel produces the gyroscopic effect onto the body. In the opposite way, the body motion induced by the external disturbances affects the gimbals motion as well. In this paper, we conduct the experimental verification of the angular rate deviations of the gimbal system interrupted by the disturbance to find the relationship between the body and the gimbal system. Experimental data are justified through the equation approximated by the least square method. Then the experimental results are compared with the derived equation.

Vibration Control of a Single-wheel Robot Using a Filter Design

In this paper, the vibration of a single-wheel mobile robot is minimized by designing a filter. An AHRS (Attitude and heading reference system) sensor is used for measuring the state of the robot. The measured signals are analyzed using the FFT method to investigate the fundamental vibrational frequency with respect to the flywheels speed of the gimbal system. The IIR notch filter is then designed to suppress the vibration at the identified frequency. After simulating the performance of the designated filter using the measured sensor data through extensive experiments, the filter is actually implemented in a single-wheel mobile robot, GYROBO. Finally, the performance of the designed filter is confirmed by performing the balancing control task of the GYROBO system.