Development And Motion Control Of Single-Axis Reaction Wheel Based Cubic Axis Case Study
Development and Motion Control of Single-Axis Reaction Wheel Based Cubic Robot by R.G. Vimukthi Pathirana
A thesis submitted in partial fulfillment of the requirements for the degree of Master of Engineering in
Mechatronics
Examination Committee: Dr. Manukid Parnichkun (Chairperson) Dr. Mongkol Ekpanyapong
Assoc. Prof. Erik L.J. Bohez
Nationality: Srilankan
Previous Degree: Bachelor of Engineering in Mechatronics
Asian Institute of Technology Thailand
Scholarship Donor: AIT Fellowship
Asian Institute of Technology School of Engineering and Technology Thailand
July 2015
ACKNOWLEDGEMENTS
I wish to express many thanks and deep gratitude to Dr. Manukid Parnichkun, who gave me immense support, guidance and encouragement throughout and also for sponsoring …show more content…
The inverted pendulum 3 The reaction wheel 4 Tilt estimation from IMU 6 I2C bus 7 Stabilization tasks of Cubli 8 Cad drawing of the Cubli 9 MeMa cube 9 Cad drawing 10 Test bed of cube 11 Wiring diagram 13 Flow chart of the program 14 Angle without filter 17 Angle with filter 17 With impulse disturbances applied randomly time to time 19 Balancing with no disturbances 19 Angular velocity of the cube around its edge with disturbances 20 Angular velocity of the cube around its edge with no disturbances 20 Angular velocity of the reaction wheel with no disturbance 21 Angular velocity of the reaction wheel with disturbance …show more content…
And the readings initially are not very accurate to use to control the movements and stabilization tasks. Therefore a simple kalman filter is used to get more accurate angle values. The resulting graphs show the angle readings with and without the filter. Without filter, the error range for the angle value of 46.30 is around 45.70 to 46.90, so it’s +/- 0.6 degree error.
Figure 4.1: Angle without filter
With filter, the error range for the same angle value of 46.30 is around 46.20 to 46.40, so just +/- 0.1 degree error. Therefore with filter, sensor manages to recover a total of +/- 0.50 errors.
Figure 4.2: Angle with filter
18
4.2 Impulse disturbance
4.2.1 Plot with and without impulse disturbances
Impulse disturbances are applied randomly time after time to the cube while it is balancing around its edge, and the angle value is being recorded.
Figure 4.3: With impulse disturbances applied randomly time to
A thesis submitted in partial fulfillment of the requirements for the degree of Master of Engineering in
Mechatronics
Examination Committee: Dr. Manukid Parnichkun (Chairperson) Dr. Mongkol Ekpanyapong
Assoc. Prof. Erik L.J. Bohez
Nationality: Srilankan
Previous Degree: Bachelor of Engineering in Mechatronics
Asian Institute of Technology Thailand
Scholarship Donor: AIT Fellowship
Asian Institute of Technology School of Engineering and Technology Thailand
July 2015
ACKNOWLEDGEMENTS
I wish to express many thanks and deep gratitude to Dr. Manukid Parnichkun, who gave me immense support, guidance and encouragement throughout and also for sponsoring …show more content…
The inverted pendulum 3 The reaction wheel 4 Tilt estimation from IMU 6 I2C bus 7 Stabilization tasks of Cubli 8 Cad drawing of the Cubli 9 MeMa cube 9 Cad drawing 10 Test bed of cube 11 Wiring diagram 13 Flow chart of the program 14 Angle without filter 17 Angle with filter 17 With impulse disturbances applied randomly time to time 19 Balancing with no disturbances 19 Angular velocity of the cube around its edge with disturbances 20 Angular velocity of the cube around its edge with no disturbances 20 Angular velocity of the reaction wheel with no disturbance 21 Angular velocity of the reaction wheel with disturbance …show more content…
And the readings initially are not very accurate to use to control the movements and stabilization tasks. Therefore a simple kalman filter is used to get more accurate angle values. The resulting graphs show the angle readings with and without the filter. Without filter, the error range for the angle value of 46.30 is around 45.70 to 46.90, so it’s +/- 0.6 degree error.
Figure 4.1: Angle without filter
With filter, the error range for the same angle value of 46.30 is around 46.20 to 46.40, so just +/- 0.1 degree error. Therefore with filter, sensor manages to recover a total of +/- 0.50 errors.
Figure 4.2: Angle with filter
18
4.2 Impulse disturbance
4.2.1 Plot with and without impulse disturbances
Impulse disturbances are applied randomly time after time to the cube while it is balancing around its edge, and the angle value is being recorded.
Figure 4.3: With impulse disturbances applied randomly time to