Low pass filter with hybrid fuzzy design for balancing the two-wheeled mobile robot application
This paper proposes a low pass filter and hybrid fuzzy design for balancing an extendable double-link two-wheel mobile robot (TWMR) system. The proposed system mimics a double inverted pendulum scenario, where the angular position of the first link (Link1) is to be varied depending on the value of t...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
UNSYS digital
2014
|
| Subjects: | |
| Online Access: | http://irep.iium.edu.my/41754/ http://irep.iium.edu.my/41754/ http://irep.iium.edu.my/41754/1/taqiIJRM2014.pdf |
| Summary: | This paper proposes a low pass filter and hybrid fuzzy design for balancing an extendable double-link two-wheel mobile robot (TWMR) system. The proposed system mimics a double inverted pendulum scenario, where the angular position of the first link (Link1) is to be varied depending on the value of the angular position of the second link (Link2) and the elongation of the extendable link (Link3) attached to Link2 and with various payloads. The two-wheeled mobile robot together with the extendable link on Link2 makes the system become more flexible but yet unstable without any controller. The inclination of the Link3 at any interest angle of Link2 will affect the centre of gravity (COG) of the system especially when the payload is having a significant weight. Although, the system can be balanced by controlling the COG analysis via the hybrid fuzzy controller, but the responses are still oscillating and not smooth. Thus, an additional low pass filter is designed and added to shape the controller outputs, before the signals are fed into the TWMR system. The results obtained show that the filter designed managed to improve the oscillating and smoothing of the angular position of Link1 and Link2 within a shorter time. The performance of the system, such as the settling time and the rise time of the responses are improved by at least 43%. Meanwhile, the root mean square error of the responses can be improved up to 60.7%. |
|---|