Design and development of hybrid controller for two-wheeled wheelchair /
Conventional wheel configuration of currently available electric wheelchair poses several constraints, such as limited mobility due to three or four wheel structures, limited maneuverability due to existence of front casters, and occupies a large space. In this thesis, an electric wheelchair capable...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
Kuala Lumpur :
Kulliyyah of Engineering, International Islamic University Malaysia,
2015
|
| Subjects: | |
| Online Access: | http://studentrepo.iium.edu.my/handle/123456789/4455 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Conventional wheel configuration of currently available electric wheelchair poses several constraints, such as limited mobility due to three or four wheel structures, limited maneuverability due to existence of front casters, and occupies a large space. In this thesis, an electric wheelchair capable of travelling on two wheels is proposed to overcome those constraints. The proposed two-wheeled wheelchair has five degree-of-freedoms (DOF), i.e. angular motion of the right and left wheels, angular motion of the motor housing and the extendable seat, and the linear motion of the extendable seat in vertical direction. These motions are confirmed by constructing a non-parameterized LEGO® model. Then, in accordance with the main objective of this thesis, which is to develop a controller for stabilizing a two-wheeled wheelchair in balancing mode, the reduced 3-DOF nonlinear mathematical model is obtained using Euler-Lagrange method. In balancing mode, the wheelchair stands on its two main wheels and the seat extension is inactive. Thus, the system becomes unstable with characteristics similar to a mobile double-inverted pendulum on two wheels. This 3-DOF nonlinear mathematical model is used in the simulation and hardware experiment. To stabilize the system to upward position, a class of controller is needed. In this thesis, a nonlinear hybrid controller based on information fusion and fuzzy logic is proposed. This nonlinear hybrid controller is achieved by first processing the state of the system using a fusion function. Then the output of this function is fed into Fuzzy-PD type controller. This method reduces the number of fuzzy inputs, which in turn reduces the number of fuzzy rules. Hence, the 'rule explosion' could be avoided. The fuzzy gains are then optimized by an online tuning method called extremum seeking (ES). Integral of squared error (ISE) is used within ES algorithm. Evaluation of the proposed hybrid controller—abbreviated as IFFES—is then conducted by simulation using MATLAB®. The performance of the proposed hybrid controller is then compared with the linear optimal controller, conventional fuzzy logic controller, and non-ES-optimized hybrid controller in terms of rise time, settling time, and peak overshoot. Simulation results show that IFFES can bring the two-wheeled wheelchair upright from its initial condition ( and ) with better performance than the optimal control, conventional FLC, and non-ES-optimized controller (IFFC). The settling time is 20%-60% faster. The rise time is 100% faster for linear displacement and 5%-14% faster for angular positions of the first and second links, and , respectively. Though there is an increase by 4% overshoot for , the peak overshoot of is lowered by 27% and zero overshoot for . These results also proved that ES can be used together with fuzzy control system. After computer simulation confirms the performance of the proposed hybrid controller, it is implemented on a third hardware prototype as a testing platform. Since the condition of the platform is initially upward, disturbances are introduced to the platform by tilting it twice manually. The result shows that IFFES is able to stabilize the platform within 10 seconds after the first disturbance and within 8 seconds after the second disturbance. The main contributions of this thesis are the 3-DOF nonlinear mathematical model of the two-wheeled wheelchair in balancing mode and the hybrid nonlinear controller for stabilizing it. |
|---|---|
| Item Description: | Abstracts in English and Arabic. "A thesis submitted in fulfilment of the requirement for the degree of Master of Science (Mechatronics Engineering)."--On t. p. |
| Physical Description: | xx,144 leaves : ill. ; 30cm. |
| Bibliography: | Includes bibliographical references (leaves 131-139). |