Discrete reference tracking control to swing up an electric wheelchair

• Published in: Asian journal of control Vol. 25; no. 5; pp. 3369 - 3383
• Main Authors: Gray, Michael, Guerra, Thierry Marie, Delprat, Sébastien, Mohammad, S.
• Format: Journal Article
• Language: English
• Published: Hoboken Wiley Subscription Services, Inc 01.09.2023; Asian Control Association (ACA) and Chinese Automatic Control Society (CACS)
• Subjects: Automatic Control Engineering; Barriers; Brushless motors; Computer Science; descriptor system; Linear matrix inequalities; Parameter varying control; People with disabilities; quasi‐LPV; reference tracking; robust control; Takagi–Sugeno; Tracking control; wheelchair; Wheelchairs; Wheels; Robust control; Descriptor system; Reference tracking; Wheelchair; Takagi Sugeno; Descriptor system Reference tracking Robust control quasi-LPV Takagi Sugeno Wheelchair; quasi-LPV
• ISSN: 1561-8625; 1934-6093
• DOI: 10.1002/asjc.3053

Mobility using wheelchair is a crucial feature for disabled people. Between a manual wheelchair and an electric wheelchair, there is a very interesting and promising solution consisting of an electrical kit that converts a manual wheelchair to an electric wheelchair. Autonomad‐Mobility is a company selling such kits, where the wheels are equipped with a brushless motor. It presents new possibilities for mobility, especially for crossing obstacles and running on different types of soil. Several methods have been tested, from linear to nonlinear ones including the linear parameter‐varying control framework. Nevertheless, the most dangerous and critical issue is to be able to produce a so‐called safe and robust swing‐up: the transition from standard mode (four wheels on the ground) to gyroscopic mode (the wheelchair balancing on the rear push‐wheels). The gyroscopic mode is a unique feature improving mobility by providing a way to overcome obstacles that would normally be unsurpassable due to the caster wheels. Whereas some results exist, the solution must be compatible with any wheelchair and any disabled person whatever is her/his height and weight. Robustness of this unstable feature is therefore the key point of the work knowing that “exact” modeling is unrealistic. With the model in descriptor form and using the quasi‐linear parameter‐varying framework new stabilization conditions formulated as linear matrix inequality constraints are developed for attaining the goal. Simulations and real‐time experiments demonstrate the interest of this approach.