M. Khalili1, H.F.M. Van der Loos1, J. Borisoff2
1University of British Columbia, Canada, 2British Columbia Institute of Technology, Canada
INTRODUCTION
A power-assisted wheelchair (PAW) consists of a manual wheelchair (MWC) and electric power add-ons (e.g., powered-wheels and/or front-/rear-end attachments) that connect to the wheelchair’s frame. These add-ons are used to provide extra power assistance to the user and mitigate or eliminate the physical load of MWC propulsion. A power-assisted pushrim-activated wheelchair (PAPAW) is an example of a PAW in which the conventional manual wheels are replaced with motorized ones. Using PAPAWs can contribute to maintaining a physically active lifestyle and provide the potential to have a more balanced sense of personal autonomy to users [1]. PAPAW motion is coordinated by both the user’s input force at the pushrims and the powered-wheels’ controllers. Due to the inherent complexities of this collaborative shared-control between the user and the powered-wheels, control of PAPAWs has become one of the highly researched topics in this field. To present an overview of these research studies, we present a summary of the existing literature related to the control of PAPAWs, briefly discuss the strengths and limitations of these studies, and propose potential research topics for future investigation in this field.
BACKGROUND
Detailed descriptions of control algorithms of commercialized powered wheels (e.g., Alber e-motion and twion, Yamaha JW Swing) are not openly available in journal literature. However, generally available information regarding their operation reveals these wheels are instrumented with sensors to determine the user’s interaction with the pushrim (e.g., measuring the user’s input force/torque). The wheels’ actuators respond to the user’s input by regulating the assistance torque or the wheels’ velocity at each stroke. It is important to note that the user’s direct and continuous interaction with the pushrims is required to initiate and maintain the wheels’ motion unless the wheels are set to operate in a cruise control mode.
Performance evaluation of some of these powered-wheels revealed two main challenges faced by PAPAW users that are: (1) coordinating the pushes on each wheel to achieve a smooth drive on different terrains [2-4]; and (2) maintaining a safe and stable ride on steep slopes [2,4]. It is hypothesized that the first limitation is related to the fact that each wheel is controlled separately and there is no communication between the two wheels. Therefore, the controller cannot compensate for an unintended user input asymmetry or unexpected road disturbances. The second limitation is related to the lack of information concerning the chair or road conditions, e.g., inclination. Therefore, the controller cannot compensate for the gravitational forces that are applied to the system. In this regard, researchers studied and developed more advanced control techniques to address these limitations. We reviewed the literature on these topics and summarize the current state of knowledge on the control of PAPAWs.
METHOD
To identify the published literature related to PAPAW control, we searched three electronic databases: Medline (Source PubMed), Google Scholar, and IEEE Xplore (Digital Library). We used different combinations of the following keywords: “power-assisted wheelchair”, “pushrim-activated power-assisted wheelchair”, “powered- wheels”, “manual wheelchair”, “controller”, “observer”, “shared-control”, “adaptive control”, and “robust control”. We examined the references of the selected literature to identify relevant articles that may have been missed in the database searches. We excluded the studies on the control of joystick-operated and assistant-propelled PAPAWs, (semi-)autonomous navigation of PAPAWs, and PAPAW mechanical design.
RESULTS
We found 40 articles that met our inclusion criteria and were therefore selected for a full review. After screening these articles, we eliminated duplicate studies. In this case, we included either the original publication or the document with the most comprehensive text. In the end, a total of 22 articles were selected for this review. Due to limited space, we excluded the information regarding the model dynamics of the systems, measured or estimated states, and observer design strategies from this text. Following the research team discussions, four main research themes emerged from the reviewed literature. We categorized the studies based on these four research themes and further classified them by their high-level research objectives and low-level controller design structure (e.g., the control objective and control technique). A summary of these findings is presented in Table 1.
Main research theme
|
Research objective(s)
|
Controller design
|
Basic wheeling assistance
|
A proportional torque controller that takes as input the user’s input torque and the wheels’ angular velocity to provide proper assistance torque [5]
|
|
A position-based trajectory controller that uses a minimum jerk trajectory planning technique to provide a smooth ride during the recovery phase [6]
|
||
A PI velocity controller that takes as input the user’s input torque and a minimum jerk and acceleration planning technique to provide a smooth ride [7]
|
||
Providing basic propulsion assistance
|
A fuzzy torque controller that uses the estimation of the user input torque and EMG signal of the hand muscles to adjust the torque assistance ratio [8]
|
|
A PI velocity controller that uses the synchronization error between the two wheels to control the velocity of each wheel [9]
|
||
Wheeling on a straight/ curved road
|
A PI velocity controller that uses the push frequency information to estimate the user intention and generate linear and yaw velocity references [10]
|
|
A two-dimensional torque controller that uses the linear and rotational input torque of the left and right wheels to adjust the assistance torque and its duration for the straight and rotational motion of the wheelchair [11]
|
||
A fuzzy torque controller that uses the wheelchair’s posture angle and angular velocity and the user input torque to adjust the assistance torque [12]
|
||
A torque controller that takes the user’s input torque and the temporal similarity between the right and left wheel input torque to generate a balanced assisted torque to compensate for user input asymmetry [13]
|
||
Wheeling on sloped surfaces
|
Two PD force and position/velocity controllers to compensate for longitudinal and lateral gravity disturbances [14-15]
|
|
A PD torque controller that uses the driving velocity feedback and the user input torque to adjust the assistant torque on sloped and rough surfaces [16]
|
||
A robust PD velocity and acceleration controller that uses linear and angular velocity (of the wheelchair) feedback to adjust the assistance torque and control the longitudinal and rotational motion on slopes [17]
|
||
Adapting to environmental conditions
|
A torque controller that uses tilt sensor data and fully compensates for gravity up-/down-hill and gear mechanism friction [18]
|
|
Road disturbance rejection
|
An adaptive, robust, acceleration controller that uses encoder data and a disturbance observer to adjust the power assistance ratio [19]
|
|
A fuzzy torque controller that uses the driving distance information to adjust the assistance torque and its duration for different terrain conditions [20]
|
||
A proportional torque controller that uses a disturbance observer to estimate external disturbances and provides a compensation torque to the right and left wheels [6]
|
||
A PD force-sensor-less controller that uses velocity feedback to adjust the assistance torque and eliminate the effects of disturbance torques [21]
|
||
Improving safety and stability
|
Backward tipping prevention and velocity limitation
|
A torque controller that uses a phase-plane analysis of the center of mass and its velocity to adjust the assistance torque on flat or sloped surfaces [14]
|
A torque controller that uses the Lagrange stability analysis approach to adjust the assistance torque and prevent backward overturn [22]
|
||
A velocity controller that uses a minimum jerk model to perform optimal velocity tracking and uses a regenerative braking system [23]
|
||
Enabling task- specific operations
|
Step climbing
|
A PI position controller that uses the inclination angular velocity feedback to raise the front wheels and perform an inverted pendulum control [24]
|
A torque controller to raise the front wheels and a PD position controller to move up a step while following a polynomial trajectory [25]
|
||
One-hand propulsion
|
A torque controller that takes the user’s input torque and its rate of change to infer the intention of the user to move straight, turn, or stop [15,26]
|
The four research themes and the proposed solutions to achieve the research objectives are: (1) providing basic propulsion control for PAPAWs by generating appropriate wheeling assistance corresponding to users’ physical characteristics or intentions and the desired motion patterns; (2) adapting to the environmental conditions by mitigating or eliminating the effects of external force or disturbances; (3) improving safety and stability of PAPAW use by introducing stability margins, implementing safe-mode operation states, and adjusting the control assistance accordingly; and (4) enabling task-specific operations by employing operation-mode switching strategies. Regarding the low-level controller structure, we identified the following topics that were examined and discussed in the selected literature: (1) the main control strategy, e.g., torque, position, velocity, or acceleration control; (2) the controller type, e.g., PI, PD, adaptive, and/or fuzzy controllers; and (3) user intention detection strategies, e.g., measured/estimated user’s input torque, upper extremity muscle activation signals, and/or kinematic characteristics of the wheelchair motion.
DISCUSSION
Overall, preliminary outcomes of the reviewed literature revealed that implementation of the proposed controllers in PAPAWs can improve user-wheelchair interaction by providing an intuitive sense of control to the users, providing a stable and smooth ride, and compensating for undesirable external disturbances.
After examining the methodologies and outcomes of the reviewed literature, we identified three main limitations. First, the controllers are designed based on a predefined model characteristic of a wheelchair and user in which the dynamic interactions between the two were disregarded and the individual characteristics of users are ignored. Second, the proposed adaptive control strategies are not activated immediately. For example, in the case of moving on an uneven surface, PAPAW users must exert higher forces at least for the first couple of strokes until the controller starts compensating for the external disturbances. Third, these studies were mainly performed with a very limited number of participants that in most cases were not expert wheelchair users. Most of these studies lack quantitative analysis or qualitative and subjective assessments of the results. Moreover, no statistically significant differences were found between the PAPAWs’ performance with conventional control techniques and the proposed advanced control strategies.
CONCLUSION
PAWs’ potential to improve the personal autonomy and quality of life of people with mobility disability have motivated research and development of power add-ons for wheelchairs. PAPAWs specifically have been the focus of a small number of studies in this field. We reviewed the published literature on the control of PAPAWs and presented a summary of their research methodologies in this paper. The findings of our research showed that studies on control of PAPAWs were primarily focused on improving the safety and quality of the ride by employing advanced adaptive and robust control strategies.
Although the outcomes of previous research have shown promising improvements in the control of PAPAWs, there are still significant limitations that restrict users’ flexible interaction with the device and the environment. Based on these limitations and the outlined shortcomings of the previously studied control techniques, we propose the following research topics for future investigations in this area: (1) personalized learning-based controllers that take into account the dynamic user-device interactions (e.g., using each individual’s physical characteristics and sensed changes over time); (2) controllers with quick adaptation capabilities (e.g., using probabilistic algorithms to determine or predict the environmental conditions and changes); and (3) systematic assessment of PAPAWs’ controller performance. Moreover, advancing modular controllers to enable the control of multiple power add-ons (e.g., to enhance environmental accessibility) and designing controllers to perform structural adaptations (e.g., changing sitting position and orientation to maintain the safety and stability of the user) could be other examples of future research directions.
REFERENCES
[1] M. Khalili, J. Borisoff, and H. F. M. Van der Loos, “Towards an Autonomy-Based Approach to Design and Develop Mobility Assistive Technologies for Spinal Cord Injury Populations,” Proceedings of the RESNA Annual Conference. Arlington; 2018.
[2] E. M. Giesbrecht, J. D. Ripat, J. E. Cooper, and A. O. Quanbury, “Experiences with using a pushrim-activated power-assisted wheelchair for community-based occupations: A qualitative exploration,” Can. J. Occup. Ther., 2011 78(2):127–136.
[3] P. R. Giacobbi, C. E. Levy, F. D. Dietrich, S. H. Winkler, M. D. Tillman, and J. W. Chow, “Wheelchair Usersʼ Perceptions of and Experiences with Power Assist Wheels,” Am. J. Phys. Med. Rehabil., 2010 89(3): 225–234.
[4] K. L. Best, R. L. Kirby, C. Smith, and D. A. Macleod, “Comparison between performance with a pushrim- activated power-assisted wheelchair and a manual wheelchair on the Wheelchair Skills Test,” Disabil. Rehabil., 2006 28(4): 213–220.
[5] R. A. Cooper et al., “Performance assessment of a pushrim-activated power-assisted wheelchair control system,” IEEE Trans. Control Syst. Technol., 2002 10(1):121–126.
[6] H. Seki, T. Sugimoto, and S. Tadakuma, “Novel straight road driving control of power assisted wheelchair based on disturbance estimation and minimum jerk control,” In 40th IAS Annual Meeting. Conference Record of the Industry Applications Conference (1711–1717); 2005.
[7] H. Seki and S. Tadakuma, “Velocity pattern generation for power assisted wheelchair based on jerk and acceleration limitation,” In 31st Annual Conference of IEEE Industrial Electronics Society; 2005.
[8] Yuusuke Oonishi, Sehoon Oh, and Yoichi Hori, “New control method for power-assisted wheelchair based on upper extremity movement using surface myoelectric signal,” In 10th IEEE International Workshop on Advanced Motion Control (498–503); 2008.
[9] P.-W. Hsueh, H.-T. Pan, A. Grandjean, and M.-C. Tsai, “Integrated Synchronized Motion Control for a Force Sensorless Power-Assisted Wheelchair,” IFAC Proc. Vol., 2011 44(1):5956–5961.
[10] G. Feng, T. M. Guerra, S. Mohammad, and L. Busoniu, “Observer-Based Assistive Control Design Under Time-Varying Sampling for Power-Assisted Wheelchairs,” IFAC-PapersOnLine, 2018 51(10):151–156.
[11] Kayoung Kim, Kanghyun Nam, Sehoon Oh, H. Fujimoto, and Y. Hori, “Two-dimensional assist control for power-assisted wheelchair considering straight and rotational motion decomposition,” In 38th Annual Conference on IEEE Industrial Electronics Society (4436–4441); 2012.
[12] H. Murakami, H. Seki, H. Minakata, and S. Tadakuma, “Operationality improvement control of Electric power assisted wheelchair by fuzzy algorithm,” In IEEE International Conference on Industrial Technology (1–6); 2009.
[13] Y. Heo, E.-P. Hong, M.-S. Mun, and T.-H. Choi, “Torque balancing for power assisted wheelchair based on torque and temporal similarity,” Int. J. Precis. Eng. Manuf., 2015 16(8):1729–1734.
[14] Sehoon Oh, N. Hata, and Y. Hori, “Control developments for wheelchairs in slope environments,” In Proceedings of the American Control Conference (739–744); 2005
[15] S. Oh and Y. Hori, “Lateral disturbance rejection and one hand propulsion control of a power assisting wheelchair,” In 31st Annual Conference of IEEE Industrial Electronics Society; 2005.
[16] H. Seki and N. Tanohata, “Fuzzy Control for Electric Power-Assisted Wheelchair Driving on Disturbance Roads,” IEEE Trans. Syst. Man, Cybern. 2012 42(6):1624–1632.
[17] Sehoon Oh and Y. Hori, “Disturbance Attenuation Control for Power-Assist Wheelchair Operation on Slopes,” IEEE Trans. Control Syst. Technol., 2014 22(3):828–837.
[18] K. Lee, C.-H. Lee, S. Hwang, J. Choi, and Y. Bang, “Power-Assisted Wheelchair With Gravity and Friction Compensation,” IEEE Trans. Ind. Electron., 2016 63(4):2203–2211.
[19] S. Katsura and K. Ohnishi, “Advanced motion control for wheelchair in unknown environment,” Conf. Proc. - IEEE Int. Conf. Syst. Man Cybern. (4926–4931); 2007.
[20] H. Seki and A. Kiso, “Disturbance road adaptive driving control of power-assisted wheelchair using fuzzy inference,” In Annual International Conference of the IEEE Engineering in Medicine and Biology Society (1594– 1599); 2011.
[21] Sehoon Oh and Y. Hori, “Sensor Free Power Assisting Control Based on Velocity Control and Disturbance Observer,” In Proceedings of the IEEE International Symposium on Industrial Electronics (1709–1714); 2005.
[22] W. Li, N. Hata, and Y. Hori, “Prevention of overturn of power assisted wheelchair using novel stability condition,” In 31st Annual Conference of IEEE Industrial Electronics Society; 2005.
[23] H. Seki, K. Ishihara, and S. Tadakuma, “Novel Regenerative Braking Control of Electric Power-Assisted Wheelchair for Safety Downhill Road Driving,” IEEE Trans. Ind. Electron., 2009 56(5):1393–1400.
[24] Y. Takahashi et al., “Analysis of front wheel raising and inverse pendulum control of power assist wheel chair robot,” In 26th Annual Conference of the IEEE Industrial Electronics Society (96–100); 2000.
[25] H. Seki et al., “Novel Step Climbing Control for Power Assisted Wheelchair Based on Driving Mode Switching,” Annual Conference on IEEE Industrial Electronics (3827–3832); 2006.
[26] K. Kim, K. Payne, S. Oh, and Y. Hori, “One-handed propulsion control of power-assisted wheelchair with advanced turning mode,” IEEE International Workshop on Advanced Motion Control (633–638); 2014.