Wearable machines that assist with therapy
by Asst/Prof Raye Yeow Chen Hua
Department of Biomedical Engineering, Faculty of Engineering, National University of Singapore
Advanced Robotics Center, Faculty of Engineering, National University of Singapore
Singapore Institute of Neurotechnology, National University of Singapore
Stroke is one of the leading causes of disability, whereby 15 million people worldwide suffer a stroke each year and nearly a third of them are left permanently disabled . Post-stroke neurological impairment often leads to paralysis of one side of the patient’s body, which can affect his/her ability to perform basic activities of daily living, such as grasping household items . This inability to independently perform daily tasks reduces the patient’s quality of life. In addition, about 60% of the stroke patients tend to experience contractures in their limbs, especially the hand and ankle, particularly due to prolonged immobility . The presence of joint contractures can further aggravate the patient’s disability conditions.
The most common method of recovering mobility functions in the limbs of the stroke patients is through regular therapist-assisted rehabilitation exercises. However, given growing manpower constraints and greying population, Singapore’s therapist-patient ratio is a low 1:9000, as compared to the ratio in other developed countries, e.g. 1:2000 in Australia . Assuming that a therapist can see at least 8 patients per day with each session lasting between 45-60mins, the use of rehabilitation robots will be able to save each therapist at least 6 workhours per day, which can be allocated to other important patient cases. With these robotic devices serving as an adjunct, the therapist can certainly improve their efficiency to manage more patients.
While there are various robotic devices available on the market that can assist with hand and ankle movements, these devices belong to the traditional class of ‘hard’ rehabilitation robots that are often bulky, uncomfortable and expensive. For example, traditional hard upper limb exoskeletons  which are able to provide assistance to the fingers often have issues with compatibility, comfortability for the patients, where the rigid components can impede the natural finger joint motion in other degrees of freedom. Similarly for hard lower limb exoskeletons [2, 6], most of these devices are bulky and difficult for use on hospital beds to provide assistive ankle exercises, especially for long-term bedridden patients.
Therefore, we envision a new class of rehabilitation robots made entirely of soft motors and materials – soft wearable machines. In this article, we introduce our soft robotic glove and sock, which are composed of only silicone rubber and fabric materials, and yet able to provide assistance to the hand and ankle movements respectively, using our unique coral-inspired soft actuator technology.
Bio-inspiration is a key element of our Evolution Innovation Laboratory at the National University of Singapore. We studied the tentacle structure and movements of Galaxea sp. corals, since these organisms are a good example of nature’s ‘soft robots’ that can bend and extend (Fig. 1) like their electromechanical ‘hard robot’ counterparts used in the industry. The coral tentacle actuates based on the muscular hydrostat mechanism, whereby the orientation and contraction of the muscle fibres can drive the fluid within the tentacle to move the tentacle structure into a particular shape.
Inspired by this elegant mechanism from nature, we adapted an actuation concept where we design an elastomeric structure with a patterned cavity and increase the internal fluid pressure to move the structure into a desired shape (Fig. 2). With this concept, we developed our soft bending and extension actuators, using a combination of 3D-printing and soft lithography techniques, and subsequently implement them in our soft robotic glove and sock respectively for assistive and rehabilitation applications (Fig. 3).
Soft Robotic Glove
Our soft robotic glove is designed to assist in the flexion and extension of the fingers. It comprises of a fabric glove embedded with soft bending actuators for each finger, and a pump-valve control system. Once donned onto the user, the control system can be programmed to bend the individual fingers to create different common hand postures, such as full hand grasp or pinching (Fig. 4).
We currently have three different control strategies for this robotic glove: (1) button control, (2) muscle activity control and (3) contralateral hand kinematic control. For the button control, the control system has a selection of push buttons, in which each button corresponds to one specific hand posture. When the user/therapist presses a particular button, the robotic glove will assist the user in performing the designated hand posture.
For the muscle activity control, the robotic glove can be activated to assist the patient in hand movement through a wireless muscle activity sensing forearm band worn on the affected arm, thus allowing the patient to use his/her residual muscle activity on the affected arm as a form of intent command to control the robotic glove to provide assistance to the hand.
(Soft Robotic Glove Video: https://www.youtube.com watch?v=Q26MCUmzxmE)
In most cases, stroke patients are typically hemiplegic, where one side of the body is paralysed and the other side is still functional. Thus, it will provide a good opportunity for the patients to take active ownership of their rehabilitation process by providing them with contralateral hand kinematic control. In this control strategy, we let the patient wear a sensor glove on the unaffected side, and the robotic glove on the affected side. By moving the fingers on the unaffected hand, we can map the exact finger motions on the robotic glove at the affected side, thus providing a beneficial form of mirror therapy (Fig. 5).
The soft robotic glove project is funded by the Ministry of Education Academic Research Fund, and is currently undergoing clinical trials where we assess the brain activity stimulation in patients using functional magnetic resonance imaging, during robot-assisted hand therapy with the robotic glove .
Soft Robotic Sock
Our soft robotic sock is designed to assist bedridden patients with ankle dorsiflexion and plantarflexion. It comprises of five modules, namely sock, knee sleeve, double-extension actuators, joint angle sensor and pump-valve control system. The modular design permits convenient donning of the device on the bedridden patients.
The system can be remotely programmed through a graphical user interface on a laptop, whereby settings such as therapy duration and speed can be adjusted, depending on the therapist’s requirements for the patient. Once the control program is activated, it will trigger the pump to inject pressurised air into the extension actuator, which then extends to push the foot into plantarflexion. When the desired ankle joint angle is achieved (detected by the on-device angle sensor), the valve is then activated to remove the pressurised air, which contracts the actuator to pull the foot into dorsiflexion. These cycles of inflation-deflation facilitate the movement of the ankle through plantarflexion-dorsiflexion exercises (Fig. 6).
The robotic sock can be incorporated early into patients’ treatments to provide early mobilisation of affected limbs, so as to prevent development of joint contractures, and also to prevent or reduce immobility-related complications to the body, such as deep vein thrombosis and urinary tract infection .
(Soft Robotic Sock Video: https://www.youtube.com/watch?v=SF6fxF7vYyY)
The soft robotic sock project is funded by the Singapore Millennium Foundation, and is currently undergoing clinical trials where we evaluate the ankle joint range-of-motion and venous blood flow in patients using the on-device joint angle sensor and Doppler ultrasound respectively, during robot-assisted ankle exercise with the robotic sock .
‘Hard’ robotics, i.e. traditional robots typically made of rigid metallic components and driven by electromechanical actuators, has been around for over six decades in various useful consumer and industrial applications. As we enter the new age of connectivity and intelligent wearables, we wonder how we can wear robots and utilise them to empower us in daily tasks. Soft wearable machines, simply made of soft materials and motors, complement traditional robotics, particularly in safer human-robot interactions. I envision soft wearable machines as an opportunity and reality in the near future, where they play an instrumental role in providing robot-assisted movement for the human user in daily activities.
We presented our flagship projects, the soft robotic glove and soft robotic sock, which demonstrated the use of soft wearable machines to provide movement assistance to patients. Apart from robot-assisted therapy, soft wearable machines can expand into the space of robotic protection and robotic augmentation. Other ongoing developments in my lab include developing soft wearable machines that can protect elderlies from fall injuries, and also soft wearable supernumerary robotic limb attachments for augmenting human task performance. These soft wearable robotics work will not only benefit patients, but will also have the potential to penetrate and impact the wider world, including the consumer market and the defence industry, as soft wearable robotic apparels that can assist and augment task performance, while providing comfort and on-demand protection.
Collectively, soft wearable machines allow for safer human-robot interactions, and are capable of providing robot-assisted, robot-protected, and robot-augmented movement to human users. In the aspect of rehabilitation, these robots provide safe robot-assisted therapy to patients; they also act as an adjunct to therapists, thereby optimizing therapy time, improving recovery rate and increasing productivity, given growing manpower constraints. Considering the global ageing population and growing healthcare issues, these soft wearable machines will save both time and cost, and ultimately save lives.
- World Stroke Campaign (2015), https://www.world-stroke.org.
- Diaz, I., Gil, J. J., Sanchez, E. (2011). Lower-Limb Robotic Rehabilitation: Literature Review and Challenges. Journal of Robotics, Article ID 759764
- Sackley C, Brittle N, Patel S (2008). The prevalence of joint contractures, pressure sores, painful shoulder, other pain, falls, and depression in the year after a severely disabling stroke. Stroke, 39(12), 3329-34.
- Lee H. C (2008). More physiotherapists needed as population greys. The Straits Times - pg B3.
- Heo, P., Gu, G., Lee, S.-j., Rhee, K., & Kim, J. (2012). Current hand exoskeleton technologies for rehabilitation and assistive engineering. International Journal of Precision Engineering and Manufacturing, 13(5), 807-824.
- Dzahir, M., & Yamamoto, S.-i. (2014). Recent Trends in Lower-Limb Robotic Rehabilitation Orthosis: Control Scheme and Strategy for Pneumatic Muscle Actuated Gait Trainers. Robotics, 3(2), 120.
- Yap, H. K., Lim, J. H., Nasrallah, F., Low, F.-Z., Goh, J. C. H., & Yeow, R. C.-H. (2015). MRC-Glove: A fMRI Compatible Soft Robotic Glove for Hand Rehabilitation Application. Paper presented at the IEEE International Conference on Rehabilitation Robotics (ICORR).
- Bernhardt, J., English, C., Johnson, L., & Cumming, T. B. (2015). Early mobilisation after stroke: early adoption but limited evidence. Stroke, 46(4), 1141-1146.
- Low, F. Z., Tan, H. H., Lim, J. H., Yeow, C. H. (2016). Development of a Soft Pneumatic Sock for Robot-assisted Ankle Exercise. ASME Journal of Medical Devices, 10(1), 014503-014503-5.
About the Author
Asst/Prof Raye Yeow Chen Hua
Raye Yeow is an Assistant Professor with the Department of Biomedical Engineering at the National University of Singapore since 2012, and an affiliated Principal Investigator with the Singapore Institute for Neurotechnology and the Advanced Robotics Center. He received his B.Eng. (2006) and PhD (2010) in Bioengineering from the National University of Singapore, and his postdoctoral training in the BioRobotics Lab at Harvard University (2012). He was awarded the NUS Overseas Postdoctoral Fellowship (2010), NUS Young Investigator Award (2014), Yamaguchi Medal (2015) and MIT Technology Review Innovators Under 35 Asia (2016). His research interest is in developing soft wearable robotics for various healthcare applications. He is also a scientific advisor to four start-up companies working on medical wearables. (Email: firstname.lastname@example.org)