Methods and Tools for Product Design
Relatore: Prof. Mario Covarrubias Rodriguez
Tutor: Prof. Francesco Braghin
Università di Provenienza: Politecnico di Milano - Ingegneria Meccanica
Cognitive and Motor Rehabilitation System for Home-Therapy Development of an Active Hand Exoskeleton and an AI-Based Virtual Assistant
Nowadays, we are experiencing a social transformation: the world population is ageing . A longer life brings new opportunities but it can cause an increased incidence of diseases such as stroke. Worldwide, stroke is the third leading cause of disability .
A stroke can be defined as a neurological deficit due to a central nervous system infarction or haemorrhage. Among its numerous consequences, there is the contralateral paralysis of the hand . This symptom hinders many activities of daily living. Many stroke survivors are involved in rehabilitation therapy with medical experts for years after the disease. It has a high social and economic impact both on patients and society. Literature review shows that moving towards a home-therapy with mechatronic devices seems promising. However, there are still numerous open challenges in literature such as the comfort, the portability/packability, the cost, and the usability for non-expert users .
Hand exoskeleton (HANDY)
- Portable, affordable, comfortable, customizable, and easy-to-use prototype
- Performing rehabilitative hand exercises for post-stroke patients defined with medicalprofessionals
3D virtual assistant (MAIA)
- Understanding human natural language using AI-based algorithms
- Interactive 3D character that offers assistantship and encouragement to patients toreduce the current therapy abandon rate
Global desktop application
- Providing an integrated desktop application for both therapists and patients thatincludes passive and active rehabilitative exercises
The first outcome of the thesis is the prototype of hand exoskeleton named HANDY.
The ‘glove’ (Fig. 2 - 3D printed flexible hand exoskeleton of HANDY) is designed to extend the user’s fingers without harming the skin, and following the natural movement.
It is a combination of different flexible and rigid parts obtained using additive manufacturing processes.
The remote actuation unit includes linear actuators, electronic components (including a PCB realized on purpose, a LiPo battery, and a dedicate router for wireless communication). Also, some security elements such as an emergency push-button are present. All are mounted inside a padded backpack to obtain a highly portable device (Fig. 3 - Global setup of HANDY).
The second result of this thesis is a 3D virtual assistant named MAIA (Fig. 4 - Final 3D model of MAIA).
It was born to deal with some therapy abandon causes: loneliness, need of help and discouragement.
Starting from the appearance decided with preliminary focus groups with therapists and patients, I created the final 3D rigged and animated character.
MAIA can interact with users using natural language. It combines some AI-based features (Fig. 5 - Natural Language Processing algorithm) to capture the user’s need, convert it into a machine-readable input, classify the context and the main concepts (intents/entities) of the request, chose and perform the appropriate actions, generate a textual response for the user, and synthesize it with a neural voice.
To combine motor and cognitive aspects of the rehabilitation system, a global desktop application has been developed. It allows therapists to define the therapy and monitor patients progresses. On the other hand, patients can perform passive, active and cognitive exercises based on activities of daily living (ADL). Fig. 6 (Cognitive and motor exercise based on ADL with the Leap Motion controller) shows an example of this last case. The Leap Motion controller tracks the user’s hand, HANDY assists the patients performing the grasping movement, and MAIA provides feedback and encouragement.
The cognitive and motor rehabilitation system for the hand has been designed to assist post-stroke patients during their rehabilitative therapy towards regaining their autonomy. Users (therapists and patients) have been actively consulted during the whole development. This led to several advantages.
- HANDY is lightweight (only 190g) on the hand, permitting repetitiverehabilitative exercises without fatigue. The glove material minimizes skinabrasion and its shape prevents misalignments and finger hyperextension.Portability is ensured by the small backpack and wireless dedicated connection
- MAIA enables some interactions using natural language (avoiding predefinedkeywords) such as asking assistantship and instructions. It provides a morehearty and non-invasive feedbacks compared to traditional ones.
- The application provides users with customizable passive, active and cognitiveexercises based on the patient’s condition.
During the preliminary tests with post-stroke patients, these aspects led to promising results in terms of usability and willingness to use of the system.
 P.D. United Nations, Department of Economic and Social Affairs, World Population Prospects 2019:Highlights. ST/ESA/SER.A/423, 2019.
 S.L. James et al., Global, regional, and nationalincidence, prevalence, and years lived with disability for 354 Diseases and Injuries for 195 countries andterritories, 1990-2017: A systematic analysis for theGlobal Burden of Disease Study 2017, Lancet, 2018.
 R.L. Sacco et al., An updated definition of stroke for the21st century: A statement for healthcare professionalsfrom the American heart association/American strokeassociation, Stroke, vol. 44, pp. 2064–2089, 2013.
 F. Aggogeri, T. Mikolajczyk, and J. O’Kane, Robotics for rehabilitation of hand movement in stroke survivors, Adv. Mech. Eng., 2019.