RICERCA ITALIANA     

Development of novel methods for the measurement of mechanical quantities to optimize the movement rehabilitation

Università degli Studi di Roma "La Sapienza"

Abstract

Research in the field of rehabilitation is expanding considerably. Recent works have shown a variety of different approaches to adapting mechatronic system for rehabilitation clinics and many other advanced technologies are being adapted and developed for this purpose. The involvement of mechanical systems in rehabilitation can be divided into three categories:
(i) to assist disabled people in special need with their daily life activities;
(ii) to support mobility;
(iii) to assist therapists performing repetitive exercises with their patients.
The aims of the present research proposal are focused on the last category, in fact we plan to develop:
1) a simple and widely sensorized - Figure 1 - (multicomponent load cells and pressure sensor matrix) moving platform remotely controlled for the rehabilitation of the patient’s legs together with
2) an instrumented helmet - Figure 2 - to measure the acceleration of subject’s head and
3) intertial systems that will sensorize articular segments of the examined subjects.

Figure 1 - Motorized platform.


Figure 2 - Instrumented helmet.


Figure 3 - Scheme of the proposed tele-rehabilitation system.


The scientific objectives are as it follows.
Scientific objective #1 – Address the shortcomings in physiotherapic activities performed by proposing a remotely web-based supervision and control performed by therapists, see Fig.3.
Scientific objective #2 – Propose novel and simple training exercises specifically trimmered for the motorized platform; the exercises must be accessible, effective and fun by simulating a large variety of exercises and motivating patients using also virtual environments; so the devices allow timely care for patients who otherwise would not have access to necessary therapy.
Scientific objective #3 – Development of: (1) an instrumented motorized platform; (2) an instrumented helmet to monitor the head stability during the conduction of the imposed exercise; and, finally, (3) inertial systems that will be located on the patient to reconstruct the cinematics without the use of expensive optoelectronic systems.
Scientific objective #4 – Identify synthetic indicators for an early detection of symptoms and signs, to postpone, by means of extensive at-home exercises, the subject’s hospitalization.
Scientific objective #5 – Propose the use of a system in such a manner as to improve efficiency and to bring the tele-rehabilitation approach into light; this approach will also help subjects in rural populations with long treatment programs to treat themselves at home, at their own time, to save the hassle of commuting everyday to the clinic, and to reduce the burden on the healthcare system.
Scientific objective #5 – Allow timely care for patients who otherwise would not have access to necessary therapy.

The technological objectives are as it follows.
Technological objective #1 – Propose an innovative design of the platform capable to perform simple tasks compatible with the specific disease of the patients.
Technological objective #2 – The design be safer and cheaper to bring devices into the mainstream as a domotic class.
Technological objective #3 – Overcome the safety constraints due to the interaction of electro-mechanical systems with humans.
Technological objective #4 – Overcome the negative effects induced by the time-delay in the control of the devices due to the remote control and supervision provided via internet.
Technological objective #5 – develop an ad-hoc instrumented helmet and extensively use already developed inertial systems to reconstruct the kinematic of the patient. <<<

Principal Investigator

Paolo Cappa Università degli Studi di ROMA "La Sapienza"

Research Objectives

The discipline of rehabilitation robotics is diverse, and although a liberal definition of both words is usual, the intersection of these two areas is still small. Rapid developments in measurement techniques, robotics and computing, together with an increasing need to develop tools for daily living tasks, especially to allow a greater independence for the elderly, are expanding the role of automatic devices in assistive technologies.
In addition the rapid growth of the nations older population has brought about an increasing awareness of the special health service needs of older adults. There are significant demands on the families, lay persons and professionals who deliver health services to this population. Their demands on their home environment will increase and change with growing age, especially when their health status starts to worsen. An important aspect for all people having the need to be supported in their daily-life-activities is to remain integrated in social life, despite of their age and existing disabilities.
The opinion of the proposers of the present research is that the development of automatic systems in-house used may significantly help not only patients affected by different pathologies but also the growing group of elderly people to live longer at the place they like most, while ensuring a high quality of life, safety, and security, including e.g. health monitoring and supportive functions.
The present research proposal is focused on the design of an sensorized moving platform remotely controlled capable to perform robot-mediated physical exercises in response to the emerging themes from physiotherapy research. In fact, a systematic review of existing physiotherapy research suggests that exercise-based interventions are the only interventions for which there is a body of positive evidence.
Thus the present proposal is characterized by the following main features and concepts:
1. the development of a rehabilitation device capable widely sensorized to impose simple training exercise directly in home environments; the device are directly and remotely programmable by physiotherapists;
2. the outcome of the chosen training exercises are on-line remotely evaluated by physiotherapist.

The approach that will be adopted in the present proposal can be schematized as it follows:
1. improve functionality, not anatomy, to assist patients, so that the technology solution focuses on tasks;
2. use industrial and consumer grade products to avoid research and development setbacks due to inadequate, unsafe, or failure-prone components;
3. keep the consumer and clinicians tightly involved in the design process, so that all eventual users of the novel rehabilitation equipment are considered from the beginning; and, finally,
4. embed the design process in an iterative design-evaluation cycle, which allows the technology and evaluation/methodological tools to be developed in parallel and in concert.
To accomplish the previously indicated objectives, many factors will be considered in the design of the robotic devices: intrinsic simplicity, easiness to assemble and use, workspace, load capacity, speed, repeatability and accuracy, volume and cost.
The proposed research will try to answer some of the questions that did not or not fully answered by the literatures. These questions are:
1. What types of exercise regime/movements can be trained by the rehabilitator with full assurance of the subjects and therapists safety;
2. The morphology and mechanical structure of the device (speed, actuator, joints, links, degrees of freedom, size, end-effector, control type);
3. How the human-machine interface will look like? How the patient will be attached and interact mechanically with the device?
4. What should be sensed as a Biofeedback to be analyzed and quantified, to give an indication of the recovery progress?
5. How can we incorporate the information from bio-feedback as input to an intelligent decision making algorithm used to increase or decrease the level of exercise difficulty according to patient performance?
6. How do patient and therapist will communicate with the device cell? The level of this communication and mode.
7. The communication between the user and therapist over the local area network, its reliability and safety.

The answers of the above mentioned questions will contribute intensively to the development and optimization of design of novel devices for tele-rehabilitation, to bring this technology into light. The proposed design of the sensorized platform will perform under four different conditions:
1. Completely passive movements;
2. Assistive, helping the patient movement;
3. Guiding the movement, neither helping nor providing resistance, but guiding the patient through a prescribed routine;
4. Providing some resistance against the patient.

The Research Units will develop the following main hardware &amp; software components:
1. a mechatronic structure and controller,
2. end effectors,
3. a computer controlled system and related software,
4. a graphical user interface,
5. a database system,
6. several, complex, articulated and integrated measurement systems, and, finally,
7. a safety system.
• The mechanical design and development including:
o System end effectors;
o tools and fixtures for different exercises that form the human-device interface;
o adaptive control;
o biomechanical measurement system and its different sensors for monitoring the performance and acquiring the working environment data,
o safety system in fact the new living environment should not generate new risks.

• The development and analysis of the software modules, which will include:
o programs that support the functions of the tele-rehabilitation system;
o modules that collect and analysis data received from different sensors;
o database design system to store and sort all the performance and treatment data;
o data analysis module;
o graphical-user interface;
o communication protocol. <<<

Timescale

24 months

National and international background

2.2.1 INTRODUCTION
Physical exercise has multiple positive effects in older adults, including those with disabilities, and in the last few years several studies involving older persons have been published about interventions that showed a significant positive effect of physical training [1-3]. More precisely, exercise prevents and reduces the risk of developing secondary conditions that arise from functional decline and physical disuse [4]. Regular exercise that focuses on functional fitness, such as walking, has been associated with significant reductions in the levels of dependence and disability in older adults. The empirical relations among physical fitness levels, specifically aerobic fitness, cognition, and physical health in older adults, are well established [5]. Simple motions imposed to the subjects can help relearn elderly people with limb impairment to be functionally capable of performing basic daily life activities. Although many of the improvements in performance declined after training stopped, differences between groups in some motor performances remained significant after two years in those geriatric patients who were restricted by sedentary lifestyle, fall-related impairment, and advanced functional decline [6]. Cross-sectional and longitudinal data have also demonstrated that physically active people have a lower risk of developing Alzheimer’s disease and related cognitive disorders when compared with sedentary people [7]. Aerobic fitness training appears to have an association with reduced brain tissue loss in aging humans. In addition there is a growing body of research suggesting that some risk factors for heart disease and stroke are associated with the development of Alzheimer’s disease and related disorders. Countries that have high dietary fat consumption also tend to have a higher prevalence of dementia; these findings have led to the hypothesis that physical activity strategies designed to prevent and manage cardiovascular disease might also be effective in the prevention and management of dementia [8]. A positive outcome can be found also in terms of a general improvement of the structure of the musculo-skeletal system.

2.2.2 TELE-REHABILITATION SYSTEMS
The findings illustrated in Section 2.2.1 boosted research and development of devices that could assist in delivering functional movement in a repeatable and repetitive fashion [9, 10]. The common denominator of these devices is that they assist a person that practises functional movement, such as balancing while standing [11] or walking on a treadmill [12, 13], with a degree of mechanical support that enables movement which otherwise would have not been possible without assistance of a couple of physiotherapists working in ergonomically unfavourable conditions.
While skilled therapists can achieve good results with even rudimentary equipment, the maximum effectiveness of their existing toolbox is rapidly being reached and to increase productivity new tools are needed. There is increasing evidence that active repetitive practice of movement can have a profound effect on recovering impaired motor functions after stroke or brain injury [14-17]. However, a long and intensive program of treatment is required. This type of treatment is costly and a major burden on the healthcare system, which has resulted in inadequate duration of rehabilitation programs. A major issue is the need for a human attendant and/or caregiver to work with these patients. On the other hand, it is widely accepted that repetitive labour-intensive processes are easily and precisely handled by an automatic system. Thus, in recent years great efforts toward developing automatic system treatments were dedicated mainly motivated by the increasing public health burden associated with an ageing society. Many automatic systems were consequently designed and manufactured:
(i) to alleviate the labor-intensive aspects of physical rehabilitation;
(ii) to enable novel modes of exercise not currently available; and, finally,
(iii) to obtain higher patient and provider satisfaction resulting from positive impact on medical outcomes [18, 19].
However, expensive automatic systems for rehabilitation have primarily been sold to sites for teaching and assessment or to Research Institutes for extended evaluation and only recently they are being used at the clinic. On the contrary simple and economical mechanical systems for at home use are not sensorized or networked, therefore, there is no remote monitoring or re-evaluation of patient progress. Thus, patients need to travel repeatedly to clinics in order to be re-evaluated, and such travel is difficult for subjects with reduced autonomy. Furthermore, those home exercises that are generally not interactive can be very repetitive and boring and, therefore, the patient may not be as motivated to do the exercises prescribed, and the lack of timely therapy aggravates the patient’s medical condition. In the attempt to take advantage of automatic systems and to allow a massive utilization at home, web based tele-rehabilitation complex systems were recently proposed without the expense of a continuous supervising therapist or sophisticated control system. In a recent paper, in fact, a tele-rehabilitation system for arm and hand therapy following a stroke was proposed; it is capable to:
(i) facilitate repetitive movement therapy;
(ii) perform a customized program of therapeutic activities;
(iii) receive quantitative feedback of their rehabilitation progress [20].
Similarly a novel device for ankle tele-rehabilitation was proposed in [21]. A remote supervising caregiver can then monitor progress, make changes to the exercise program, and provide information and encouragement.
In spite of the potential advantages of rehabilitation devices, such systems are not successfully commercialized yet. Market place failures can be attributed to the following reasons:
1. the subjects perceived potential barriers if tele-health devices themselves were complicated or there was lack of training [22];
2. the physiotherapists perceive an automatic system as a “problem maker” and not as a “problem solver” [23];
3. automatic systems, unlike their human counterpart, lack the cognitive capacity to prescribe, observe and alter treatment according to patient performance [23];
4. safety of the patient is another important technical issue [23].

2.2.2.a MOTORIZED PLATFORM
Maintaining balance while standing upright is a prerequisite for successful performance of many daily activities. Recognising the essential role of human upright stance, many studies have been carried out to elucidate the physiological mechanisms underlying normal postural control [24, 25] and the pathophysiology of balance disorders [26]. These insights have increased noticeably with the advent of dynamic posturography: the assessment of balance correcting responses following controlled postural perturbations. A common type of dynamic posturography is to perturb upright stance by sudden movements of a supporting platform upon which the subject is standing [27]. Examples of commonly used platform movements include horizontal translations [28] and dorsiflexion or plantarflexion rotations about the ankle joint [29]. Studies using these movable platforms have revealed valuable information about postural control mechanisms in healthy subjects and patients with various balance disorders [29]. However, most currently available platforms have shortcomings. The first drawback relates to the size of the support surface, which for some platforms is too small to allow subjects to take corrective steps. For example, Allum and colleagues [30] use a relatively small-sized platform to which the feet of their subjects are strapped to prevent them from stepping off the platform. Studies that used larger support surfaces have stressed the importance of compensatory stepping responses, not only when balance is truly jeopardised but also under less threatening conditions [31]. Second, for some platforms, even the largest or fastest motions generated are insufficiently destabilising to actually bring subjects beyond their stability limits. For example, the most destabilising rotations supplied by the commonly used Neuro-Com platform (10° rotation amplitude, 50°/s rotation velocity) rarely cause serious balance problems for young persons, while they bring about moderate balance problems in elderly persons and patients with Parkinson’s disease [32]. We speculate that bringing subjects beyond their stability limits is required to further increase insights into the (patho-) physiological processes causing falls in daily life. Third, many movable platforms can only produce perturbations in a single direction, typically the pitch plane. Multidirectional perturbations may be more informative, because in daily life falls may occur in any given direction. Indeed, studies using multidirectional perturbations have unveiled abnormalities in patients that would have been missed using strictly unidirectional perturbations [33]. Furthermore, a multidirectional protocol reduces habituation effects by diminishing the predictability of the upcoming perturbation direction [34]. Multidirectional perturbations thus correspond better to falls in daily life, which are predominantly unexpected events.
A new apparatus (CAREN, Motek, Amsterdam) based on a Stewart platform [35, 36] has recently become available which consists of a 2 m diameter platform that can be moved by six computer driven hydraulic actuators in six degrees of freedom. The CAREN system’s graphical user interface allows the user to set the platform’s position and orientation by specifying its three translational displacements (sway, surge, heave) and three rotational displacements (pitch, roll, yaw). In contrast to other apparatus, the platform can, in principle, be programmed to rotate around any origin, and so its movement can be made to vary between or within measurement trials. The ability to control or restrict the platform’s rotation to coincide with a joint axis when a perturbing motion is used is likely to be of value in investigations that attempt to explore, for example, the relative importance of proprioceptive signals from the lower leg compared to sensory inputs produced by neck, trunk and thigh muscles.

2.2.2.b INERTIAL SYSTEMS FOR THE EVALUATION OF THE SEGMENTAL KINEMATIC
Head attitude, velocity and acceleration are relevant variables in the estimation of the capability of the equilibrium [37-39], because they represent the actual inputs to the vestibular system. The measurement of head motion, and in general of kinematics of the entire human body, is normally carried out by using stereo-photogrammetric systems. However, the complexity and costs of those apparatus can represent a “problem” when reduced amount of data are sufficient to supply valid clinical assessment; in addition, optoelectronic systems cannot be used outside the laboratory in real life situations. Similarly it appears of interest the study of segmental kinematics. The decrease in the costs of electronic components and devices, conjugated to the advancement of the miniaturization technologies, has promoted a large profusion of the use of motion sensors for human movement analysis in clinics. Thus many studies were devoted to the development of less sophisticated and expensive systems, capable of collecting only the kinetic variables requested in unrestrained working volumes. To overcome this limitation body mounted sensors, usable in ambulatory analysis and in outdoor measurements, have been proposed. They can be categorized depending in the used sensing principle: techniques based on the use of one type of transducer i.e. accelerometers (ACs) [40-41], magnetic track-ing devices [42], miniature gyroscopes [43], GPS detectors [44]; data fusion tech-niques based on a combination of ACs and miniature gyroscopes [45] or ACs and earth-magnetic field sensors [46]. The present research proposal focuses on systems entirely equipped with ACs because they are becoming cheaper and are generally characterized by low weight and reduced dimensions. Literature indicated that attempts were generally conducted to find the optimal AC configuration and post-processing algorithm, in order to extract the maximum information possible on the motion of the rigid body. Morris [47] proposed the use of six single-axis ACs, successively, in order to overcome the inherent instability of the six-AC model [48] and also to reduce the effects caused by measurement errors, Padgaonkar et Al. [49] developed a system equipped with nine linear axes, four accelerometers in a so-called “3-2-2-2 configuration”; however, the set-up requires a careful positioning of the three bi-axial ACs to use the analytical post processing associated. Scientific literature reveals relevant practical usability of such a system; in fact the “3-2-2-2 configuration” is still effectively and widely used in two separate fields: (a) evaluations of high acceleration level causing mild brain injury, (b) measurements of motion of the body.
As regards the first topic (a), the above nine-axis scheme was used to study: the impact on crash test dummies [50] and [51], the acceleration of the skull of merinos ewes induced by controlled impacts [52], the effectiveness of football helmets in vibration dumping [53]. Crisco et Al. [54] proposed a novel algorithm for calculating linear acceleration and impact location on a headset using single-axis ACs; the main limitation of that study relatively to the present topic, is that the proposed methodology permits the evaluation only of linear accelerations.
Regarding the second main research topic (b) Giansanti et Al. [55] numerically analyzed the performances of the six axes AC and the nine axes AC schemes indicated above for biomechanical applications. The outcome was that, in the presence of even small experimental errors, both schemes are unsuitable, due to the numerical drift generated from the post processing of data.
In general the previously mentioned schemes formed by six- and nine-axis ACs are prone to the effects induced during AC assembling - i.e. by inaccuracy in pointing, positioning and alignment - which could only be compensated by using additional ACs. In fact Baselli et Al. [56] utilized twelve linear axes by equip-ping a helmet with four tri-axial ACs; the helmet is capable of determining the head’s angular acceleration and the (a-g) component relative to the vestibular system. The same scheme is indicated by Zappa et Al. [57] as the minimal experimental set-up capable to evaluate the complete acceleration state of a rigid body and to ensure a unique solution. Parsa et Al. [58] and [59] proposed a general algorithm for a generic number of tri-axial ACs capable of evaluating attitude, angular velocity and acceleration of a rigid body adopting a stable data fusion technique. While planar motion was recently investigated by Williams and Fyfe [60], with the goal of determining the geometrical conditions on the minimal configurations of ACs, the extension to the 3D case has not been exploited so far. Additionally, no indication was given of an algorithm able to solve the general redundant scheme of n single-axis or bi-axial ACs, randomly oriented and fixed over a rigid body. Such schemes, compared with tri-axial AC schemes, could be more viable and less inexpensive solutions and may potentially determine an improvement of the overall characteristics thanks to the redundancy of information. <<<