Announcement

Collapse
No announcement yet.

Two funded PhD positions (University of Salford, UK)

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Two funded PhD positions (University of Salford, UK)

    Two fully-funded PhD studentships(UK/EU fees, plus bursary) at the University of Salford. Both studentships will be based within the Rehabilitation Technologies and Biomedical Engineering research group, led by Professor David Howard (School of Computing Science and Engineering) and Professor Laurence Kenney (School of Health Sciences). The group is a friendly and active research environment and currently holds over £1.5 million in external funding from EPSRC, NIHR and charities. For more details see:
    http://www.salford.ac.uk/health-scie...al-engineering
    http://www.salford.ac.uk/computing-s...al-engineering

    Studentship 1: Adaptive control of lower-limb prostheses (http://www.salford.ac.uk/study/postg...nd-engineering)
    Supervisors: Professor David Howard (CSE), Dr James Gardiner (CSE), Professor Laurence Kenney (School of Health Sciences)

    Modern lower-limb prostheses only store and return significant energy below the knee, and energy is not returned in a controlled manner. For these reasons, we believe there is an opportunity for truly transformative research leading to a step change in the performance of lower limb prostheses. This requires advances in: a) mechanical design to provide flexible energy storage and return; and b) intelligent adaptive control to match the behaviour of the prosthesis to the terrain being crossed and various gait variables.
    The first aspect is being tackled in a current EPSRC funded project involving the Universities of Salford and Manchester. The PhD proposed here will tackle the second aspect (intelligent adaptive control) and will complement the work being undertaken in the EPSRC project.
    Context
    Unilateral trans-femoral amputee gait consumes up to 60% more energy than able-bodied gait [1-3]. For higher level amputees, research suggests that energy efficiency drops by well over 80% [4]. Recently it has been shown that energy consumption in high level amputees increases significantly when walking on slopes, suggesting studies in level walking may underestimate the extent of the problem [5]. The negative effects of high energy consumption are compounded by reductions in walking speed of typically 40% for trans-femoral amputees [6] with associated low activity levels, particularly in elderly amputees [7]. These deficits are even greater in bilateral amputees[8]. This has a tremendous impact on what amputees can achieve and the consequences for their quality of life.
    The energy storage and return capabilities of prostheses are crucial to improving the situation and yet modern prostheses only store and return significant energy below the knee, and energy is not returned in a controlled manner. For example, stored energy is not available for plantar-flexion (push-off) at the end of stance. Furthermore, modern prosthetic systems donít transfer energy between joints, which is a lost opportunity as, for example, the excess of eccentric work at the knee could be stored and used in a controlled manner at other joints.
    Research problem

    As mentioned above, modern prostheses only store and return significant energy below the knee, and energy is not returned in a controlled manner. For these reasons, we believe there is an opportunity for truly transformative research leading to a step change in the performance of lower limb prostheses.
    This requires advances in:

    • Mechanical design to provide flexible energy storage and return
    • Intelligent adaptive control to match the behaviour of the prosthesis to the terrain (e.g. ramp gradient, step dimensions etc.) and gait characteristics such as speed, stride length, double-stance period etc.

    The first aspect is being tackled in a current EPSRC funded project involving the Universities of Salford and Manchester. The PhD proposed here will tackle the second aspect (intelligent adaptive control) and will complement the work being undertaken in the EPSRC project.
    Approach

    The aim will be to use sensors integrated into the prosthesis to derive measures describing the terrain being crossed (e.g. ramp gradient, step dimensions etc.) and also various gait characteristics such as speed, stride length, double-stance period etc. To achieve this, powerful regression algorithms will be used to map the raw sensor signals onto the required terrain and gait variables. To improve the performance of the regression algorithms and increase their resilience to noise and gait variability, fast signal pre-processing techniques will be employed prior to the regression stage.
    The terrain and gait measures will then be used to adapt the way that the prosthesis stores and returns energy and, hence, to optimise amputee performance. The adaptive control rules will be obtained by undertaking simulation-based optimisation studies. This will involve the mathematical modelling and simulation of amputee gait and prosthesis dynamics. A combination of multi-body dynamics and empirical muscle models will be used to represent the musculoskeletal system. The prosthesis will be modelled as a combination of passive viscoelastic elements and active energy storage and return elements (e.g. miniature hydraulic systems).
    References


    1. Chin T et al, 2006, Am J Phys Med & Rehab.
    2. Schmalz T et al, 2002, Gait & Posture.
    3. Seymour et al, 2007, P&O Int.
    4. Chin T et al, 2009, Am J Phys Med & Rehab.
    5. Starholm et al, 2010, P&O Int.
    6. Walters et al, 1999, Gait & Posture.
    7. Fortington et al, 2012, J Am Med Dir Assoc.
    8. Perry J et al, 2004, Arch Phys Med & Rehab.


    For informal queries please e-mail Professor David Howard (d.howard@salford.ac.uk)

    Applications should be made via the Universityís online application system: http://www.salford.ac.uk/study/postg...g-for-research All candidates must submit a research proposal with their application.






    Studentship 2: Adaptive control of functional electrical stimulation(FES) (http://www.salford.ac.uk/study/postgraduate/fees-and-funding/funded-phd-studentship/school-of-health-sciences)

    Supervisory team: Prof Laurence Kenney; Prof David Howard; Ms Christine Smith; Dr Mingxu Sun

    A home-based FES system for upper limb rehabilitation could benefit many stroke survivors who currently have little or no hand and arm function on their affected side. The challenge for the future is to enable home use without supervision by a therapist, which would require a system that could, to some extent, replace the therapistís role of: monitoring the patientís short term progress; adapting the exercise regime accordingly; and providing the patient with real-time feedback on their performance. This is particularly challenging and will need further research to understand how best to apply the necessary signal processing, machine learning, and adaptive control methods.
    Context
    Stroke affects approximately 150,000 people in the UK each year, leaving over 300,000 people living with moderate to severe disabilities. After stroke, many people cannot use their affected arm, and this has considerable impact on their quality of life. A major problem for stroke rehabilitation is the limited availability of physiotherapists. Therefore, home-based rehabilitation systems that donít require the presence of a therapist are needed to give the best chance of recovery through motor relearning.
    Functional Electrical Stimulation (FES) of muscles is a low cost solution which directly activates paralysed muscles through electrical stimulation via skin-surface electrodes. It has great potential as a stroke rehabilitation tool [1-4], and can even help patients with severe hand arm paralysis [5]. In contrast to traditional physiotherapy, FES provides a means of directly tapping into the nervous system, actively producing muscle contraction and movement, exciting many of the associated neural pathways. If this is synchronised with the patientís efforts to carry out meaningful tasks, it provides afferent inputs associated with the intention to create functional movement. This provides the most appropriate set of neural inputs to promote learning [2, 3]. Although recent studies have reported significant success [5, 6], the problem with existing FES systems is that they require specialist skills to set up, particularly for upper limb rehabilitation, and therefore require clinical engineering involvement for each patient, negating the potential benefits mentioned above.
    Research problem
    An Advanced FES Rehabilitation Tool has been conceived to create a user-friendly system that enables physiotherapists with no software skills to quickly and easily set up functional electrical stimulation (FES) supported reaching tasks, together with the corresponding bespoke FES controllers. These will be individually tailored to suit each patient and used in their therapy sessions to promote recovery of arm and hand function. This work is ongoing in a Salford-led, Department of Health (i4i) funded project (http://www.seek.salford.ac.uk/data/p...7397&version=1). This will lead to a commercial prototype for a clinic based system which will enable stroke physiotherapists to look after several patients simultaneously, thus increasing therapy time without increasing the burden on therapists.
    However the real challenge for the future is to enable home use without supervision by a therapist. So a home-based FES Rehabilitation Tool would, to some extent, have to replace the therapistís role of:
    a) Intelligently monitoring the patientís short term progress and providing them with real-time feedback on their performance;
    b) Adapting the FES control as the patient improves or fatigues.
    This is particularly challenging and will need further research to understand how best to apply the necessary signal processing, machine learning, and adaptive control methods.
    Approach
    Intelligent monitoring of patient task performance
    The aim will be to use body-worn sensors during FES-supported therapy sessions to derive measures describing task performance, including movement deficiencies (poor coordination), speed of task execution, smoothness of movement, and movement variability between task repetitions (which has been shown to be a good measure of the quality of motor control). These measures will then be used for real-time biofeedback purposes, providing the patient with information on their task performance during the therapy session and, hence, to some extent replacing the therapistís role of providing feedback.
    To achieve this, powerful regression algorithms will be used to map the body-worn sensor signals onto the aforementioned variables of interest. To allow the regression algorithms to operate more successfully, be resilient to noise and variability, and adapt to different patients, fast signal pre-processing techniques will be employed prior to the regression stage.
    Adaptive control of FES support so that patients are always being challenged
    As the patient improves, the FES controller should adapt in real-time so that the patient must still strive to achieve the task goals. At present FES control parameters are adjusted manually, by trial and error, and it is unclear how this can be formalised so that it can be automated. Therefore, machine leaning techniques will be applied that will use patient task performance information to make in-session patient-specific decisions on FES control parameter adjustment and, hence, to some extent replaces the therapistís role of adapting the exercise regime.
    To achieve this two approaches will be investigated:
    i. Developing a rule-based system based on the results of knowledge elicitation from a group of FES specialists.
    ii. Using machine learning methods that can automatically learn from FES specialists.
    The latter has the advantage that automated learning could continue as the FES Rehabilitation Tool is being used in clinical settings by FES trained physiotherapists.
    The proposed expert system will be based on one or more of the following decision making approaches: neuro-fuzzy, statistical decision trees, or Bayesian networks. It will combine the available expert knowledge and initial patient model (FES setup and performance) to track and adapt to the patientís changing task performance. To make the system robust and capable of refining the underlying expert knowledge over time, global optimisation and real-time signal processing methods will be applied to make the relationships between sensor data and FES control parameters more direct, sensitive and straight-forward to calibrate.
    References

    1. Alon, G., A.F. Levitt, and P.A. McCarthy, Functional electrical stimulation enhancement of upper extremity functional recovery during stroke rehabilitation: a pilot study. Neurorehabil Neural Repair, 2007. 21(3): p. 207-15.
    2. Sheffler, L.R. and J. Chae, Neuromuscular electrical stimulation in neurorehabilitation. Muscle Nerve, 2007. 35(5): p. 562-90.
    3. Rushton, D.N., Functional Electrical Stimulation and rehabilitation-an hypothesis. Med Eng Phys, 2003. 25(1): p. 75-78.
    4. Spiegel, J., et al., Functional MRI of human primary somatosensory and motor cortex during median nerve stimulation. Clin Neurophysiol, 1999. 110(1): p. 47-52.
    5. Popovic, M.R., et al., Neuroprosthesis for retraining reaching and grasping functions in severe hemiplegic patients. Neuromodulation, 2005. 8(1): p. 58-72.
    6. Mann, G., R. Lane, and P. Taylor, A study to assess the feasibility of accelerometer triggered electrical stimulation on recovery of upper limb function in chronic stroke patients, in UK Stroke Forum. 2007: Harrogate, UK.

    For informal queries please e-mail Professor Laurence Kenney (l.p.j.kenney@salford.ac.uk) or Professor David Howard (d.howard@salford.ac.uk)

    Applications should be made via the Universityís online application system: http://www.salford.ac.uk/study/postg...g-for-research All candidates must submit a research proposal with their application.
Working...
X