Prof Paul Stewart and Prof Jill Stewart: University of Derby
Prof Fran Game: Royal Derby Hospital
Following open competition for call-for-proposals after a grand challenge workshop in January 2018, Cyclops Healthcare Network, funded by EPSRC network grant (reference EP/N026985/1), has approved an award for our feasibility study project entitled
“Smart Active Footbed for Wound Prevention and Management”.
|Background: There are around 3.5 million people living with diabetes in the UK . Data suggests that around 15% of patients will experience foot ulcers  at some stage in their lives with 15 – 20% of these incidents resulting in lower limb amputation . Diabetic foot disease costs the NHS approximately £1 billion/annum . Current preventative management includes the provision of insoles which are often “off the shelf” or moulded on anatomy, which allow for no fluctuations in pressure or foot shape.
Aims: The aim of this work is to prototype a smart, active footbed that can decrease the frequency of wound development by early detection of potential risk factors for ulceration and alerting the patient to seek help when necessary. By embedding pressure/ force/ temperature sensors on the plantar aspect of the footbed, ulcer pressure points can be detected, and temperature monitoring can predict imminent skin breakdown . Some ‘smart’ materials possess an ability to change their shape and recent work has demonstrated that these materials can be manufactured with embedded sensors and actuators . A Genetic Algorithm driven [8-10] adaptive neural-network model-based closed loop controller will be developed to control the shape-change of the footbed to deliver optimal force distribution via an iterative learning algorithm. Finally, the feasibility of emerging pathology management can be assessed.
To evaluate potential sensor, actuator and material technologies to develop a proof of concept demonstrator test rig
To establish what data is available, necessary and sufficient to diagnose an emerging problem in combination with a model of the patient’s foot.
To integrate the test rig and model into a closed-loop intelligent controller to demonstrate that changing the shape of the smart material footbed can manage an injury, and collected data can indicate that referral to a specialist is required.
The algorithms, control systems and hardware developed in this project will demonstrate the feasibility of monitoring, prognostics and diagnostics in an intelligent footbed, with shape adaptation dependent on pressure sensing.
Current preventative management relies on regular podiatry monitoring. Preventative footbeds are supplied to high-risk patients but their makeup is usually shape- rather than pressure-derived. The proposed feasibility study will, for the first time integrate real-time monitoring and management of diabetic patients’ feet by multi-sensor array. This array will be closed-loop coupled to an adaptive intelligent footbed via a model-based controller supported by an AI based learning algorithm. The combination of sensors and smart materials will perform two functions; firstly, alerting patients and clinicians to areas of skin under threat of breakdown, and secondly adapting the shape-profile of the footbed to redistribute pressure and minimise wound development.
Significant advancements expected from this feasibility study are:
- Persistent monitoring of an at risk diabetic foot
- Adaptive physical structure of orthotic devices to reduce wound development.
- Early warning to patient and clinic of imminent skin breakdown.
|This feasibility study consists of 3 concurrent work packages:
1. Sensor/Actuator/Smart Materials footbed integration
This work package consists of the development of a proof of concept active footbed demonstrator test rig which integrates a matrix of temperature and force sensors with active materials. This includes Dielectric Elastomers (capable of producing large strains) and Temperature Responsive Polymers, into a series of prototype active orthotic footbeds. Utilising IISE (Institute for Innovation – University of Derby) comprehensive Additive Layer Manufacturing, High Accuracy Metrology, and Mechanical Testing facilities, we will carry out an assessment of the resulting footbeds to identify the most promising combination of technologies.
Investigators: Prof Paul Stewart, Prof Jill Stewart, ECR
2. Temperature, Force and Gait Modelling
Utilising IISE design and analysis software and high-performance computing capability, physiological/expert system models will be combined with data from the force and temperature sensor matrix developed in (1), acquired via National Instrument data acquisition devices and integrated into a combined MATLAB system analysis model. The model will demonstrate the feasibility of the real-time measurement and analysis platform, incorporating specialist knowledge and data from Prof Frances Game.
Investigators: Prof Jill Stewart, Prof Frances Game, ECR
3. Closed Loop Intelligent Control
This work package involves the design of a closed loop multivariable controller with inputs from sensor part (1) and modelling part (2) via a model-based adaptive controller [8-10] derived by neural-network learned modelling of the sensor-wound/gait analysis-actuator hyper-surface.
Significant amounts of local processing will be required to perform intelligent data analysis and reduce the bandwidth of data sent externally. A sub-work package will prototype a hardware package containing Bluetooth communications capability and local processing. Depending upon the outcome of the feasibility programme in terms of processing power, this will either be on Arduino /Raspberry Pi  platform, NVIDIA Jetson TX1  as appropriate for lab-based testing.
Investigators: Prof Paul Stewart, Prof Jill Stewart, ECR
There are 7 substantive demonstration outputs connected with this feasibility study:
- Basic, active footbed prototype demonstrator
Constructed via additive layer manufacturing techniques, integrating sensor and actuation materials and arrays with local signal and control processing and communication capability.
- Demonstrate data acquisition
Wifi communications, and pre-processing via on-board device embedded within the footbed. Full local processing will, by necessity, form part of a subsequent project.
- Demonstrate data analysis
Developed model reference based analysis and integration to physiological and expert medical analytics to give pressure and temperature distribution maps and analysis of active adaptation.
- Demonstrate smart materials application
A laboratory demonstration of the capability of the active elements within the footbed.
- Demonstrate modelling, simulation and analysis
AI analysis via Neural Network model which integrates data, physical analysis and predictive model. This is likely to be at prototype level, taking data from the embedded sensor network, analysing it, and passing it through a neural network model reference, and generating commands for tracking controllers in the footbed, and simple decision support information to be sent to patient and clinic.
- Demonstrate prototype controller
Simulated physical events which emulate wound development through heat and/or pressure distribution will be applied to the developed footbed. The control system will generate appropriate decision support output, and where appropriate will deploy active materials in the footbed to apply compensation force.
- Demonstrate foundation for further funding
A road-map for future research and development will be defined.