|Work package No.||Work package title||Type of activity||Lead Participant Short name||Lead Participant no.||Person- months||Start month||End month|
|Task 1.1 Overall Project Coordination (OBHP (8), M01 - M48):
This task includes management of all administrative issues and monitoring of the overall project progress (including the timely achievement of deliverables and milestones and proper assessment of project risks). It also involves the coordination of the contractual reporting, organization and chairing of the GA and PB meetings and project-wide meetings (including yearly review meeting with EC).
Task 1.2 Scientific Project Coordination (OBHP (4.5), OBG (1), UMG-GOE (1), M01 - M48):
This task includes overall scientific coordination of the WPs and technical activities of the project (ensuring project’s intermediate and final results are produced timely with an adequate quality level) and decision-making regarding the overall policy and technical strategy of the project (approving work plan and financial plan for further periods, high-level conflict resolution, knowledge management, IPR and ethical issues, participation of new partners, ensuring promotion of gender equity). Coordination will be led by the Project Coordinator and the Scientific Coordinator and involves the General Assembly (GA), the Project Board (PB) and Work Package Leaders (WPLs).
Task 1.3 Work package Management (OBHP (4.5), OBG (1), M01 - M48):
This task involves monitoring of the WP progress, including timely achievement of deliverables and milestones, assessment of WP risks, coordination of WP reporting, contribution to PB and GA meetings, organisation of technical and WP meetings and contribution to project-wide meetings (including yearly review meeting with EC).
Task 1.4 Quality Assessment (OBHP (4.5), M01 - M48):
This activity regards quality assurance and assessment on the fulfilment of the objectives of the project, including: risk management, comprising of risk identification, evaluation, follow-up and contingency plans and also includes progress measurement against the estimated baseline as regards fulfilment of the proposed objectives according to the quantitative and qualitative indicators of the project.
Task 1.5 Communication Management (OBHP (4.5), M01 - M48):
This task comprises of structuring, coordination and stimulation effective communication between the project partners and with the European Commission. It will involve the installation and maintenance of an external and internal project web site, mailing groups and document repository based on a collaborative online platform.
|WP 2||Dissemination, Communication and Exploitation||IA||OBHP||P1||28||M01||M48|
|Task 2.1 Dissemination (OBHP (3), UMG-GOE (2), SUPSI-IDSIA (2), UMCG (0.5), M01 - M48):
Together with the partners, OBHP will investigate the possible intellectual property rights within this project. A patent review will be performed periodically and possible new patent applications will be investigated, identified and if suitable filed together with the partners. We envisage that mutual agreements about the created intellectual property will be set up between different partners, as will also be detailed in the Consortium Agreement. Once patent protection has been looked after and the consortium agrees upon dissemination of obtained results, publishing will be done through several activities that will be accurately planned and targeted to the different stakeholders.
Project identity: The development of a common public identity for all public communication, including logo, presentation template and a general information brochure.
Website: Preparation, launch and continuous update of the project Website, which will be fed by all partners with contents and news about the progress of work including the motivation, objectives and methodology.
Scientific dissemination of the results: Suitable conferences and meetings mainly within Europe but also worldwide will be identified along with the most appropriate journals to publish results.
Conventional communication channels: Contacts with the media (press, TV) will be organized to raise awareness of the civil society and inform the wide public about the progressive achievements of the project and the benefits it can take at societal level for amputees. Patients’ association will be mobilized in this regard.
Social networks: an extensive presence of the project in the most known social networks (Facebook, LinkedIn, Twitter) will allow to reach out the widest possible audience, including those stakeholders that are not directly involved as patients, relatives, carers, researchers, developers, etc. For instance, indirectly, they will raise attention and curiosity of young people and get them closer to engineering and scientific academic careers.
Task 2.2 Communication (OBHP (3), M01 - M48):
Communication within the project will be based on video conferences, email, phone calls, collaborative online document repositories and personal meetings. OBHP will mediate the communication among project partners via these media and provide the necessary infrastructure (document server, mailing list) in addition to the communication with external
parties as described in Task 2.1. Furthermore, as project coordinator, OBHP will also maintain communication with the European Commission.
Task 2.3 Exploitation (OBHP (17.5), M01 - M48):
In order to identify the project exploitable results and ensure their future exploitation, the following activities will be carried out:
Elaboration of an exploitation plan
Coordination and implementation of the exploitation plan
Regular revision of project development to identify exploitable results
Protection of results (see Task 2.1)
These activities will be running for the full duration of the project. The exploitation plan will be reviewed and updated every year. Throughout the project, OBHP will assist all partners (but especially technical partners) in documenting the development steps according to the guidelines and processes used at OBHP. Aim of this step is to ease and support later development of a commercial product. Depending on the results and together with the partners, OBHP will progress the business plan throughout the project. This will strongly take into account the results and experiences made by the clinical partners. The clinical work carried out in this project will form a firm basis for an informed plan of exploitation. Both dissemination and exploitation activities performed during the project will be reported in the plan for use and
dissemination of foreground to be delivered to the European Commission at the end of the project.
|WP 3||System Integration||IA||OBHP||P1||41||M01||M48|
|Task 3.1 System integration lead (OBHP (7), M01 - M48):
OBHP will channel all efforts made within the project into a clinically viable solution by:
Ensuring compatibility of all developed components throughout the entire project to guarantee clinically viable solutions in all sub-results and the final result of the project
Providing a platform for communication and interaction between all partners to direct efforts to a common and unified system throughout the entire project duration
Coordinating and integrating end user tests in the development to ensure clinical viability of the conceived systems
Task 3.2 Unified interface descriptions (OBHP (1), OBG (1), UMG-GOE (3), M01 - M06):
In order to allow successful interplay of all elements developed within this project, project internally standardized interfaces for hardware and software have to be defined and enforced. The enabling subtasks of OBHP are:
To provide discussion platforms and data sheet repositories for interface agreement, documentation and consultation
To motivate and monitor adherence to the agreed interfaces from the earliest stage on of the project
Together with all project partners, OBHP will mediate the definition of:
interfaces between data acquisition and signal processing packages
interfaces between signal processing packages and prosthetic hardware
Task 3.3 Clinical test coordination (OBHP (5), M01 - M48):
Since clinical applicability and achieving high TRL through tests in relevant environments are the main focuses and goals of this project, the devised solutions and prototypes will routinely be evaluated by the medical research partners (OSS, UMCG). The results of these tests will in turn influence the design progression of the developed socket materials and software (agile project management). OBHP will mediate and support the exchange between the clinical and technical
working groups by:
Providing a template for instructions of the developer to the clinical staff on how to use the tested system and targeted questions to invoke helpful suggestions.
Providing a template for the clinical staff to give informative feedback of the results to the developing group (laudation, problems, suggested solutions)
Task 3.4 Central prototype assembly (OBHP (14), OBG (7), M01 - M48):
Due to its geographical vicinity to the clinical partner OSS, its engineering knowledge and it orthopaedic technical facilities, OBHP is designated to cover physical assembly of testable prototypes for the clinical investigations as well as to provide electromechanical assistance throughout these tests. OBHP will be assisted by OBG, the only partner in the consortium developing new hardware for INPUT. OBHP will receive prototypes of signal acquisition and signal processing from other partners (OBG and SUPSI-IDSIA) and standard prosthetic hardware from OBHP itself and assemble these to stable, functioning prototypes for testing on end-users. In order to provide these services for the project, OBHP will
Ensure adherence to predefined interfaces by frequent checks and discussions with all partners, see Task 2.2
Organize hands-on workshops for all partners to increase system understanding
Provide the location, tools and personnel for prototype assembly
Support the clinical testing team during the experiments in:
Regular application: initializing, controlling, handling and operating the device
Troubleshooting: on-the-spot error correction in case of system failure as far as possible
Assist the clinical partner in providing technical feedback to the designing partners after testing for constructivefeedback to promote development progress
OBHP will also ensure that all necessary documentation for the tested prototypes is available to support and enable clinical testing with respect to regulation compliance and ethics.
Task 3.5 Providing test platform setup (OBHP (3), M01 - M06):
In order to enable rapid interface conformity testing on the spot for all partners, OBHP will provide each partner, as far as needed, with a device to allow testing from an early stage on. To this end, either full arm prostheses, hardware emulators or software emulators thereof will be developed if needed and distributed to the partners requiring such a device for testing prototype systems of signal acquisition and processing. The goal of this task is to facilitate easy transition from local developing to controlling a physical arm for clinical testing for each partner. OBHP will:
Assess the feasibility and practicability of providing each partner with a physical Otto Bock arm prosthesis or a physical emulator
In case of decision for full arm prostheses, OBHP will distribute these among the partners as loan devices.
In case of decision for a physical emulator, the device will be designed, manufactured and distributed to the partners by OBHP
In case of decision against a physical test device, a virtual arm software that emulates a physical arm prosthesis will be developed and provided
Provide each partner with an abstraction layer software that receives abstract prosthetic movement commands and translates them into direct control commands for the prosthesis or emulator
|WP 4||Signal Acquisition Hardware||IA||OBG||P2||82||M01||M48|
Task 4.1 Electronics for EMG recording (OBG (24), M01 - M27):
This task deals with the development of multichannel EMG amplifiers (circuit design, layout etc.) which are small enough to integrate them in a prosthetic system. One possibility could be that those amplifiers will be integrated in the distal end of a liner. It needs to be investigated how this could be done. The connection of the electronics will be a main topic here.
Another important work will be the development of shielding concepts, as the separation of electronics from electrodes could lead to a system which is more prone to electromagnetic interferences and cable movement artefacts. This includes noise sources such as power line noise, cell phones or electronic anti-theft devices. E.g. active shielding of the cables has been applied successfully in preliminary tests to reduce the influence of power line noise. This work is based on an already existing desktop amplifier which was developed to record EMG from dry electrodes in the EU funded project AMYO.
Task 4.2 Liner/Socket for multichannel EMG recording (OBG (22), UMG-GOE (2), M01 - M27):
A new liner/socket interface for EMG recording will be developed. Focus is on the combination of conductive and non- conductive materials. Experience showed that this is a critical point to record high quality EMG which is not prone to artefacts such as electrostatic noise
Conductive polymers to have a consistent liner/socket surface to the skin will be investigated. One key is to find a polymer which does not tend to movement artefacts
Moreover the likelihood of lift offs will be decreased by exploring special shape and/or flexibility of electrode material and new socket designs.
Task 4.3 Design validation and iterative development (OBG (34), M06 - M48):
All activities in this work package will be carried out according to an iterative, agile research and development process. This task was specifically introduced with a run time over almost the entire project duration to reflect this agile development, while keeping the other Tasks of this WP focused and concise. Already starting with the first design concepts and mock-ups we will integrate orthopaedic technicians and end-users (amputees) into the process in order to
understand their needs. The evaluations comprise laboratory as well as clinical tests and the results will continuously guide the development progression of the electrode liner/socket interface.
|WP 5||EMG Models||IA||UMG-GOE / ICL||P3||43||M01||M24|
|Task 5.1 Influence of electrode characteristics on robustness in myocontrol (UMG-GOE (11), OBG (2), SUPSI-IDSIA (2), M01 - M09):
The effect of the recoding system on the degree of variability of EMG features under non-ideal conditions (such as arm posture, electrode shift, etc.) has not been extensively investigated in the literature. With advanced biophysical EMG generation models, it will be possible to simulate electrode systems by varying the electrode size, spacing, location, and derivation (e.g., monopolar, bipolar). The validity of the simulations will then be tested, in a selected number of conditions, with experimental recordings. In this way, it will be possible to study the changes in EMG characteristics (e.g., power spectrum, amplitude, higher-order moments) for changes in influencing factors depending on the specific electrode system used. Specifically, it is expected that the redundancy of information introduced by a large number of electrodes may make the extraction of some EMG features more robust. For example, preliminary data indicate that EMG factorization is more robust to electrode shift when computed from a large number of electrodes rather than from the minimum number theoretically needed. This subtask will aim at finding a compromise between robustness and number of electrodes to provide clear and specific information to WP4 for the development of the multi-channel electrodes integrated into the socket.
Task 5.2 Modelling factors of influence on the surface EMG (UMG-GOE (11), SUPSI-IDSIA (4), M01 - M24):
Although it is well known that EMG is sensitive to many factors of influence, the way in which the characteristics of the EMG change with these factors has still to be clarified. In this work-package, a large set of simulations will be used to describe in detail the changes in EMG features with changes in electrode location, muscle fatigue, etc. The modelling work will be complemented by experiments that will cover some of the validation aspects. Finally, each factor of influence will be investigated with various electrode configurations, linking this subtask to the task 5.1. The main factors of influence that will be modeled are electrode shift, change in the spatial distribution of muscle activity, changes in electrode impedance, and muscle fatigue. Among these, experimental validation will focus only on the change in the spatial distribution of muscle activity and fatigue, which can be re-created in the lab
Task 5.3 Linearity in muscle activity for multi-DoF control (UMG-GOE (11), SUPSI-IDSIA (2), M01 - M18):
One of the main methods tested for prosthesis control will be regression of the EMG into forces or kinematics. For this purpose, a mapping is necessary and this is built usually on labelled data. However, it is not possible to calibrate/train the system on all possible combinations of DOFs across the full range of motions. To overcome this problem, the linear summation property in muscular activations will be tested extensively. The main rationale for such investigation is that single DOF training may be sufficient for multi-DOF control if the activations of each DOF sum linearly. This hypothesis is equivalent to the modularity or muscle synergy hypothesis. Moreover, within this linear framework, it is reasonable to expect that different postures can be compensated for by linear terms in the summation (this would be equivalent to identify a gravity-related synergy and some gravity-independent synergies). In addition to experiments in healthy volunteers as well as in amputees, motor control and EMG models will be used to interpret the experimental findings.
|WP 6||Machine Learning||IA||SUPSI-IDSIA||P5||91||M01||M48|
|Task 6.1 Integrated Software Suite (SUPSI-IDSIA (12), UMG-GOE (2), M01 - M08):
An integrated software analysis suite will be developed and compiled, containing implementations of state-of-the-art data analysis and machine learning methods, together with features for processing data in the formats generated by the project beneficiaries. During this stage, UMG-GOE will provide sample EMG data collected during previous projects (AMYO, MYOSENS) to test and evolve these algorithms and methods early on in the project, even before the finalization of the signal acquisition hardware (WP4).
Task 6.2 Algorithm Development (SUPSI-IDSIA (25), UMG-GOE (15), M01 - M36):
UMG-GOE will provide EMG data collected with prototype prostheses. SUPSI-IDSIA will investigate which methods can be used to extract the desired prosthesis control signals with a high level of robustness observing real-time constraints, using the software analysis suite developed in task 6.1. During this stage, the nature of the control commands to be extracted from EMG data is to be determined, based on the results of the initial machine learning experiments. In particular, UMG-GOE and SUPSI-IDSIA will practice tight collaboration for the challenging task of controlling multiple DOF while system training data will only be available from single DOF motions (WP5, Task 5.3).
The main technology investigated will be artificial neural networks with particular architectures, such as deep learning or LSTM networks. Advanced neural network learning methods have only rarely been applied to biosignal processing tasks, e.g. Nielsen et al. (2011) (see table 6) use neural networks, but with only one hidden layer. Using more complex networks for EMG signal processing is of highest interest, both practically and scientifically. In accordance with the goal of market-readiness of the prosthesis at the end of INPUT, the main work for this task will be done during the first two years of the project. Around 90% of the development work should be finished at the end of year 2. This would allow WP8 and WP9 to use the device in the clinical tests that commence at this stage of the project.
Task 6.3 Implementation on Portable Hardware (SUPSI-IDSIA (22), OBG (2), M09 - M40):
During this stage, suitable machine learning-based classification and regression methods, as determined in Task 6.2, will be implemented on a microcontroller integrated into a wearable prosthesis, and evaluated in a series of end-user tests (WP9). For local tests without end-users, OBHP will provide SUPSI-IDSIA with test hardware. For this implementation, the quality of the recognizer and the hardware constraints have to be carefully balanced, furthermore observations obtained during the user tests will be integrated into the development loop. We note that we expect even relatively complex neural network based processing (see task 6.2) to be feasible on a microcontroller, as soon as the training process has been completed.
As with task 6.2, the main work for this task is to be finished at the end of year two. The remainder of this task will consist of an integrated loop of user tests and algorithmic improvements, tightly integrated with task 6.4.
Task 6.4 Algorithm Life Cycle Management (SUPSI-IDSIA (12), UMG-GOE (1), M06 - M48):
This task deals with the continuous management of the machine learning methods employed in the INPUT project. In particular, necessary adjustments to the implementation in terms of speed, robustness, and user experience will be made, especially based on the results of the end-user testing (WP9). While this task runs over almost the entire duration of the project, the focus lies on the later project stages, when the core algorithm development work will have been done.
|WP 7||Evaluation Paradigm||IA||OSS||P4||58||M01||M48|
|Task 7.1 Devise Test Protocol (OSS (9), OBHP (0.5), UMG-GOE (2), UMCG (1), M01 - M18):
Defining several test protocols will be the main part in this work task.
setup a protocol for testing non-stationarities such as electrode shift
dual task tests
define functional ADL movements to compare with SHAP and the shelf test
Set of functional ADL movements
Task 7.2 Comparison tests between standard tests (SHAP, shelf test) and complex ADL test (OSS (8), UMCG (1), M08 - M22):
Well excepted tests like the SHAP and the shelf test describe the overall function of the patient with his prosthesis. A shorter execution time will describe a better functional use of the prosthetic limb. However, these tests do not describe in more detail how time differences were accomplished. Compensation movements which may have an adverse long term influence on the patient are not considered. The main goal of this work task will be to compare the outcome of the traditional tests to the performance in functional ADL movements, captured by a motion capture system. The correlation between the different tests and how compensation movements can be quantified will be elaborated.
Task 7.3 Testing amputees (OSS (31), OBHP (1), UMG-GOE (1), UMCG (1), M10 - M30):
Using a motion capture system to quantify and verify the movements of amputees by using the different test protocols.
Testing the influence of simultaneous and sequential control strategies on the quality of prosthesis utilization.
A comparative evaluation of the chosen tests between healthy subjects and amputees to quantify the motions which are possible and the necessary compensation movements.
Testing the robustness of the system using repeatable tests. Doffing/donning of the prosthesis, influence of additional weight etc. on the quality of movement.
Task 7.4 Support UMCG with clinical test setup (OSS (0.5), UMG-GOE (2), M20 - M48):
The main goal of this subtask will be to support the transfer of the final testing protocols to UMCG, help in setting up the test protocols and during the time of analysing the output.
|WP 8||Patient Training ||IA||UMCG||P6||55.5||M01||M48|
|Task 8.1 Developing computer-based rehabilitation game (UMCG (13.5), M01 - M12):
The game to be developed in this work package uses an avatar that is controlled by the same algorithms that control a prosthetic hand in a simultaneous and proportional way. The avatar will have as many DOF controllable as the prosthesis, although the avatar movements do not necessarily have to be the same as those of a prosthesis (e.g., a model airplane accelerating/decelerating, turning right and left and going up and down could reflect hand open/close, wrist pronation/ supination and wrist flexion/extension of a prosthesis). The avatar can move in all directions, because movement directions can be activated simultaneously. The proportional aspect of the control allows changing the movement speed of the avatar. Therefore, in such a way, the control of the prosthesis could be trained in a fun and motivating way. In this task, the game will be developed and implemented on a computer. The game will be optimized using behavioural experiments in an iterative process. During the iterative process decisions have to be made about with which game implementations to proceed. The decisions about the game that are taken further will be based on a test that contains a certain level of the game that has to be performed by people who have trained with the game and a control group. In this task the game will be developed with able-bodied users controlling the game with their myosignals.
Task 8.2 Establish which feedback to use for in-game learning for different types of users (UMCG (13.5), M13 - M24):
Users differ in their capacity of proportional myocontrol. WP9 establishes whether this difference in capacity also exists for users of simultaneous proportional control exploited in the prosthetic control device developed in INPUT. Task 9.1 develops a test for this purpose.
Participants in the experiments in this task first perform the test developed in Task 9.1 to classify the participants in high capacity and low capacity performers. Task 8.2 uses the game developed in Task 8.1 and varies different types of feedback in the game. It will be examined which type of feedback works best to support low capacity performers and which feedback supports high capacity performers best in order to improve their muscle signals and control quality. This part of the project will be executed with able-bodied and amputee participants.
Task 8.3 Establish whether transfer to daily life tasks is different for game-based learning than for training methods based on conventional training (UMCG (26.5), OSS (2), M24 - M48):
In this task it will be established whether game-based training of simultaneous proportional control has a better result than a training based on methods based on conventional training of prosthesis use. To test the difference between these two training methods, the test devised in Task 9.1 will be used. In cooperation with OSS we will devise additional tests and measurements to evaluate the differences between conventional training and training with the game developed in Tasks 8.1/8.2.
In this task two groups of experienced prosthetic users (using a prosthesis for more than 1 year, mean wearing time more than 4 hours a day), each group comprising 4 participants, are equipped with a prototype of the simultaneous prosthetic control device. Note that because these users have experience with myocontrol, they can start at a more advanced level of myocontrol. One of the groups receives a training with conventional performance feedback (“trial and error”). The other group receives training with the developed computer-game in which feedback is individualized and structured to help users to produce the appropriate myosignals.
At the beginning of the training, each patient will be tested with the test developed in Task 9.1, so that each participant can be scored on his/her level of capacity of simultaneous proportional control without training. Then participants receive the training. After the training participants perform the test of Task 9.1 again together with a series of tests performed in the lab, based on ADL tasks.
With this setup, three questions will be answered:
Is there an advantage for training with a game versus training with prosthetic functioning feedback on a screen?
Is the efficiency different for high capacity performers and low capacity performers of simultaneous proportional myocontrol?
Does game-based training with simultaneous proportional control transfer to the use of simultaneous proportional control in a daily living task performed in the lab?
The answers to these questions will be disseminated to the literature in scientific papers.
|WP 9||Clinical Testing||IA||UMCG||P6||66.5||M01||M48|
|Task 9.1 Developing a test of use of the simultaneous proportional prosthetic control device (UMCG (14), OBG (2), UMG-GOE (2), SUPSI-IDSIA (3), M01 - M24):
The test to be developed aims to test the performance of users with the simultaneous proportional prosthetic control device. For this test the algorithms of the simultaneous proportional prosthetic control device need to be ready. This algorithm will be provided by tasks 5.2 and 6.2. The hardware to measure myosignals and to connect to the computer program will be provided by tasks 3.5 and 4.1. Because both hardware and software already exist from the previous projects and unified interface descriptions (WP3), the development of the test can start before all the new equipment is finished. Therefore, this project can start at M01, together with Task 9.2.
The test to be developed here is inspired by earlier work done in the EU FP7 IAPP project AMYO (Grant No. 251555, 2011-2014) in a collaboration with OBHP and UMG-GOE (Ge et al, 2013). Ge et al developed a classification score of the ability to control the prosthesis in a traditional pattern recognition paradigm. The current project builds on this work by applying this method to the algorithms used to simultaneously and proportionally control a multi-DoF prosthesis. In the method, Item-Response Theory is applied to develop a score that reflects the ability to control a simultaneous proportional prosthetic control device. This score will be different for different participants, therefore, we will develop a test that enables us to score participants on different levels of usage of a simultaneous proportional prosthetic control device.
To perform the test, multi-channel EMG is recorded from participants and the signals are decoded using the algorithms developed in WP4, WP5, and WP6. The signals control a prosthetic hand and arm on the screen. The test will be composed of several stages. In stage 1, participants produce different grips of the hands and movements of the wrist. The time it takes to produce the posture will be measured. In stage 2, the selection of the grips is tested in a more functional way. During functional use of the prosthesis in daily life, the position and orientation of the prosthesis will depend on the task that is performed. This implies that the posture of the arm in which the control algorithm has to be robust needs to be taken into account in the test. Therefore, the selection of the grips and wrist orientations will be tested in different arm postures, that is, the arm hanging down, pointing horizontal, and pointing upwards in the sagittal and frontal plane. Stage 3 tests the selection of prosthetic postures in dynamic conditions. The same will be done as in stage 2, but now the arm will move between hanging down and pointing horizontal, pointing horizontal in the sagittal plane and pointing horizontal in the frontal plane, etc. In this stage the production of the prosthetic postures is tested during a movement.The robustness of prosthetic posture selection will be measured, and transformed into a score at each stage. Moreover, in a different part of the test the cognitive load of operating the device will be assessed to evaluate the extent to which the control of the new device is natural and intuitive. To this end, well known tasks such as counting backwards or reciting the alphabet in reversed order will have to be performed by the subjects while accomplishing prosthetic dexterity tests (Box&Blocks test, Clothes Pin Relocation test, SHAP test) or the test described in the above. The achieved performance will be compared to that achieved when the subject can dedicate full attention to the prosthetic task. A large sample of able-bodied participants (n = 50) will be trained extensively (exact duration will be based on piloting), in this test starting in Stage 1 and progressing to Stage 3 if individual skills allow. Based on these data a gold standard will be developed for ability to select a prosthetic posture simultaneously and proportionally. Based on these measurements, also a training time will be derived that participants need so that a stable score can be derived. Then four patients will be tested and it will be examined whether the test can distinguish skill level between good and bad performers.
Task 9.2 Develop an activity monitoring system to use in daily life situations (UMCG (13), OSS (4.5), M01 - M24):
Activity monitoring is getting more and more important to assess the contribution of assistive technology, such as prosthetic devices, to the patient’s quality of life. The current task develops an activity monitoring system to get insight in the use of the prosthesis in daily life (type of grips and motions, number of actuations) and the body movement used to handle the prosthesis during the day.
The monitoring of the prosthesis will be done by reading out the on-board microprocessors of the prosthesis. The posture of the arm during the handling of the prosthesis will be measured with body-worn sensors which are at UMCG’s disposal. To develop the system, able-bodied participants and prosthetic users will wear these sensors while performing ADL tasks in the lab. Using these data, algorithms will be developed to extract and classify the relevant behaviours and characteristics of the body angles used. In this work package, the analyses will be done always offline.
The activity monitoring system developed here is aimed to measure daily life behaviour in a lab situation. Because behaviour in daily life situations shows a lot of variability in behaviour and ranges of body postures usually not used in standardized lab-tasks. This makes that standard movement registration systems (opto-electrical systems like Optotrak or Vicon) cannot be applied here because markers get out of view of the cameras. Therefore, to measure behaviour in daily life situations we need to rely on alternative measurement devices. Note that the activity monitoring system as developed here should provide a basis to devise a system (e.g., downsizing, optimizing computational algorithms) that is capable to measure quality of behaviour in the end-user’s operational environment. As such the knowledge and insight developed here will set the stage for the following step to be taken in the future, which is measuring in real daily life. Note however, that although the developed prosthetic control device of INPUT will be used at home in Task 9.3, the activity monitoring device will only be used in the lab.
Task 9.3 Test advancement of simultaneous proportional control device over conventionally controlling algorithms (UMCG (27), OSS (1), M25 - M48):
In this task we will determine to what extent the performance with the newly developed simultaneous proportional prosthetic device advances over the conventionally controlled multi-DOF prostheses. Therefore, a group of 8 patients is examined. The experiment is started with conducting a semi-structured interview in which advantages, disadvantages of their current prosthesis and control method, and future wishes of the patients are inventoried. Each patient performs a set of daily life tasks in the lab. This set is developed in close collaboration between UMCG and OSS. Participants wear the activity monitoring system developed in 9.2 to measure the quality of the behaviour. Then each patient is tested with the test developed in task 9.1, to gauge capabilities of simultaneous, proportional prosthetic control. After that, patients are fitted a prototype version of the simultaneous proportional prosthetic control device using what can be provided at this stage by Tasks 4.2, 4.3, 5.3, 6.2, and 6.3. The patients will use the new device for two to three weeks in their daily life. During the wearing period short interviews will be conducted on a regular basis, inquiring after advantages and disadvantages of the provided system. After the try-out period, the tests and interviews that were performed with their own prosthesis are repeated with and for the new device. Moreover several tests from the literature will be used (Deijs et al, submitted; Van der Niet et al, 2010; 2014) to be able to compare performance with the new device with scores of conventional prosthetic control devices.
In the analyses, performance between conventional prosthetic control devices and simultaneous proportional prosthetic control devices will be examined. The focus will be on robustness of prosthetic postures during performance in daily life. Moreover, the quality of the movements, such as range of elbow and shoulder angles, and smoothness of angular trajectories will be assessed. Moreover, the focus is on how controlling the prosthesis affects the movements in the arm. Together, with these data the question can be answered whether the simultaneous proportional prosthetic control device allows for smoother control of the prosthesis and higher quality of movements in ADL use of a prosthesis. The results will give a good indication about the improvements made with the simultaneous proportional prosthetic control device developed in INPUT compared to other devices. The findings of this task will be fed back in the development of the new prosthetic control device WP4, WP5, and WP6. In this way, the last year of the project can be devoted to fine-tuning the hardware devices and the software algorithms.