Definition of the second scenario for the clinical evaluations

1. Part-financed by the European Regional Development Fund Definition of the second scenariofor the clinical evaluations Annemarie Kokosy and Matthew Pepper ISEN / EKHUFT April 24th, 2012 University of Essex

2. Plan of presentation I. Why we must define and agree the scenario? II. Project plan III. Clinical goals IV. Future challenges and the work to be done V. Main goals

3. Why we must define and agree the scenario? • For a collaborative cross border work, it is necessary to have a common objective • For a more effective work accelerating the implementation of the algorithms on the powered wheelchair from ISEN and University of Kent, we must define the inputs and outputs of each partner • We must take into account the user’s needs expressed through the European survey (end date: April 30th for the French side) • We must work in collaboration with the medical staff from EKHUFT and GHICL

4. Why we must define and agree the scenario? • At the end of the project we must have implemented on the powered wheelchair from ISEN or University of Kent the results from each activity

5. Project plan – Time line End 31Dec/2013 Feb/2012 Aug/2012 Feb/2013 Aug/2013 4/12 6/12 10/12 12/12 4/13 6/13 10/13 12/13 1st scenario (collision avoidance) Technical work Clinical trials April 24th, 2012 2nd scenario (assisted navigation in semi-autonomous way) Technical work Clinical trials April 24th, 2012 Ethics Approval 3rd scenario (autonomous navigation) Technical work April 24th, 2012 Clinical trials Ethics Approval 5

6. Clinical goals • Second scenario – Assisted/Semi-autonomous Navigation • First evaluation: environment without obstacles - pass along a corridor and through a doorway and stop in front of a table/desk (user in the loop); • Second evaluation: environment with obstacles - pass along a corridor and through a doorway and stop in front of a table/desk (user in the loop); • Evaluations at University of Kent and EKHUFT with healthy people • Clinical evaluations at GHICL The project will be a success if this is achieved 6

7. Future challenges and the work to be done Step 1: Develop the hardware and software for the detection of the doors (minimal dimension of the door is 90cm) and tables • ISEN: system with IR and US (works only if the wheelchair is in front of door) • UoK - Michel: will try to detect by using a camera (to be developed) Step 2: Two switches: one to turn on/off the interface device (called SYSIASS) and the second to turn on/off the wheelchair 7

8. Future challenges and the work to be done Step 3: Develop software for assisted navigation – along a corridor and through a doorway • ISEN: potential field (in progress) • UoK - Michel: “localised potential field” (to be developed) Step 4: Develop software to store, analyse and present that data. • Data to store: what the user doing, what the device doing, sensors, time • What user doing: direction and amplitude • What the device doing: modified amplitude, direction • All sensors: to be done • Time: to be done 8

9. Future challenges and the work to be done Step 5: Visual feedback (how the device is intervening) or voice or vibrations • Jean Marc: visual feedback and vibration • Michel: the voice approach (especially the sentences) if he wants Step 6: Secure Data Transmition (John, Gareth, Jenya, Klaus) Step 7: Drive the wheelchair using the multi-modal HMI (Huosheng, Lei Wei, Dongbing Gu, Theo and Ericka Rechy-Ramirez) 9

10. Future challenges and the work to be done • Work within the requirements of the Medical Devices Directive - MDD 93/42/EEC [rev 2007] • Get permission from University of Kent for evaluations with healthy people • Clinical evaluations at GHICL (the Hospital of Garches has not funds) • Sign a licence agreement withDynamicControl • Wheelchair Users Questionnaire for the English side 10

11. Future challenges and the work to be done • Project strategy (decision on March 29th, 2012 – University of Kent) Joystick CAN-DX BUS DynamicControl electronics Analog signals Joystick board Will be replaced by Serial GPSB Module from DynamicControl Digital signals Wheelchair “intelligence” board Sensors 11

12. Future challenges and the work to be done • Project strategy (decision on March 29th, 2012 – University of Kent) CAN-DX BUS Joystick DynamicControl electronics Serial GPSB Module Sensors board Sensors Wheelchair “intelligence board” (PC or mother board)

13. Future challenges and the work to be done • Project strategy (decision on March 29th, 2012 – University of Kent) CAN-DX BUS Multi-modal HMI DynamicControl electronics Serial GPSB Module Sensors board Sensors Wheelchair “intelligence board” (PC or mother board)

14. Main goals Activity 1 • Localization and perception •  using monocular camera/Kinect/odometers/US/IR • System modeling and identification •  using the dynamic model of powered wheelchair • Navigation •  using A*/potential field/nonlinear dynamic optimization • Robust control •  using sliding mode

15. Main goals Activity 2 • Development of ICMetrics boards •  using PIC or Arduino & ARM • Development of an experimental platform for feature collection and analysis • Design of a flexible and secure channel •  using Shamir secret sharing technology in conjunction with the ICmetric system Activity 3 • Development of a multi-modality HMI •  using facial muscle signals (EMG) and face image information

16. Main goals Activity 3: powered wheelchair prototype  clinical trials of the first scenario at the Hospital of Garches  second powered wheelchair equipped with US/IR/Arduino at ISEN  third powered wheelchair equipped with PC/IR/US at University of Kent  own powered wheelchair equipped with PC/camera/sensors at University of Essex