Evaluation of Interactive Systems. Human-body Measurement devices (adapted from Halla Olafsdottir s class)

Transcription

1 Evaluation of Interactive Systems Human-body Measurement devices (adapted from Halla Olafsdottir s class) Caroline Appert - 8/9

2 Human-body measurements Devices Skeletal Muscles EMG = ELECTROMYOGRAPHY Heart EKG/ECG = ELECTROCARDIOGRAM Heart Rate Blood Pressure Brain EEG = ELECTROENCEPHALOGRAM (electrical activity) fnir = Functional near-infrared spectroscopy (blood composition) Skin Skin Conductance

3 Skeletal muscles ElectroMyoGraphy (EMG) Electrical activity of skeletal muscles recorded Electrode placement very important for a good recording Surface Intramuscular (needle) Measures can indicate stress, physical and mental. 3

4 EMG Device example: MyoBand 4

5 Heart ElectroCardioGram (EKG/ECG) Heart s electrical activity Heart Rate Beats/min Blood Pressure Pressure in the arteries When heart is pumping & filling (/9) Measures that can indicate physical and mental stress. 5

6 Brain ElectroEncephaloGram (EEG) Measures brain s spontaneous electrical activity Can indicate Abnormal brain activity Alertness (sleepy vs. alert) Stages of sleep. 6

7 EEG Device Example: MindSet neurosky.com 7

8 Brain functional Near InfraRed Spectroscopy (fnir) Measures oxygen levels changes in the prefrontal cortex Used to estimate amount of brain activity for example during cognitive tasks biopac.com 8

9 Skin Measure electrical conductance of the skin which varies with moisture level (galvanic skin response - GSR) Sweat is controlled by the sympathetic nervous system (reflex in case of harmful event) Used for stress estimation! External factors such as temperature and humidity affect GSR measurements 9

10 Kinematics Kinematics is the science that describes movements with respect to time and space Measures: Position, displacement, velocity, acceleration

11 Kinematics - references Body Planes Axes of Motion translations along the x, y, z axes rotations along the x, y, z axes Every movement can be described by those 6 degrees

12 Acceleration An accelerometer measures the acceleration it experiences relative to free fall (translational movements) Can measure angle relative to earth and direction of movement Can be uniaxial, axial, 3 axial

13 Orientation A gyroscope measures relative changes in orientation (rotational movements) 3

14 Joint angles A goniometer measures an angle or allows an object to be rotated to a precise angular position Can be Manual, Electric Uniaxial, -axial 4

15 D Positioning Photogrammetric & Cinematography (Muybridge 88 s) 5

16 3D Positioning Optical systems utilize data captured from image sensors to triangulate the 3D position of a marker Markers can active (Diode s/led s), or passive (retroreflective) rotational information must be inferred from the relative orientation of three or more markers 6

17 3D Positioning Microsoft Kinect

18 3D Positioning Electromagnetic-based systems calculate position and orientation by the relative magnetic flux of three orthogonal coils on both the transmitter and each receiver (6DOF) potential interferences with magnetic and electrical objects in the environment 8

19 3D Positioning Inertial systems Combo of Accelerometers (linear acceleration), Gyroscopes (rotation) and Magnetometers (direction of the earths magnetic field) Mobile systems 9 xsens.com

20 3D Positioning Mechanical Systems Exoskeleton equipped with Goniometers metamotion.co m

21 EYE TRACKERS Eye movements Saccade (fast eye movement to move focus to object) Smooth pursuit (following moving objects) Fixation (movement stops and eye acquires content) We perceive world during fixations

22 EYE TRACKERS Infrared light illuminates eyes & sensor/ camera captures the reflection of the eyes. Record: Movement & fixations Types: Head mounted & External

23 Kinematics and evaluation e.g., analyze physical navigation when interacting with a very large display Figure. Pair behaviors with different collaboration styles. (a) Participants working on separate discs. (b) One participant shows his partner the target container while moving his own disc. (c) Under CloseComm condition, the participant on the right is telling his partner move this one there. Layout Locality are tightly bound to the articulatory level of collaboration. By (c) the decision-making process choosing an abstract class where is minimal, we limit the higher-level cognitive processes such as discussing or arguing about a choice. Yet, at the interaction level, the task is representative of a large class of real-world CloseTech Local Time as those outlined Time situations such in the4previous section Each user can spell out the label of the disc they have picked (a) user may indicate a destination container (b) if so that the other they see or remember one. With drop-for-partner, instead of indicating the destination container, the other user can finish the pick-and-drop action on behalf of the first user by dropclosecomm Local Time ping her disc,divi&conq Local e.g. with a4 dedicated button on his device. In other words, a user can drop the disc picked by her partner (e) at the location of the user s cursor. (f) Liu et al. [3] controlled information density by the size of the (d) is a simple shared interaction technique that Drop-for-partner textual labels. We use the smallest size ( pt) corresponding requires minimal technical support and that can be used in to the higher information density. Participants must be close CloseComm Distant CloseTech Distant Time Time a variety of tasks. The purpose is to compare collaboration 4 to the 6 8 and 4 the labels display to read must 6move8 around to (h)are further away than the adjawith shared interaction support vs. without it, i.e. with verbal read the labels of discs that Divi&Conq Distant Time or gestural communication only. Both cases require the cent columns (Figure ). Liu et al. [3] operationalized the (g)users to coordinate and synchronize their actions, so the benefit of difficulty of the task by using different numbers of categories saving time by having the other user look for a container or ( for Easy tasks and 4 for Hard tasks). We use 8 categories move the disc to its destination may be offset by theloosecomm Distant coorditime a task Time which, combined with theloosetech Distant smalllabel4size, 6results in nation overhead. hard enough that collaboration is likely to be beneficial. Figure 7. Navigation paths of eight pairs in different STYLE LAYOUT conditions. LooseComm-Local and LooseTech-Local are omitted as they are very similarthe to Divide&Conquer-Local (top-left Loose graph). Each the movement of both participants in a bird s eye view of the wall Crossing above two dimensions, vs. graph Closeshows: Col- on the left, In order operationalize distribution of information room with the wall on the left (unit is meter); on the right, the same paths stretched over to a normalized timelinethe (x-axis) to help understand the pair s on laboration and Communication Only vs. Shared Interaction, the wall, we use types of layouts, Local and Distant, by navigation patterns. In addition, picks ( ), drops ( ) and drop-for-partner () actions are plotted on two the paths. leads to four collaboration styles. We add a baseline condicontrolling the distance of misclassified discs to the closest tion, Divide&Conquer, to contrast with the other four explicit appropriate With Local layouts, all pick-and-drops Can Liu, Olivier Chapuis, Michel Beaudouin-Lafon & Eric Lecolinet. Shared Interaction oncontainer. atight Wall-Sized Display a Data Task. In CHI '6: niquebeenables collaboration a distancein when the datamanipulation collaborative styles: Local Distant can done between adjacent at containers: the distance beto systems, be manipulated is scattered (Figure 7-f).May 6. Proceedings of the 34th international conference on Human factors in computing 75-86, ACM, tween a pick and a drop is short enough that users only need Divide&Conquer: the task is performed in parallel and the 5 ce (m)

24 Kinematics and evaluation e.g., analyze visual navigation when interacting with a graphical interface Figure 6. Panopticon (top) and VideoBoard (bottom) heat maps for surveillance video, with key areas (red boxes) and mouse clicks (red dots) shown. Figure 7. Panopticon (top) and VideoBoard (bottom) heat maps for edited video, with key areas (red boxes) and mouse clicks (red dots) shown. Dan Jackson, James Nicholson, Gerrit Stoeckigt, Rebecca Wrobel, Anja Thieme, and Patrick Olivier. 3. Panopticon: a parallel video overview system. UIST '3. ACM,

25 Kinetic Measurement Devices Kinetics is the science that examines the forces that produce the movement and result from the movement Variables: Forces & moments 5

26 KINETICS Force transducers Force measuring device that gives an electrical signal proportional to the force applied to it. axis & 3 axis 6

27 KINETICS Force Plates 4 corner type: 4 tri-axial transducers mounted at each corner Central support type force plate 7

28 Human-body measurements Sensors RAW signal is usually very noisy Valuable Information in a practically useless format e.g., EMG Is negative or positive Can not be used to quantify force Can not be used for comparison Needs to be processed Rectified Filtered Normalized 8