Human Emotion Recognition System Using Smart Sensors Subhas Mukhopadhyay FIEEE, FIEE Distinguished Lecturer, IEEE Sensors Council

Transcription

1 Human Emotion Recognition System Using Smart Sensors Subhas Mukhopadhyay FIEEE, FIEE Distinguished Lecturer, IEEE Sensors Council

2 Sensor A sensoris a device that measures a physical quantity and converts it into a signal which can be read by an observer or by an instrument. A sensor generates an electrical signal related to a physical, biological or chemical parameter. A good sensor obeys the following rules: 1.Is sensitive to the measured property 2.Is insensitive to any other property 3. Does not influence the measured property

3 Many Sensors Measurand Transduction Principle Physical Properties Pressure Piezoresistive Temperature Humidity Flow Thermistor, thermo-mechanical, Thermocouple Resistive, capacitive Pressure change, thermistor Motion Properties Position E-mag, GPS, contact sensor Velocity Angular velocity Acceleration Contact Properties Strain Piezoresistive Force Torque Slip Vibration Doppler, Hall effect, optoelectronic Optical encoder Piezoresistive, piezoelectric, optical fiber Piezoelectric, piezoresistive Piezoresistive, optoelectronic Dual torque Piezoresistive, piezoelectric, optical fiber, sound, ultrasound

4 Presence Tactile/contact Contact switch, capacitive Proximity Hall effect, capacitive, magnetic, seismic, acoustic, RF Distance/range Motion E-mag (sonar, radar), magnetic, tunneling E-mag, IR, acoustic, seismic (vibration) Biochemical Biochemical agents Identification Personal features Personal ID Biochemical transduction Vision Fingerprints, retinal scan, voice, vision motion analysis

5 Sensor Output Analog Digital 4-20 ma current loop -Parallel(bytes, words with +- 10V DC hand- shaking), TTL, Open collector mv Tristate, line driver/receiver interface +5 V, +10 V devices Audio (0-20 khz) AC Discrete (5V, 24 V, differential - Ultrasonic (20 khz-1 MHz) AC line driver logic) - RS-232C RS-422 RS-485 IEEE-488 (GPIB) Ethernet USB Firewire FieldBus

6 Smart Sensors: Smart sensors are an extension of traditional sensors with advanced learning and adaptation capabilities. The system must also be reconfigurable and perform the necessary data interpretation, fusion of data from multiple sensors and the validation of local and remotely collected data. Smart sensors therefore contain embedded processing functionality that provides the computational resources to perform complex sensing and actuating tasks along with high level applications. The functions of a smart sensor system can be described in termsof compensation, information processing, communications and integration.

7 Wireless Sensors Network -WSN Wireless Sensor Network User Gateway Node Wireless Sensor Node 7

8 Precision agriculture Environment comfort & efficiency Smart homes Alarms, security, surveillance. Disaster management Health Care Traffic Management Transportation safety Land mine Detection Applications of WSN Sensor Augmented Fire Response Earthquake Response Manufacturing Wind Response Elder Care

9 Involved Technologies Network Technology Computational Power Sensor Network Sensor Technology

10 Challenges in WSN s Energy Computation Communication Scalability Fault Tolerance Power Consumption

11 Future Directions of Sensors Research: Wireless Sensors Network Cognitive Sensors Network Body Area Network Internet Sensing Applications: Environment and Climate Monitoring Health Monitoring Structural Health Monitoring Smart Grids

12 Sensors research: Sensing System and Sensor Fabrication Sensors Interfacing Sensors Fabrication

13 Selection of sensors Operating Principle Availability (Sources location, delivery schedule, payment options, continuation of supply) Cost cost of sensor itself, delivery cost Performance Figures >> Range >> Ease of use >> Power supply requirements >> Accuracy >> Drift >> Hysteresis effect >> Effect of other parameters >> Any special arrangement >> Any other.

14 Sensor Technologies Comparisons Temperature Sensing Technologies Principle Thermocouple RTD Thermistor IC's Temperature Range (-233 C C) (-205 C C) (-38 C C) (-40 C ) Long-term Stability Good Excellent Poor to Fair Good Accuracy Medium High Medium Medium Repeatibility Fair Excellent Fair to Good Excellent Sensitivity Low Medium Very High High Response Medium to Fast Medium Medium to Fast Fast Linearity Fair Good Poor Excellent Self Heating No Very Low to Low High Medium Size Small to Large Small to Medium Small to medium Small Cost Relativly Cheap Expensive Cheap Relatively cheap Humidity Sensing Technologies Principle RHS TCHS CHS Humidity Range 5%-95% 0-100% 5%-95 % Long-term Stability Good Excellent Good Accuracy High High High Repeatibility Fair to Good Good Excellent Sensitivity Medium Medium High Response Medium Medium to Fast Fast Linearity Good Good Excellent Size Small Small to Medium Small Cost Cheap Expensive Relatively cheap

15 Summary of Wireless Protocols Standard ZigBee BlueTooth WiFi WiMax (IEEE ) (IEEE (IEEE (IEEE WPAN) WLAN) WWAN) Range 100 m 10 m 5 km 15 km Data rate kbps 1 Mbps-3 Mbps 1Mbps-450 Mbps 75 Mbps Band-width 2.4 GHz 2.4 GHz 2.4, 3.7, and 5 GHz 2.3, 2.5 and 3.5 GHz Network Topology Star, Mesh, Cluster Tress Star Star, Tree, P2P Star, Tree, P2P Applications Wireless Sensors (Monitoring and Control) Wireless Sensors (Monitoring and Control) PC based Data acquisition, Mobile Internet Mobile internet

16 Zigbee Communication Protocol C8051F020 and PC connected with XBee through Serial Interface

17 Xbee -Modules Low Cost Low Power (~50mA Current) Easy to connect & configure Up to bps Range: 100m (XBee) 1600m (XBee Pro)

18 Testing of ZigBee Modules XBee Module Series 2 An extended-range RF module designed for ZigBee mesh networking. Low cost, low-power wireless applications

19

20 IEEE 1451 In General Defines a physical connection between analog signals and TEDS (Transducer Electronic Data Sheet) data. TEDS allows the use of a very small memories through the use of templates. The templates define the significance and units associated with the stored data and mapping of the data in memory Templates is written using Template Description Language (TDL)

21 IEEE Usage Data contained in the TEDS includes transducer type and manufacturer identification to choose the correct template for the interpretation of the stored data. It may include calibration information, sensitivity and so many other details.

22 Emotion Recognition: Emotion is the complex psychophysiological experience of an individual s state of mind as interacting with biochemical (internal) and environmental (external) influences. In humans, emotion fundamentally involves physiological arousals, expressive behaviours and conscious experience. Emotion is associated with mood, temperament, personality and disposition and motivation. Motivations direct and energise behaviour, while emotions provide the affective component to motivation, positive or negative.

23 Human Emotions: Make your emotions truly human by understanding them and controlling them. Let them motivate you to true goodness. Categories of Emotions: Ordinary Emotions: Good that is easy to obtain (desire, joy, Love) or evil that is easy to avoid (disgust, sorrow, hate). Emergency emotions: Good that is difficult to obtain or an evil that is difficult to avoid.

24 Basic Emotions: Anger Five Physiological Signals Courage Desire Despair Disgust Fear Hate Hope Joy Love Sorrow Skin Temperature Skin Conductance Blood Volume Pulse Respiration Eletromyography (EMG)

25 WSN Based smart home for elderly people

26 Population Aging There are currently 510,000people over the age of 65 yrs in New Zealand. 90 millions people over 60 years in India. In the United States alone, the number of people over age 65 yrsis expected to hit 70 million by 2030, almost doubling from 35 million in 2000.

27 Consequences Are.. Expenditures of the US for health care will project to rise to 17.9% of the GDP ($2.9 trillion) by Many elderly people are forced to consigned to expensive retirement homes. Many elderly people choose to stay at home also for privacy/dignity issues.

28 Why Do We Concern About Elderly Care? Because our parents are the next in line to be qualified as elderly, and then we are next to the next in line. Can advances in sensing cum instrumentation technology, embedded controller, wireless communications enable elderly people to regain their capability of independent living? We believe the answers are Yes!

29 Functional Block Diagram Device #1 Sensor Unit #1 Cellular Modem RS232 Device #2 Sensor Unit #2 Central Controller Unit RS232 PC Device #n Sensor Unit #n

30 The Fabricated System SAM: Nana hasn t used the kettle all morning. Panic Button

31 Electrical Appliance Monitoring Unit Sensor Unit (SU) Power Supply Current Transformer & circuitry Microcontroller RF Module LED Display

32 Future Concept of Embedded Smart Sensor A self-powered smart sensor module with built-in communication electronics

33 Detection of daily life pattern and abnormal behavior

34 CONCLUSIONS A smart home to care elderly people based on wireless sensors. The system doesn t use camera or vision based system and thus acceptable to elderly community. The integrated system is able to support people who wish to live independently.

35 Zigbee Based Wearable Physiological Parameter Monitoring System

36 Functional Block Diagram Wearable Physiological Parameter Monitoring System

37 Fabricated System and experimental results

38 Temperature Measurement Maxim DS600 IC Special Presentation at ITRC, Taiwan

39 Temperature Sensor Maxim DS600 IC-Analog output Very simple no external circuitry Exposed Pad 8-Pin for Quick thermal response ±0.5 C Accuracy (-20 C to +100 C) output in the form of mv, fed to an analog (port) input pin of the microcontroller

40 Temperature Measurement VOUT = Device Temperature ( C) x ΔV/ΔT + VOS Device Temperature ( C) = (VOUT- VOS)/ (ΔV/ΔT) Device Temp. =(VOUT-509mV)/ 6.45mV/ C for ADC conversion Temperature(Vout) = ((Result/16) *2430/4095 Temperature = ((Temperature(Vout)- 509)*10000/645) DS600 Temperature Sensor(in Lab) Voltage o/p (V) Temperature

41 Interfacing-Microcontroller Analog signal from temp. sensor connected to ADC0, pin.1 of SiLabs C8051F020 microcontroller ADC0 conversion initialized using Timer3 overflow Timer3 interrupt (EIE2.0) enabled Reading ADC on interrupt and collects few values and average them to get Displayed Temp ( C) : Temperature = ((Temperature(Vout)-509)*10000/645)

42 Results 34.0 Make new result graph with values as well refen ce (actu al) tem p [ C ] (Julia) temperature male measured temp [ C] w rsit ellbow neck reference (actual) temp [ C] female 1 temp measured temp [ C] w rist ellbow neck temperature lab temperature [ C*100] time [sec]

43 Heart Rate Sensor Transmitter: GaAs infrared LED with a wavelength of 940nm Receiver: Infra red sensor-phototransistor (SDP8406) The near-infrared spectroscopy method Transmitter-sensor pair is clipped on tip of Index finger of the person

44 This signal, which is in the form of pulses is then amplified and is fed to a low-cost microcontroller for analysis and display Circuit Diagram

45 Interfacing with microcontroller Output signal-pulses is fed to pin 0.2 of port0(digital I/P port) of C8051F020 Timer0 generates pulses and time0 ISR (using interrupt(0) INT0 )is used to count them Time between pulses counted and are stored in BPM_T_count Frequency = /BPM_T_count_value BPM = (Frequency*6)/10)

46 Results Male Heart Rate Reference: pulse watch Error M M M M M M M M M M M M Female Heart rate Reference: pulse watch Error F F F F F F F F F F F F

47 Accuracy Change of heart rate can be detected faster with developed sensor Accuracy of Heart Rate Sensor reference [bpm] sensor [bpm]

48 Impact Sensor Sensor is the ADXL213 Accelerometer, requires low energy and Dual axis One axis is used for monitoring movements The range of duty cycle varies between 0-100% 50% is horizontal position w.r.t earth surface

49 Impact measurement provides a digital signal detected by interrupt(it1) on PORT0 pin 0.3 Timer0 generates pulses and time0 ISR (using interrupt(1) INT1 )is used to count them(t_count_y) calculates the width of the pulse (Duty_count_Y) Time between pulses counted and are stored in T_count_Y and width in Duty_count_Y Duty_y = (D_count_Y *100)/T_count_Y;

50 Result

51 Typical results of the Impact sensor for different condition Special Presentation at ITRC, Taiwan

52 Conclusions and Future Work System with three sensors is developed Small and Light Weight Smart monitoring and easy to use On the wrist Individual sensor results and all sensors in one unit obtained and compared with references Future work More results for each sensor Adding Emergency button Improving the PCB and software part for reliability Make a wearable unit for wrist Collecting data for three units

53 Thank You