survey and implementation of smart space scenarios
play

Survey and Implementation of Smart Space Scenarios Philipp H - PowerPoint PPT Presentation

Survey and Implementation of Smart Space Scenarios Philipp H artinger Advisors: Marc-Oliver Pahl Benjamin Hof Chair for Network Architectures and Services Department for Computer Science Technische Universit at M unchen September


  1. Survey and Implementation of Smart Space Scenarios Philipp H¨ artinger Advisors: Marc-Oliver Pahl Benjamin Hof Chair for Network Architectures and Services Department for Computer Science Technische Universit¨ at M¨ unchen September 29, 2014 Philipp H¨ artinger: Survey and Implementation of Smart Space Scenarios 1

  2. Outline 1 Motivation 2 Survey of Smart Space Scenarios Sources & Selection Criteria Results 3 Implementation of Occupancy Detection System Design Evaluation & Results 4 Sources Philipp H¨ artinger: Survey and Implementation of Smart Space Scenarios 2

  3. Motivation What are we talking about? Pervasive Computing Mark Weiser [Wei91] computing in everything Smart Space physical space with sensors and actuators gather state of user and environment = context reason about context and act appropriately Philipp H¨ artinger: Survey and Implementation of Smart Space Scenarios 3

  4. Table of Contents 1 Motivation 2 Survey of Smart Space Scenarios Sources & Selection Criteria Results 3 Implementation of Occupancy Detection System Design Evaluation & Results 4 Sources Philipp H¨ artinger: Survey and Implementation of Smart Space Scenarios 4

  5. Survey Goal overview of implemented smart space scenarios used hardware (sensors, actuators) used software (middleware) Sources HUC 1999–2000 UbiComp 2001–2013 Pervasive 2002–2012 Selection Criteria sensors and actuators in a well-defined physical space implementation in reality or simulation Philipp H¨ artinger: Survey and Implementation of Smart Space Scenarios 5

  6. Survey Results contact switch microphone location camera motion RFID light Category Project Title Other Sensors Middleware AAL, Health Care Monitoring Behavior in Home [Nou+00] � � acceleration, tilt, vibration Activity Recognition in Home [TIL04] � Health-Status Monitoring [BBA05] � STAR [WA05] � � pressure, break-beam Diet-Aware Dining Table [Cha+06] weight � CAMP [Lun+07] � � Height Sensors for Biometric Identification [SSW10] distance Convenience AwareMirror [FKN05] � � temperature � Audio Location [SD05] � ReflectiveSigns [M¨ ul+09] � Kitchen Activity Recognition [LRF12] � Detecting Cooking State [Hir+13] gas Break-Time Barometer [Kir+13] electrical current, humidity � � � Find My Stuff (FiMS) [Nic+13] � Energy, Lighting ViridiScope [Kim+09] � � � GasSense [Coh+10] � SunCast [LW12] � AgentSwitch [Fis+13] electrical current Intelligent Luminaires [M¨ ak+13] � HVAC Management of Energy & Thermal Comfort [BK01] CO2, mixed gas GPS-Thermostat [GIL09] � temperature Smart Thermostat [Lu+10] � � PreHeat [Sco+11] � � � temperature TherML [Koe+13] � Office Smart Classroom [Xie+01] � � � EasyMeeting [Che+04] bluetooth sensing agent � � Detect User Activity and Availability [M¨ uh+04] � � � Detection of Interaction Groups [BMR05] � Virtual Secretary [DS08] � � Weightless Walls [Tak10] � � ReticularSpaces [Bar+12] � � Philipp H¨ artinger: Survey and Implementation of Smart Space Scenarios 6

  7. Survey Results Occupancy Detection contact switch microphone location camera motion RFID light Category Project Title Other Sensors Middleware AAL, Health Care Monitoring Behavior in Home [Nou+00] � � acceleration, tilt, vibration Activity Recognition in Home [TIL04] � Health-Status Monitoring [BBA05] � STAR [WA05] � � pressure, break-beam Diet-Aware Dining Table [Cha+06] weight � CAMP [Lun+07] � � Height Sensors for Biometric Identification [SSW10] distance Convenience AwareMirror [FKN05] � � temperature � Audio Location [SD05] � ReflectiveSigns [M¨ ul+09] � Kitchen Activity Recognition [LRF12] � Detecting Cooking State [Hir+13] gas Break-Time Barometer [Kir+13] electrical current, humidity � � � Find My Stuff (FiMS) [Nic+13] � Energy, Lighting ViridiScope [Kim+09] � � � GasSense [Coh+10] � SunCast [LW12] � AgentSwitch [Fis+13] electrical current Intelligent Luminaires [M¨ ak+13] � HVAC Management of Energy & Thermal Comfort [BK01] CO2, mixed gas GPS-Thermostat [GIL09] � temperature Smart Thermostat [Lu+10] � � PreHeat [Sco+11] � � � temperature TherML [Koe+13] � Office Smart Classroom [Xie+01] � � � EasyMeeting [Che+04] bluetooth sensing agent � � Detect User Activity and Availability [M¨ uh+04] � � � Detection of Interaction Groups [BMR05] � Virtual Secretary [DS08] � � Weightless Walls [Tak10] � � ReticularSpaces [Bar+12] � � Philipp H¨ artinger: Survey and Implementation of Smart Space Scenarios 6

  8. Table of Contents 1 Motivation 2 Survey of Smart Space Scenarios Sources & Selection Criteria Results 3 Implementation of Occupancy Detection System Design Evaluation & Results 4 Sources Philipp H¨ artinger: Survey and Implementation of Smart Space Scenarios 7

  9. Implementation Overview Hardware and Software Design DS2OS OccupancyDetection RoomManager motion (PIR) Service Service sensor ------------------------------ -------------------- analyzes manages sensor readings associations temperature to infer occupancy RoomID↔Galileo Intel Galileo sensor Node.js brightness Webserver sensor SensorGateway Service --------------------- pulls sensor readings over HTTP TCP/IP Network Philipp H¨ artinger: Survey and Implementation of Smart Space Scenarios 8

  10. Implementation: Hardware Intel Galileo Ethernet RESTful Web Interface (Node.js) Sensors Motion Temperature Brightness Figure created with Philipp H¨ artinger: Survey and Implementation of Smart Space Scenarios 9

  11. Implementation: Software (DS2OS) DS2OS DS2OS Services OccupancyDetection RoomManager Service Service ------------------------------ -------------------- GalileoGatewayService analyzes manages sensor readings associations to infer occupancy RoomID↔Galileo OccupancyDetectionService timer threshold parameter T if no motion for T seconds SensorGateway Service --------------------- ⇒ unoccupied pulls sensor readings over HTTP RoomManagerService facilitates access to sensors and occupancy state of a room get roomManager/office/occupancy/isOccupied/value Philipp H¨ artinger: Survey and Implementation of Smart Space Scenarios 10

  12. Implementation: Software (UI) Web User Interface HTML & JavaScript queries occupancy states periodically (via RoomManager) colors rooms on map accordingly green = unoccupied red = occupied Philipp H¨ artinger: Survey and Implementation of Smart Space Scenarios 11

  13. Evaluation Experiment 1: controlled experiment in 1 room over 4 hours Expected Results: unoccupied or high activity ⇒ high detection rate occupied, but low activity ⇒ lower detection rate detection rate depends on distance Experiment 2: measurements in 3 rooms over 5 days Expected Results: typical usage patterns of the rooms become inferable Middleware Usage fast and easy implementation due to DS2OS middleware Philipp H¨ artinger: Survey and Implementation of Smart Space Scenarios 12

  14. Experiment 1: Setup 4 hours 0–3 persons 3 Galileo boards hand logged data 2 different situations: desk A desk B 0.5 m empty low motion 2 2.5 m . 2 m 5 m high motion 3 1 Philipp H¨ artinger: Survey and Implementation of Smart Space Scenarios 13

  15. 2 desk A desk B Experiment 1: Results 3 1 Motion sensor data Galileo1 1 0 OccupancyDetectionService with T=60s 1 0 Difference between service output and hand logged data 1 false positive 0 false negative −1 Hand logged occupancy 1 low motion high motion 0 10 46 80 92 109 133 150 181 195 218 240 time in minutes Galileo 1 (middle) Figure created with Matlab Philipp H¨ artinger: Survey and Implementation of Smart Space Scenarios 14

  16. 2 desk A desk B Experiment 1: Results 3 1 Motion sensor data Galileo2 1 0 OccupancyDetectionService with T=60s 1 0 Difference between service output and hand logged data 1 false positive 0 false negative −1 Hand logged occupancy 1 low motion high motion 0 10 46 80 92 109 133 150 181 195 218 240 time in minutes Galileo 2 (desk) Figure created with Matlab Philipp H¨ artinger: Survey and Implementation of Smart Space Scenarios 15

  17. 2 desk A desk B Experiment 1: Results 3 1 Motion sensor data Galileo3 1 0 OccupancyDetectionService with T=60s 1 0 Difference between service output and hand logged data 1 false positive 0 false negative −1 Hand logged occupancy 1 low motion high motion 0 0 10 46 80 92 109 133 150 181 195 218 240 time in minutes Galileo 3 (door) Figure created with Matlab Philipp H¨ artinger: Survey and Implementation of Smart Space Scenarios 16

Recommend


More recommend