Platzhalter für Bild, Bild auf Titelfolie hinter das Logo einsetzen Institute of Operating Systems and Computer Networks Stephan Sigg, Shuyu Shi, Felix Büsching, Yusheng Ji, Lars Wolf Leveraging RF-channel fluctuation for activity recognition On active and passive systems; Continuous and RSSI-based signal features
21
Embed
Leveraging RF-channel fluctuation for activity recognition
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Platzhalter für Bild, Bild auf Titelfolie hinter das Logo einsetzen
Institute of Operating Systemsand Computer Networks
Stephan Sigg, Shuyu Shi, Felix Büsching, Yusheng Ji, Lars Wolf
Leveraging RF-channel fluctuation for activity recognitionOn active and passive systems; Continuous and RSSI-based signal features
Felix Büsching: Leveraging RF-channel fluctuation for activity recognition | 2 Institute of Operating Systemsand Computer Networks
Outline
Motivation: Why should anyone want to do this? Test SetupResultsConclusion
Felix Büsching: Leveraging RF-channel fluctuation for activity recognition | 3 Institute of Operating Systemsand Computer Networks
Personal Emergency Response System (PERS)
Observed behavior 1: Normal usageWorks – since several years
Fall occurs Button pressed Alarm initiated
Felix Büsching: Leveraging RF-channel fluctuation for activity recognition | 4 Institute of Operating Systemsand Computer Networks
Personal Emergency Response System (PERS) – Non-Usage
Observed behavior 2: Non-usage Forgets to wear senderDoesn’t even know where it is
No button to press No alarmFall occurs
Felix Büsching: Leveraging RF-channel fluctuation for activity recognition | 5 Institute of Operating Systemsand Computer Networks
Personal Emergency Response System (PERS) – Refused Usage
Observed behavior 3: Refused usage “The mobile nursing service will come anyway – in 2-4 hours” “I’ll wait here on the floor – don’t want to bother anyone”
Fall occurs Button not pressed No alarm
Felix Büsching: Leveraging RF-channel fluctuation for activity recognition | 6 Institute of Operating Systemsand Computer Networks
Personal Emergency Response System (PERS) – Anxious Usage
Observed behavior 4: Anxious usage “I feel safe – I better wear that thing 24/7”Many false alarms during nighttime
Bad dream occursButton accidentally pressed
False Alarm
Felix Büsching: Leveraging RF-channel fluctuation for activity recognition | 7 Institute of Operating Systemsand Computer Networks
Summary: Personal Emergency Response System (PERS)
1. Normal usage OK, this is boring…
2. Non-usage Additional stationary sensors for activity detection
3. Refused usage Additional stationary sensors for activity detection
4. Anxious usage Only stationary sensors for activity detection
Using RF-channel fluctuation for activity recognition may be a good idea!
Felix Büsching: Leveraging RF-channel fluctuation for activity recognition | 8 Institute of Operating Systemsand Computer Networks
Motivation for RF-based Activity Detection
Worn devices are not always present when neededMost IoT-devices include a radio transceiverNo interaction is requiredNot intrusive like camerasHigher acceptance
Felix Büsching: Leveraging RF-channel fluctuation for activity recognition | 9 Institute of Operating Systemsand Computer Networks
Outline
Motivation Test SetupResultsConclusion
Felix Büsching: Leveraging RF-channel fluctuation for activity recognition | 10 Institute of Operating Systemsand Computer Networks
1/4 different devices
USRP1: Active SDR-based DFAR 900 MHz (RFX900 board)Vert900 Antenna Sine signal, continuously modulated on carrier 80 Hz sample rate
Felix Büsching: Leveraging RF-channel fluctuation for activity recognition | 11 Institute of Operating Systemsand Computer Networks
2/4 different devices
USPR N210: Passive SDR-based DFAR 82.5 MHz (WBX board), Vert900 AntennaEnvironmental FM radio from nearby radio station 64 Hz sample rate
Felix Büsching: Leveraging RF-channel fluctuation for activity recognition | 12 Institute of Operating Systemsand Computer Networks
3/4 different devices
INGA WSN-Node: Active RSSI-based DFAR 2.4 GHz IEEE802.15.4 radio PCB-AntennaUDP-Packets 100 packets per secondRSSI-values from radio (received MAC-Frames)
Felix Büsching: Leveraging RF-channel fluctuation for activity recognition | 13 Institute of Operating Systemsand Computer Networks
4/4 different devices
Iphone 4: Accelerometer-based activity recognition 3 axis accelerometerAttached to body 40 Hz sampling rateRecorded simultaneously to other experiments
x
yz
Felix Büsching: Leveraging RF-channel fluctuation for activity recognition | 14 Institute of Operating Systemsand Computer Networks
Test Setup: Scenario 1
Felix Büsching: Leveraging RF-channel fluctuation for activity recognition | 15 Institute of Operating Systemsand Computer Networks
Test Setup: Scenario 2
Felix Büsching: Leveraging RF-channel fluctuation for activity recognition | 16 Institute of Operating Systemsand Computer Networks
Performed Activities
Empty room
Standing At 3 different locations
Lying At 3 different locations
Walking At 3 different speeds
Crawling
Felix Büsching: Leveraging RF-channel fluctuation for activity recognition | 17 Institute of Operating Systemsand Computer Networks
Felix Büsching: Leveraging RF-channel fluctuation for activity recognition | 21 Institute of Operating Systemsand Computer Networks
Summary, Conclusion & Outlook
Device-free, no worn sensors, no interaction requiredMonitoring of elderly by RF-Channel Activity Recognition works!… nearly as good as accelerometer based (Passive is not that good)
Monitoring smart spaces works for simple activities.
OutlookOnly RSSI-based information will be needed for activity detection