SLIDE 1
Computer Science@UMASS Amherst
Emerging ecosystem of on-body sensors
activity monitoring wristband GPS, Accelerometer, Gyroscope, Camera, Microphone blood pressure glucose monitors implantable “nanosensors” sleepshirt sleep monitor
Backscatter communication for wearables/IoT Deepak Ganesan Associate - - PowerPoint PPT Presentation
Backscatter communication for wearables/IoT Deepak Ganesan Associate - - PowerPoint PPT Presentation
Backscatter communication for wearables/IoT Deepak Ganesan Associate Professor Computer Science UMass Amherst Emerging ecosystem of on-body sensors GPS, Accelerometer, Gyroscope, activity monitoring Camera, Microphone wristband blood pressure
SLIDE 2
SLIDE 3
Computer Science@UMASS Amherst
Why is communication a problem?
Communica)on ¡is ¡a ¡major ¡bo/leneck ¡for ¡low-‑power ¡opera)on
Power Consumption (uW)
1 1000 1000000
Communication Sensing
802.15.4 (Zigbee) Bluetooth Low Energy Ultra-Wide Band Camera (Stonyman) ECG Microphone Accelerometer WiFi
SLIDE 4
Computer Science@UMASS Amherst
Backscatter communication
Backscatter reader Sensor
Backscatter is extremely efficient because the device is reflecting the signal rather than generating a signal.
Transistor
Carrier Wave Reflected Signal
SLIDE 5
Computer Science@UMASS Amherst
Why is backscatter a challenging domain?
Carrier Wave Reflected Signal
SLIDE 6
Computer Science@UMASS Amherst
Why is backscatter a challenging domain?
Carrier Wave Reflected Signal
Energy
…
available ¡energy ¡is ¡miniscule
distance from reader
SLIDE 7
Computer Science@UMASS Amherst
Why is backscatter a challenging domain?
Carrier Wave Reflected Signal
- 0.3
- 0.2
- 0.1
0.1 0.2 0.3 0.4 2 4 6 8 10 12
Amplitude Time (seconds)
I Channel Q Channel
Antenna coupling
- 1
- 0.8
- 0.6
- 0.4
- 0.2
0.2 0.4 0.6 2 4 6 8 10 12
Amplitude Time (seconds)
I Channel Q Channel
Antenna rotation
SLIDE 8
Computer Science@UMASS Amherst
Why is backscatter a challenging domain?
Carrier Wave Reflected Signal
Human activity
- 0.4
- 0.2
0.2 0.4 2 4 6 8 10
Signal Strength Transmission Time (ms)
transmitted signal decoding fails
- 0.6
- 0.4
- 0.2
0.2 0.4 0.6 2 4 6 8 10 12
Amplitude Time (seconds)
I Channel Q Channel
Capacitor charge decay
SLIDE 9
Computer Science@UMASS Amherst
Why is backscatter a challenging domain?
Carrier Wave Reflected Signal
RX
0° 90° ADC PLL VCO ADC Antenna Low Noise Amplifier Mixer Mixer Low-pass Filter Low-pass Filter Base Band Base Band
TX
0° 90° DAC Base Band PLL VCO DAC Antenna Power Amplifier Mixer Mixer Low-pass Filter Base Band Low-pass Filter
TX RX
Sensor Antenna Switch Z0 Z1 Antenna Comparator D1 C1 C2 R1 R2 +
- Reader
Message Processing
Asymmetry
SLIDE 10
Computer Science@UMASS Amherst
Why is backscatter a challenging domain?
Carrier Wave Reflected Signal
Range Throughput Power
bit-by-bit backscatter
[NSDI 2014]
Re-thinking tag architecture
[Mobicom 2014]
Hybrid harvesting
[MobiSys 2012]
Concurrent backscatter
[HotWireless 2014]
Asymmetry
SLIDE 11
Computer Science@UMASS Amherst
Taking stock…
Major step is to measure/evaluate/iterate under uncontrolled environments Many promising techniques for solving hard problems in backscatter: range (50 ft), throughput (Mbps), harvesting (Reader/Ambient/WiFi)
SLIDE 12
Computer Science@UMASS Amherst
Case for a backscatter testbed
- Testbed #1: Backscatter-instrumented house
- Reader/WiFi/Ambient Backscatter for IoT/wearables
- Testbed #2: Backscatter-instrumented human
- Mobile backscatter reader +“bandaid” sensors
- Why? design complete solutions
- Robustness in indoor environments
- Perpetual link while user is mobile
SLIDE 13
Computer Science@UMASS Amherst