quartz/content/notes/07-location-sensors.md
2023-03-21 12:15:32 +13:00

8.2 KiB
Raw Blame History

title tags
07-location-sensors
lecture
info305

Terminology

  • Precision vs. Accuracy

  • How practical are systems that are:

    • Precise but not accurate?
    • Accurate but not precise? Accuracy vs. Precision
  • Degrees of Freedom (DoF):

  • How many translational and rotational movements are possible?

  • How many degrees of freedom in 3D space?

    • 2 DoF?
    • 3 DoF?
    • 6 DoF?
    • 9 DoF?
    • Unlimited DoF?
  • How many degrees of freedom in 2D space? Degrees of Freedom

  • (Tri)angulation:

    • Based on angles between signals
      • Referred to as angles of arrival
      • Angle to reference point determined by means of special antennas
    • Positions of reference points have to be known
    • 3D-positioning requires three reference points (triangulation) Angulation

[!INFO] if you have two reference points and the angle towards them, you can get the location

  • (Tri)lateration
    • Based on distances between device and reference points
    • Calculated e.g. by time of arrival (ToA) or signal strength
    • Positions of reference points have to be known
    • 3D-positioning requires three reference points (trilateration) Lateration

[!INFO] siimilra to triangulation but you need three points and its calculated using distancees not angles. the distances can be calculated by time of arrival of a signal

Cell-based

  • Positioning based on GSM cell ID
  • Mobile Positioning System (MPS) by Ericsson to improve accuracy
    • Few changes of infrastructure
    • No changes of user devices
  1. Cell Global Identity (CGI) for identification of the cell
    • Uses GSM cell ID
    • If sector-antennas used position can be narrowed to segment of the circle
    • Angulation

[!INFO] compute your position based on which cell tower you are connected to. each tower has a unique ID. used by emergency providers, they know where you last logged in

  1. GSM based on FDMA (Frequency-division multiple access) and TDMA (Time-division multiple access)
    • Exact timing required for the synchronization of uplink and downlink
    • Timing Advance (TA)
    • Device calculates distance based on signal propagation time
    • Can be used for positioning in combination with CGI |200

[!INFO] "towers" use FDMA or TDMA 21-data-link-layer#multiple access links and protocols Timing advance used for multiple access can be used to calculate time

  1. Uplink-Time Difference of Arrival (UL-TOA)
    • Four base stations needed
    • Measurement of signal propagation time at base stations
    • Calculation of position based on lateration
    • Accuracy 50 150m
    • Network-based location technology,
      • Can locate any type of mobile phone.
      • Only available to the owner of the sensor network Uplink-Time Difference of Arrival

[!INFO]

  • Integrated approach based on existing infrastructure

  • Low costs (for user)

  • Low power draw

  • High availability

  • Works indoor and outdoor (no direct line of sight needed)

  • Information not always available to the user (often only network provider)

    • Used for “Enhanced 911”
  • But relatively low accuracy

  • Other cell-based location approaches: WiFi/WLan-cells

    • Integrated infrastructure approach
    • WiFi/WLAN (cell-id, lateration, fingerprinting)
  • Based on already existing WLAN infrastructure, primarily installed for communication purposes

  • Cell ID requires database with mapping WiFi cell ID to GPS or …. (e.g. created by Apple, Google)

  • Lateration

    • Requires accurate information about access point positions
    • Measurement of signal strength of various access points
      • At least three access points have to be available
      • Measurements are influenced by obstacles like walls
  • Improved approach: Fingerprinting

  • Fingerprinting:

    • 1st Phase: Radio Map Creation (Offline):
      • Measurements of fingerprints at reference points -> stored in fingerprint DB
    • Definition of reference points according to accuracy needs and building structure
  • 2nd Phase: Radio Map Usage (Online):

    • Fingerprint of current position is taken
    • Search for closest matching reference point on radio map Position taken from that reference point or interpolated
  • Fingerprint

    • Location: lat,lon,floor
    • VisibleAccessPoints{
      • Fingerprint DB
      • AP1:
        • SSID: eduroam
        • MAC: 08:17:35:33:5f:80
        • signal strength: -64
      • AP2
        • SSID: VPN/WEB
        • MAC: 08:17:35:33:5f:81
        • signal strength: -61
      • APn o...

Fingerprinting services using Wifi

  • ~99% of queries use Google location database (Google Maps Geolocation API)
  • Alternatives (less coverage & accuracy)
    • OpenCellID
    • Mozilla Location DB
  • Not available when offline, cell DB too large
  • WLAN cells can move/change - frequent update required

  • Cell structure: circular (theory), hexagonal (planning), irregular (reality)
  • Neighbouring cells use different frequencies to minimise interference
    • Some overlap inevitable
  • Typical cell sizes:
    • WiFi: 10m - 100m
    • 3G (UMTS): 100m - 5km
    • 2G (GSM): 100m - 35km

7-cell cluster (commonly used) cell characteristics

Proximity Sensors & Near Field Communication

  • Different technologies for sensing proximity or exchanging data (often dual purpose)
    • Stand-alone infrastructure approaches
    • Not widely accepted due to special hardware requirements infrastructure costs (for tracking)
  • Conceptually often similar to cell-based approaches but require extra infrastructure

  • WIPS (Infrared)
    • Beacons installed in the rooms sending unique ID
    • Users badges receive signals of local beacons
    • Received beacon ID is sent to location server via WLAN
    • Server maps received beacon ID to semantic location which is sent back to the user
  • Active Badge (Infrared)
    • Users carry badge sending specific user ID
    • IR-receivers in the rooms receive those signals
    • Position of user is tracked by central server
    • Energy-efficient badges (sending short signals of 0,1s each 15s)

  • Bluetooth Low Energy LE (Bluetooth specification 4.0)
  • Not to confuse with “classic” Bluetooth (e.g. used in headsets or for I/O devices)
    • Optimized for battery powered sensors
    • Months or years of battery liftetime
    • Lower data rate (1MBIT/sec)
    • Backwards compatible (same frequency and modulation)

  • Two classes for peripherals:

    • Beacons (pure static broadcasts:
      • iBeacon simply broadcasts UUID
      • Additional data possible on request
    • Sensors (broad/unicast with sensor data)
      • Different profiles temperature, gravity, hear rate, pressure, …
      • Notification possible to avoid polling
    • Bidirectional communication
      • Not primarily focus, mostly for setting parameters
  • Location and Proximity Sensing using Bluetooth LE

    • One beacon (e.g. per room)
      • Pure proximity sensing, looking for strongest signal
      • E.g. Region monitoring to detect Beacon presence
    • Many beacons (e.g. per room)
      • Range calculation using Received Signal Strength Indicator (RSSI) and calibrated transmitter power (txPower = RSSI at 1m)
      • Trilateration using several beacons but not very accurate
  • Range different for beacons but typically between 20-50m (rarely 100m)

  • Different protocols on top of Bluetooth LE
    • iBeacon (Apple)
      • Broadcasts a UUID
      • ID is used with database integrated in the app
      • Further information on request (e.g. range information)
    • Eddystone (Google)
      • Beacons broacasts information about the beacon (telemetry frame e.g. battery or sensor information)
      • Beacons broadcasts and redirects to an URL (physical web)

|200