FineGrained AdHoc Localization in Wireless Sensor Networks

1
Fine-Grained Ad-Hoc Localization in Wireless
Sensor Networks

• Andreas Savvides
• Center for Embedded Networked Sensing (CENS)
• Networked and Embedded Systems Lab (NESL)
• http//nesl.ee.ucla.edu/projects/ahlos
• http//nesl.ee.ucla.edu/projects/smartkg

2
Location Awareness in Sensor Networks
Ad-Hoc Localization
Random Deployment
Operate in the presence of obstacles
Rapid Infrastructure Setup

• Multihop networks
• may span of over large
• geographical regions
• Not always easy to provision
• for the proper placement of
• landmarks

• Form a multihop network
• to avoid obstacles
• Landmarks may not always
• be within range of all nodes

• Reduce the cost
• and time overhead of
• installing new systems

3
Localization in Smart Kindergarten

• Derive locations of students and objects
• Track head motion patterns
• Use ultrasonic Time-of-Flight
• Requirements
• Unobtrusive operation
• Low power consumption
• High degree of accuracy
• Ease of deployment
• Smart beacon calibration
• Communicate the locations back to the
  infrastructure

4
Platforms Medusa MK-2

• Medusa MK-2 Node
• For localization experiments
• 40MHz ARM THUMB
• 1MB FLASH, 136KB RAM
• 0.9MIPS/MHz
• 480MIPS/mW (ATMega 242MIPS/mW)
• can run eCos, uCLinux
• RS-485 bus
• Out of band data collection
• Formation of arrays
• 3 current monitors (Radio, Thumb, rest of the
  system)
• 540mAh Rechargeable Li-Ion battery

 
5
Ultrasonic Ranging Latency
USND TX Start
Ultrasound Detected
Transmitter
Receiver
6
Ranging Characterization

• Lab characterization of ranging module, at 25
  pulses (temperature 21.4 C)

7
Localization Algorithms

• Platforms are computationally constrained
• Incomplete beacon node information
• Nodes need to collaborate to jointly estimate
  their locations -gt collaborative multilateration
• Need to avoid error propagation
• Distributed operation to avoid node failures
• Lightweight processing and efficient
  communication to preserve power

8
Collaborative Multilateration
beacon nodes

• Utilize measurement information over multiple
  hops
• Solve the problem in a fully distributed manner

9
Centralized Collaborative Multilateration
1
5
4
3
6
2
The objective function is
Can be solved using iterative least squares
utilizing the initial Estimates from phase 2 -
solve with an Extended Kalman Filter
10
Distributed Collaborative Multilateration

• Instead, we propose a simple approximation
• Each node performs a multilateration using only
  next-hop neighbor information in the context of a
  collaborative subtree
• If multilaterations follow a consistent pattern
  then a global gradient with respect to the whole
  collaborative subtree is established (driven
  using Distributed Depth First Search)
• Much less computation, similar result, fully
  distributed operation with desirable side effects

11
Distributed Collaborative Multilateration
2
5
3
4
1
The unknown nodes need to perform their atomic
multilateration in the same order, driven by a
Distributed Depth First Search algorithm gt
local computations, follow a global gradient
12
Distributed Collaborative Multilateration
2
5
3
Error is reduced at each iteration, because we
are operating in an over-constrained setup
4
1
The unknown nodes need to perform their atomic
multilateration in the same order, driven by a
Distributed Depth First Search algorithm gt
local computations, follow a global gradient
13
Convergence Process

• From SensorSim
• simulation
• 40 nodes, 4 beacons
• IEEE 802.11 MAC
• 10Kbps radio
• Average 6 neighbors
• per node

14
Gains in Computation Overhead

• Computation cost based on MATLAB FLOPS outputs
• Result difference between centralized and
  distributed is very small
• Mean 0.015 mm, Standard Deviation 0.0054mm
• A group of nodes can collectively solve a
  non-linear optimization problem than none of the
  nodes can solve individually.
• Distributed computation cost between 3-4 MFLOPS
  per node

15
Communication Cost and Latency

• Convergence time increases
• with group size
• Similar trend in the
• communication cost

• Communication cost evenly
• distributed across all nodes
• Communication cost can be further
• reduced by reducing group size

16
Error Behavior of Multihop Localization

• Many sources of error
• Channel error, algorithmic and computation error
  and setup error
• Setup error is associated with design-time and
  deployment time parameters
• Deployment geometry
• Network density
• Beacon concentration
• Measurement technology accuracy
• Certainty in beacon locations
• Cramer-Bound Analysis to show how the setup error
  behaves as the network scales

17
Node and Beacon Density Effects
RMS Error(m)
RMS Error(m)
Node density (nodes/m2)
Number of beacons
100 Node network, 4 20 beacons
200 Node network, 10 beacons
18
Smart Beacon Calibration
19
Host PC Controller Software
3D GUI Client(s)

• Manager
• Packet based
• One to many switching
• Facilitate online
• processing of incoming
• data
• Allows direct use of
• MATLAB code

TCP Server
Other SW
Localizer
Calibration SW
Gateway Node
Serial I/O
20
Conclusions

• Collaborative Multilateration is a feasible
  solution
• Works with binary obstacles
• Distributed localization is feasible in some
  scenarios
• An initial testbed is there, working on
  completion.
• Geometry is a problem
• Ready to generate larger traces of measurement
  data for further study
• More experiments using the testbed
• Moving from advance hardware testing to a more
  complete system that provides localization and
  tracking services