RVSN Group > Distributed Sensor Network Localization and Tracking
Distributed Sensor Network Localization and Tracking
Summary
Given a wireless sensor network consisting of low-power devices, localization is the task of discovering the 2D or 3D positions of the sensor nodes. We have devised a robust distributed localization algorithm that makes use of distance estimates between nodes to compute 2D positions. In addition, the motion of mobile nodes can be reliably tracked. Our experimental testbed consists of the Cricket Location System designed by the Networks and Mobile Systems Group at MIT.
Movies
The following videos showcase some of the work that we have been doing:
- Mobility Sequence 1 [AVI w/ audio] [MOV w/ audio] [AVI no audio]
- Shows a sensor network being deployed, followed by the automatic localization of the static nodes. A mobile node is introduced and smoothly tracked by the network of static nodes.
- Mobility Sequence 2 [AVI no audio]
- Same setup as sequence 1, showing a slightly different path for the mobile node.
- Static Sequence [AVI]
- This video shows the physical deployment of a static network on the left and the visualization of the localization algorithm running in real-time on the right. After the 5-node network is deployed, two nodes are moved to show that the algorithm can recover from this situation.
- Static Sequence showing robust quads [AVI]
- The same sequence as above, also showing "robust quads" which help to achieve a unique localization. This is the novel innovation of our algorithm. See the paper below for more information.
The AVI files should play fine with the latest version of Windows Media Player and under Linux using mplayer. Mac OS X users may need to download the MS-MPEG4v2 codec.
People
| Name | Position | |
|---|---|---|
| David Moore | dcm@mit.edu | Alumni |
| John Leonard | jleonard@mit.edu | Faculty |
| Daniela Rus | rus@csail.mit.edu | Faculty |
| Seth Teller | teller@csail.mit.edu | Faculty |
Publications
- D. Moore, J. Leonard, D. Rus, S. Teller. "Robust distributed network localization with noisy range measurements." In Proceedings of the Second ACM Conference on Embedded Networked Sensor Systems (SenSys '04). Baltimore, MD. November 3-5, 2004. pp. 50–61. Winner of the Best Paper Award. [PDF]
Software
Our algorithm has been implemented to run on-board the Cricket hardware platform. In addition, we have a suite of tools for visualization and simulation on a host PC. The embedded software was developed on top of the TinyOS software framework.
This software is made available under the GPL license.
Phase I (cluster localization) implementation
Runs on-board crickets using TinyOS or on a PC using TOSSIM. Also includes a visualization tool, glcricket, that runs under X-Windows with OpenGL.
Software Download:
cricket.tar.gz
Installation Instructions: README
Phase III (cluster stitching) implementation
Runs on a PC using Matlab. Note that the Cricket platform does not provide enough memory to implement Phase III on-board the hardware. Instructions for this Matlab software is included in the tarball. Updated March 16, 2005: Now includes scripts to localize large-scale simulated networks.
Software Download:
phase3-20050316.tar.gz
This material is based upon work supported under a National Science Foundation Graduate Research Fellowship. Any opinions, findings, conclusions or recommendations expressed in this publication are those of the author and do not necessarily reflect the views of the National Science Foundation.
|
Robotics, Vision, and Sensor Networks Group
32 Vassar Street, 32-33xCambridge, MA 02139 Tel: 617-253-6583 Fax: 617-258-7413 |
 |