Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
eScholarship provides open access, scholarly publishingservices to the University of California and delivers a dynamicresearch platform to scholars worldwide.
Center for Embedded Network SensingUniversity of California
Title:GeoNet: A Platform for Rapid Ditributed Geophysical Sensing (SYS 12)
Author:Martin Lukac; Vinayak Naik; Allen Husker; Igor Stubailo; Paul Davis; Deborah Estrin
Publication Date:01-01-2006
Series:Posters
Permalink:https://escholarship.org/uc/item/7bx706ft
Additional Info:BibTex @misc{ title={GeoNet: A Platform for Rapid Ditributed Geophysical Sensing (SYS 12)},author={Martin Lukac, Vinayak Naik, Allen Husker, Igor Stubailo, Paul Davis, Deborah Estrin},abstract={We will provide TDMA-based state-synchronization service that will provide low latencyand high reliability. Our TDMA service will be built on top of the existing MAC found inthe radios for Aensbox. We will provide deployment tools, which will enable the deployer togauge the quality of the deployed network. Our tools will use above mentioned efficient andreliable state-synchronization service and will work without assuming any end-to-end routingservice. We will provide system software for Aensbox, which a multi-processor and multi-radio platform. The desired system software must utilize dual processor and radio dependingupon the energy consumption of the operation, e.g. a continuous battery monitoring operationcan be delegated to a low power processor (MSP), while analysis of a large data-set willbe done by high power processor (XScale). Although, comprehensive evaluation of the entiresystem can be done only at the real scale, it is necessary to gain confidence at early stagesof development where real deployment is not possible.}, url={http://research.cens.ucla.edu/pls/portal/url/item/20A7A437D3C08C53E0406180528D49CF}, year={2006}, }
Abstract:We will provide TDMA-based state-synchronization service that will provide low latency and highreliability. Our TDMA service will be built on top of the existing MAC found in the radios for Aensbox.We will provide deployment tools, which will enable the deployer to gauge the quality of thedeployed network. Our tools will use above mentioned efficient and reliable state-synchronizationservice and will work without assuming any end-to-end routing service. We will provide systemsoftware for Aensbox, which a multi-processor and multi-radio platform. The desired systemsoftware must utilize dual processor and radio depending upon the energy consumption of theoperation, e.g. a continuous battery monitoring operation can be delegated to a low powerprocessor (MSP), while analysis of a large data-set will be done by high power processor (XScale).Although, comprehensive evaluation of the entire system can be done only at the real scale, itis necessary to gain confidence at early stages of development where real deployment is notpossible.
eScholarship provides open access, scholarly publishingservices to the University of California and delivers a dynamicresearch platform to scholars worldwide.
Copyright Information:All rights reserved unless otherwise indicated. Contact the author or original publisher for anynecessary permissions. eScholarship is not the copyright owner for deposited works. Learn moreat http://www.escholarship.org/help_copyright.html#reuse
Problem Description: Problem Description: Construct a hardware-software platform for Construct a hardware-software platform for GeoNetGeoNet
Proposed Solution:Proposed Solution: A platform composed of suitable hardware and software for A platform composed of suitable hardware and software for GeoNetGeoNet
GeoNetGeoNet: A Platform for Rapid: A Platform for RapidDistributed Geophysical SensingDistributed Geophysical Sensing
Martin Lukac, Vinayak Naik, Allen Husker, Igor Stubailo, Richard Guy, Paul Davis, Deborah EstrinUCLA – CENS – http://research.cens.ucla.edu
Introduction:Introduction: GeoNetGeoNet to predict time and locations of earthquakes to predict time and locations of earthquakesAnatomy of an earthquake
• An earthquake = (Foreshocks , Main earthquake , Aftershocks)– Foreshocks: may occur before an earthquake, low value on the Richter scale– Main earthquake: identified when the vibrations reach a peak value– Aftershocks: a series of shocks following the main earthquake– The magnitude of the aftershocks reduce as time passes
Existing models of earthquakes
Requirements of the GeoNet platform
AENSbox from Aevena A service to scheduletransmissions in time
UCLA UCLA –– UCR UCR –– Caltech Caltech –– USC USC –– CSU CSU –– JPL JPL –– UC Merced UC Merced
Center for Embedded Networked SensingCenter for Embedded Networked Sensing
• Concentric circle model– Adequate for large-field motions
• Tree (Markov chain) model– More accurate for near-field motions: suitable for earthquake predictions– However, yet to be corroborated by spatially extensive measurements
Use of WSN to study tree model• Rapidly deployed large-scale WSN to measure vibrations (GeoNet)
– Quickly deploy nodes after foreshocks– Deploy 100-1000 nodes at a separation of 0.5-1Km
• Technology applicable to structural monitoring
• A need for tools to enable rapid deployment of WSN– Limited time: A limited time is available to deploy node after noticing foreshocks– No routing: Tools such ping and iperf do not work without routing in a multi-hop network– No fairness: Existing tools do not guarantee fairness while collecting network statistics
• A need of power-efficient wireless sensor node– Earthquakes can cause damage: GeoNet nodes need to last at least for a week on batteries– Existing nodes are not energy efficient: they use high power processor for simple operations
and high power radios for small control packets• A need of network-time in the absence of GPS reception
– GPS is unavailable in tunnels, canyons, and in buildings– A network time with accuracy greater than 1msec is required
• AENSbox– Multi-radio and multi-processor based node– Power-efficiency via selection of optimal radio
and processor• Radios:
– IEEE 802.11 and IEEE 802.15.4– Use of 802.15.4 radio for short beacon packets– Use of 802.11 radio for long data packets
• Processors:– Intel PXA255 XScale: for analyzing,
compressing, and transmitting sensory data tosink node
– 2x Texas Instrument MSP430F1611: control802.15.4 radio and sample seismic sensor data
– Combination of processors and radios allows forsleep scheduling of XScale processor for greaterpower savings
• Transmissions scheduling is suitable to carrytraffic generated by rapid deployment tools– All nodes need to forward link and state information– All the transmissions occur in a periodic manner– Provides fairness: each node receives equal share of
medium access
Rapid deployment tools forlarge-scale WSN
• Tools to visualize network statistics– From all nodes in real time– Statistics such as ETT, ETX, node health,
estimated bandwidth, etc.
System software for Aensbox• Duty-cycle processors and radios for energy
efficiency, e.g. duty cycling Xscale processor– XScale processor to sleeps, while MSP processor
sampls and stores data in serial flash– Wakeup XScale processor when serial flash is full– XScale processor transfers data to disk or forwards
data towards sink
Time synchronization service• Provides network-wide time
– Use of synchronous reception property ofwireless communication to synchronize time
– Useful where GPS is unavailable
Concentric circle model Tree model
Deployed AENSbox
• Data sampling and synchronizationbetween two AENSboxes:
Geophone
AENSbox GPS
WIFI802.15.4
AENSbox Data Sampling
Stargate-2 as analternative hardware
• Multiple radios viz. 802.11, 802.15.4,and bluetooth