Telecons

From NESTFE Wiki

Jump to: navigation, search
[edit]

All Teams Telecon for NEST Final Experiment

  • Who: ALL (DARPA, Kestrel, MIT, OSU, SRI, UCB, UIUC, UVa, and VU)
  • What: Weekly telecon to ensure smooth progress
  • When: Every FRIDAY, 11AM PST
    • July 29, 2005
    • August 5, 2005
    • August 12, 2005
    • August 19, 2005
    • August 26, 2005
  • Where: 410 Soda (UCB)
  • How:
    • Dial: (800) 310-8221
    • Passcode: 277532


[edit]

July 25, 2005 - Telecon Highlights

  • XSM Patch A: UVa will deploy along North/South road (purple box)
    • UVa: 85K; 20 nodes
    • CMU: ??K+flash; 20 nodes
  • XSM Demo B: South/East corner (blue box)
    • Kestrel: ??K; 100 nodes
    • Illinois: <64K; 100 nodes
    • MIT: 60K-80K; 64 nodes
    • SRI: <64K; 50 nodes
    • VU: 45K; 16-30 nodes
  • TRIO Patch A:
    • OSU: 21x7 grid of nodes needed
    • UCB: Part of larger UCB
    • Notes: Berkeley will provide OSU with a Deluge slot for their code image.
  • Programming Issues
    • ALL TEAMS MUST USE THE TinyOS 1.1.14 RELEASE FOR THE NETWORK STACK AND DELUGE FOR COMPATIBILITY REASONS
    • Subset Activation Strategy: Each team is responsible for identifying the subset of nodes it wants to use out of the 100 deployed XSMs and include code to read the TOS local address and decide if that node is to participate.
  • Action Items
    • UCB: Deployment area survey (grass height, etc.)
    • UCB: Radio results for Trio to OSU
    • UCB: Details on synching to TinyOS 1.1.14
    • ALL: Use the Wiki to update details pertaining to your demos BY 2005-07-29 Telecon
    • ALL: Next telecon is Friday, July 29, 2005. Call in details above.
    • OSU is encouraged to load Berkeley's Kraken library, which pulls in all code used on the Trio. Description of Kraken can be found at http://today.cs.berkeley.edu/nestfe/index.php/Software_Architecture
    • Localization Teams: August 15 Telecon with VU to work out details.
[edit]

June 6, 2005 - Telecon Highlights

Going over demo plans

  • Vanderbilt radio ranging
    • Branislav (Brano) Kusy
    • 4" tall grass is best, but untested in larger grass
    • One base station vs many base station
    • ... on average 30m radio range, placed randomly / casually / arbitrarily
    • Planned deployment in the sothern green area
    • Nodes on tripod another option
    • How accurate does ground truth need to be?
      • Their ranging accuracy right now is a few centimeters
      • We would need to do this with tape measurements?
      • One-meter accuracy for ground truth may be sufficient
      • Measure some subset very accuractely
    • Just demoing ranging itself
    • Produces a list of ranges per node
    • It's 4-wise ranging not pairwise ranging,
  • Vanderbilt GRATIS not on-field, just in a room
    • They will bring their own nodes
  • UIUC
    • Distirbuted localization
    • ~200 XSM's v3
    • ... on average 30m radio range, placed randomly / casually / arbitrarily
    • 420 meters square on edge, not too much grass
    • No sensing?
    • UIUC and Kestrel doing localization based on Vanderbilt ranging
    • Will run Vanderbilt code (not use pre-acquired data by Vanderbilt)
      • Really? Discussion on this.
      • Takes time to do ranging, and is redundant for centralized localization systems
      • But a "whole system" is a compelling demo
  • Kestrel
    • Distributed localization
    • UIUC setup works for them as well
    • Want GPS ground truth per node
    • Seeding with anchors (GPS? Use ground truth?)
  • SRI
    • Base-station authentication scheme
    • ... for a base station to disseminate commands and programs
    • Only need a multihop XSM network
  • MIT
    • Tracking
    • Working with Mitre for acoustic ranging
    • XSM's
    • Looking at 8x8 with 20m spacing
    • Grass height?
      • "factor of 3 range difference between parking lot and high grass" -Akos
  • UVA
    • Not demoing spotlight
    • Classification, power management, hardware "in the loop"
    • Low grassy area, at most 10m spacing, a road to deploy around is good
    • Bring their own nodes
    • Need a large and small vehicle
    • Walking GPS for ground truth
  • OSU
    • Use a portion of the Trio deployment
    • Want ~200 Trio's, something like a grid
    • 10m spacing requirement
    • Comm-site testbed
      • acquiring data remotely and sending it to the far-remote base station
    • Cocktail of sensing, need vehicles
  • Vijay prefers
    • One XSM testbed, not many
    • Ten minutes transition time between demos
      • David recommends a pipeline, means we need two testbeds
      • Demonstrate Deluge dissemination if it fits witihn the timeline
  • Testbeds
    • Trio testbed (Berkeley, OSU)
    • VU XSM testbed (VU, UIUC, Kestrel)
    • UVa XSM testbed (Virginia)

Other questions

  • Do we need portable power?
  • Tents?
  • Dry run the weekend before for a dress rehersal
  • Leave out XSM's and Trio's (XSM power?)
  • Berkeley and OSU will have two-way telecons about Trio and Tier-2, etc

Agenda

  1. Vanderbilt - ~200 XSM, radio ranging
  2. UIUC - VU XSM, localization
  3. Kestrel - VU XSM, localization
  4. SRI - whatever XSM, secure multihop command dissemination with authentication
  5. MIT - ~60 XSM, tracking, acoustic ranging

Order

  1. Berkeley
  2. Vanderbilt
  3. Kestrel
  4. Virginia
  5. UIUC
  6. OSU
  7. SRI
  8. MIT

--- Cory Sharp 07:35, 6 Jun 2005 (PDT)

Retrieved from "http://nest.cs.berkeley.edu/nestfe/index.php/Telecons"