# Sastry NEST meeting notes
# Date: 2/20/2004
# Present: Phoebus, Tanya, Cory, Bruno, Bonnie, Mike, Sastry, Luca
#   Massimo, Dave Molnar, Terrence Tong, Yang? (undergrad working with Joe 
#   Polastre, I'm sorry I missed the name), Mark Myers and Clas Jacobson 
#   from United Technologies

- Cory and Bruno gave a dry run of the presentation for VJ on 2/23 next Monday
  * They summarized the status of the NEST project to the UTC visitors and
    the new students.
  * Will need to do the presentation on March 2nd at Stanford.

- Bald Camel Test Update (professor Sastry)
  * Purpose of demo: To assess how to integrate MIR sensor with sensor networks
    ~ Crossbow had difficulty resolving the eletromagnetic issues of the MIR 
      sensors with the radio interference, etc.
    ~ Advantaca modified the mica mote and integrated it with GPS and the MIR 
      sensor, then packaged it nicely to be deployed from flying vehicles
      (foam covering, strong plastic encasing, gimbal system to ensure antenna
      was always upright).
    ~ Test was also meant to assess issues with flight deployment of the sensors
  * Sensor calibration and multitarget tracking testing on Monday and Tuesday
    ~ Song managed to run target tracking successfully for one object.
      There were problems with two objects crossing paths - not enough resolution
      for sensor report time stamps and  tracking
  * Flight drop demo took place on Thursday 2/19
    ~ interesting to see helicopter drop two line of motes... unfortunately,
      the motes were deployed around 400 meters off target, and was out of radio
      range with the base station.
    ~ after moving the base station to within range, only half the sensors 
      reported back.  This was because some of the wiring on flexible circuit
      connectors was sheared off from the impact of the sensors hitting the floor.
      (Apparently, the nodes bounced 10-15 feet in the air after impact)  After 
      fixing those torn connections, those silent nodes started reporting again.

- Meeting time will be moved to Friday, 2 PM in the Robotics Lab



Post Meeting notes:
*******************
  The meeting ran out of time to do status updates.  This is from Phoebus's
understanding of the situation.

- Testbed Hardware update (Phoebus)
  * Received enough parts to make prototype board.  Built and tested Prototype.
    ~ Found wiring bugs to 6 pin header
    ~ Must remove unnecessary 1 mOhm resistor on reset pin to program with 
      eMote
  * Fixed design. Ready to fabricate boards, but put on hold because will be
    doing Telos mote integration as well.
  * Meeting with Joe Polastre on Telos accomodations.
    ~ Will need to add additional 10 pin molex header to connect to the Telos mote.
      Possibly will need 12 pin connector (trying to minimize since Telos mote
      is running out of space for connectors). Joe will talk to Rob and get back to
      me.
    ~ Will also need to add 5V booster circuitry to pinOut board to bypass the
      mica2dot power board.  It is necessary to move the 5V booster circuitry
      if we wish to run the pinOut board entirely using the battery on the Telos.
    ~ Kamin is apparently trying to change the Ultrasound board to use I2C instead
      of UART for data transmission.


- Testbed Hardware update (Tanya)
   * will draw up slides of plan for Bruno's presentation on Monday.


- Setting up Base Station (Mike)
  * Debian Linux has been installed on Fauvre.eecs


- Robot Localization (Mike)
  * Found two potential "canned solutions" to getting the ground truth on robot locations using
    cameras.
    1) http://playerstage.sourceforge.net/mezzanine/mezzanine.html
       ~ 1 camera, no accuracy claims
    2) http://www-ra.informatik.uni-tuebingen.de/publikationen/2003/lilien03-rose2003a_col.pdf
       ~ 3 cameras, claims accuracy up to 1 cm
  * Mike will pursue further... ask for source code from those projects.


- Testbed Software
  * First step is to get a baseline implementation of 1 hop logging and aggregation of 
    magnetometer readings back to the base station.  This will allow us to figure out the
    distance to place the steel battery from the mica2dot sensors.
    ~ need to integrate Sukun's software in tinyos-1.x/contrib/GGB with magnetometer stack in
      tinyos-1.x/contrib/PEGSensor