vigilnet an integrated sensor
play

VigilNet: An Integrated Sensor 70 Mica2 motes along 280 foot long - PDF document

Prototype System VigilNet: An Integrated Sensor 70 Mica2 motes along 280 foot long perimeter Network System for Energy- Equipped with 433 Mhz. Chipcon short range Efficient Surveillance (<1000 ft) radio, sensor board with magnetic


  1. Prototype System VigilNet: An Integrated Sensor � 70 Mica2 motes along 280 foot long perimeter Network System for Energy- � Equipped with 433 Mhz. Chipcon short range Efficient Surveillance (<1000 ft) radio, sensor board with magnetic sensor � Laptop base station for visualization Eric Nicks � TinyOS running on each mote � Entire system consists of 40,000 lines of NesC and Java code allegedly downloadable from Sourceforge Requirements Architecture � Longevity � Confidential nature � Inaccessibility � Adjustable sensitivity � Environmental, security, criticality � Stealthiness � Size, radio frequency detection � Effectiveness � Location, latency Execution Cycle Phase II: Time Sync � GPS obvious solution, but is large and expensive � Reference broadcast scheme –Phase and frequency difference between clocks shared between motes by message sharing � Not selected due to overhead and energy use � Flooding time synchronization protocol selected � modified to run during phase II only 1

  2. Phase III: Localization Phase IV: Asymmetric Detection � Walking GPS � Asymmetric links – Different reception rates along different directions between same pair of motes � Installer wears GPS device � Link Symmetry Detection � Motes infer position based on beacons sent out by � Designed to reduce irregularity for upper network layers wearable GPS device � Local Beaconing � VigilNet created prototype system Each node has an ID � � Accurate up to 0.8 +/- 0.5 meters, studied to be Beaconing node adds all discovered neighbors to beacon � sufficient for routing, sensing, and tracking When node receives beacon, it adds the node to its list � Checks to see whether its own ID is on the list, if so this is a � symmetric link, otherwise asymmetric Phase V: Backbone Creation Phase VI: Sentry Selection � Sentry mote performs reconnaissance duty � Diffusion Tree � 2 conditions for sentry selection � Time synchronization messages used to distribute information needed to create routes back to � Internal node of diffusion tree basestation � No neighboring sentries � Creates network from base station to leaf motes � Contention: Sentry declaration message with � Updates to the tree allow for mote failure delay on each mote detection � Energy balancing: Delay based on energy � New routes updated during each execution � Sensing coverage: Declaration message range smaller than magnetic sensor range Phase VIII: Phase VIII: Event Tracking and Power Reporting Management � Problem: Many motes are going to track the vehicle as it passes through the mote path, creating increased traffic and energy � Proactive consumption � Periodic sleep commands sent by sentry to each neighboring mote with timer associated with data � Solution: Report Aggregation � Non-sentry motes receive the sleep commands, � Groups of motes formed dynamically in response and sleep the length of the timer to an event, and migrate when an event moves � Reactive � A group leader reports to the base station. � Transition between sleep and wakeup controlled by timer. Chosen for VigilNet 2

  3. Evaluation of In-Network Aggregation Critique � Still in a semi-controlled environment � Moisture and temp � Base Station not accounted for in experiment � Animal reclamation � Need to develop protective housing � Gateway to long distance command and control not implemented � No dynamic human configuration performed � Magnetometer sensor unable to sense human presence, recommend additional RF, sound, or visualization (activated based on other sensors) system as a backup � Has to be a better more dynamic solution to localization than walking GPS. Improper results could be found if the sensor is moved by wind, water, animal, etc. � No real study for effects to performance when multiple sensors “die” Comparison � Still a study, not a real world application like Great Duck Island � Energy savings in Great Duck Island could have been quite significant with the reactive power scheme of this experiment � 34 day median life for burrow motes in Great Duck Island will not work for military � External tools mentioned as need in Great Duck Island paper to allow non-experts to install the motes. Same things will be needed for these motes if 18 year old soldiers will be installing these surveillance networks in field 3

Recommend


More recommend