Potential Applications Examples of Nodes Architecture Issues Protocols Early Ideas Summary EECS 122 University of California Berkeley EECS122 Contents Index SENSORS Slide from D Estrin Slide from D Estrin Micro sensors onboard processing and wireless interfaces all feasible at very small scale Seismic Structure response Marine Microorganisms EECS122 Contents Index can monitor phenomena up close Will enable spatially and temporally dense environmental monitoring Embedded Networked Sensing will reveal previously unobservable phenomena Middleware Framework Network Sensor Sensor Speech Database Management Management Fusion Recognizer Data Miner WLAN Access Wired Network Point High speed WLAN Piconet Bridge Piconet Contaminant Transport Wireless LAN WLAN WLAN Piconet Bridge Sensors Modules Piconet Sensor Badge Ecosystems Biocomplexity Networked Toys Disaster Response SENSORS Applications EECS122 Contents Index SENSORS Applications Slide from Mani Srivastava Slide from Mani Srivastava Rockwell Hidra 3 5 x3 5 x3 StrongARM 1100 processor 133 MHz Connexant s RDSSS9M Radio 100 kbps 1 100 mW 40 channels Various sensors on factory floor machinery Monitor Vibrations temperature etc http wins rsc rockwell com EECS122 Contents Index Berkeley Motes Atmel microcontroller temperature light humidity pressure 3 axis magnetometers 3 axis accelerometers 10kbps 20m SENSORS Nodes UCLA iBadge Wearable Sensor Badge acoustic in out DSP temperature pressure humidity magnetometer accelerometer ultrasound localization orientation via magnetometer and accelerometer bluetooth radio Sylph Middleware EECS122 Contents Index UCLA Medusa Localizer Node 40MHz ARM THUMB 1MB FLASH 136KB RAM 0 9MIPS MHz 480MIPS W RS 485 bus Out of band data collection formation of arrays 540mAh Rechargeable Li Ion battery SENSORS Nodes Slide from Mani Srivastava BWRC s PicoNode TripWire Sensor Node Slide from Jan Rabaey Solar Cell PCB 1 0 5 mm Battery 3 6 mm mm IPaq Sensor Node Chip encapsulation 1 5 mm 7 6 mm m 5c 3c m 5 66 0 8 0 Version 1 Light Powered 1 0 0 9 0 10 0 0 9 80 01 23 4 0 1 27 Components and battery mounted on back Version 2 Vibration Powered EECS122 Contents Index SENSORS Nodes EECS122 Contents Index Sensor Network architecture 0 0 SENSORS Nodes Key tasks performed by Sensor Nodes Sensing Data Processing Communication Networking is a distinguishing feature A Survey on Sensor Networks I Akyildiz W Su Y Sankarasubramaniam and E Cayirci IEEE Comm Magazine August ANSI IEEE Std 802 11 1999 Edition EECS122 Contents Index A Survey on Sensor Networks I Akyildiz W Su Y Sankarasubramaniam and E Cayirci IEEE Comm Magazine August ANSI IEEE Std 802 11 1999 Edition EECS122 Contents Index Low Power special protocols Ex Wake up periodic sleep Differences between Sensor Networks and traditional Mobile Ad Hoc Networks MANETs Much lager number of nodes Dense deployment Limited in power Mainly use broadcast paradigm Higher failure rate No global identifiers Small Memory specialized code Ex TinyOS Sylph Localization GPS triangulation Synchronization GPS NTP Addressing Naming Location based Function based Directory Hash Routing Broadcast multicast anycast mobility Transport Intermittent connectivity noisy links error tolerance Observation Protocols should be application specific not generic EECS122 Contents Index EECS122 Contents Index SENSORS Issues Slide from D Estrin Slide from D Estrin User Queries External Database Resource constraints call for more tightly integrated layers In network Application processing Data aggregation Query processing Open Question Data dissemination storage caching Can we define an Internet like architecture for such applicationspecific systems Adaptive topology Geo Routing MAC Time Location SENSORS Protocols Extent Spatial Density of sensors relative to stimulus Data rate of stimulii Frequency spatial and temporal density of events Efficiency System lifetime System resources Locality spatial temporal correlation Variability Ad hoc vs engineered system structure System task variability Mobility variability in space Resolution Fidelity Detection Identification Mobility Rate and pattern Latency Response time Robustness Vulnerability to node failure and environmental dynamics Autonomy Multiple sensor modalities Computational model complexity Scalability Resource constraints Over space and time Energy BW Storage Computation EECS122 Contents Index SENSORS Protocols Special Features in Sensor Networks Event and Query Find data somewhere in network Broadcast a request Collect data from all nodes Examples of routing algorithms Grab Gossip Ant Directed Diffusion Data Centric Rumor Most slides from D Estrin 00 Routing Finding Energy Location EECS122 Contents Index Spatial and Temporal Scale Phy comm sensing actuation SP EECS122 Contents Index SENSORS Ideas PA Power Available at a node i Energy used to transmit over link i Energy efficient routing Maximum PA route SinkA B C T Minimum energy route Sink A B T Minimum hop route SinkD T Maximum minimum PA node route Sink D T EECS122 Contents Index 1 2 SENSORS Ideas Routing 31 24 Builds a cost field toward a particular node then reliably routing queries across a limited size mesh toward that node Overhead of network flood Queries route along short paths Delivered cheaply and reliably Not designed specifically to support in network processing A Survey on Sensor Networks I Akyildiz W Su Y Sankarasubramaniam and E Cayirci IEEE Comm Magazine August ANSI IEEE Std 802 11 1999 Edition EECS122 Contents Index EECS122 Contents Index SENSORS Ideas Routing Grab 1 Nodes flood by sending message to some of neighbors By the redundancy in the link most nodes receive the flooded packed Used to deliver query or flood events Less overhead than conventional flooding Not be designed specifically for energy constrained contexts EECS122 Contents Index SENSORS Ideas Routing 00 Gossip 1 5 Provides a mechanism for doing a limited flooding of a query toward the event Set reverse gradients to send data back along the best route Results in high quality paths But requires an initial flooding of the query for exploration EECS122 Contents Index SENSORS Ideas Routing Directed Diffusion EECS122 Contents Index SENSORS Ideas Agent traverses the network encoding the quality of the path they have traveled and leave the encoded path as state in the nodes At every node an agent picks its next hop probabilistically biased toward already known good paths Very effective in dealing with failure because always some amount
View Full Document
Unlocking...