Introduction to Wireless Sensor Networks Smart Dust 4

Introduction to Wireless Sensor Networks Smart Dust 4 April 2005 The University of Iowa. Copyright© 2005 1 A. Kruger, R. Abel, C. Mueller, M. Karson

Imagine if you will… • Two opposing military forces, Alpha and Omega, are separated by a portion of jungle. • Each wants to locate and identify enemy positions and movements. • Alpha wants a safer, more efficient means of performing reconnaissance α – Human resources for intelligence gathering are non-optimal • Costly Ω – Money – Human life • Human error • Non-persistent The University of Iowa. Copyright© 2005 2 A. Kruger, R. Abel, C. Mueller, M. Karson

Deployment • Army Alpha deploys an unmanned aerial vehicle – Ejects tens of thousands of various kinds of rice sized motes • Terrestrial based • Air based • Water based • Motes automatically form a sensor field – Light, temperature, vibration, radar, magnetic, acoustic, seismic or a miniature camera. The University of Iowa. Copyright© 2005 3 α α α Ω A. Kruger, R. Abel, C. Mueller, M. Karson

Effect of Sensor Network • Army Omega dispatches intelligence officers and equipment into sensor field αα α α α α α α Ω Ω The University of Iowa. Copyright© 2005 4 α αΩ αα α α α Ω A. Kruger, R. Abel, C. Mueller, M. Karson

What is Smart Dust? • Cute name for a network of miniscule motes – Term “smart” comes from abilities of individual motes as well as overall function of network – Term “dust” comes from the goal of packaging a fully functional mote in a 1 mm 3 package • Project started at the University of California at Berkeley – Funded by DARPA (Defense Advanced Research Projects Agency) • Most research aimed at military and defense applications The University of Iowa. Copyright© 2005 5 A. Kruger, R. Abel, C. Mueller, M. Karson

Vision • Think pixie dust - Scatter hundreds of sensors which are nearly un-noticeable • The size of a grain of sand complete with sensors, CPU, receiver, transmitter, antenna and a power supply • Communication ranges of 1000 ft. or more The University of Iowa. Copyright© 2005 6 A. Kruger, R. Abel, C. Mueller, M. Karson

Long Term Goals of Project • Autonomous sensing and communications in 1 mm 3 • Optimize every aspect of WSN – Battery life ( several years ~5 ) – Size (1 mm 3) • Contains all elements of the mote – Range • Some sources predict up to 1 km – Processing power • On board motes • In networking messaging • Billions of computations requiring only picowatts (10 -12) – Communications • Laser – Power Consumption – Deployment • “Floating” motes • UAV deployment The University of Iowa. Copyright© 2005 7 A. Kruger, R. Abel, C. Mueller, M. Karson

History • Invented by Kris Pister (University of California, Berkley) in 1992 • Smart Dust started as a joke when everyone was talking about smart homes, smart buildings, smart bombs… • Smart Dust was the start of WSNs – In 1994 Pister started his research on Smart Dust and began developing Motes (Hardware) – ~2001, Jason Hill, and David Culler (both at Berkley) worked together to develop Tiny. OS for Pisters hardware. The resulting mote was called: MICA – [Tiny. OS let] the mote’s hardware perform only critical functions, which in turn extends the mote’s lifetime – “It’s all about energy. ” (Pister) • Partner in Dust Inc with Jason Hill (2002). The University of Iowa. Copyright© 2005 8 A. Kruger, R. Abel, C. Mueller, M. Karson

Are we there yet? • Short answer, not quite – Minute motes have been developed in academic labs – Larger motes have been used in WSNs • How close? – Dust™ Networks is trying to produce practical motes that are approaching the size of an Aspirin pill – Package size seems to be main hurdle The University of Iowa. Copyright© 2005 9 A. Kruger, R. Abel, C. Mueller, M. Karson

Problems with Size • Package size – Need to integrate sensor, CPU, transmitter, receiver, antenna onto a single chip – Currently size is about 5 mm cube – Dust Inc mote is 1 inch square The University of Iowa. Copyright© 2005 10 A. Kruger, R. Abel, C. Mueller, M. Karson

A case study SPEC • The first Single Chip Mote – – – 2 mm 2. 5 mm AVR-like RISC core 3 k memory 8 -bit on chip ADC FSK transmitter (19, 200 kbps @ 40 ft) SPI programming • Serial Peripheral Interface (For in-system programming) – RS 232 compatible UART – 4 -bit input port, 4 -bit output port – $0. 30 in quantity The University of Iowa. Copyright© 2005 11 A. Kruger, R. Abel, C. Mueller, M. Karson

CPU Size/Power Considerations • RISC processors – employed due to their small die size, and their ability to run in low power modes. – Code density is of crucial importance • The ARM 7 TDMI is a 32 bit processor with an additional 16 bit instruction set – The instruction set can be switched by the software to adapt to current circumstances. – Power Saving Solutions • Active – Fixed Frequency – Frequency Scaling – Dynamic Voltage Scaling (DVS) • Power Saving (i. e. sleep, hibernate…) The University of Iowa. Copyright© 2005 12 A. Kruger, R. Abel, C. Mueller, M. Karson

Problems with Programming • Mass programming – Smart Dust networks may involve thousands of nodes – Programming them individually is not practical • Embedded systems solution – Update firmware • Wirelessly • Automatically • When update available The University of Iowa. Copyright© 2005 13 A. Kruger, R. Abel, C. Mueller, M. Karson

Problems with Cost • Manufacturing costs increase as size decreases with computer chips • Large scale networks – The cost of each mote must be very small for costs of a practical system to remain realistic – Predictions are $1/mote within 5 years The University of Iowa. Copyright© 2005 14 A. Kruger, R. Abel, C. Mueller, M. Karson

Power Consumption Solutions • Ultralow-Energy ADC – Sampling Rate of 100 k. Hz – Power dissipation is 3. 1 μW – Standby power is 70 p. W – Energy per 8 -bit sample is 31 p. J • 1 k. WH = 3. 6 million J – Die area is 0. 053 mm 2 • Used onboard mote shown in previous The University of Iowa. Copyright© 2005 15 A. Kruger, R. Abel, C. Mueller, M. Karson

Zero Power Communication • Optical communication is possible using Microactuators (MEMS) (Karakehayov). – Active-Steered Laser Systems • Needs power to generate a beam – Passive reflective systems • Can modulate an existing beam using very little power • Can be done with a Corner Cube Retroreflector (CCR), three mutually orthogonal mirrors • Modulation is accomplished by slightly turning a mirror such that the light is no longer reflected towards the information sink • Mirror rotation can be accomplished 1000 times per second at a cost of less than one nano. Joule per transition. • CCRs can be roughly oriented using a magnetic compass The University of Iowa. Copyright© 2005 16 A. Kruger, R. Abel, C. Mueller, M. Karson

The Sleep-Awake Protocol • Uses 2 Modes of Laser Communication – Broadcast Beacon Mode (low energy short length communication) – Point Directed Mode (data transmission) • Assumptions – No geolocation capabilities assumed (GPS) – No communication (transmitted or received) during sleep cycle, sensors may be active The University of Iowa. Copyright© 2005 17 A. Kruger, R. Abel, C. Mueller, M. Karson

The Protocol • Search Phase: Uses a periodic low energy broadcast of a beacon of angle towards the wall in order to discover a particle nearer to the Wall than itself. • Direct Transmission Phase: 2 Sends info( ) to 3 via a direct line (laser) and sends a success message to 1 (i. e. the particle that it received the information from). • Backtrack Phase: If the Search Phase fails to discover a particle nearer to , then sends a fail message to. The University of Iowa. Copyright© 2005 18 A. Kruger, R. Abel, C. Mueller, M. Karson

Analysis • This Technique is quite new and a thorough comparison is not available. • BUT – Sparse Topology and Energy Management (STEM) uses a similar technique (actively puts nodes to sleep) and performs nearly two orders of magnitude better then Sensor Networks without Topology Management The University of Iowa. Copyright© 2005 19 A. Kruger, R. Abel, C. Mueller, M. Karson

Possible Applications • Military applications – Remote vehicle & personnel sensing/monitoring – Missile guidance • Civilian applications – Ambient environment monitoring – Long range, ubiquitous communications – Power grid monitoring and maintenance • Boost power transmission The University of Iowa. Copyright© 2005 20 A. Kruger, R. Abel, C. Mueller, M. Karson

Sources • Scott, M. D. , Boser, B. E. , Pister, K. S. J. , “An ultralow-energy ADC for Smart Dust”, IEEE Journal of Solid-State Circuits, V. 38, Issue 7, July 2003, pgs 1123 -1129 • Karakehayov, Z. ; “Zero-power design for Smart Dust networks”, Intelligent Systems, 2002. Proceedings. 2002 First International IEEE Symposium, Volume 1, 10 -12 Sept. 2002 Page(s): 302 - 305 vol. 1 • Chatzigiannakis, I. ; Nikoletseas, S. , “A sleep-awake protocol for information propagation in smart dust networks”, Parallel and Distributed Processing Symposium, 2003. Proceedings. International 22 -26 April 2003. • Frost Gorder, P. , “Sizing up smart dust”, Computing in Science & Engineering, Volume 5, Issue 6, Nov. -Dec. 2003 Page(s): 6 - 9 The University of Iowa. Copyright© 2005 21 A. Kruger, R. Abel, C. Mueller, M. Karson

Thank You The University of Iowa. Copyright© 2005 22 A. Kruger, R. Abel, C. Mueller, M. Karson