The presentation was "Ubiquitous Networked Sensing and Society" by Gregory J. Pottie of the UCLA Electrical Engineering Department. This was a joint meeting with the Los Angeles Chapters of IEEE Computer Society, IEEE Engineering Management Society, IEEE Society on the Social Implications of Technology (SSIT), and Computer Professionals for Social Responsibility (CPSR).

CENS is the Center for Embedded Network Sensors at UCLA, USC, UCR, CALTECH and CSU and is the leader in the field of embedded network sensor research.

Professor Pottie started out with a description of sensor network development in the past with LWIM III at UCLA in 1996, AWAIRS I at UCLA/RSC in 1998, Sensor Mote at UCB in 2000 and Medusa, MK-2, at UCLA NESL in 2002. LWIM III featured Geophone, RFM radio, PIC in a star network. AWAIRS I a Geophone, DS/SS radio, strongARM and multi-hop networks, Sensor mote has a RFM radio, Atmel, and TinyOS.

Geophones use accelerometers to sense vibrations in the ground (seismogram). Researchers are moving toward the use of Linux networks with heterogeneous nodes because TinyOS on its own is not sufficient for all applications of interest and is difficult to program. Professor Pottie presented a diagram of a sensor network where the sensor obtains on site readings and may do significant local processing and store some data there. Information is cued to be transmitted back to a human user who can do high resolution processing and maintain a data base. The user can revise queries and contact the sensor for more information. Low power sensors can provide a high duty cycle but with a high failure rate. Higher powers provide a low false alarm rate at a lower duty cycle as there are energy constraints on the system. Some users have relatively low data requirements, but science users usually want all the data.

Embedded networked sensors can use micro-sensors to do onboard processing, with wireless reporting feasible at very small scale and can monitor phenomena "up close". They enable spatially and temporally dense environmental monitoring. They can monitor the biocomplexity of ecosystems, contaminant transport, marine microorganisms and seismic structure response. Embedded network sensing will reveal previously unobservable phenomena.

The mission is to address scientific issues of natural and global priority through pioneering research and education, to develop and demonstrate architectural principles and methodologies for deeply embedded, massively distributed, sensor-rich distributed systems, to apply and disseminate these systems in support of scientific research critical to social and environmental concerns, and to create meaningful inquiry-based science instruction using embedded networked sensing technology.

Development of new embedded systems has proceeded as CENS Compatible chemical sensors are needed for soil/water quality monitoring, security, and precision agriculture. Current chemical-sensor development uses potentiometric or chronocoulometry electrochemical nitrate sensors, MEMS liquid chromatograph systems and mass spectrometers. Nitrate sensor developments in the short term use potentiometric nitrate sensors with detection limits in parts per million (ppm) and Amperometric nitrate sensors with detection limits in parts per billion (ppb). In the medium term Liquid Chromotology LC-on-a-chip provides separation and identification of ions. In the long term surface plasma resonance provides ultra sensitive detection.

Error resilient contaminant monitoring requires sensor network error reliability in complex media (air, water, soil). There has been a real life problem monitoring contaminants in water in Palmdale. There is a requirement for "real time" analysis and reporting locally instead of "logging". This requires model calibration and verification of forecasting capability. Wide area contaminant transport monitoring is required for larger scale, multimedia problems linking remote and in situ sensing over multiple scales. Complete watershed management requires visualization and exploration of massive heterogeneous data streams. Some of this will be covered by the NSF CLEANER initiative.

The Networked Info-mechanical Systems (NIMS) architecture uses Robotic aerial access to a full 3-D environment; enables sample acquisition and self awareness of obstacles; and provides for reconfiguration of sensors to calibrate and reduce model uncertainty. NIMS (also referred to as a "tree bot") enables speed and efficiency and also provides energy transport for a sustainable presence. The Infrastructure for sustained observations enables sustained surveillance or study; logistics are vastly simplified and it has been easy to install even in early generations. There has been continuous operation since late March 2004 (Piecewise continuous as there have been some failures causing interruptions).

Data Management is required to overcome multimedia, multiscale problems in time and space. The physical world can generate infinite data. Management, visualization, exploration of massive heterogeneous data stream is required and there are data integrity issues. There are few standards and they tend to be discipline/problem specific. Tools are being developed that enable many different types of users; engineers, scientists, teachers and students, to access the same basic data and to make the data extensible to new applications.

There are new directions to future projects in agriculture, security, theatre, biology, coral reef, global seismic grids/facilities and gaming. There are new types of macro programming and use of RFIDs. There is now robo gaming using real objects in motion which is beyond computer graphics as the play is done with autonomous robots. Physical capabilities and restraints replace simulated effects making the game more interesting.

ENS provides a revolutionary tool for basic science. Environmental science includes global warming, watershed management, effects of human activity on particular environments (including farms and forests). Better public policy can be provided for farm and forest management and better use of water resources. Contaminant monitoring is improved as air, soil and water environmental samples can be brought to biochemical analysis engines. There is the potential for improved public health and for lower cost remediation. Earth sciences provide for the study of seismic activity at multiple sites and determination of ground/structure interaction so improved building codes can be developed.

ENS supports a broad set of applications including interactive public spaces-interactive games and other entertainment experiences, automation of retail and many other service functions. Industrial automated tracking using RFID. Remote medical monitoring to provide better diagnosis and reduced cost. Security at borders, in urban areas and in homes. However: automated surveillance raises very large privacy concerns. Educational/ public awareness can be improved by providing the physical world as a set of various www sites. Providing individuals with information resources comparable to those of governments will have profound regulatory consequences.

Sensor Networks can cover a vast range of sizes from local to worldwide. The first phases of RFID tags are passive and the RFID readers are the sensor nodes of networks. However, RFID nodes with sensors and active communication are classic sensor networks. Some RFID capabilities are already be supplied by earlier technologies such as optical scanning of bar code for the type of goods and getting frequent shopper information. The stored database is used for inventory, tracking of buying habits, and the cash register. Credit and debt cards use a swipe of a magnetic strip on the card. Both bar codes and magnetic strips can be replaced by RFID.

There are social implications to RFID both good and bad. On the good side, consumers may be equipped with readers, may get additional data, price, consumer reports, manufacturing history, ingredient lists and be referred to alternative products, possibly with targeted advertising. They can act as triggers for service preferences and provide web access with higher data rates in obtaining information. RFID provides the opportunity to put shopping information in walls, including popup ads per "Minority Report" where the trigger was the iris of the eye.

RFID expands the possibilities of pervasive computing. RFID binds information to individual items (including people). It is not the first technology to do this as credit and debit cards do it now and it won't be the last. RFID simplifies collection of diverse information about the object and can distribute it widely with a communication infrastructure. Needs can be denoted by the services that are enabled by proximity, rather than merely reading the tags. Who controls the data is very important.

Tags and sensors can be used for ecology monitoring with tags on plants or animals. Cameras can be used for studying growth and sensors to determine environmental conditions. There is a need to be sure that readings are obtained on the same object so that accurate measurements over time can be accumulated.

The availability of RFID provides new capabilities and concerns about cameras. Cameras usually are not networked; most observational cameras have their tapes written over after short time periods. Reporting combined with tags has vastly different privacy implications. Face ID software is in infancy, but is definitely improving. Tag record identity can readily accumulate a record of activities with cameras by identifying persons so they can easily be tracked by multiple cameras. Information connected at many points tied to ID may provide massive surveillance at low cost, reducing barriers to use.

Embedded networks require responsibility in their use so that true public interest will be served. Information technology can be structured to make certain uses more probable, e.g., Internet privacy is a consequence of the original protocol produced by researchers. It is important that the technology to allow good privacy choices be designed into it at the beginning because later changes are difficult and expensive. RFID will require strict security of information to prevent unauthorized uses. Other uses will require a warrant or public debate. Kill switches under the control of customers should be provided for RFID tags on commercial products. Cameras could remove "bystander" information at source by providing silhouettes rather than images.

There is no doubt that pervasive monitoring is coming because there are powerful economic incentives (such as Wal-Mart) and there are legitimate government and private security concerns. The question is how will it be regulated. If data privacy is inherent in architecture, third party access will require more effort. We need public discussion to set these limits, but so far we haven’t had public discussion so decisions are not likely to serve the broader public interest. State and Federal regulators have already held hearings (a good first step) but will legislation be reactive or proactive?

In conclusion, the connection of the physical world to computer networks will have profound societal consequences and adding actuation enables remote control. Tools developed for basic science investigation will result in not only new science but support for other applications. Robustness, sustainability, data management, platform and software architecture issues are similar. Now is the time to consider the regulatory regime as changes to assure desired societal outcomes become increasingly expensive after systems are widely deployed.

You can find out more about networked sensors at: www.cens.ucla.edu

There are many paths at this site that lead to detailed discussion of sensors, networks and their applications. In particular the CENS Annual Progress Report for May 1, 2003 to April 30, 2004 contains much interesting information.

Dr. Pottie provided an excellent description of networked sensors and the uses of RFID tags. Our LA ACM meeting length is too short to do justice to the questions raised by his presentation. We could have used about a two hour discussion period on the many points developed. The meeting was enjoyable and very thought provoking.