Title

A Study of Vocalization and Social Behavior of the Acorn Woodpecker (Melanerpes Formicivorus) Based on the Remote Sensor Network

Authors

Yuan Yao, Ying Lin, Hanbiao Wang, Edward Stabler, Kung Yao, & Charles Taylor

Adaptive Language Group

Abstract

The acorn woodpecker (melanerpes formicivorus) is a highly social species with co-breeding behavior. The vocalization of the acorn woodpecker is thought to be complex due to the complexity of its social system. This research aims to study the vocalization and relevant social behaviors of the acorn woodpecker through various aspects such as the vocal features, the functions of different call types, general patterns of vocal communication, intraspecific /interspecific responses to vocal signals. In the study the distributed wireless sensor network will be applied to collect and monitor the vocal signals of acorn woodpeckers. Furthermore, this system could identify and localize individuals from their calls. After the sensor network is built, the information collected by the network will be used in bioacoustic and behavioral analysis.

CENS Area

Habitat

Title

Impact of Climate Change on the Spatial Distribution of Alpine Vegetation in the White Mountains, California

Authors

B. R. Zutta & P. W. Rundel

Abstract

No abstract provided.

CENS Area

Habitat

Title

Web Tasking Interface of the Extensible Sensing System

Authors

Poonam Jolly, Ben Huang, Andrew Wu, & Deborah Estrin

Abstract

Current developments in sensing technology will allow sensor networks to be embedded in diverse environments to allow for the collection, manipulation, and transmission of experimental data.    One such application of this technology is the Extensible Sensing System, which is designed to be the testbed for sensors and user interfaces to be used in environment sensing.  The ultimate goals of the ESS are to provide a means of data flow out of a sensor network, remote sensor tasking, and
network monitoring. Within this extensive project is the design and implementation of a webtasking interface to allow the end-user to interact with the network of sensors. The Web Tasking Interface is expected to provide an external user with the ability to monitor and access data being collected by the sensor network from a remote location by means of a webserver. Various webpages are created to allow the user to modify the nodes in the network with regard to the type of data collected, the number of nodes deployed, the data itself, etc.  With this type of system design, this external user is capable of requesting changes to the network itself, thus making the management of such a network a much simpler task.  In this system, the user makes a data request which travels over a “Publish-Subscribe” bus to the intended target where the message is received by an agent.  The ESS Agent is expected to receive serialized messages that are sent over the publish-subscribe bus, unpack the necessary data they carry, and provide for the appropriate execution of the tasks specified within the message.  Through this implementation, the agent is equipped to provide a means of effective tasking and communication between a remote user and the sensor network. Through the various components of the ESS, sensor networks will have immense capabilities that will revolutionize the face of technology and expose the scientific community to vast amounts of data that was previously unattainable.  This project seeks to allow greater access to scientific data and bridge the gap between the virtual and physical implementations of sensor technology.

CENS Area

Habitat

Title

The Extensible Sensing System**

Authors

Eric Osterweil, Mohan Mysore, Mohamad Rahimi, & Andrew Wu

Abstract

The ESS is designed as a suite of components that enable and facilitate sensor networks, their deployment, management, and usage. In meeting its charter, the ESS spans a comprehensive set of platforms and operates in a diverse set of environments. It is, therefore, based on a tiered architecture comprised of; motes (mica2, tos, chipcon, sensor interface board), microservers (strongarm/xscale, linux, motenic, 802.11), commodity platforms (x86, linux, 802.11) platforms, and enterprise platforms (such as solaris and oracle).  Within the ESS, the motes are responsible for:

·         Sensor data acquisition, conversion, and aggregation

Each mote is equipped (or able to be equipped) with a sensor interface board. These boards are designed with an array of channels ranging from analog to digital I/O to onboard sensors. Data from sensor board channels is converted from raw data values into standard units.

·         Simple temporal filtering and triggering

The motes are programmed with logic that allows them to sample channels at application specific intervals. The motes are also able to react to channels that fire at irregular intervals in response to stimuli. These abilities are currently being augmented to include simple temporal pattern matching.

·         Online tasking

Tasking is supported by sending messages to the mote network that specify the desired sampling and filtering behavior of motes.

·         Duty cycling (S-MAC and CPU idle)

Node-lifetime is a primary concern in long-lived deployments. By duty-cycling the CPU and the radio (via S-MAC), nodes are able to save power and extend their own lifetime in a manner that supports bursty traffic.

Microservers offload processing and work from the motes. Their role includes:

·         Data aggregation/caching

The more capable microservers transport data traffic out of the mote-network and into the rest of the system. To do this, they aggregate data from motes and buffer it before sending bursts to the rest of the ESS.

·         Tasking and communication bridge

While residing on the mote network and the internet, the microservers are able to forward data to the motes (as well as aggregate from them). Data such as configuration and tasking requests are relayed to motes through microservers. When dealing with facilities above the sensor network and microserver layer, typical server environments such as commodity platforms and enterprise systems (x86 running linux or sun solaris, equipped with 802.11 or wired ethernet) facilitate more resource intensive operations.

·         Subject servers

Data are multicast from arbitrary ESS components across a publish and subscribe middleware bus platform called subject servers.

·         Subject clients

ESS components receive data from the subject server bus by registering as subject clients.

·         Analysis tools (matlab, labview)

Data that exist in the ESS are processed and viewed through standard data acquisition tools by connecting to the subject server bus as subject clients.

·         Visualization (Web GUI, Labview)

Data are also visualized, and the system is managed through Web GUIs that also act as subject clients.

·         Database (Oracle)

Sensor data and configuration data is archived in and analyzed from an oracle database that runs on a sparc solaris server.

We will demonstrate the ESS in support of a microclimate monitoring system for use in ecosystems research. The following two architectural features will be demonstrated:

·         Mote DAQ system based on sensor interface board that supports a wide array of sensors and sensor types.

·         A pub/sub bus which supports robustness and flexibility over local and wide areas and is for intermittent connectivity. The bus supports flexible registration and presence, robust management using interrogation and re-tasking remote devices, and heartbeats for system health.

CENS Area

Systems

Title

Triggering on Area: A Systems Approach**

Authors

Benjamin Greenstein, Deborah Estrin, Eddie Kohler, & David Culler

Abstract

We study and devise methods for detecting spatially distributed features, and in particular, for detecting significant regions of homogeneity. We propose and develop a flexible architecture and macro-programming environment for distributed detection and triggering in sensor networks.

CENS Area

Systems

Title

Packet Delivery Performance in Dense Wireless Sensor Networks

Authors

Jerry Zhao & Ramesh Govindan

Abstract

Wireless sensor networks promise fine-grain monitoring in a wide variety of environments. Many of these environments (indoor environments or habitats) can be harsh for wireless communication. From a networking perspective, the most basic aspect of wireless communication is the packet delivery performance: the spatio-temporal characteristics of packet loss, and its environmental dependence. These factors will deeply impact the performance of data acquisition from these networks. In this presentation, we report on a systematic medium-scale (up to sixty nodes) measurement of packet delivery in three different environments: an indoor office building, a habitat with moderate foliage, and an open parking lot. Our findings have interesting implications for the design and evaluation of routing and medium-access protocols for sensor networks.

CENS Area

Systems

Title

Multi-hop Code Distribution for Sensor Networks

Authors

Thanos Stathopoulos, John Heidemann, & Deborah Estrin

Abstract

The multicast reliability problem has been studied extensively in the wired internet domain, and, more recently, in wireless and ad-hoc networks as well. In the context of wireless sensor networks, comprised of small, low power and resource-constrained nodes that interact with the physical environment, it remains an active research area. A special case of multicast reliability that is of particular interest in the sensor network domain is code distribution. The ability to add new functionality or perform software maintenance without having to physically reach each individual node is an important piece of a sensor network infrastructure. It becomes critical as sensor networks mature and move toward larger deployment sizes. We present eemoap, an energy-efficient multihop over the air programming mechanism. We discuss the design goals, choices and optimizations for building an efficient code distribution scheme on mica2 motes. The mechanism is evaluated in simulation as well as in an actual mote implementation.

CENS Area

Systems

Title

Environmental Energy Management in Sensor Networks**

Authors

Aman Kansal, Dunny Potter, & Mani Srivastava

Abstract

Energy is a major constraint in the design of embedded sensor networks. The energy supplied by batteries with acceptable form factor may not be sufficient for several applications. Extracting energy from the environment is a feasible alternative in many practical scenarios. We present methods to make the most of the available energy in a distributed system and to operate reliably from highly variable energy sources. We also demonstrate the first version of our hardware which is compatible with several existing sensor nodes for enabling energy harvesting.

CENS Area

Systems

Title

“Do Not Disturb”: An Application Leveraging Heterogeneous Sensor Network**

Authors

Omprakash Gnawali & Mark Yarvis

Abstract

A modern corporate workspace, consisting of matrix of cubicles on a large floor, does not provide isolation from noise from other employees. Noise becomes a distraction when loud sound reaches a person who is working in a cubicle. "Do not disturb" (DND) is a sensor networking application designed to address distracting noise in a corporate workspace. The system alerts people who are talking loud and may be causing distraction to co-workers in neighboring cubicles. DND is an example application for which we designed and built technologies to leverage processing heterogeneity in a sensor network.

CENS Area

Systems

Title

Em View: The Em* Visualizer**

Authors

Lewis Girod, Alberto Cerpa, & Henri Dubois-Ferriere

Abstract

Em View is an extensible visualization system that integrates to emstar-based systems, simulators, and emulators. It can also be connected via gateways and other means to other systems. Em View grew out of a need to develop a unified visualization system to replace a host of individually developed single-application visualizers. Thus, Em View is designed specifically to visualize a wide range of applications from a single reusable framework. A highly customizable interface enables multiple visualization schemes to coexist in a single display, while a simple API significantly reduces the coding required to develop new application specific visualization modules.

CENS Area

Systems

Title

Hierarchical In-Network Processing

Authors

Ram Kumar, Vlasios Tsiatsis, & Mani Srivastava

Abstract

There is a rich diversity of sensor platforms that are currently available. The platforms cover a large range of MIPS, which is a metric that measures the rate of instruction execution in processors. Also, some of them have specialized architectures (for e.g. DSP, custom H/W on FPGA) which make them efficient for a certain class of applications. However, a single platform alone is not scalable to the large dynamic range of the computational complexity of the sensor network applications. This is mainly because, the processors with high MIPS have higher power consumption which make them inefficient for performing computations that are not very demanding. Conversely, the processors with low MIPS are not suitable for demanding applications. In order to make the entire networked system of sensors to be scalable to the computation load, we advocate a hierarchical organization of the heterogeneous nodes. In such a network, any application is broken down into sub-tasks which are mapped onto the nodes (at different levels of hierarchy) that are most efficient for executing it. In order to provide more fine grained input to the system, we deploy the resource constrained nodes (called micro-node from now on) densely throughout the network. The larger nodes with more computation capability (called macro-node) are deployed sparsely to assist the micro-nodes in performing their computation.

CENS Area

Systems

Title

Integrating Wireless Sensor Networks

Authors

Chih-Chieh Han, Roy Shea, Athanassios Boulis, & Mani Srivastava

Abstract

None provided.

CENS Area

Systems

Title

SensorWare in Use

Authors

Chih-Chieh Han, Athanassios Boulis, Roy Shea, & Mani Srivastava

Abstract

None provided.

CENS Area

Systems

Title

Scalable Multi-Resolution Storage and Search in Sensor Networks

Authors

Deepak Ganesan, Ben Greenstein, Denis Perelyubskiy, Deborah Estrin, & John Heidemann

Abstract

Wireless sensor networks enable dense sensing of the environment, offering unprecedented opportunities for observing the physical world. Centralized data collection and analysis adversely impact sensor node lifetime. Previous sensor network research has, therefore, focused on in network aggregation and query processing, but has done so for applications where the features of interest are known a priori. When features are not known a priori, as is the case with many scientific applications in dense sensor arrays, efficient support for multi-resolution storage and iterative, drill-down queries is essential. Our system demonstrates the use of in-network wavelet-based summarization and progressive aging of summaries in support of long-term querying in storage and communication-constrained networks. We evaluate the performance of our linux implementation and show that:

a)       wavelet-based hierarchical summarization provides accurate responses to a broad spectrum of spatio-temporal queries with low communication overhead; and

b)       in a storage-constrained environment, graceful query degradation over time is achieved by networked aging of summaries, such that more useful summaries are retained longer. To our knowledge, no one has examined data aging in the context of highly distributed sensor systems.

CENS Area

Systems

Title

Matching Data Dissemination Algorithms to Application Requirements

Authors

Fabio Silva, John Heidemann, & Deborah Estrin

Abstract

None provided.

CENS Area

Systems

Title

Density, Accuracy, Delay and Lifetime Tradeoffs in Wireless Sensor Networks: A Multidimensional Design Perspective

Authors

Sachin Adlakha, Saurabh Ganeriwal, Curt Schurgers, & Mani Srivastava

Abstract

With the growing interest in wireless sensor networks, techniques for their systematic analysis design and optimization are essential. Despite numerous research efforts in optimizing hardware, algorithms and protocols for these networks, it remains largely unexplored how these innovations can be all tied together to design a sensor network for a specific practical application. We propose a methodology that starts from four independent quality of service (QOS) parameters and allows the user to completely and unambiguously describe the desired performance, without having to deal with the details of individual devices or protocols. By making appropriate choices in terms of device capabilities and run-time techniques, a design can be positioned in this four-dimensional QOS space. Furthermore, we describe a technique to explore the associated tradeoffs at design time, using both analytical expressions and simulations. To illustrate the benefits of our approach, a design example is worked out, which shows a five fold improvement in network operational lifetime by adapting the event reporting delay.

CENS Area

Systems

Title

The Impact of Spatial Correlation on Routing with Compression in Wireless Sensor Networks

Authors

Sundeep Pattem, Bhaskar Krishnamachari, & Ramesh Govindan

Abstract

While several data aggregation techniques have been proposed for sensor networks, an understanding of the performance of various data aggregation schemes across the range of spatial correlations is lacking. We analyze the performance of routing with compression in sensor networks using an application-independent measure of data compression (an empirically obtained approximation for the joint entropy of sources as a function of the distance between them) to quantify the size of compressed information, and a bit-hop metric to quantify the total cost of joint routing with compression. Analytical modeling and simulations reveal that while the nature of optimal routing with compression does depend on the correlation level, surprisingly, there exists a static clustering scheme which can provide near-optimal performance for a wide range of spatial correlations. The implication, that there exist relatively simple energy-efficient aggregation schemes for correlated sources has much practical importance.

CENS Area

Systems

Title

The Acquire Mechanism for Efficient Querying in Sensor Networks

Authors

Narayanan Sadagopan, Bhaskar Krishnamachari, & Helmy

Abstract

We propose a novel and efficient mechanism for obtaining information in sensor networks which we refer to as acquire. In acquire an active query is forwarded through the network, and intermediate nodes use cached local information (within a look-ahead of d hops) in order to partially resolve the query. When the query is fully resolved, a completed response is sent directly back to the querying node. We take a mathematical modeling approach to calculate the energy costs associated with acquire. The models permit us to characterize analytically the impact of critical parameters, and compare the performance of acquire with respect to alternatives such as flooding-based querying (FBQ) and expanding ring search (ERS). We show that with optimal parameter settings, depending on the update frequency, acquire obtains order of magnitude reduction over FBQ and a potential reduction of 60% over ERS in consumed energy.

CENS Area

Systems

Title

Distributed Index for Multi-dimensional Data in Sensor Networks**

Authors

Xin Li, Young Jin Kim, Ramesh Govindan, & Wei Hong

Abstract

In many sensor network applications, data or events are named by attributes. Many of these attributes have scalar values, so one natural way to query events of interest is to use a multi-dimensional range query. An example of such a range query is: “list all events whose temperature lies between 50 degree and 60 degree, and whose light levels lie between 10 and 15.” Such queries are useful for correlating events occurring within the network represented by different attribute values. In this demo, we will demonstrate an implementation of a distributed index that scalably supports multi-dimensional range queries. Our distributed index for multi-dimensional data} (or dim) uses a novel geographic embedding of a classical index data structure, and is built upon the GPSR geographic routing algorithm.

CENS Area

Systems

Title

SCALE: A Tool for Connectivity Assessment in Lossy Environments

Authors

Alberto Cerpa, Naim Busek, & Deborah Estrin

Abstract

Wireless sensor networks will allow fine-grained monitoring in a wide range of environment (indoor and outdoor). Many of these environments, present very harsh conditions for wireless communication using low-power radios, including multipath/fading effects, reflections from obstacles, and attenuation from foliage. In this work, we introduce scale, a network wireless measurement tool that uses packet delivery as the basic application-level metric. Scale facilitates the gathering of packet delivery statistics using the same hardware platform and in the same environment targeted for deployment. Using up to 55 nodes, we were able to measure and study the connectivity conditions of two hardware platforms, mica 1 and 2 motes, in three different environments: an outdoor habitat reserve, an urban outdoor environment in a university campus, and an office building, under systematically varied conditions. Among other things, we found that there is no clear correlation between packet delivery and distance in an area of more than 50% of the communication range, temporal variations of packet delivery are correlated with mean reception rate of each link, and the percentage of asymmetric links varies from 5% to 30%. Data collected using scale have interesting implications in the design, evaluation, and parameter tuning of sensor network protocols and algorithms.

CENS Area

Systems

Title

Voronoi Scoping in Sensor Networks**

Authors

Henri Dubois-Ferriere, Lewis Girod, & Deborah Estrin

Abstract

A common scenario for sensor networks is data-gathering: nodes gather data (periodically, or triggered by certain conditions) and send it back to a sink (or basestation). Data is sent over multiple hops, using a tree topology. Such a system with a single sink presents scaling difficulties as the network grows: the average distance to the sink increases, and nodes near the top of the tree must relay a disproportionate amount of traffic. Placing multiple sinks (basestations) can alleviate this, but on the other hand introduces additional overhead if all sinks flood the entire network. Voronoi scoping is a simple distributed technique to limit floods so that each flood reaches only the subset of nodes in the originating sinks's region; that is the nodes who have the originator as closest sink. We have implemented the algorithm in one-phase pull; this demo uses emview to visualize in real-time the clusters formed in the LECS lab ceiling array of motes.

CENS Area

Systems

Title

Schedule and Latency Control in S-MAC**

Authors

Yuan Li, Wei Ye, & John Heidemann

Abstract

None provided.

CENS Area

Systems

Title

Monitoring with a Mix of IR Sensors and Cameras**

Authors

Abhishek Rajgarhia, Fred Stann, & John Heidemann

Abstract

None provided.

CENS Area

Systems

Title

Evaluation of GPSR in the Diffusion Filter Framework

Authors

Xi Wang, Fabio Silva, & John Heidemann

Abstract

Directed diffusion is a data-centric communication mechanism for sensor networks. It is supported by a flexible diffusion filter framework. GPSR (greedy perimeter stateless routing) is a routing mechanism commonly used in sensor networks. This project is an effort to connect GPSR with directed diffusion. Compared with existing routing mechanisms in the diffusion platform, including push, pull and gear, GPSR has quite different qualities in performance. It is an importation addition to the diffusion platform. As a widely known protocol, GPSR is also a good benchmark for ongoing and future research works. In this project, GPSR is implemented as a filter in the diffusion platform, and we performed simulations to compare the performance of different routing mechanisms. There are two findings from simulations results: a) overall GPSR is a rather efficient routing protocol. B) the behavior of GPSR is quite sensitive to node density. The performance will degrade significantly for sparse (barely connected) wireless networks. The poster will show simulation results with analysis. It will also show the internal structure and interfaces of the GPSR filter, as the implementation of the GPSR filter is good example of integrating new routing filters to the diffusion platform.

CENS Area

Systems

Title

Wireless Seismic Data Collection

Authors

Paul Davis, Jeremy Elson, Deborah Estrin, Allen Husker, & Igor Stubailo

Abstract

None provided.

CENS Area

Seismic

Title

A Wireless Seismic Sensing Array**

Authors

Ning Xu & Ramesh Govindan

Abstract

We will demonstrate the collection of accelerometer reading by motes in a multi-hop manner, we have implemented data compression to reduce dataset and a NACK based protocol to achieve reliability.

CENS Area

Seismic

Title

Seismic Network Deployment Preparations

Authors

Allen Husker, Igor Stubailo, Monica Kohler, & Paul Davis

Abstract

Technological and scientific preparations are occurring for the development of a multi-hop radio-linked seismic array (MHRLSA) of 50 broadband stations (GURALP 3ts) and its first few deployments. A ruggedized data relay device (DRD) is being fabricated using Intel’s new low power, small form factor stargate motherboard. A DRD will be placed at each node of the array and configured as a local area network (LAN) with station spacing up to 10 km. The objective is to use protocols that have been developed for the world wide web to optimize the acquisition of real time seismic array data. The array DRD’s will self configure to optimize RF transmission from the nodes to a data concentrator using dynamic routing, from where it will be transmitted real time over a long link back to the laboratory. The MHRLSA will have the ability to generate and communicate network time, and compare local GPS time with network time, allowing the opportunity to switch to network time if GPS time has failed or no GPS is available. In this way robust real time will be maintained avoiding needs for timing corrections after recording. Scientific preparation is being done on two fronts: factor data analysis and a Los Angeles basin seismic study. Twenty four single component seismometers are continuously recording data within the UCLA factor building. A model of seismic amplitudes is developed using simple scattering and attenuation theory and fit to data from the 1997 Los Angeles basin passive seismic experiment (LABPSE) using a non-linear least squares inversion. LABPSE ran roughly north to south through the San Gabriel and Los Angeles basins about 60 km with an average station spacing of less than 5 km. The model takes into account site amplification due to soil type, geometrical spreading, earthquake magnitude, and attenuation. The results indicate that for most stations the model is a good first order approximation of the amplitude variation. However events from a restricted range of azimuths for stations near the basin edge have amplitudes that are several times larger than the model predicts. One interpretation of these results is that the Whittier fault at the edge of the Los Angeles basin creates a zone of increased amplification for earthquakes along the path of the fault. This effect has subsequently been included in the model. The model will be applied to and updated with data from the UCLA campus deployment which is located at a basin edge. The dense (~100m) spacing will allow for waveform correlation leading to more exact deep structure determination and hazard analysis.

CENS Area

Seismic

Title

Vision Based Navigation**

Authors

Jason Meltzer & Stefano Soatto

Abstract

We present a method for inferring the location of a robot relative to a three-dimensional map of its environment. The map, created off-line, consists of image patches, their locations in space, and their associated normal vectors. Observational data consist of one or many images taken from the robot’s current viewpoint (a point and orientation in space). We develop a framework for matching images of a scene (observations) to a map and show how this can be applied to the task of robot localization. Localization is posed as an optimization problem, where the observed data and the map are aligned to produce an estimate of the current pose of the robot. We provide a formalization of our model and demonstrate experimental results in unstructured environments.

CENS Area

Signal Processing

Title

Ad-Hoc Localization Infrastructure**

Authors

Yen-Cheng Kuan, Andres Savvides, & Mani Srivastava

Abstract

None provided.

CENS Area

Signal Processing

Title

Collaborative Data Compression in Sensor Networks

Authors

Ameesh Pandya, Aman Kansal, Greg Pottie, & Mani Srivastava

Abstract

We consider the network information theoretic problem of finding the rate distortion bound for separate coding of multiple correlated sources. First, we find the outer region for separate coding with distortion when only one of the sources is to be reproduced and the other sources are used as helpers. We then find the outer region for separate coding when all the sources are to be reproduced with specified distortions. The rate