SensorNet Paper Reviews by rc496
From Ece
- Return to the paper reading schedule or the Sensor Networks homepage.
Presentation January 22, 2007
WIRELESS INTEGRATED NETWORK SENSORS
Presentation file is here
Review Jan 29,2007
Just-In-Time Sensor Networks
G. Yee, B. Shucker, J. Dunn, A Sheth, R Han, "Just-In-Time Sensor Networks", IEEE Workshop on Embedded Networked Sensors (EmNets) 2006, pp. 6-10.
This paper discusses a new idea about real time deployment for a node or nodes can be deployed from cannon when and where additional nodes are needed with any reasons within seconds after request is received. The cannon have to have capability to execute deployment application, have rapid response, good precision, on-demand activation and multiple ammunition inputs. The writers well suggest this idea to solve problems that happened in sensor networks such as bad communications due to distance or geographical, optimization of the network it self that can not be handled by software only, and optimization of power consumption.
The writers present several advantages that really helpful during applied sensor networks in the field such as improving sensing capabilities with incremental search which also has a significant improvement in target tracking application, improve robustness of routes where it is needed by deploying additional nodes to that area, improving the efficiency of power consumption of the network when an additional node that is equipped with radio and cheap power supply is deployed and this node does the broadcasting process while the rest of the network unaware with this process which is not wasting their power. No hardware details or testing result of this method are presented in this paper.
MANTIS OS: An Embedded Multithreaded Operating System for Wireless Micro Sensor Platforms
S. Bhatti, J. Carlson, H. Dai, J. Deng, J. Rose, A. Sheth, B. Shucker, C. Gruenwald, A. Torgerson, R. Han, "MANTIS OS: An Embedded Multithreaded Operating System for Wireless Micro Sensor Platforms, " ACM/Kluwer Mobile Networks & Applications (MONET), Special Issue on Wireless Sensor Networks, vol. 10, no. 4, August 2005
This paper discusses about new operating system in network sensor called MANTIS (MultimodAl system for NeTworks of In-situ wireless Sensors) which for MANTIS sensor OS (MOS) can be implemented in less than 500 bytes of memory which is very impressive from memory efficiency point of view and also can manage efficiency of power consumption. It has capability for multithreaded executing program which is improve the execution time and it can support more complex applications.
Writers explain about the architecture of MANTIS operating system from Application and APIs, Kernel and scheduler, network stack and “comm” layer. And writers also discuss about threads and event for MANTIS operating system. The description about MANTIS hardware as the hardware tools when doing the experiment with this operating system also well discuss.
This paper also includes the experiment result from applying MANTIS operating system, such as result in execution time and power consumption in sleep mode and awake mode. It would be much better if the writers also include the result of using another OS as comparison with the result from using MANTIS OS and it will point where MANTIS OS is better than other operating systems.
The Emergence of Networking Abstractions and Techniques in Tiny OS
P. Levis, S. Madden, D. Gay, J. Polastre, R. Szewczyk, A. Woo, E. Brewer and D. Culler, The Emergence of Networking Abstractions and Techniques in TinyOS, USENIX/ACM Symposium on Networked Systems Design and Implementation (NSDI). 2004.
This paper discusses about Tiny OS as the operating system for sensor network and divides the network abstraction into 4 categories such as general abstraction, specialized abstraction, in-flux abstraction and absent abstraction based on the structure of the program inside Tiny OS. From communication point of view, additional information about single hop communication and multi hop communication and how the implementation are given to inform about capabilities of this operating system to control and to execute protocols that is used in the communication in the network. Examples are provided in this paper according to the abstraction categorize. Beside communication part, Tiny OS also manage the efficiency of power consumption and time synchronization which is important during communication process and sensing process.
A Dynamic Operating System for Sensor Networks
C.-C. Han, R.K. Rengaswamy, R. Shea, E. Kohler and M. Srivastava. A Dynamic Operating System for Sensor Networks, International Conference on Mobile Systems, Applications, and Services (MobiSys), 2005.
This paper discusses a new operating system called SOS, which is created based on the writers’ experiences with Tiny OS, that has dynamic ability and general purpose operating system without decreasing power consumption efficiency significantly. It consists of common kernel and dynamic application modules (not in Tiny OS) which implements messaging, dynamic memory, and module loading and unloading among other services. Operating system architecture is also described into several parts that significant such as modules structure, modules insertion, modules removal, message scheduling and dynamic memory.
Writers also include information of evaluation SOS compare with tiny OS and Maté Bombilla virtual machine in several parameters such as surge tree formation latency, surge forwarding delay, and surge packet delivery ratio. Writers also include power consumption and code size for each operating system of certain application. No hardware information and description is discussed in this paper.
Review, February 5,2007
Medium Access Control with Coordinated Adaptive Sleeping for Wireless Sensor Networks
W. Ye, J. Heidemann and D. Estrin, Medium Access Control with Coordinated, Adaptive Sleeping for Wireless Sensor Networks, IEEE/ACM Transactions on Networking, June 2004.
Authors of this paper are explaining about new medium access control (MAC) special for wireless sensor network because the method that exist such as IEEE 802.11 is for general purpose wireless device. In wireless sensor network, a new method should be discovered especially for energy consumption point of view. S-MAC is designed specially for wireless sensor network. The idea from this method is how to cut down power consumption during communication period, especially while in idle mode, because in idle mode, the power consumption is the biggest compare with other process.
To cut down power consumption, this method is putting nodes into periodic sleep state. For example sleep, awake, listen, and sleep again. It adapts IEEE 802.11 method especially for collision avoidance, where 802.11 is good at. This method also does synchronization where nodes exchange the schedule with neighbors. With this exchange schedule, a node knows when the neighbor will awake for a certain time to listen and when the neighbor is sleeping. Another important thing that this method proposes is adaptive listening, where a node that overhears the communication wakes up at the end of transmission so that the data can be transferred faster. If this data is not for that node, then it will go to sleep state.
Overall, this method is a good method for sensor networks MAC, but weaknesses are appeared. First, is about the synchronization method, where several things can happen such as clock drifting while the network is busy and for large scale sensor networks, it's not a good way to send SYNC packet one at the time before transmitting the data.Probably this SYNC packet can be sent with the data.
MACAW
V. Bharghavan, A. Demers, S. Shenker, and L. Zhang, MACAW: A Media Access Protocol for Wireless LANs , SIGCOMM Symposium on Communications Architectures and Protocols (SIGCOMM'94), September 1994.
This paper is discussing about a new protocol of wireless communication. MACAW (Multiple Access Collision Avoidance for Wireless) is designed to improve wireless communication better than any other protocol at already exist at that time, where wireless device was starting growing. The main problem that this method handles is the collision data during communication.
Authors give information about another protocol at that time such as CSMA and MACA as the base of MACAW design. CSMA suggest method of carrier sense before doing data transmission. The collision avoidance is done by testing the signal strength of the transmitter. But in this method, no appropriate information for collision avoidance. MACA is proposed a packet radio as alternative for old CSMA, which is contained CTS and RTS signal.
MACAW method is a combination between CSMA and MACA methods. Not only sending RTS and CTS, this method also using ACK signal as indicator ending of transmission. Another signal that is used in this method is DS signal as indicator that RTS and CTS exchange is successful and a data transmission is about to occur. This signals is for overhearing station. RRTS is a signal for Request RTS signal which is used when a station receives RTS and can't response. There is a backoff algorithm that is used in this method where the value is changing because of transmission effect. It decreases when RTS-CTS exchange successful and increases when collision happens. With this method, a retransmission can be managed. But it does not manage very well because once the station is success, then most likely it will be used all the time, while other stations won't be used anymore. To prevent this from happening, a modified backoff algorithm is introduced, where backoff counter also being sent and the station that receives this packet, will copy the value into its backoff counter. This method also handles multiple destination data by using backoff counter for how long to wait before sending RTS. It will start sending the data for the shortest waiting period.
This method might be the base idea of any new methods or protocols in wireless communication. But it is not specifically for wireless sensor network usage, but it is more general purpose.
Ultra-Low Duty Cycle MAC with Scheduled Channel Polling
W. Ye, and J. Heidemann, Ultra-Low Duty Cycle MAC with Scheduled Channel Polling, ACM Conference on Embedded Networked Sensor Systems (SenSys'06), November 2006.
A method for sensor networks is designed for maximum efficiency of power consumption. In this paper, authors discuss about a method that can cut power consumption by factor 10-100, with using Ultra-low duty cycle. Most of the method that exists is low duty cycle but to achieve that low, it needs a method that can run with ultra low duty cycle.
Authors introduce SCP-MAC (Schedule Channel Protocol - MAC) which the design use 2 approaches such as protocol scheduling method and Low Power Listening (LPL). This method puts nodes into periodic sleep state when no traffic in the network and it will perform channel polling periodically. The polling time is synchronized to all neighbor nodes and with this method, one short wakeup tone can wake the node up. This also fixes LPL method weakness for waking up the nodes, which used to send long preambles. The synchronization part is essential for this method. It does the synchronization while sending the data, which is from the testing result shows that this synchronization method is the best way to do synchronization because the energy of transmitting data is lot more than transmitting synchronization packet, and it seems zero. Adaptive listening also being used in this method for busy traffic.
The result of this method testing shows that this method can be used for ultra-low duty cycle and it also decreases significantly compare with other method with the same testing. Any further optimization method that is used in this method is not discussed in this paper.
Presentation, February 12, 2007
ScatterWeb
ScatterWeb - Low Power Sensor Nodes and Energy Aware Routing. Schiller, J., Liers, A., Ritter, H., Winter, R., Voigt, T., Freie Universität Berlin, Germany; System Sciences, 2005. HICSS '05. Proceedings of the 38th Annual Hawaii International Conference on, 03-06 Jan. 2005, pp. 286c-286c
Presentation is here
Review February 19,2007
Tenet
O. Gnawali, B. Greenstein, K.-Y. Jang, A. Joki, J. Paek, M. Vieira, D. Estrin, R. Govindan, and E. Kohler, The TENET Architecture for Tiered Sensor Networks, ACM Conference on Embedded Networked Sensor Systems (SenSys), November 2006.
Writers introduce Tenet, the architecture of tiered sensor networks as a solution to handle any constrains in the network sensor. They are divided into 2 tiers, lower tier consist of more motes, and higher tier consist of less nodes with high bandwidth radios which is called as Master. Certain nodes will handle certain amount of motes and motes do not communicate each other except for passing the beacon signal. The nodes are interconnecting with each other and form a network, which make a node can talk to mote from different nodes. This architecture uses novel tiered routing mechanism where a mote will transmit the respond to the nearest node, which is the purpose of the beacon signal. And the node will pass on the information to the node that sends task. The communication between node and mote is form of task, which is a request to do activity. And without any task command, motes will not do any process which will cut down the power consumption point of view.
It is a good idea where the network will become more manageable compare with network sensor where every nodes has to handle networking and data.It is fit for detection or monitoring application but it will depend on what application that is going to be applied.
SATIRE
R. Ganti, P. Jayachandran, T.F. Abdelzaher, and J.A. Stankovic, SATIRE: A Software Architecture for Smart AtTIRE, International Conference on Mobile Systems, Applications, and Services (MobiSys), June 2006.
This paper is discussing about SATIRE, an example of how sensor network being used for human monitoring device to record owner’s activity and location as personal archive. Several problems are discussed in this paper such as data collection, data upload, data synchronization, reconstruction of activity, and user’s privacy. It is straight forward type of paper where writers give the solution to the problems. The architecture of the software that is needed for this application is divided into several layers, such as application layer for user interface application, parsing layer to process raw data from sensors and interpretation layer to handle any algorithm to process the data.
To handle the data collection and storage problem, the writers suggest to truncate the data from motes and sensor such that it still have the significant data but it will be consume less space compare with the original data. For data upload, writers suggest to use beaconing system as one of method that can be used. Data synchronization can be done by simple beaconing which has number represents specific mote. Stillness detection can be used to reduce power consumption and data collection which is a significant technique for this application. Any algorithm can be used to reconstruct the data to show the activity, but in this paper, writers use exponential weighted moving average (EWMA) method to do this. For user’s privacy, writers suggest that the data only can be uploaded if the mote is close enough with user’s PC and it will upload wirelessly or wired.
TSAR
P. Desnoyers, D. Ganesan and P. Shenoy, TSAR: A Two Tier Storage Architecture Using Interval Skip Graphs, ACM Conference on Embedded Networked Sensor Systems (SenSys), November 2005.
Writers introduce TSAR, as storage architecture that is suitable for sensor network application which has constraints such as data space, power consumption, time efficiency, etc. They put the architecture into 3 tiers. Lowest tier is assumed to form a dense deployment of sensor, Middle tier consist of power-rich sensor proxies that has significant computation, memory and storage resource and can be used continuously. Highest tier is the application that queries the sensor network through query interface.
In TSAR,sensor stores the data in its own memory (local storage) and periodically send the summary, called metadata, to the proxy, which will be put index for each summary using a novel index structure called interval skip graph by the proxy .
Those indexes are distributed through the proxies that make application can easily to search through the index to find a specific data to request. And with this index system, any proxy can access data from sensor which is not in its coverage.
One thing that this paper need is the example of summary, metadata and the index, which will be interesting because the idea of locally store and generally access is one aspect that is needed in sensor networks.
Review February 26, 2007
EnviroSuite
L. Luo, T.F. Abdelzaher, T. He, and J.A. Stankovic. EnviroSuite: An Environmentally Immersive Programming Framework for Sensor Networks. ACM Transaction on Embedded Computing System (TECS), 2006.
In this paper, writers introduce EnviroSuite, object based programming system which is the programming framework for sensor networks. With using EnviroSuite, programmers need to do Environmentally Immersive Programming (EIP). This EnviroSuite is depending on 2 things, the library (EIPLib) which contains algorithms for specific implementations or services, and the compiler (EIPLC) that configures and restructures from EIPLib to produce the implementation in nesC language. Using EnviroSuite, programmers recreate the application in virtual world with every aspect that is needed for this application, and with the library and the compiler, and produce a set of command for the hardware for that specific application, which the writer called classical programming as too low level.
From the evaluation result, the writers show the difference of code line comparing EnviroSuite and nesC and it shows that EnviroSuite generates less code line compare with writing the program itself in nesC language. Problems that might occurs from this EnviroSuite is that it depends on the programmers do the EIP and any mistake on that part, can cause the application will not work properly. Beside human error, nature also has significant effects for the hardware. Any small adjustment is necessary when the application is running on the field.
There will be a time where no older function or service can be used or not in the library for certain application might be a problem for this method because it a library is added, the compiler also needs to be updated because the compiler interpret the command from library to nesC language.
nesC
D. Gay, P. Levis, R. von Behren, M. Welsh, E. Brewer, and D. Culler, The nesC Language: A Holistic Approach to Networked Embedded Systems, ACM Conference on Programming Language Design and Implementation (PLDI), June 2003.
This paper is discussing about nesC programming language that is created for networking embedded system in sensor networks application. This programming language is implemented in Tiny OS, an operating system for sensor networks and any others program in sensor networks area. There are constraints that nesC tries to overcome and also become the base of the language design. Size is one constraint that lies in sensor network area. It is not only the program memory size but also external memory or flash memory that is used to store the data or doing calculation. The program needs to be able to operate in such small space. Writers state in this paper that the core of Tiny OS only requires 400 bytes of code and data memory, combined, which is very compact. nesC does not use any dynamic memory allocation which makes it as a static language. All needs for dynamic allocation most of them are handled from the component models and the interface parameter from nesC. The application runs base on the event, where one event triggers an interrupt and then the program will be executed, which make nesC can handle asynchronous program and also synchronous program with scheduling.
One main thing that writers address in this paper is about concurrency in the programming. nesC has tools to handle the concurrency with atomic section and task where the program handles any necessary action to maintain the concurrency of the program. Writers do not discuss too much about the compiler, which might be interesting because the program will depend on it to generate small size of codes to be stored in the hardware.
Mate
P. Levis, N. Patel, S. Shenker, and D. Culler, Mate: a Virtual Machine for Tiny Networked Sensors, International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS), Dec 2002.
This paper introduces Maté, a tiny virtual machine for sensor networks where it breaks the code into small capsules of 24 instructions, reducing the energy cost of transmission and reprogramming and for some complex program, it can compose into several capsules and represent it under 100 bytes. And it runs on motes.
It is a Tiny OS component that sits on top of system components and it controls the flow of operation where it will not need any buffer during the operation, which conserves RAM. Each code has type and version information and Maté has 4 kinds of capsules: message send capsules, message receive capsules, timer capsules and subroutine capsules. It will execute the capsules based on the event and it runs concurrently. These capsules are used to do the reprogramming using the information version that contains in the code. As the evaluation that is covered by writers showed that using Maté, the cost of installation is less than binary code which shows that Maté is more efficient in the transmission cost and time compare with transmission binary code.
Active Sensor Networks
P. Levis, D. Gay, and D. Culler, Active Sensor Networks, USENIX/ACM Symposium on Networked Systems Design and Implementation (NSDI), 2005.
This paper is introducing ASVMs (Application Specific Virtual Machines) as a way to reprogram wireless sensor networks. It is the general form from Maté which is designed to support higher level programming for active sensor networking and it is running on Tiny OS. Writers’ evaluation in this paper satisfies 3 main things that writer’s statement: concurrency on running the program with using implicit concurrency management, propagation of the data which is the statistical result of data communication, and flexibility. The flexibility of programming language for this VM can be showed by supporting 3 language currently such as: TinyScript, motlle and TinySQL queries. Flexibility of Applications can be shown by 2 examples of VM that writers built. As comparison to Maté, writers should also present the data from using Maté as the comparison and readers can clearly see the advantage of ASVM.
Presentation, March 5, 2007
FireWxNet
C. Hartung, S. Holbrook, R. Han, C. Seielstad, “FireWxNet: A Multi-Tiered Portable Wireless System for Monitoring Weather Conditions in Wildland Fire Environments”, Fourth International Conference on Mobile Systems, Applications and Services (MobiSys) 2006, pp. 28-41
Presentation is here
Paper Review, March 19, 2007
Beyond average: Toward Sophisticated Sensing with Queries
Joseph M. Hellerstein, Wei Hong, Samuel Madden, and Kyle Stanek, Beyond Average: Towards Sophisticated Sensing with Queries, IPSN'03.
This is a status report of the writer’s effort for more complex sensing task in TinyDB. As the background information, writes provide brief information about TAG framework for handling queries with SQL-like language and how the aggregate queries are handled. There are 4 points that writes discuss as the progress from their research. First one is the new language feature such as query that is automatically posed based on external action (events) and storage points. Second progress is the capability to generate a map based on data queries by using several algorithms such as naïve algorithm, in-network algorithm, Lossy algorithm, and sparse grid. Third progress that is discussed in this paper is the application of wavelet techniques to generate histogram of statistical information that is gathered from the SQL’s built-in aggregates. Fourth progress is the implementation of TinyDB for vehicle tracking application. Suggestions for this paper are: 1. Need to put information about power consumption from the nodes in the application, which will give better information about power consumption for the database application in sensor network and the system itself 2. Need to put information about the time that is spent for querying data from the network. It will give better picture about the advantage of using database. 3. Show the result of wavelet histogram which writers did not show in the paper.
Synopsis Diffusion
Synopsis Diffusion for Robust Aggregation in Sensor Networks. S. Nath, P. Gibbons, S. Seshan and Z. Anderson ACM SenSys, Baltimore, MD, November 2004
<p>This paper is introducing a new method, Synopsis Diffusion, for achieving more accurate and reliable result of the query. In this case, the query is the result of data operation that is gathered from nodes. Any failure of data transmission from one node will affect the answer of the query which might be happened with Spanning Based Tree aggregation, and miscalculation also might be happened from multi path routing protocol. To handle this problem, Synopsis Diffusion combines routing scheme “Ring” and aggregation scheme “Count” which will guarantee there will be no miscalculation due to receiving same data multiple times and it will be robust. </p> <p>In this paper, writers did not explain much about how the “Ring” routing scheme handles the changes of nodes in the network, and what the “Ring” characteristic, which make us the reader still not quite understand how this “Ring” method works in that situation. This paper has mathematical proof of their method to handle data duplication for the multi-path routing. Writers give a good evaluation report of this method’s performance by doing the comparison with other methods’ performance in a specific area with graphs. One thing that might improve this paper if the writers also give some information about the real world application that uses this method and the node information where this method is applied. </p>
Chord: A Scalable peer to peer lookup service for internet applications
I. Stoica, R. Morris, D. Karger, M. F. Kaashoek, and H. Balakrishnan. Chord: A scalable peer-to-peer lookup service for Internet applications. Technical Report TR-819, MIT, March 2001.
<p> This paper is discussing about a method called Chord for managing file storage and retrieval in the network. For this paper, it is used in the internet application. Chord system has certain number of nodes that involve, and each node will have some information of other nodes, but not all of them. The system is constructed as a circle contains number of nodes depend on how many bit the hash result. In this paper, it is using SHA-1 as the hash function that produces 160bits number. So the capacity is 2^160. Each node has an IP address that will be hashed as the node identifier. The file name also through hashing process and it is the key. Then each node will have information about their neighbor node stored in a table called finger table and this information will be used for the searching optimization. With using this finger table, each node does not have to store all nodes information in the system, but it stores some of information about the neighbor and it helps to speed up the searching process of file and the IP address. It also will be adapted to the changes whether there is node joins the system or leaves the system. The table will be corrected by background process that runs periodically to check the table. </p> <p> This is a computer network method and it is interesting method. It has high possibility to be able to apply some characteristic of this method, not all of them in the sensor network but it might be need further research on the data from the nodes, the communication efficiency, storage, power consumption, and processing time due to constraints that sensor network has. </p>
Sensor Networks as a Database
R. Govindan, J. M. Hellerstein, W. Hong, S. Madden, M. Franklin, S. Shenker, "The Sensor Network as a Database," USC Technical Report No. 02-771, September 2002.
<p>In this paper, the writers point out ideas about sensor network database. They suggest that there are 2 based ideas for sensor network database architecture which are in-network implementation of database query operators such as grouping, aggregation and joins, and allowing approximate result for database query. Writers discuss about join operator and aggregation operator in this paper with any possible way to implement it to the sensor network which always have constraints in power and storage. </p> <p> There are no real application of database is discussed in this paper, there are no real new method that is created for any issue that come up with database set up, and there are no further information about the query process method that might be the base method of sensor network database. Basically, it is an open area that writer want to show to the reader about possibility of database set up in the sensor network or any process that uses database as the part of the application of the sensor network and any possible way to support this database function. </p>
Paper Review, April 2, 2007
The Semantic Web
Tim Berners-Lee, James Hendler and Ora Lassila, "The Semantic Web", Scientific American, May 2001
This article might be the first article that talks about semantic web. In semantic web, information is given well defined meaning and it will allow human and computer to use that information together. Most of the information at that time is for human only, which means that it can not be used by computer to get information that is needed for another process or function. It is not a separate part of the web but it is an extension. There are 2 technologies that already appeared XML (eXtensible Markup Language) and RDF (Resource Description Framework). XML allows user to add additional structure or tag to the documents and RDF expresses the meaning of those additional structure or tags from XML. The third technology that helps development of semantic web is ontologies which is the collection of information that defines the relations of the terms. Another term that is covered in this article is Agents, which is the software. This software acts as the engine side that process information and produces an output as the user assigned. With semantic data, it allows the agent can get information from the data and by combining all data that is required; the agent can produce the result. It is a breakthrough idea at that time where internet is still new and it is described with an activity that makes reader be able to visualize how this is working. This concept is becoming the base of current web application and the basic idea which is semantic, also being used not only for the web, but also the database application especially in the query process.
The Semantic Web Revisited
Nigel Shadbolt, Tim Berners-Lee and Wendy Hall, "The Semantic Web Revisited", IEEE Intelligent Systems, 21(3) pp. 96-101, May/June 2006
The improvement of semantic web usage can be seen in the example that is given in the article. A life science research that needs data sets from scientist communities from different area is one of the possibilities that can be achieved through semantic web. It makes a document into a data. Another improvement that happens to semantic web is universal resource identifiers (USI). Any resource that associate with USI can be processed by anyone (link, refer to or retrieve a representation). Then there is RDF translation which can extract RDF from XML and XHTML document using XSLT language and associated with the original content. For ontology method, no discussion about new method but writers covers the use of OWL (web Ontology Language) and further discussion about rules and interference also support the semantic web. The progress that has been achieved until this time might keep progressing in the future because more complex application needs improvement of the system and further exploration of each part that supports the semantic web. Another progress that is stated in the article is Folksonomies. It represents a structure that emerges when individuals manage their requirement themselves. This article shows and discusses more about the path that improved around semantic web compare with the first article about Semantic Web, 2001.
Plug and Play Sensors in Wireless Networks
Dunbar, M., Plug-and-play sensors in wireless networks, IEEE Instrumentation & Measurement Magazine, Vol.4, No.1, pp.19-23, March 2001
This article is discussing about plug and play sensors in wireless sensor networks. It takes CrossNet as the example of this discussion. To be a plug and play sensor, IEEE 1451 is the standard interface that makes it happened. Following this standard for any sensors development will make sensors can be used in any system that supports this standard. This paper also gives an example of TEDS (Transducer Electronic Data Sheet) so that readers know what TEDS is. The way writer discuss this mater is very attractive, where there are pictures about the diagram of representation of the CrossNet structure. But, there are no further example of this type plug and play sensor in application case or any real world hardware description that is support plug and play. Any information about CrossNet will make this article more informative. From user point of view, a plug and play sensor will be very attractive and very useful in the application of sensor network because there is no need to change any parameter manually if the design is supporting IEEE 1451. Any sensors will be easily connected to the network and directly can be used and it cuts the time cost for any adjustment that may be needed for any changes in sensor deployment.
IEEE 1451: A Standard in Support of Smart Transducer Networking
Kang Lee, IEEE 1451: A Standard in Support of Smart Transducer Networking, IEEE Instr. and Meas. Tech. Conference, May 2000
This paper covers a general description of IEEE P1451 and its sub projects which all are related to the common transducer interface for sensors and actuators with the system. In this paper, there are 4 sub projects that are covered such as IEEE 1451.2, IEEE 1451.1, IEEE 1451.3 and IEEE 1451.4 IEEE 1451.2 is a project that defines TEDS (Transducer Electronic Data Sheet) and its data format and communication protocol between transducer and microprocessor. The idea of this project is that a transducer has its information inside itself and it is transferable electronically to Network Capable Application Process (NCAP) and it means that the transducer has capability for plug and play, without need to set any parameter manually. IEEE 1451.1 is a project that defines an object model for network smart transducer and software interface specification for each model. With an object model, the programming of the interface to hardware will not be complex by using input/output driver. IEEE 1451.3 is a project that is related to standardization of interface with large number of smart transducer for synchronization. IEEE 1451.4 is a project that is related to interface of different kind of transducer and nodes to the system which is using a small TEDS for analog transducer and a node list. With those descriptions, the writer gives a summary of IEEE 1451, standardization support of transducer. This is a good paper for beginner reader where it describes IEEE 1451 in global way, but it will be more interesting if there is any hardware or transducer example in this explanation. The writer discusses each project with a block diagram which gives reader more visual understanding how each mater construct. And with this standard, I think it is very useful and it is the base of any smart transducer development and many smart sensors can be produced in any area of application and still can be used together.
IEEE 1451 Standard and Frequency Output Sensors
Sergey Y. Yurish, IEEE 1451 Standard and Frequency Output Sensors: How to Obtain a Broad-Based Industry Adoption?, Sensors & Transducers Magazine, Vol. 59, Issue 9, September 2005, pp. 412-418
In this paper, writer talks about standardization of frequency output sensors based on IEEE 1451 standard for intelligent transducer. It is stated in the paper that there is no standard of frequency output sensor which might caused some difficulties in the application as plug and play sensors. To bridge this difference and make it as a standard, writer discusses about the useful of UFDC-1 (Universal Frequency to Digital Converter) which is an integrated converter for frequency, ratio, duty cycle and phase-shift, and applies UFDC-1 to the IEEE 1451 family. With using IEEE 1451-2 as the standard of transducer and interface which involves new TEDS for frequency output sensors that contains parameters that are needed in the deployment and IEEE 1451-4 as the standard of multi mode transducer that allows this sensor to connect with a system that supports this standard, especially for data acquisition system. In my opinion, the useful of using frequency output sensors is depend on the application and the resource of node or system that is going to use this sensor. It is very useful if there is a standard for the frequency output because it will really help on the design of a system that is going to use it.
Semantics-enabled framework for knowledge discovery from Earth observation data archives
Durbha, S.S.; King, R.L., Semantics-enabled framework for knowledge discovery from Earth observation data archives, IEEE Trans. Geoscience and Remote Sensing, Vol. 43, no.11, pp.2563-2572, Nov. 2005
This paper is discussing about a framework calls Intelligent Interactive Image Knowledge Retrieval (I3KR) that is successfully used in data retrieval from earth observation data archives, which is very large amount of data. Problem that occurs was a limitation on retrieving image information based on query on structural metadata. That is why a new method needs to be adopted such that it has capability to access the data based on content and semantics. The purpose of this framework is to allow context sensitive query processing over the archives. It is designed in semantic way and it is using domain-specific ontologies as the model of the information and it makes the query process much more efficient and allow user to find the information from such huge amount of data. In this paper, writers also describe about the ontologies method that is used in this framework design and it has block diagram that explain about the ontologies. There is an integrated environment available for the user with possibility to interact with the system and allows the user to search or find information that is needed from the archives. It is a great application of semantic ways of the data query processing especially in big archives which become the necessity of any content management in the world.
SensorML
Open Geospatial Consortium, OpenGIS Sensor Model Language (SensorML) Implementation Specification, Document #OGC 05-086r2, Feb. 2006
This document is describing about SensorML, one of main components from SWE (Sensor Web Enablement) where in SensorML, all activities that are conducted by sensors are modeled as processes and any of this process has its own description depend on what the behavior of the sensor in the system. By doing this, it will allow another application to access any sensor in the system. It uses XML Schema for the design of every model of sensor that is created inside SensorML. The document contains all details about the structure of sensorML and it also has the xml code for sensorML. I believe the idea of having a platform specific for the sensors and model each sensor as a process will become the basic design of complete sensor network for the application because it will be very helpful in managing the sensor network especially large sensor networks. It also can be connected to the web and each sensor is accessible remotely through internet.
Presentation April 9, 2007
Referenced Broadcast Synchronization
J. Elson, L. Girod and D. Estrin, "Fine-Grained Network Time Synchronization using Reference Broadcasts", In Proceedings of the Fifth Symposium on Operating Systems Design and Implementation (OSDI 2002) , Boston, MA. December 2002.
Presentation is here
Review April 16,2007
HTM
Jeff Hawkins and Dileep George, "Hierarchical Temporal Memory: Concepts, Theory, and Terminology" from Numenta
This paper introduces Hierarchical Temporal Memory (HTM) that is built based on how neocortex works. Neocortex is part of human brain that works with human’s sensor and it makes human to know what human’s sensor detects, such as Hot, Cool, Warm, or from eye, human can see tree, cat and so on. And the idea to mimic this neocortex to sensor network is to achieve capability like human’s brain in detecting some specific thing. With some certain algorithm that is stored in the node and with specific sensor to detect it, the network can take those data from sensors and make a conclusion what the meaning of those reading and then send that information to the higher level node. With conclusions from all lower nodes, higher level nodes also can make certain decision or conclusion to what they detected and further, higher level nodes also can make a prediction of what will happen next and send this information to the higher level nodes. And it keeps going to the user’s place. To be able to do this, the network needs to be trained first before can make any conclusion. It is a big step for sensor network to implement HTM. It is similar with how Neural Network works in concepts but most of the application of neural network is done by computer. In sensor networks, to be able to implement HTM and has capability to retrain itself, it will depends on the node itself. It needs a computer like node which has faster processor and bigger memory and a certainty in the communication because it will not work without certainty of communication link and it also might cost more than regular sensor networks. It will not be practical in the application if the user has to train the network for periodic time to adjust the capability of the network and I believe that the training process will not be enough if it only happens just before deployment. It is an interesting paper to read and the writers well explain about HTM concept and put some examples about how HTM works in sensor network and the application from it. One suggestion for this paper is to include the power consumption from the previous HTM application to show whether this new technology consumes more power or more efficient in the power consumption.
UWB(Ultra Wideband)
UWB (Ultra Wideband) is a name that is given to radio technology that has bandwidth larger than 500 MHz. The application for UWB is approved for high frequency (3.1 – 10.6 GHz). There are several examples of device that use UWB such as radar, positioning device, imaging device, etc. These devices use pulses in UWB for the operation and this pulses can be in low rate or high rate. With high rate pulses, in communication system, provide a high speed communication (GHz) which is very useful in some applications that rely on communication. The difference between UWB and lower band communication is in the method of transmitting information. In lower band, amplitude/frequency/phase modulation is the technique, but in UWB, modulation can also be applied by using orthogonal pulses, or by generating radio energy for pulse position or time modulation. More technique can be applied in UWB compare with lower band because it is needed to control or to maintain the communication in due to noise that involves in high frequency communication and lots of information that spread in the band. It is a good article for knowledge about UWB and it will be better if the writer adds some information about the differences between UWB and lower bandwidth beside the modulation technique and maybe also provide an example list of frequency that is used by devices that already registered.
