A wireless sensor network For structural monitoring
read more
Citations
Health monitoring of civil infrastructures using wireless sensor networks
Fidelity and yield in a volcano monitoring sensor network
Data Streams: Models and Algorithms
Environmental Sensor Networks: A revolution in the earth system science?
Environmental Sensor Networks: A revolution in Earth System Science?
References
Directed diffusion: a scalable and robust communication paradigm for sensor networks
Wireless sensor networks for habitat monitoring
System architecture directions for networked sensors
JPEG2000 : image compression fundamentals, standards, and practice
Wavelets and Subband Coding
Related Papers (5)
Frequently Asked Questions (16)
Q2. What is the a priori analyis of training data?
The choice of signal threshold and number of quantization bins is assumed to be determined by a priori analyis of training data to obtain maximum compression benefit within the specified error bounds.
Q3. How did the authors perform the wavelet lifting transform?
The authors were able to perform sensor data sampling, 128 sample CDF(2,2) wavelet lifting transform as well as writing the decomposed buffer to the EEPROM for sampling rates upto 250Hz.
Q4. How can the authors reduce the current draw of the vibration card?
The authors believe the authors can reduce the current draw by customizing the board to their application using various tricks: removing the on-board microprocessor and the SRAM, and multiplexing the signal conditioning circuitry across the different channels.
Q5. What is the role of wireless sensor networks in structural engineering?
Advances in structural engineering depend upon the availability of many detailed data sets that record the response of different structures to ambient vibration (caused, for example, by earthquakes, wind, or passing vehicles) or forced excitation (delivered by large special-purpose shakers).
Q6. How long does it take to declare a bridge open?
A transportation agency is ready to declare a newly built bridge open, but allows a team of structural engineers one or two days to measure structural properties by, for example, driving a large truck through the bridge.
Q7. How does the slope of the line correspond to the drift?
The slope of the line corresponds to a drift of 10ppm, which matches the relative drift the authors measured usingbelieve that because their timestamping requirements are relatively coarse (order of ms), they will not affect Wisden significantly.
Q8. What are the two factors that influenced the choice of the above lifting transform over other kernels?
The predict and update operations for the CDF(2,2) lifting transform are:di ← di − 12 (si + si+1)si ← si − 14 (−di−1 − di)The choice of the above lifting transform over other kernels was based on two factors: computation overhead and compression performance.
Q9. What is the need to timestamp the samples?
Even if the data sets are used only for frequency analysis, it is necessary to timestamp the samples in order to distinguish responses due to different events.
Q10. How many packets are sent per node?
if the authors set the sending rate to 2 packet/sec per node, their network essentially collapses and very few of the packets are received.
Q11. What is the main reason for collecting structural response to ambient vibrations?
systems that collect structural response to ambient vibrations are generally long-running, since the occurrence of significant ambients may be unpredictable.
Q12. How did the authors extract the clock signal from the Mica-2s?
The authors used an oscilloscope to extract the clock signal from 7 different Mica-2s, then analyzed the dominant frequency in the signal.
Q13. How can the authors reduce the acquisition latency of vibration data?
As the authors show below, such an approach can compress vibration data by a factor of 20; when coupled with event detection, it can reduce the acquisition latency to less than a minute in many cases.
Q14. What is the overall data rate required to transmit the samples?
the overall data rate required to transmit the samples is a function of the duty-cycle of the vibrations, and directly affects7
Q15. Why do the authors perform the normalization operations during the wavelet thresholding step?
The authors perform the normalization operations during the wavelet thresholding step rather than during wavelet decomposition to be more computationally efficient.
Q16. How can the authors measure packet loss performance?
This packet loss performance can be measured both passively (using actual data transmissions) and actively (using probes sent by nodes).