Accomplishing Better than Human Design in Detecting Events of Interest in Bandwidth Constrained Sensor Networks .


49 views
Uploaded on:
Description
Achieving Better than Human Design in Detecting Events of Interest in Bandwidth Constrained Sensor Networks. Kalyan Veeramachaneni, Lisa Osadciw Development and Research in Evolutionary Algorithms for Multi sensor Smart Networks (DREAMSNet)
Transcripts
Slide 1

Accomplishing Better than Human Design in Detecting Events of Interest in Bandwidth Constrained Sensor Networks Kalyan Veeramachaneni, Lisa Osadciw Development and Research in Evolutionary Algorithms for Multi sensor Smart Networks (DREAMSNet) Department of Electrical Engineering and Computer Science Syracuse University, New York GECCO 2008, Human Competitive Results Awards, July 14, Atlanta, U.S.A

Slide 2

What are we identifying? Advanced society depends on discovery or deciding the importance of the nearness or nonappearance of a flag Digital Communications Pipeline/Bridges break identification Genuine User recognition utilizing biometrics Presence of flying machine, ships, or engine vehicles Locating crisis work force Weather Phenomena Building Security Sensors are situated in remote territories settling on choices utilizing an assortment of criteria Maximum A-Posteriori Criterion Maximum Likelihood Criterion Minimum Error Criterion

Slide 3

Sensor1 u 1 X 1 Fusion Rule Sensor2 X 2 u 2 AND OR Second Classifier Only First Sensor Only Bandwidth Constrained Detection Networks Noise just Event Likelihood thickness demonstrate for a sensor

Slide 4

Event is announced just in this quadrant, i.e. Furthermore, run False Alarms: recognizing an occasion that did not happen Threshold on Sensor 2 Noise Misses: Fail to identify an occasion * Event Threshold on Sensor 1 Bandwidth Constrained Detection Networks Two sorts of Errors should be diminished If the whole perception esteem is transmitted to a focal handling hub, an effective machine learning system can be intended to accomplish better exactness Shown beneath are 20000 specimens of perceptions, 10000 have a place with occasions, 10000 to commotion. 9 to 32 bits required per test if all bits are transmitted Reduces to 1 bit choice if choice is transmitted rather

Slide 5

Region where an occasion is proclaimed Region where an occasion is pronounced LRT (Human) Based Design: 2 limits on every sensor 2 Sensor just combination control PSO Based Design: Simple 1 Threshold for every sensor AND combination lead Very couple of blunders (E)Competitive Result: Correlated Sensors Designs for 0.1 Correlation

Slide 6

Humies Categories Covered (G) The outcome tackles an issue of undeniable trouble in its field. (B) The outcome is equivalent to or superior to an outcome that was acknowledged as another logical result when it was distributed in an associate checked on logical diary. (D) The outcome is publishable in its own perfectly fine new logical result - autonomous of the way that the outcome was mechanically made. (E) The outcome is equivalent to or superior to anything the latest human-made answer for a long-standing issue for which there has been a progression of progressively better human-made arrangements.

Slide 7

HUMIES Category G The outcome takes care of an issue of unquestionable trouble in its field. Measure of Research and Publications on Topic Indicates Complexity Quick Check Research Publications 120 Journal Articles with Approximately 45 Discussing Similar Design Issues 48 Textbooks At Least Currently On Sale In This Area 5 Dissertations manage same issue and give human created outlines Paper Published that Addresses the Difficulty John N Tsitsiklis, Michael Athans, "On Complexity of Decentralized Decision making and location issues" 23rd IEEE Conference on Decision and Control, 1984 Optimizing Distributed Detection for 2 Sensors Independent sensors: Intractable Correlated sensors: NP Complete -

Slide 8

HUMIES Category B The outcome is equivalent to or superior to an outcome that was acknowledged as another logical result when it was distributed in an associate inspected logical diary Much Research Published in Area Since the 50s/60s Beginning in Radar Type I/Type II Errors Fail to recognize the occasion Detect an Event that did not happen Decouple the two issues: advance limits and plan best combination administer independently When just marked preparing datasets are accessible execution is delicate to edge look exactness When probability models are accessible Optimize Threshold Exceed Ratio of Conditional Distributions

Slide 9

Human Design Solution: Person-by-Person Optimal (PBPO) for Independent Sensors Human Design Solution: Likelihood Ratio Test (LRT) Design Human Competitive Result: Particle Swarm Optimization (PSO) Based Design Sensor1 u 1 X 1 Fusion Rule Sensor2 X 2 u 2 Optimize edges exclusively by keeping different edges and combination govern consistent Use LRT for autonomous or connected inferring combination run Joint improvement of edges and Fusion Rule No shut frame arrangement exists

Slide 10

HUMIES Category B (cont) Particle Swarm Optimized Detector Simplifies Sensor Network Adaptation Able to consolidate execution parameters to at the same time handle an assortment of circumstances consider assets (vitality and correspondence transfer speed) Reduce sort I or sort II blunders crosswise over various degrees of relationship Simpler Receiver Single edge plan contrasted with LRT based plans that can prompt numerous edges as the Likelihood proportion gets to be non straight Adapt to plan for either likelihood thickness models or a named preparing dataset gave Automatically Handles the Heterogeneity in Practical Sensor Networks

Slide 11

HUMIES Category D The outcome is publishable in its own great new logical result - free of the way that the outcome was mechanically made . 5 Papers Published Including a Best Paper Correlated Sensors Kalyan Veeramachaneni and Lisa Osadciw , "Plan of Distributed Detection Systems with Heterogeneous Correlated Sensors," 44th Annual Allerton Conference on Communications and Control , Allerton Park, Illinois, September, 2007. Autonomous Sensors Kalyan Veeramachaneni, Lisa Osadciw , Pramod Varshney"Adaptive Multimodal Biometric Management Algorithm," IEEE Transactions on Systems Man and Cybernatics : Part C: Applications and Reviews , Vol. 35, No. 3 August 2005. Applications Biometrics : Kalyan Veeramachaneni , Nisha Srinivas, Lisa Osadciw , and Arun Ross, "Planning Optimal Fusion Strategies for Correlated Biometric Classifiers", IEEE CVPR Conference, Anchorage , Alaska, June, 2008. (Best Paper Award) Pipeline Crack Detection : Kalyan Veeramachaneni , Weizhong Yan, Kai Goebel, and Lisa Osadciw , "Enhancing Classifier Fusion Using Particle Swarm Optimization", IEEE Multi-Criteria Decision Making (MCDM) Symposium , Honolulu, Hawaii, April, 2007. Versatile Sensor Management Lisa Osadciw and Kalyan Veeramachaneni , "Sensor Management through Efficient Fitness Function Design," Proceedings of 41st Annual Asilomar Conference on Signals, Systems, and Computers , Asilomar, CA, November, 2007.

Slide 12

HUMIES Category E The outcome is equivalent to or superior to anything the latest human-made answer for a long-standing issue for which there has been a progression of progressively better human-made arrangements. Long-standing issue since the 1950s in Radar Research Succession of better arrangements as examined in Category B First Single Detectors Derived for the Following Criterion Maximum A-Posteriori Criterion – augment the posteriori likelihood of having a place with one occasion to the next conceivable occasion Maximum Likelihood Criterion – expands likelihood of having a place (probability) to occasion Minimum Error Criterion – minimize the quantity of blunders in choices

Slide 13

HUMIES Category E (cont): Matched Filter Designed in 50s and 60s from Radar Maximum Signal to Noise Criterion – amplify motion over the commotion foundation to help recognition by coordinated channel (North, Van Vleck, Middleton) Inverse Probability Criterion – (Wald, Neyman, Pearson) Likelihood Ratio - in view of Shannon\'s data hypothesis (Woodward & Davis) Distributed Detection (Tenney & Sandell-1979 through today) Chair Z., P. K. Varshney, "Optimal Data Fusion in Multiple Sensor Detection Systems," IEEE Trans. on Aerospace and Elect. Frameworks, Vol. AES-22, No. 1, pp. 98-101, Jan. 1986 . Tang, Z. - B., K. R. Pattipati, and D. L. Kleinman, "An Algorithm for Determining the Decision Thresholds in a Distributed Detection Problem," IEEE Trans. on Systems, Man, and Cybernetics, Vol. SMC-21, pp. 231-237, Jan./Feb. 1991. Kam., M., Q. Zhu., and W. S. Dim, "Optimal Data Fusion of Correlated Local Decisions in Multiple Sensor Detection Systems," IEEE Transactions on Aerospace and Elect. Syst., Vol. 28, pp. 916-920, July 1992. Subside Willet, Peter F. Swaszek, Rick S. Blum, "The Good, Bad, and Ugly : Distributed Detection of Known Signal in Dependent Gaussian Noise," IEEE Transactions on Signal Processing, Vol. 48, No. 12, December 2000. Kalyan Veeramachaneni, Lisa Ann Osadciw, Pramod K Varshney, "An Adaptive Multimodal Biometric Fusion Algorithm Using Particle Swarm", Proceedings of SPIE, Aerosense, April 21-25, 2003, Orlando. Saeed A. Aldosari, Jose M. F. Moura, "Combination in Sensor Networks with Communication Constraints", International Symposium on Information Processing in Sensor Networks, April 26-27, 2004, Berkeley, CA.

Slide 14

HUMIES Category E (cont): Swarm Solution: Type I/Type II Errors are Balanced in Real Time Based on Current System Needs Simultaneously lessen, "Inability to identify the occasion", "Distinguish an Event that did not happen " Reduce Communication Bandwidth Decisions at Sensor to Reduce Message Size Saving Bandwidth Fusion Architecture Can Be Modified in Real-Time Based on Bandwidth and Energy Needs Minimize Energy Save in interchanges with Smaller Messages and Fewer Through Fusion Reduce calculations with Simpler plans and Fusion Rules Ease of Adaptation to Other Applications Communication Management for Any Wireless Sensor Network and Architecture Various Sensor Networks for Aircraft Routing at Airports, First Response Networks, Large, Remote Sensor Networks, Health Monitoring Sensor Networks, and so on

Recommended
View more...