We consider the problem of estimating the location of a target in 2D using random measurements of... more We consider the problem of estimating the location of a target in 2D using random measurements of sensorspsila distance-differences to it. The linear algebraic solution to this problem becomes incomplete under random measurement errors, while the geometric method that always returns a solution has been developed only for 3-sensor configurations. We present three methods to extend the geometric method to a larger number of sensors, and compare their performance using simulations. We then consider two cases where measurements are collected over the same time-window at all sensors: (a) measurements are time-synchronized, and (b) measurements are independently collected within the time-window. We present the localization methods, mean of estimators and mean of measurements for case (a) and (b), respectively, and compare them using simulations.We describe a method for localizing point radiation sources from intensity measurements which uses the localization as a component. Based on simul...
Proceedings of the 6th ACM conference on Embedded network sensor systems, 2008
The localization of a radioactive source can be solved in closed-form using 4 ideal sensors and t... more The localization of a radioactive source can be solved in closed-form using 4 ideal sensors and the Apollonius circle in a noise-and error-free environment. When measurement errors and noise such as background radiation are considered, a larger number of sensors is needed to produce accurate results, particularly for extremely low source intensities. In this paper, we present an efficient fusion algorithm that can exploit measurements from n sensors to improve the localization accuracy, and show how the accuracy scales with n. We report testbed results for a 0.911 μCi source to illustrate the effectiveness of our algorithm, in particular performance comparisons with state-of-the-art fusion algorithms based on Mean of Estimates (MoE) and Maximum Likelihood Estimation (MLE). We show that ITP is more accurate than MoE, whereas the choice between ITP and MLE is generally a tradeoff between accuracy and run time efficiency. Higher-intensity radioactive sources are not safe for actual experiments. In this case, we present simulation results based on a validated simulation model. We show that a low-intensity 400 μCi source, similar to the radioactivity of a concealed dirty bomb, can be localized to within 32.5 m using a sensor density of about 1 per 1100 m 2 in a surveillance area.
Ad-hoc networks of devices and sensors with (limited) sensing and wireless communication capabili... more Ad-hoc networks of devices and sensors with (limited) sensing and wireless communication capabilities are becoming increasingly available for commercial and military applications. The first step in deploying these wireless sensor networks is to determine, with respect to application-specific performance criteria, (i) in the case that the sensors are static, where to deploy or activate them; and (ii) in the case that (a subset of) the sensors are mobile, how to plan the trajectory of the mobile sensors. These two cases are collectively termed as the coverage problem in wireless sensor networks. In this book chapter, we give a comprehensive treatment of the coverage problem. Specifically, we first introduce several fundamental properties of coverage that have been derived in the literature and the corresponding algorithms that will realize these properties. While giving insights on how optimal operations can be devised, most of the properties are derived (and hence their corresponding algorithms are constructed) under the perfect disk assumption. Hence, we consider in the second part of the book chapter coverage in a more realistic setting, and allow (i) the sensing area of a sensor to be anisotropic and of arbitrary shape, depending on the terrain and the meteorological conditions, and (ii) the utilities of coverage in different parts of the monitoring area to be non-uniform, in order to account for the impact of a threat on the population, or the likelihood of a threat taking place at certain locations. Finally, in the third part of the book chapter, we consider mobile sensor coverage, and study how mobile sensors may navigate in a deployment area in order to maximize threat-based coverage.
The sheer number and sophistication of cyber attacks are making our nation's critical comput... more The sheer number and sophistication of cyber attacks are making our nation's critical computer networks increasingly vulnerable. At the same time, these networks are being called upon to play a key role in processing, data storage, monitoring and control of critical ...
We analyze the quality of monitoring (QoM) of stochastic events by a periodic sensor which monito... more We analyze the quality of monitoring (QoM) of stochastic events by a periodic sensor which monitors a point of interest (PoI) for q time every p time. We show how the amount of information captured at a PoI is affected by the proportion q/p , the time interval p over which the proportion is achieved, the event type in terms of its stochastic arrival dynamics and staying times and the utility function. The periodic PoI sensor schedule happens in two broad contexts. In the case of static sensors, a sensor monitoring a PoI may be periodically turned off to conserve energy, thereby extending the lifetime of the monitoring until the sensor can be recharged or replaced. In the case of mobile sensors, a sensor may move between the PoIs in a repeating visit schedule. In this case, the PoIs may vary in importance, and the scheduling objective is to distribute the sensor's coverage time in proportion to the importance levels of the PoIs. Based on our QoM analysis, we optimize a class of p...
Identification of a low-level point radioactive source amidst background radiation is achieved by... more Identification of a low-level point radioactive source amidst background radiation is achieved by a network of radiation sensors using a two-step approach. Based on measurements from three or more sensors, a geometric difference triangulation method or an N -sensor localization method is used to estimate the location and strength of the source. Then a sequential probability ratio test based on current measurements and estimated parameters is employed to finally decide: (1) the presence of a source with the estimated parameters, or (2) the absence of the source, or (3) the insufficiency of measurements to make a decision. This method achieves specified levels of false alarm and missed detection probabilities, while ensuring a close-to-minimal number of measurements for reaching a decision. This method minimizes the ghost-source problem of current estimation methods, and achieves a lower false alarm rate compared with current detection methods. This method is tested and demonstrated u...
distributed sensor network (MADSN) distributes the computation into the participating leaf nodes.... more distributed sensor network (MADSN) distributes the computation into the participating leaf nodes. And this approach hence makes it possible to significantly reduce the consumption of communication power and bandwidth, while lowering the risk of being spied with hostile intent. A MADSN is usually divided into an appropriate number of subtasks, each carried out by a mobile agent carrying the executable instructions for data integration dispatched by the processing element. The agents selectively visit the leaf sensors along a certain path to fuse the data incrementally on a sequential basis. A final data fusion is performed when all mobile agents return to the processing element. Three technical issues associated with MADSN are addressed in [5]: mobile agent routing, data integration, and optimum performance.
We propose a circuit-switched high-speed end-to-end transport architecture (CHEETAH) as a network... more We propose a circuit-switched high-speed end-to-end transport architecture (CHEETAH) as a networking solution to provide high-speed end-to-end circuit connectivity to end hosts on a dynamic call-by-call basis. Not only is it envisioned as a complementary service to the basic connectionless service provided by today's Internet; it also relies on and leverages the presence of this service. Noting the dominance of Ethernet in LANs and SONET/SDH in WANs, CHEETAH circuits will consist of Ethernet segments at the ends and Ethernet-over-SONET segments in the wide area. In this article we explain the CHEETAH concept and describe a wide-area experimental network testbed we have deployed based on this concept. The network testbed currently extends between Raleigh, North Carolina, Atlanta, Georgia, and Oak Ridge, Tennessee, and uses off-the-shelf switches. We have created CHEETAH software to run on end hosts to enable automated use of this network by applications. Our first users of this n...
Data transport methods with end-to-end performance guarantees are required by the next generation... more Data transport methods with end-to-end performance guarantees are required by the next generation distributed computing applications. Transport methods based on default best-effort IP networks are not capable of providing such guarantees due to complex traffic distributions in wide-area networks and highly dynamic network conditions. In this paper, we present a measurement-based transport method using source routing strategy to minimize message delays via multiple quickest paths. This method is based on a linear approximation using effective bandwidth, and is computationally efficient and also analytically tractable under fairly general conditions. This method is implemented using application-level daemons and tested at Internet nodes. The results illustrate its superior performances over the default TCP/IP method.
This paper discusses algorithmic and implementation aspects of a distributed remote visualization... more This paper discusses algorithmic and implementation aspects of a distributed remote visualization system that optimally decomposes and adaptively maps the visualization pipeline to a wide-area shared or dedicated network. The first node of the system typically generates or stores raw data sets, and a remote client resides on the last node equipped with a display device ranging from a personal desktop to a powerwall. Intermediate nodes include workstations, clusters, or rendering engines, which can be located anywhere on the network. We employ a regression method to estimate the effective bandwidth of a transport path. Based on link measurements, node characteristics, and module properties, we strategically organize visualization pipeline modules into groups and dynamically assign the groups to various network nodes to achieve minimal total delay or maximal frame rate. We propose polynomial-time algorithms using the dynamic programming method to compute optimal solutions for the problems of pipeline decomposition and network mapping under different constraint conditions. The proposed remote visualization system is implemented and deployed at several geographically distributed nodes for experimental testing. The proposed decomposition and mapping scheme is generic and can be applied to other distributed applications whose computing components form a linear arrangement.
We consider a network of sensors distributed in a target area providing environmental measurement... more We consider a network of sensors distributed in a target area providing environmental measurements that are subject to normally distributed, independent additive noise. Each sensor node applies a threshold rule to the measurements to decide the presence of a target; the distance to the target together with the threshold value de- termines its hit and false alarm probabilities or rates using a signal attenuation model. We propose a centralized threshold-OR fusion rule for combining the individual sensor node decisions. Under the statistical independence of sensor measurements, we derive fusion threshold bounds using Chebyshev's inequality based on individual hit and false alarm probabilities but without requiring a priori knowledge of the underlying probability distribu- tions. We derive conditions to ensure that the fused method achieves a higher hit rate and lower false alarm rate compared to the weighted averages of individual sensor parameters. The simulations using Monte Car...
ABSTRACT Networks of radiation counters have been recently developed for detecting low-level, haz... more ABSTRACT Networks of radiation counters have been recently developed for detecting low-level, hazardous radiation sources, and they have been utilized in indoor and outdoor characterization tests. Subsequently, the test measurements have been “replayed” using multiple sub-networks, which enabled the analysis of various scenarios beyond the tests. We present a particle filter algorithm that combines measurements from gamma counters across the network to detect radiation sources. Using replays from an outdoor test, we construct a border monitoring scenario that consists of twelve 2 in.×2 in. NaI detectors or counters deployed on the periphery to monitor a 42×42 m2 region. A 137Cs source is moved across this region, starting several meters outside and finally moving away from it. The measurements from individual, pairs and boundary detectors are replayed using the particle filter algorithm. The algorithm outputs demonstrate, both quantitatively and qualitatively, the benefits of networking all boundary counters: the source is detected meters before it enters the region, while being inside, and until moving several meters away. On the other hand, when counters are used individually or in pairs, the source is detected for much shorter durations, and sometimes not detected at all while inside the region.
A cloud computing infrastructure typically consists of a number of sites that house servers and a... more A cloud computing infrastructure typically consists of a number of sites that house servers and are connected to the Internet. Its operation critically depends both on cyber components, including servers and routers, and physical components, including fiber and power routes. Both types of components are subject to attacks of different kinds and frequencies, which must be accounted for the initial provisioning and subsequent operation of the infrastructure. The cyber and physical components may be individually attacked and defended, and the infrastructure is required to provide an aggregate computational capacity C. We present a game-theoretic approach for the provisioning and operation of the infrastructure under uniform cost models. We first show that the Nash Equilibrium under different formulations to be computable in polynomial time, and derive provisioning choices to ensure the capacity C with probability PS. Then, we derive conditions for reinforcing the infrastructure, and sh...
We investigate the resilience of cyber physical systems by modeling the interaction between provi... more We investigate the resilience of cyber physical systems by modeling the interaction between provider and attacker as a simultaneous game that incorporates cyber and physical spaces. Both the provider and attacker aim to maximize their individual utility, which is determined by a trade-off between target revenue and investment cost. The system resilience function is formulated as a power-form product of the survival probabilities of cyber and physical spaces, each with a corresponding correlation coefficient. The contest success functions based on the reinforcement and attack levels are used to estimate the survival probabilities of cyber and physical spaces. We present the provider strategies based on the Nash equilibrium of the game, and analyze the sensitivities with respect to cyber and physical correlation coefficients, target revenues and costs. The results show that these correlation coefficients affect the cyber and physical reinforcement strategies, and also provide new insi...
We consider the problem of estimating the location of a target in 2D using random measurements of... more We consider the problem of estimating the location of a target in 2D using random measurements of sensorspsila distance-differences to it. The linear algebraic solution to this problem becomes incomplete under random measurement errors, while the geometric method that always returns a solution has been developed only for 3-sensor configurations. We present three methods to extend the geometric method to a larger number of sensors, and compare their performance using simulations. We then consider two cases where measurements are collected over the same time-window at all sensors: (a) measurements are time-synchronized, and (b) measurements are independently collected within the time-window. We present the localization methods, mean of estimators and mean of measurements for case (a) and (b), respectively, and compare them using simulations.We describe a method for localizing point radiation sources from intensity measurements which uses the localization as a component. Based on simul...
Proceedings of the 6th ACM conference on Embedded network sensor systems, 2008
The localization of a radioactive source can be solved in closed-form using 4 ideal sensors and t... more The localization of a radioactive source can be solved in closed-form using 4 ideal sensors and the Apollonius circle in a noise-and error-free environment. When measurement errors and noise such as background radiation are considered, a larger number of sensors is needed to produce accurate results, particularly for extremely low source intensities. In this paper, we present an efficient fusion algorithm that can exploit measurements from n sensors to improve the localization accuracy, and show how the accuracy scales with n. We report testbed results for a 0.911 μCi source to illustrate the effectiveness of our algorithm, in particular performance comparisons with state-of-the-art fusion algorithms based on Mean of Estimates (MoE) and Maximum Likelihood Estimation (MLE). We show that ITP is more accurate than MoE, whereas the choice between ITP and MLE is generally a tradeoff between accuracy and run time efficiency. Higher-intensity radioactive sources are not safe for actual experiments. In this case, we present simulation results based on a validated simulation model. We show that a low-intensity 400 μCi source, similar to the radioactivity of a concealed dirty bomb, can be localized to within 32.5 m using a sensor density of about 1 per 1100 m 2 in a surveillance area.
Ad-hoc networks of devices and sensors with (limited) sensing and wireless communication capabili... more Ad-hoc networks of devices and sensors with (limited) sensing and wireless communication capabilities are becoming increasingly available for commercial and military applications. The first step in deploying these wireless sensor networks is to determine, with respect to application-specific performance criteria, (i) in the case that the sensors are static, where to deploy or activate them; and (ii) in the case that (a subset of) the sensors are mobile, how to plan the trajectory of the mobile sensors. These two cases are collectively termed as the coverage problem in wireless sensor networks. In this book chapter, we give a comprehensive treatment of the coverage problem. Specifically, we first introduce several fundamental properties of coverage that have been derived in the literature and the corresponding algorithms that will realize these properties. While giving insights on how optimal operations can be devised, most of the properties are derived (and hence their corresponding algorithms are constructed) under the perfect disk assumption. Hence, we consider in the second part of the book chapter coverage in a more realistic setting, and allow (i) the sensing area of a sensor to be anisotropic and of arbitrary shape, depending on the terrain and the meteorological conditions, and (ii) the utilities of coverage in different parts of the monitoring area to be non-uniform, in order to account for the impact of a threat on the population, or the likelihood of a threat taking place at certain locations. Finally, in the third part of the book chapter, we consider mobile sensor coverage, and study how mobile sensors may navigate in a deployment area in order to maximize threat-based coverage.
The sheer number and sophistication of cyber attacks are making our nation's critical comput... more The sheer number and sophistication of cyber attacks are making our nation's critical computer networks increasingly vulnerable. At the same time, these networks are being called upon to play a key role in processing, data storage, monitoring and control of critical ...
We analyze the quality of monitoring (QoM) of stochastic events by a periodic sensor which monito... more We analyze the quality of monitoring (QoM) of stochastic events by a periodic sensor which monitors a point of interest (PoI) for q time every p time. We show how the amount of information captured at a PoI is affected by the proportion q/p , the time interval p over which the proportion is achieved, the event type in terms of its stochastic arrival dynamics and staying times and the utility function. The periodic PoI sensor schedule happens in two broad contexts. In the case of static sensors, a sensor monitoring a PoI may be periodically turned off to conserve energy, thereby extending the lifetime of the monitoring until the sensor can be recharged or replaced. In the case of mobile sensors, a sensor may move between the PoIs in a repeating visit schedule. In this case, the PoIs may vary in importance, and the scheduling objective is to distribute the sensor's coverage time in proportion to the importance levels of the PoIs. Based on our QoM analysis, we optimize a class of p...
Identification of a low-level point radioactive source amidst background radiation is achieved by... more Identification of a low-level point radioactive source amidst background radiation is achieved by a network of radiation sensors using a two-step approach. Based on measurements from three or more sensors, a geometric difference triangulation method or an N -sensor localization method is used to estimate the location and strength of the source. Then a sequential probability ratio test based on current measurements and estimated parameters is employed to finally decide: (1) the presence of a source with the estimated parameters, or (2) the absence of the source, or (3) the insufficiency of measurements to make a decision. This method achieves specified levels of false alarm and missed detection probabilities, while ensuring a close-to-minimal number of measurements for reaching a decision. This method minimizes the ghost-source problem of current estimation methods, and achieves a lower false alarm rate compared with current detection methods. This method is tested and demonstrated u...
distributed sensor network (MADSN) distributes the computation into the participating leaf nodes.... more distributed sensor network (MADSN) distributes the computation into the participating leaf nodes. And this approach hence makes it possible to significantly reduce the consumption of communication power and bandwidth, while lowering the risk of being spied with hostile intent. A MADSN is usually divided into an appropriate number of subtasks, each carried out by a mobile agent carrying the executable instructions for data integration dispatched by the processing element. The agents selectively visit the leaf sensors along a certain path to fuse the data incrementally on a sequential basis. A final data fusion is performed when all mobile agents return to the processing element. Three technical issues associated with MADSN are addressed in [5]: mobile agent routing, data integration, and optimum performance.
We propose a circuit-switched high-speed end-to-end transport architecture (CHEETAH) as a network... more We propose a circuit-switched high-speed end-to-end transport architecture (CHEETAH) as a networking solution to provide high-speed end-to-end circuit connectivity to end hosts on a dynamic call-by-call basis. Not only is it envisioned as a complementary service to the basic connectionless service provided by today's Internet; it also relies on and leverages the presence of this service. Noting the dominance of Ethernet in LANs and SONET/SDH in WANs, CHEETAH circuits will consist of Ethernet segments at the ends and Ethernet-over-SONET segments in the wide area. In this article we explain the CHEETAH concept and describe a wide-area experimental network testbed we have deployed based on this concept. The network testbed currently extends between Raleigh, North Carolina, Atlanta, Georgia, and Oak Ridge, Tennessee, and uses off-the-shelf switches. We have created CHEETAH software to run on end hosts to enable automated use of this network by applications. Our first users of this n...
Data transport methods with end-to-end performance guarantees are required by the next generation... more Data transport methods with end-to-end performance guarantees are required by the next generation distributed computing applications. Transport methods based on default best-effort IP networks are not capable of providing such guarantees due to complex traffic distributions in wide-area networks and highly dynamic network conditions. In this paper, we present a measurement-based transport method using source routing strategy to minimize message delays via multiple quickest paths. This method is based on a linear approximation using effective bandwidth, and is computationally efficient and also analytically tractable under fairly general conditions. This method is implemented using application-level daemons and tested at Internet nodes. The results illustrate its superior performances over the default TCP/IP method.
This paper discusses algorithmic and implementation aspects of a distributed remote visualization... more This paper discusses algorithmic and implementation aspects of a distributed remote visualization system that optimally decomposes and adaptively maps the visualization pipeline to a wide-area shared or dedicated network. The first node of the system typically generates or stores raw data sets, and a remote client resides on the last node equipped with a display device ranging from a personal desktop to a powerwall. Intermediate nodes include workstations, clusters, or rendering engines, which can be located anywhere on the network. We employ a regression method to estimate the effective bandwidth of a transport path. Based on link measurements, node characteristics, and module properties, we strategically organize visualization pipeline modules into groups and dynamically assign the groups to various network nodes to achieve minimal total delay or maximal frame rate. We propose polynomial-time algorithms using the dynamic programming method to compute optimal solutions for the problems of pipeline decomposition and network mapping under different constraint conditions. The proposed remote visualization system is implemented and deployed at several geographically distributed nodes for experimental testing. The proposed decomposition and mapping scheme is generic and can be applied to other distributed applications whose computing components form a linear arrangement.
We consider a network of sensors distributed in a target area providing environmental measurement... more We consider a network of sensors distributed in a target area providing environmental measurements that are subject to normally distributed, independent additive noise. Each sensor node applies a threshold rule to the measurements to decide the presence of a target; the distance to the target together with the threshold value de- termines its hit and false alarm probabilities or rates using a signal attenuation model. We propose a centralized threshold-OR fusion rule for combining the individual sensor node decisions. Under the statistical independence of sensor measurements, we derive fusion threshold bounds using Chebyshev's inequality based on individual hit and false alarm probabilities but without requiring a priori knowledge of the underlying probability distribu- tions. We derive conditions to ensure that the fused method achieves a higher hit rate and lower false alarm rate compared to the weighted averages of individual sensor parameters. The simulations using Monte Car...
ABSTRACT Networks of radiation counters have been recently developed for detecting low-level, haz... more ABSTRACT Networks of radiation counters have been recently developed for detecting low-level, hazardous radiation sources, and they have been utilized in indoor and outdoor characterization tests. Subsequently, the test measurements have been “replayed” using multiple sub-networks, which enabled the analysis of various scenarios beyond the tests. We present a particle filter algorithm that combines measurements from gamma counters across the network to detect radiation sources. Using replays from an outdoor test, we construct a border monitoring scenario that consists of twelve 2 in.×2 in. NaI detectors or counters deployed on the periphery to monitor a 42×42 m2 region. A 137Cs source is moved across this region, starting several meters outside and finally moving away from it. The measurements from individual, pairs and boundary detectors are replayed using the particle filter algorithm. The algorithm outputs demonstrate, both quantitatively and qualitatively, the benefits of networking all boundary counters: the source is detected meters before it enters the region, while being inside, and until moving several meters away. On the other hand, when counters are used individually or in pairs, the source is detected for much shorter durations, and sometimes not detected at all while inside the region.
A cloud computing infrastructure typically consists of a number of sites that house servers and a... more A cloud computing infrastructure typically consists of a number of sites that house servers and are connected to the Internet. Its operation critically depends both on cyber components, including servers and routers, and physical components, including fiber and power routes. Both types of components are subject to attacks of different kinds and frequencies, which must be accounted for the initial provisioning and subsequent operation of the infrastructure. The cyber and physical components may be individually attacked and defended, and the infrastructure is required to provide an aggregate computational capacity C. We present a game-theoretic approach for the provisioning and operation of the infrastructure under uniform cost models. We first show that the Nash Equilibrium under different formulations to be computable in polynomial time, and derive provisioning choices to ensure the capacity C with probability PS. Then, we derive conditions for reinforcing the infrastructure, and sh...
We investigate the resilience of cyber physical systems by modeling the interaction between provi... more We investigate the resilience of cyber physical systems by modeling the interaction between provider and attacker as a simultaneous game that incorporates cyber and physical spaces. Both the provider and attacker aim to maximize their individual utility, which is determined by a trade-off between target revenue and investment cost. The system resilience function is formulated as a power-form product of the survival probabilities of cyber and physical spaces, each with a corresponding correlation coefficient. The contest success functions based on the reinforcement and attack levels are used to estimate the survival probabilities of cyber and physical spaces. We present the provider strategies based on the Nash equilibrium of the game, and analyze the sensitivities with respect to cyber and physical correlation coefficients, target revenues and costs. The results show that these correlation coefficients affect the cyber and physical reinforcement strategies, and also provide new insi...
Uploads
Papers by Nagi Rao