Pub Date : 2014-06-24DOI: 10.1109/NAECON.2014.7045790
Erik Blasch
Information fusion consists of organizing a set of data into meaningful reports to answer queries, forge a consistency story, and determine situation awareness. To provide situation understanding requires context both in information estimation and data management. In this paper, we highlight the importance of context estimation, assessment, and management to support information fusion analysis. A demonstrated example for multimodal human-based text analysis and video-based sensing and tracking is shown where context provides the basis for associating the multimodal data correlated in space and time. The use of context is demonstrated as (1) semantic text call-outs from users monitoring a video following a target for classification, (2) geographical road information from a database to locate a target, and (3) sensor-based estimation for simultaneous target tracking and classification. Context assessment and management from human-based and sensor-based sources is shown for information fusion situation awareness.
{"title":"Context aided sensor and human-based information fusion","authors":"Erik Blasch","doi":"10.1109/NAECON.2014.7045790","DOIUrl":"https://doi.org/10.1109/NAECON.2014.7045790","url":null,"abstract":"Information fusion consists of organizing a set of data into meaningful reports to answer queries, forge a consistency story, and determine situation awareness. To provide situation understanding requires context both in information estimation and data management. In this paper, we highlight the importance of context estimation, assessment, and management to support information fusion analysis. A demonstrated example for multimodal human-based text analysis and video-based sensing and tracking is shown where context provides the basis for associating the multimodal data correlated in space and time. The use of context is demonstrated as (1) semantic text call-outs from users monitoring a video following a target for classification, (2) geographical road information from a database to locate a target, and (3) sensor-based estimation for simultaneous target tracking and classification. Context assessment and management from human-based and sensor-based sources is shown for information fusion situation awareness.","PeriodicalId":318539,"journal":{"name":"NAECON 2014 - IEEE National Aerospace and Electronics Conference","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-06-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127891750","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2014-06-24DOI: 10.1109/NAECON.2014.7045812
Yangjie Qi, Bin Zhang, T. Taha, Hua Chen, Raqibul Hasan
The Interest in specialized neuromorphic computing architectures has been increasing recently, and several applications have been shown to be capable of being accelerated on such a platform. This paper describes the implementation of a multicore digital neuromorphic processing system on an Altera Quartus II FPGA. Static routing was used to allow communication between the cores on the FPGA. Two applications were mapped to the system: image edge detection and ECG. Compared to an Intel processor implementation of these applications, the FPGA based neural implementations provided about 3× and 127× speedup for the edge detection and ECG applications. Given that both applications were implemented with the same base Verilog code, with only a change in the synaptic weights and number of neurons utilized, the system has the capability to accelerate a broad range of applications.
最近，人们对专门的神经形态计算架构越来越感兴趣，并且有几个应用程序已经被证明能够在这样的平台上加速。本文介绍了一个多核数字神经形态处理系统在Altera Quartus II FPGA上的实现。静态路由用于允许FPGA上的内核之间的通信。该系统主要应用于图像边缘检测和心电检测。与英特尔处理器实现的这些应用相比，基于FPGA的神经网络实现为边缘检测和心电应用提供了大约3倍和127倍的加速。考虑到这两个应用程序都是用相同的Verilog基本代码实现的，只有突触权重和所用神经元数量的变化，该系统有能力加速广泛的应用程序。
{"title":"FPGA design of a multicore neuromorphic processing system","authors":"Yangjie Qi, Bin Zhang, T. Taha, Hua Chen, Raqibul Hasan","doi":"10.1109/NAECON.2014.7045812","DOIUrl":"https://doi.org/10.1109/NAECON.2014.7045812","url":null,"abstract":"The Interest in specialized neuromorphic computing architectures has been increasing recently, and several applications have been shown to be capable of being accelerated on such a platform. This paper describes the implementation of a multicore digital neuromorphic processing system on an Altera Quartus II FPGA. Static routing was used to allow communication between the cores on the FPGA. Two applications were mapped to the system: image edge detection and ECG. Compared to an Intel processor implementation of these applications, the FPGA based neural implementations provided about 3× and 127× speedup for the edge detection and ECG applications. Given that both applications were implemented with the same base Verilog code, with only a change in the synaptic weights and number of neurons utilized, the system has the capability to accelerate a broad range of applications.","PeriodicalId":318539,"journal":{"name":"NAECON 2014 - IEEE National Aerospace and Electronics Conference","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-06-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125780571","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2014-06-24DOI: 10.1109/NAECON.2014.7045774
C. Devlin, J. Vella, D. Walker, J. Lombardi, N. Limberopoulos, J. Derov
Gradient Index Media have multiple applications for controlling the wave propagation and harvesting its received energy or processing its embedded data. A nanofabrication process was developed to produce a Luneburg lens on silicon, to operate in the optical regime, with feature sizes smaller than 100nm. The focused energy of the Luneburg lens is compared with lenses of non-spatially-varying index as well as a linear spatial variation, and its enhanced focusing is clearly demonstrated. In addition, a mechanism of thermal tunability of such devices is proposed and is supported by our results.
{"title":"Nanoscale gradient index media fabrication for extreme control and tunability of optical wave propagation","authors":"C. Devlin, J. Vella, D. Walker, J. Lombardi, N. Limberopoulos, J. Derov","doi":"10.1109/NAECON.2014.7045774","DOIUrl":"https://doi.org/10.1109/NAECON.2014.7045774","url":null,"abstract":"Gradient Index Media have multiple applications for controlling the wave propagation and harvesting its received energy or processing its embedded data. A nanofabrication process was developed to produce a Luneburg lens on silicon, to operate in the optical regime, with feature sizes smaller than 100nm. The focused energy of the Luneburg lens is compared with lenses of non-spatially-varying index as well as a linear spatial variation, and its enhanced focusing is clearly demonstrated. In addition, a mechanism of thermal tunability of such devices is proposed and is supported by our results.","PeriodicalId":318539,"journal":{"name":"NAECON 2014 - IEEE National Aerospace and Electronics Conference","volume":"114 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-06-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130093892","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2014-06-24DOI: 10.1109/NAECON.2014.7045779
M. I. Vakil, J. A. Malas, D. Megherbi
This work presents a two-phase image registration technique utilizing a hybrid feature-based and an area-based similarity measure of partially overlapped aerial imagery in presence of affine translation and rotation transformations. The resulting selectively guided execution of similarity measures provides a reduction in search space, reducing the computational cost of the proposed algorithm. This multi-stage approach enhances the capability to perform image registration of low resolution imagery where scenes may have many structures but lack well defined structures for conventional feature extraction or lack to have enough variations in the intensity values to diminish statistical dependencies. The inherent statistical attributes of area-based methods are exploited through the sequential use of complex correlation and mutual information on physics-based features.
{"title":"Guided execution of hybrid similarity-measures for registration of partially overlapped aerial imagery","authors":"M. I. Vakil, J. A. Malas, D. Megherbi","doi":"10.1109/NAECON.2014.7045779","DOIUrl":"https://doi.org/10.1109/NAECON.2014.7045779","url":null,"abstract":"This work presents a two-phase image registration technique utilizing a hybrid feature-based and an area-based similarity measure of partially overlapped aerial imagery in presence of affine translation and rotation transformations. The resulting selectively guided execution of similarity measures provides a reduction in search space, reducing the computational cost of the proposed algorithm. This multi-stage approach enhances the capability to perform image registration of low resolution imagery where scenes may have many structures but lack well defined structures for conventional feature extraction or lack to have enough variations in the intensity values to diminish statistical dependencies. The inherent statistical attributes of area-based methods are exploited through the sequential use of complex correlation and mutual information on physics-based features.","PeriodicalId":318539,"journal":{"name":"NAECON 2014 - IEEE National Aerospace and Electronics Conference","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-06-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131117427","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2014-06-24DOI: 10.1109/NAECON.2014.7045813
Weisong Wang, C. Yakopcic, E. Shin, K. Leedy, T. Taha, G. Subramanyam
This paper describes the fabrication of memristor devices based on titanium and hafnium oxides. The device cross sectional area is varied to observe the impact this has on the current-voltage characteristic. A modeling technique is then utilized that is capable of matching the current-voltage characteristics of memristor devices. The model was able to match the titanium oxide device described in this paper with 13.58% error. The device model was then used in a neuromorphic simulation showing that a circuit based on this device is capable of learning logic functions.
{"title":"Fabrication, characterization, and modeling of memristor devices","authors":"Weisong Wang, C. Yakopcic, E. Shin, K. Leedy, T. Taha, G. Subramanyam","doi":"10.1109/NAECON.2014.7045813","DOIUrl":"https://doi.org/10.1109/NAECON.2014.7045813","url":null,"abstract":"This paper describes the fabrication of memristor devices based on titanium and hafnium oxides. The device cross sectional area is varied to observe the impact this has on the current-voltage characteristic. A modeling technique is then utilized that is capable of matching the current-voltage characteristics of memristor devices. The model was able to match the titanium oxide device described in this paper with 13.58% error. The device model was then used in a neuromorphic simulation showing that a circuit based on this device is capable of learning logic functions.","PeriodicalId":318539,"journal":{"name":"NAECON 2014 - IEEE National Aerospace and Electronics Conference","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-06-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121491082","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2014-06-24DOI: 10.1109/NAECON.2014.7045802
R. Ilin
This work utilizes high resolution images in order to improve the classification accuracy on low resolution images. The approach is based on the machine learning paradigm called LUPI - “Learning Using Privileged Information”. In this contribution, the LUPI paradigm is demonstrated on images from the Caltech 101 dataset.
{"title":"Machine learning approach to fusion of high and low resolution imagery for improved target classification","authors":"R. Ilin","doi":"10.1109/NAECON.2014.7045802","DOIUrl":"https://doi.org/10.1109/NAECON.2014.7045802","url":null,"abstract":"This work utilizes high resolution images in order to improve the classification accuracy on low resolution images. The approach is based on the machine learning paradigm called LUPI - “Learning Using Privileged Information”. In this contribution, the LUPI paradigm is demonstrated on images from the Caltech 101 dataset.","PeriodicalId":318539,"journal":{"name":"NAECON 2014 - IEEE National Aerospace and Electronics Conference","volume":"90 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-06-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130951699","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2014-06-24DOI: 10.1109/NAECON.2014.7045811
Venkataramesh Bontupalli, Raqibul Hasan, T. Taha
With the rise in cyber-attacks on computing systems and the rapid increase in use of mobile systems, it is essential to secure these mobile devices. Given that these systems can roam on multiple networks, with no guarantee on security adopted on each network, including Intrusion Detection Systems (IDS) on the mobile platforms can be beneficial in preventing cyber-attacks. One of the key problems with implementing IDS on mobile platforms is the increased power consumption. This paper presents low power circuits that implement the string matching tasks within the Snort IDS. These tasks can take up to 80% of the power consumed for Snort. The circuit presented is based on memristor crossbars and evaluate Snort rules at 0.013mW per signature. The circuits are easy to program, utilizing only two resistance states for the memristors. They are highly parallel and dense, utilizing a brute-force string matching algorithm. These circuits could additionally be utilized for other string matching operations.
{"title":"Power efficient architecture for network intrusion detection system","authors":"Venkataramesh Bontupalli, Raqibul Hasan, T. Taha","doi":"10.1109/NAECON.2014.7045811","DOIUrl":"https://doi.org/10.1109/NAECON.2014.7045811","url":null,"abstract":"With the rise in cyber-attacks on computing systems and the rapid increase in use of mobile systems, it is essential to secure these mobile devices. Given that these systems can roam on multiple networks, with no guarantee on security adopted on each network, including Intrusion Detection Systems (IDS) on the mobile platforms can be beneficial in preventing cyber-attacks. One of the key problems with implementing IDS on mobile platforms is the increased power consumption. This paper presents low power circuits that implement the string matching tasks within the Snort IDS. These tasks can take up to 80% of the power consumed for Snort. The circuit presented is based on memristor crossbars and evaluate Snort rules at 0.013mW per signature. The circuits are easy to program, utilizing only two resistance states for the memristors. They are highly parallel and dense, utilizing a brute-force string matching algorithm. These circuits could additionally be utilized for other string matching operations.","PeriodicalId":318539,"journal":{"name":"NAECON 2014 - IEEE National Aerospace and Electronics Conference","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-06-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117077137","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2014-06-24DOI: 10.1109/NAECON.2014.7045765
Kristian E. Warner, M. Lanzerotti, C. Bartsch, J. Lombardi
This paper presents a method of characterizing single-walled carbon nanotube (SWCNT) ink and discusses preliminary research results of our characterization efforts. First, the process of ink jet printing SWCNT ink onto organic and inorganic substrates is discussed. Next, the tests for measuring sheet resistance, conductance, thickness, roll-off, and S-parameters of the ink are described. Future research directions will be discussed, including the additional characterization of SWCNT ink.
{"title":"Characterization of electrical and physical properties of single-walled carbon nanotube ink","authors":"Kristian E. Warner, M. Lanzerotti, C. Bartsch, J. Lombardi","doi":"10.1109/NAECON.2014.7045765","DOIUrl":"https://doi.org/10.1109/NAECON.2014.7045765","url":null,"abstract":"This paper presents a method of characterizing single-walled carbon nanotube (SWCNT) ink and discusses preliminary research results of our characterization efforts. First, the process of ink jet printing SWCNT ink onto organic and inorganic substrates is discussed. Next, the tests for measuring sheet resistance, conductance, thickness, roll-off, and S-parameters of the ink are described. Future research directions will be discussed, including the additional characterization of SWCNT ink.","PeriodicalId":318539,"journal":{"name":"NAECON 2014 - IEEE National Aerospace and Electronics Conference","volume":"59 10","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-06-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131993003","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2014-06-24DOI: 10.1109/NAECON.2014.7045810
C. Yakopcic, Raqibul Hasan, T. Taha
This paper describes a memristor based neuromorphic circuit that is capable of learning. Target memristors within the crossbar circuit were set to be stuck in either high or low resistance states to observe fault tolerance within the memristor crossbar. The simulations are carried out in SPICE using a detailed memristor model so that the crossbar is simulated as accurately as possible. In some cases the circuit was able to successfully learn when half of the memristors in the crossbar were set to be defective. Due to additional bias circuitry, this neuromorphic memristive learning circuit appears to be more tolerant to error than alternative designs.
{"title":"Tolerance to defective memristors in a neuromorphic learning circuit","authors":"C. Yakopcic, Raqibul Hasan, T. Taha","doi":"10.1109/NAECON.2014.7045810","DOIUrl":"https://doi.org/10.1109/NAECON.2014.7045810","url":null,"abstract":"This paper describes a memristor based neuromorphic circuit that is capable of learning. Target memristors within the crossbar circuit were set to be stuck in either high or low resistance states to observe fault tolerance within the memristor crossbar. The simulations are carried out in SPICE using a detailed memristor model so that the crossbar is simulated as accurately as possible. In some cases the circuit was able to successfully learn when half of the memristors in the crossbar were set to be defective. Due to additional bias circuitry, this neuromorphic memristive learning circuit appears to be more tolerant to error than alternative designs.","PeriodicalId":318539,"journal":{"name":"NAECON 2014 - IEEE National Aerospace and Electronics Conference","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-06-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126577810","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2014-06-24DOI: 10.1109/NAECON.2014.7045799
Marc R. Ward, Trevor J. Bihl, K. Bauer
This research considers simulated laser radar (LADAR) vibrometry for vehicle identification. Time sampled data is considered for developing multiple nonlinear autoregressive neural network (NARNet) classifier models. Emphasis is placed on robustness to sensor location and using small amounts of data. Decision level fusion is used to combine results from multiple classifiers. Results offer improved classification performance as compared to the literature.
{"title":"Vibrometry-based vehicle identification framework using nonlinear autoregressive neural networks and decision fusion","authors":"Marc R. Ward, Trevor J. Bihl, K. Bauer","doi":"10.1109/NAECON.2014.7045799","DOIUrl":"https://doi.org/10.1109/NAECON.2014.7045799","url":null,"abstract":"This research considers simulated laser radar (LADAR) vibrometry for vehicle identification. Time sampled data is considered for developing multiple nonlinear autoregressive neural network (NARNet) classifier models. Emphasis is placed on robustness to sensor location and using small amounts of data. Decision level fusion is used to combine results from multiple classifiers. Results offer improved classification performance as compared to the literature.","PeriodicalId":318539,"journal":{"name":"NAECON 2014 - IEEE National Aerospace and Electronics Conference","volume":"52 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-06-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128484081","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：