Pub Date : 2006-09-01DOI: 10.1109/NSSPW.2006.4378809
Vijayendra Mohan Roy, G. V. Anand
In this paper we introduce a nonlinear detector based on the phenomenon of suprathreshold stochastic resonance (SSR). We first present a model (an array of 1-bit quantizers) that demonstrates the SSR phenomenon. We then use this as a pre-processor to the conventional matched filter. We employ the Neyman-Pearson(NP) detection strategy and compare the performances of the matched filter, the SSR-based detector and the optimal detector. Although the proposed detector is non-optimal, for non-Gaussian noises with heavy tails (leptokurtic) it shows better performance than the matched filter. In situations where the noise is known to be leptokurtic without the availability of the exact knowledge of its distribution, the proposed detector turns out to be a better choice than the matched filter.
{"title":"Performance Analysis of a Suprathreshold Stochastic Resonance Based Nonlinear Detector","authors":"Vijayendra Mohan Roy, G. V. Anand","doi":"10.1109/NSSPW.2006.4378809","DOIUrl":"https://doi.org/10.1109/NSSPW.2006.4378809","url":null,"abstract":"In this paper we introduce a nonlinear detector based on the phenomenon of suprathreshold stochastic resonance (SSR). We first present a model (an array of 1-bit quantizers) that demonstrates the SSR phenomenon. We then use this as a pre-processor to the conventional matched filter. We employ the Neyman-Pearson(NP) detection strategy and compare the performances of the matched filter, the SSR-based detector and the optimal detector. Although the proposed detector is non-optimal, for non-Gaussian noises with heavy tails (leptokurtic) it shows better performance than the matched filter. In situations where the noise is known to be leptokurtic without the availability of the exact knowledge of its distribution, the proposed detector turns out to be a better choice than the matched filter.","PeriodicalId":388611,"journal":{"name":"2006 IEEE Nonlinear Statistical Signal Processing Workshop","volume":"134 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116588232","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 : 2006-09-01DOI: 10.1109/NSSPW.2006.4378854
E. Wan
In this presentation, we first provide an overview of Sigma-Point filtering methods, which include the Unscented Kalman Filter (UKF), Central Difference Kalman Filter (CDKF), and several variants with hybrid extensions to sequential Monte Carlo filtering (e.g., particle filtering). In the second half, we focus on recent applications to integrated navigation systems (INS), which provide state-estimation by combining GPS and inertial measurements. In addition, we present new work on using video data to extract the equivalent state-information (i.e., replacing the INS) for use in closed-loop control of an Unmanned Aerial Vehicle (UAV).
{"title":"Sigma-Point Filters: An Overview with Applications to Integrated Navigation and Vision Assisted Control","authors":"E. Wan","doi":"10.1109/NSSPW.2006.4378854","DOIUrl":"https://doi.org/10.1109/NSSPW.2006.4378854","url":null,"abstract":"In this presentation, we first provide an overview of Sigma-Point filtering methods, which include the Unscented Kalman Filter (UKF), Central Difference Kalman Filter (CDKF), and several variants with hybrid extensions to sequential Monte Carlo filtering (e.g., particle filtering). In the second half, we focus on recent applications to integrated navigation systems (INS), which provide state-estimation by combining GPS and inertial measurements. In addition, we present new work on using video data to extract the equivalent state-information (i.e., replacing the INS) for use in closed-loop control of an Unmanned Aerial Vehicle (UAV).","PeriodicalId":388611,"journal":{"name":"2006 IEEE Nonlinear Statistical Signal Processing Workshop","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131404034","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 : 2006-09-01DOI: 10.1109/NSSPW.2006.4378828
J. Gunnarsson, L. Svensson, E. Bengtsson, L. Danielsson
In this paper we present and validate a new modelling frame-work for joint driver intention classification and tracking of vehicles, a framework derived for automotive active safety systems. Such systems require reliable predictions of the traffic situation to act in time when a dangerous situation occur. Our proposal has two main benefits. First, it incorporates the intention of the driver into the vehicle motion model and thereby improves the prediction capability. The result is a multiple motion model where each model corresponds to a specific driver intent. Second, the connection between different driver plans and corresponding motion model enables a formal classification of the most likely driver intention. To validate our concept, we apply the motion model on real data using a particle filter implementation. Initial studies indicate promising performance.
{"title":"Joint Driver Intention Classification and Tracking of Vehicles","authors":"J. Gunnarsson, L. Svensson, E. Bengtsson, L. Danielsson","doi":"10.1109/NSSPW.2006.4378828","DOIUrl":"https://doi.org/10.1109/NSSPW.2006.4378828","url":null,"abstract":"In this paper we present and validate a new modelling frame-work for joint driver intention classification and tracking of vehicles, a framework derived for automotive active safety systems. Such systems require reliable predictions of the traffic situation to act in time when a dangerous situation occur. Our proposal has two main benefits. First, it incorporates the intention of the driver into the vehicle motion model and thereby improves the prediction capability. The result is a multiple motion model where each model corresponds to a specific driver intent. Second, the connection between different driver plans and corresponding motion model enables a formal classification of the most likely driver intention. To validate our concept, we apply the motion model on real data using a particle filter implementation. Initial studies indicate promising performance.","PeriodicalId":388611,"journal":{"name":"2006 IEEE Nonlinear Statistical Signal Processing Workshop","volume":"140 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121281635","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 : 2006-09-01DOI: 10.1109/NSSPW.2006.4378819
Jack Li, W. Ng, S. Godsill
In tracking applications, the target state (e.g., position, velocity) can be estimated by processing the measurements collected from all deployed sensors at a central node. The estimation performance significantly relies on the accuracy of the sensor positions/rotations when data fusion is conducted. Since in practice precise knowledge of this sensor information is seldom available, in this paper we propose a Sequential Monte Carlo (SMC) approach to jointly estimate the target state and resolve the sensor position uncertainty.
{"title":"Online Target Tracking and Sensor Registration using Sequential Monte Carlo Methods","authors":"Jack Li, W. Ng, S. Godsill","doi":"10.1109/NSSPW.2006.4378819","DOIUrl":"https://doi.org/10.1109/NSSPW.2006.4378819","url":null,"abstract":"In tracking applications, the target state (e.g., position, velocity) can be estimated by processing the measurements collected from all deployed sensors at a central node. The estimation performance significantly relies on the accuracy of the sensor positions/rotations when data fusion is conducted. Since in practice precise knowledge of this sensor information is seldom available, in this paper we propose a Sequential Monte Carlo (SMC) approach to jointly estimate the target state and resolve the sensor position uncertainty.","PeriodicalId":388611,"journal":{"name":"2006 IEEE Nonlinear Statistical Signal Processing Workshop","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132694189","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 : 2006-09-01DOI: 10.1109/NSSPW.2006.4378845
N. Kantas, Sumeetpal S. Singh, A. Doucet
We describe how a completely decentralized version of Recursive Maximum Likelihood (RML) can be implemented in dynamic graphical models through the propagation of suitable messages that are exchanged between neighbouring nodes of the graph. The resulting algorithm can be interpreted as a generalization of the celebrated belief propagation algorithm to compute likelihood gradients. This algorithm is applied to solve the sensor localisation problem for distributed trackers forming a sensor networks. An implementation is given for dynamic nonlinear model without loops using Sequential Monte Carlo (SMC) or particle
{"title":"Distributed Self Localisation of Sensor Networks using Particle Methods","authors":"N. Kantas, Sumeetpal S. Singh, A. Doucet","doi":"10.1109/NSSPW.2006.4378845","DOIUrl":"https://doi.org/10.1109/NSSPW.2006.4378845","url":null,"abstract":"We describe how a completely decentralized version of Recursive Maximum Likelihood (RML) can be implemented in dynamic graphical models through the propagation of suitable messages that are exchanged between neighbouring nodes of the graph. The resulting algorithm can be interpreted as a generalization of the celebrated belief propagation algorithm to compute likelihood gradients. This algorithm is applied to solve the sensor localisation problem for distributed trackers forming a sensor networks. An implementation is given for dynamic nonlinear model without loops using Sequential Monte Carlo (SMC) or particle","PeriodicalId":388611,"journal":{"name":"2006 IEEE Nonlinear Statistical Signal Processing Workshop","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131820238","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 : 2006-09-01DOI: 10.1109/NSSPW.2006.4378851
V. Asirvadam, S. McLoone
This paper investigates sequential learning method with new form of weight update applied on a decomposed form of training algorithms using Radial Basis Function (RBF) network. Adding each basis function to the hidden layer during the course of training facilitate the weight update to be decomposed on neuron by neuron basis. A new form weight update is introduced where the weight update is based on minimal displacement of the current input elements to the elements of the nearest centre of the Gaussian neuron.
{"title":"Sequential Learning Methods on RBF with Novel Approach of Minimal Weight Update","authors":"V. Asirvadam, S. McLoone","doi":"10.1109/NSSPW.2006.4378851","DOIUrl":"https://doi.org/10.1109/NSSPW.2006.4378851","url":null,"abstract":"This paper investigates sequential learning method with new form of weight update applied on a decomposed form of training algorithms using Radial Basis Function (RBF) network. Adding each basis function to the hidden layer during the course of training facilitate the weight update to be decomposed on neuron by neuron basis. A new form weight update is introduced where the weight update is based on minimal displacement of the current input elements to the elements of the nearest centre of the Gaussian neuron.","PeriodicalId":388611,"journal":{"name":"2006 IEEE Nonlinear Statistical Signal Processing Workshop","volume":"51 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115143021","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 : 2006-09-01DOI: 10.1109/NSSPW.2006.4378846
N. Ikoma, Masahide Sakata, M. Doi
Aiming at human friendly user interface using omnidirection camera by holding it with fingers, we propose a method to track fingers in the image using particle filters. Mixture tracking is employed as a particle filtering method since it can track multiple modes of posterior distribution effectively leading to the ability to track multiple fingers in the image. We also propose a sequential labeling technique for clustering the particles in the framework of mixture tracking. This effectively classifies the particles into each finger in the image. Tracking performance has been demonstrated in real images of omnidirection camera.
{"title":"Mixture Tracking of Multiple Fingers Image in Omnidirection Camera for Human Friendly Interface","authors":"N. Ikoma, Masahide Sakata, M. Doi","doi":"10.1109/NSSPW.2006.4378846","DOIUrl":"https://doi.org/10.1109/NSSPW.2006.4378846","url":null,"abstract":"Aiming at human friendly user interface using omnidirection camera by holding it with fingers, we propose a method to track fingers in the image using particle filters. Mixture tracking is employed as a particle filtering method since it can track multiple modes of posterior distribution effectively leading to the ability to track multiple fingers in the image. We also propose a sequential labeling technique for clustering the particles in the framework of mixture tracking. This effectively classifies the particles into each finger in the image. Tracking performance has been demonstrated in real images of omnidirection camera.","PeriodicalId":388611,"journal":{"name":"2006 IEEE Nonlinear Statistical Signal Processing Workshop","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125968022","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 : 2006-09-01DOI: 10.1109/NSSPW.2006.4378816
E. Sullivan, J. D. Holmes, W. M. Carey
Model-based signal processing techniques for a short towed array are experimentally demonstrated for both broad band and narrow-band signals. It is shown that accurate bearing and horizontal wavenumber estimations, for frequencies ranging from 220 Hz to 1228 Hz, can be made for acoustic apertures as short as 0.5 wavelength. The data were the result of an experiment carried out in Nantucket sound, where an autonomous underwater vehicle towed a hydrophone array for the purpose of forming a long synthetic aperture. The AUV towed the array outward from a fixed narrow-band source on a straight radial track out to 4 km. Processing of the signal over the entire 4 km synthetic aperture, based on a normal-mode propagation mode, allowed for characterization of the channel. During the experiment a ferry passed through the area. This "target of opportunity" provided a basis for testing model-based passive synthetic aperture bearing estimation routines.
{"title":"Model-Based Processing for a Short Towed Array","authors":"E. Sullivan, J. D. Holmes, W. M. Carey","doi":"10.1109/NSSPW.2006.4378816","DOIUrl":"https://doi.org/10.1109/NSSPW.2006.4378816","url":null,"abstract":"Model-based signal processing techniques for a short towed array are experimentally demonstrated for both broad band and narrow-band signals. It is shown that accurate bearing and horizontal wavenumber estimations, for frequencies ranging from 220 Hz to 1228 Hz, can be made for acoustic apertures as short as 0.5 wavelength. The data were the result of an experiment carried out in Nantucket sound, where an autonomous underwater vehicle towed a hydrophone array for the purpose of forming a long synthetic aperture. The AUV towed the array outward from a fixed narrow-band source on a straight radial track out to 4 km. Processing of the signal over the entire 4 km synthetic aperture, based on a normal-mode propagation mode, allowed for characterization of the channel. During the experiment a ferry passed through the area. This \"target of opportunity\" provided a basis for testing model-based passive synthetic aperture bearing estimation routines.","PeriodicalId":388611,"journal":{"name":"2006 IEEE Nonlinear Statistical Signal Processing Workshop","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130826882","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 : 2006-09-01DOI: 10.1109/NSSPW.2006.4378829
Hendrik Kuck, N. de Freitas, A. Doucet
Experimental design is a fundamental problem in science. It arises in the planning of medical trials, sensor network deployment and control as well as in costly data gathering in physics, chemistry and biology. Bayesian decision theory provides a principled way of treating this problem, but leads to an intractable joint optimization and integration problem. Here, we propose a viable solution to this hard computational problem using sequential Monte Carlo samplers.
{"title":"SMC Samplers for Bayesian Optimal Nonlinear Design","authors":"Hendrik Kuck, N. de Freitas, A. Doucet","doi":"10.1109/NSSPW.2006.4378829","DOIUrl":"https://doi.org/10.1109/NSSPW.2006.4378829","url":null,"abstract":"Experimental design is a fundamental problem in science. It arises in the planning of medical trials, sensor network deployment and control as well as in costly data gathering in physics, chemistry and biology. Bayesian decision theory provides a principled way of treating this problem, but leads to an intractable joint optimization and integration problem. Here, we propose a viable solution to this hard computational problem using sequential Monte Carlo samplers.","PeriodicalId":388611,"journal":{"name":"2006 IEEE Nonlinear Statistical Signal Processing Workshop","volume":"83 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114752717","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 : 2006-09-01DOI: 10.1109/NSSPW.2006.4378806
S. Bhaumik, M. Srinivasan, S. Sadhu, T. Ghoshal
An Adaptive Grid Method based on the well-known point-mass approximation has been developed for computation of risk-sensitive state estimates in non-linear non-Gaussian problems. Risk-sensitive estimators, believed to have increased robustness compared to their risk neutral counterparts, admit closed form expressions only for a very limited class of models including linear Gaussian models. The Extended Risk-Sensitive Filter (ERSF) which uses an EKF-like approach fails to take care of non-Gaussian problems or severe non-linearities. Recently, a particle-filter based approach has been proposed for extending the range of applications of risk-sensitive techniques. The present authors have developed the adaptive grid risk-sensitive filter (AGRSF), which was partially motivated by the need to validate the particle filter based risk-sensitive filter and uses a set of heuristics for the adaptive choice of grid points to improve the numerical efficiency. The AGRSF has been cross-validated against closed-form solutions for the linear Gaussian case and against the risk-sensitive particle filter (RSPF) for fairly severe non-linear problems which create a multi-modal posterior distribution. Root mean square error and computational cost of the AGRSF and the RSPF have been compared.
{"title":"A Risk Sensitive Estimator for Nonlinear Problems using the Adaptive Grid Method","authors":"S. Bhaumik, M. Srinivasan, S. Sadhu, T. Ghoshal","doi":"10.1109/NSSPW.2006.4378806","DOIUrl":"https://doi.org/10.1109/NSSPW.2006.4378806","url":null,"abstract":"An Adaptive Grid Method based on the well-known point-mass approximation has been developed for computation of risk-sensitive state estimates in non-linear non-Gaussian problems. Risk-sensitive estimators, believed to have increased robustness compared to their risk neutral counterparts, admit closed form expressions only for a very limited class of models including linear Gaussian models. The Extended Risk-Sensitive Filter (ERSF) which uses an EKF-like approach fails to take care of non-Gaussian problems or severe non-linearities. Recently, a particle-filter based approach has been proposed for extending the range of applications of risk-sensitive techniques. The present authors have developed the adaptive grid risk-sensitive filter (AGRSF), which was partially motivated by the need to validate the particle filter based risk-sensitive filter and uses a set of heuristics for the adaptive choice of grid points to improve the numerical efficiency. The AGRSF has been cross-validated against closed-form solutions for the linear Gaussian case and against the risk-sensitive particle filter (RSPF) for fairly severe non-linear problems which create a multi-modal posterior distribution. Root mean square error and computational cost of the AGRSF and the RSPF have been compared.","PeriodicalId":388611,"journal":{"name":"2006 IEEE Nonlinear Statistical Signal Processing Workshop","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2006-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126293219","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：