ro Approved for public release; distribution is unlimited. l/'llu C^üALiflr IB f7.:~: ifM&ap... more ro Approved for public release; distribution is unlimited. l/'llu C^üALiflr IB f7.:~: ifM'."f--s REPORT DOCUMENTATION PAGE Form Approved OMB No. 0704-0188 Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instruction, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this
World Academy of Science, Engineering and Technology, International Journal of Computer and Information Engineering, 2017
In this work, a design of effective Bearing-only TMA Algorithm is done. The measured bearing angl... more In this work, a design of effective Bearing-only TMA Algorithm is done. The measured bearing angles are very noisy. Moreover, for multi-beam sonar, the measurements is quantized due to the sonar beam width. To deal with this, modified gain extended Kalman filter algorithm is used. The algorithm is fine-tuned, and many modules are added to improve the performance. A special validation gate module is used to insure stability of the algorithm. Many indicators of the performance and confidence level measurement are designed and tested. A new method to detect if the target is maneuvering is proposed. Moreover, a reactive optimal observer maneuver based on bearing measurements is proposed, which insure converging to the right solution all of the times. To test the performance of the proposed TMA algorithm a simulation is done with a MATLAB program. The simulator program tries to model a discrete scenario for an observer and a target. The simulator takes into consideration all the practical aspects of the problem such as a smooth transition in the speed, a circular turn of the ship, noisy measurements, and a quantized bearing measurement come for multi-beam sonar. The tests are done for a lot of given test scenarios. For all the tests, full tracking is achieved within 10 minutes with very little error. The range estimation error was less than 5%, speed error less than 5% and heading error less than 2 degree. For the online performance estimator, it is mostly aligned with the real performance. The range estimation confidence level gives a value equal to 90% when the range error less than 10%. The experiments show that the proposed TMA algorithm is very robust and has low estimation error. However, the converging time of the algorithm is needed to be improved.
: This thesis investigates the application of wavelet decompositions to classification applicatio... more : This thesis investigates the application of wavelet decompositions to classification applications. Two feature extraction tools are considered: Local Discriminant Bases scheme (LDB) and Power method. Several dimension reduction schemes including a newly proposed one called the Mean Separator neural network (MS NN) are discussed. Two types of classifiers are investigated and compared: Classification Trees (CT) and Back-propagation neural network (BP NN). Classification experiments conducted on synthetic and real-world underwater signals show that: (1) the Power feature extraction method is more robust to time synchronization issues than the LDB scheme is; (2) the MS NN scheme is a successful dimension reduction scheme that may be used with both LDB and Power feature extraction methods; and (3) the BP NN is a more powerful classifier than CT as it has fewer constraints than CT in partitioning the feature input space.
Conference Record of Thirty Second Asilomar Conference on Signals Systems and Computers, Dec 1, 1998
A simple type of projection pursuit scheme is discussed and applied to classification application... more A simple type of projection pursuit scheme is discussed and applied to classification applications. The proposed scheme is used to significantly reduce the number of class features obtained from the wavelet packet decomposition of the signals to be classified. Results show this scheme can be used to classify underwater data without significant loss of performance.
Conference Record of Thirty Second Asilomar Conference on Signals Systems and Computers, Dec 1, 1998
ABSTRACT A simple type of projection pursuit scheme is discussed and applied to classification ap... more ABSTRACT A simple type of projection pursuit scheme is discussed and applied to classification applications. The proposed scheme is used to significantly reduce the number of class features obtained from the wavelet packet decomposition of the signals to be classified. Results show this scheme can be used to classify underwater data without significant loss of performance.
A simple type of projection pursuit scheme is discussed and applied to classification application... more A simple type of projection pursuit scheme is discussed and applied to classification applications. The proposed scheme is used to significantly reduce the number of class features obtained from the wavelet packet decomposition of the signals to be classified. Results show this scheme can be used to classify underwater data without significant loss of performance.
ABSTRACT This work discusses feature extraction and dimension-reduction issues in the context of ... more ABSTRACT This work discusses feature extraction and dimension-reduction issues in the context of classification applications. A new type of projection pursuit algorithm is presented. This scheme is designed to reduce the number of class features obtained from the wavelet packet decomposition of the signals to be classified. Results show the simple scheme can be used to classify various types of signals in a noisy environment
ro Approved for public release; distribution is unlimited. l/'llu C^üALiflr IB f7.:~: ifM&ap... more ro Approved for public release; distribution is unlimited. l/'llu C^üALiflr IB f7.:~: ifM'."f--s REPORT DOCUMENTATION PAGE Form Approved OMB No. 0704-0188 Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instruction, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this
World Academy of Science, Engineering and Technology, International Journal of Computer and Information Engineering, 2017
In this work, a design of effective Bearing-only TMA Algorithm is done. The measured bearing angl... more In this work, a design of effective Bearing-only TMA Algorithm is done. The measured bearing angles are very noisy. Moreover, for multi-beam sonar, the measurements is quantized due to the sonar beam width. To deal with this, modified gain extended Kalman filter algorithm is used. The algorithm is fine-tuned, and many modules are added to improve the performance. A special validation gate module is used to insure stability of the algorithm. Many indicators of the performance and confidence level measurement are designed and tested. A new method to detect if the target is maneuvering is proposed. Moreover, a reactive optimal observer maneuver based on bearing measurements is proposed, which insure converging to the right solution all of the times. To test the performance of the proposed TMA algorithm a simulation is done with a MATLAB program. The simulator program tries to model a discrete scenario for an observer and a target. The simulator takes into consideration all the practical aspects of the problem such as a smooth transition in the speed, a circular turn of the ship, noisy measurements, and a quantized bearing measurement come for multi-beam sonar. The tests are done for a lot of given test scenarios. For all the tests, full tracking is achieved within 10 minutes with very little error. The range estimation error was less than 5%, speed error less than 5% and heading error less than 2 degree. For the online performance estimator, it is mostly aligned with the real performance. The range estimation confidence level gives a value equal to 90% when the range error less than 10%. The experiments show that the proposed TMA algorithm is very robust and has low estimation error. However, the converging time of the algorithm is needed to be improved.
: This thesis investigates the application of wavelet decompositions to classification applicatio... more : This thesis investigates the application of wavelet decompositions to classification applications. Two feature extraction tools are considered: Local Discriminant Bases scheme (LDB) and Power method. Several dimension reduction schemes including a newly proposed one called the Mean Separator neural network (MS NN) are discussed. Two types of classifiers are investigated and compared: Classification Trees (CT) and Back-propagation neural network (BP NN). Classification experiments conducted on synthetic and real-world underwater signals show that: (1) the Power feature extraction method is more robust to time synchronization issues than the LDB scheme is; (2) the MS NN scheme is a successful dimension reduction scheme that may be used with both LDB and Power feature extraction methods; and (3) the BP NN is a more powerful classifier than CT as it has fewer constraints than CT in partitioning the feature input space.
Conference Record of Thirty Second Asilomar Conference on Signals Systems and Computers, Dec 1, 1998
A simple type of projection pursuit scheme is discussed and applied to classification application... more A simple type of projection pursuit scheme is discussed and applied to classification applications. The proposed scheme is used to significantly reduce the number of class features obtained from the wavelet packet decomposition of the signals to be classified. Results show this scheme can be used to classify underwater data without significant loss of performance.
Conference Record of Thirty Second Asilomar Conference on Signals Systems and Computers, Dec 1, 1998
ABSTRACT A simple type of projection pursuit scheme is discussed and applied to classification ap... more ABSTRACT A simple type of projection pursuit scheme is discussed and applied to classification applications. The proposed scheme is used to significantly reduce the number of class features obtained from the wavelet packet decomposition of the signals to be classified. Results show this scheme can be used to classify underwater data without significant loss of performance.
A simple type of projection pursuit scheme is discussed and applied to classification application... more A simple type of projection pursuit scheme is discussed and applied to classification applications. The proposed scheme is used to significantly reduce the number of class features obtained from the wavelet packet decomposition of the signals to be classified. Results show this scheme can be used to classify underwater data without significant loss of performance.
ABSTRACT This work discusses feature extraction and dimension-reduction issues in the context of ... more ABSTRACT This work discusses feature extraction and dimension-reduction issues in the context of classification applications. A new type of projection pursuit algorithm is presented. This scheme is designed to reduce the number of class features obtained from the wavelet packet decomposition of the signals to be classified. Results show the simple scheme can be used to classify various types of signals in a noisy environment
Uploads