-
Siamese Learning-based Monarch Butterfly Localization
Authors:
Sara Shoouri,
Mingyu Yang,
Gordy Carichner,
Yuyang Li,
Ehab A. Hamed,
Angela Deng,
Delbert A. Green II,
Inhee Lee,
David Blaauw,
Hun-Seok Kim
Abstract:
A new GPS-less, daily localization method is proposed with deep learning sensor fusion that uses daylight intensity and temperature sensor data for Monarch butterfly tracking. Prior methods suffer from the location-independent day length during the equinox, resulting in high localization errors around that date. This work proposes a new Siamese learning-based localization model that improves the a…
▽ More
A new GPS-less, daily localization method is proposed with deep learning sensor fusion that uses daylight intensity and temperature sensor data for Monarch butterfly tracking. Prior methods suffer from the location-independent day length during the equinox, resulting in high localization errors around that date. This work proposes a new Siamese learning-based localization model that improves the accuracy and reduces the bias of daily Monarch butterfly localization using light and temperature measurements. To train and test the proposed algorithm, we use $5658$ daily measurement records collected through a data measurement campaign involving 306 volunteers across the U.S., Canada, and Mexico from 2018 to 2020. This model achieves a mean absolute error of $1.416^\circ$ in latitude and $0.393^\circ$ in longitude coordinates outperforming the prior method.
△ Less
Submitted 4 July, 2023;
originally announced July 2023.
-
High-resolution and reliable automatic target recognition based on photonic ISAR imaging system with explainable deep learning
Authors:
Xiuting Zou,
Anyi Deng,
Yiheng Hu,
Shiyu Hua,
Linbo Zhang,
Shaofu Xu,
Weiwen Zou
Abstract:
Automatic target recognition (ATR) based on inverse synthetic aperture radar (ISAR) images, which is extensively utilized to surveil environment in military and civil fields, must be high-precision and reliable. Photonic technologies' advantage of broad bandwidth enables ISAR systems to realize high-resolution imaging, which is in favor of achieving high-performance ATR. Deep learning (DL) algorit…
▽ More
Automatic target recognition (ATR) based on inverse synthetic aperture radar (ISAR) images, which is extensively utilized to surveil environment in military and civil fields, must be high-precision and reliable. Photonic technologies' advantage of broad bandwidth enables ISAR systems to realize high-resolution imaging, which is in favor of achieving high-performance ATR. Deep learning (DL) algorithms have achieved excellent recognition accuracies. However, the lack of interpretability of DL algorithms causes the head-scratching problem of credibility. In this paper, we exploit the inner relationship between a photonic ISAR imaging system and behaviors of a convolutional neural network (CNN) to deeply comprehend the intelligent recognition. Specifically, we manipulate imaging physical process and analyze network outputs, the relevance between the ISAR image and network output, and the visualization of features in the network output layer. Consequently, the broader imaging bandwidths and appropriate imaging angles lead to more detailed structural and contour features and the bigger discrepancy among ISAR images of different targets, which contributes to the CNN recognizing and distinguishing objects according to physical laws. Then, based on the photonic ISAR imaging system and the explainable CNN, we accomplish a high-accuracy and reliable ATR. To the best of our knowledge, there is no precedent of explaining the DL algorithms by exploring the influence of the physical process of data generation on network behaviors. It is anticipated that this work can not only inspire the accomplishment of a high-performance ATR but also bring new insights to explore network behaviors and thus achieve better intelligent abilities.
△ Less
Submitted 3 December, 2022;
originally announced December 2022.
-
Hardware-irrelevant parallel processing system
Authors:
Xiuting Zou,
Shaofu Xu,
Anyi Deng,
Rui Wang,
Weiwen Zou
Abstract:
Parallel processing technology has been a primary tool for achieving high-speed, high-accuracy, and broadband processing for many years across modern information systems and data processing such as optical and radar, synthetic aperture radar imaging, digital beam forming, and digital filtering systems. However, hardware deviations in a parallel processing system (PPS) severely degrade system perfo…
▽ More
Parallel processing technology has been a primary tool for achieving high-speed, high-accuracy, and broadband processing for many years across modern information systems and data processing such as optical and radar, synthetic aperture radar imaging, digital beam forming, and digital filtering systems. However, hardware deviations in a parallel processing system (PPS) severely degrade system performance and pose an urgent challenge. We propose a hardware-irrelevant PPS of which the performance is unaffected by hardware deviations. In this system, an embedded convolutional recurrent autoencoder (CRAE), which learns inherent system patterns as well as acquires and removes adverse effects brought by hardware deviations, is adopted. We implement a hardware-irrelevant PPS into a parallel photonic sampling system to accomplish a high-performance analog-to-digital conversion for microwave signals with high frequency and broad bandwidth. Under one system state, a category of signals with two different mismatch degrees is utilized to train the CRAE, which can then compensate for mismatches in various categories of signals with multiple mismatch degrees under random system states. Our approach is extensively applicable to achieving hardware-irrelevant PPSs which are either discrete or integrated in photonic, electric, and other fields.
△ Less
Submitted 23 June, 2020;
originally announced June 2020.
