{"title":"Sea Ice Image Semantic Segmentation Using Deep Neural Networks","authors":"Benjamin Dowden, O. de Silva, Weimin Huang","doi":"10.1109/IEEECONF38699.2020.9389229","DOIUrl":"https://doi.org/10.1109/IEEECONF38699.2020.9389229","url":null,"abstract":"Semantic segmentation is the process of classifying pixels in an image into different classes. This method is quite established in autonomous vehicles, biomedical image processing, and remote sensing applications. In this paper, we evaluate the applicability of semantic segmentation for sea ice classification using image feeds from on-board an icebreaker. For this purpose, we evaluate SegNet and PSPNet101 neural network architectures to segment images into four classes: ice, ocean, vessel, and sky. The Nathaniel B. Palmer dataset, which captures 2-month footage of the icebreaker completing an Antarctic expedition was used. A subset of the dataset was labeled to generate a 240-image dataset achieving an accuracy of 97.8% classification for the 26 image test dataset. These results validate the applicability of deep learning methods for sea ice detection using images, which can be further improved by classifying the ice type to support marine navigation and mapping applications.","PeriodicalId":198531,"journal":{"name":"Global Oceans 2020: Singapore – U.S. Gulf Coast","volume":"16 5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126961717","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}
Yongqing Wu, Shefeng Yan, Xiaochuan Ma, Dongyu Yuan, Jilong Li
{"title":"Underwater Long Baseline Localization based on Doppler Measurements for a Fast-Moving Underwater Vehicle","authors":"Yongqing Wu, Shefeng Yan, Xiaochuan Ma, Dongyu Yuan, Jilong Li","doi":"10.1109/IEEECONF38699.2020.9389332","DOIUrl":"https://doi.org/10.1109/IEEECONF38699.2020.9389332","url":null,"abstract":"Underwater navigation that relies solely on dead reckoning suffers from unbounded position error growth. In traditional positioning method based on TDOA, usual interference from multipath reflections and background noise leads to the choice of some fault correlation peaks under low SNR condition, especially in LBL. In the paper, the combination of LBL and Doppler shift estimation of an AUV will be expected to enable survey accuracy at any water depth where the position of an underwater acoustic transmitter equipped on the moving platform can be determined. And based on pattern time delay shift coding pulse, the positioning solution method about frequency estimation of Doppler shift in LBL could be described with Doppler effects from AUV moving. And a computation simulation, including three beacons in LBL, was performed to verify the performance of the proposed algorithm.","PeriodicalId":198531,"journal":{"name":"Global Oceans 2020: Singapore – U.S. Gulf Coast","volume":"113 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126962946","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}
{"title":"Designing a Universal Control Module for Unmanned Underwater Vehicles*","authors":"Thilo Müller, Jonas Pabst, T. Jeinsch","doi":"10.1109/IEEECONF38699.2020.9389161","DOIUrl":"https://doi.org/10.1109/IEEECONF38699.2020.9389161","url":null,"abstract":"This paper discusses the design of a universal control module for modular built unmanned underwater vehicles. Modular built in this context means that the designed control module has to be able to deal with different actuator configurations, sensor suits and general vehicle and -hull design. To be able to design such a module, two assumptions are made: Firstly, that the used actuators are limited to thrusters. And secondly, that the hydrodynamic parameters of the vehicle at hand are roughly known. To deal with this modularity, the design of the control module is built on two principles: By usage of the well known techniques of force allocation and sensor fusion a separation of the control layer from the actual vehicle configuration is achieved. And as knowledge of the vehicle's dynamics is essential for the control, the algorithms and parameters of the module are based on the known dynamic model of the vehicle. The paper concludes with the description of the design of a nonlinear controller for the motion of a vehicle as described, based on feedback linearisation and LQI-control. It is shown that by choosing the parameters of the linearised system to equal the linear parts of the vehicle's dynamics, this approach yields the same structure as a two degrees of freedom control structure that is using the nonlinear parts of the vehicle's dynamics as feedforward terms. For this controller, simulation results are given, showing its ability to control different vehicles.","PeriodicalId":198531,"journal":{"name":"Global Oceans 2020: Singapore – U.S. Gulf Coast","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115043363","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}
{"title":"Task Assignment of Multi-AUVs Based on Artificial Bee Colony Genetic Hybrid Algorithm","authors":"Di Zhang, Jun-Jia He, Sai Liu, Di Wang","doi":"10.1109/IEEECONF38699.2020.9389097","DOIUrl":"https://doi.org/10.1109/IEEECONF38699.2020.9389097","url":null,"abstract":"In this paper, the Artificial Bee Colony(ABC) Algorithm is used to study the multi-autonomous underwater vehicle(AUV) missions allocation scheme of a heterogeneous net. An improved bee colony-niched genetic algorithm (IABC-NGA) is proposed to solve the multi-AUV task allocation scheme in the late evolutionary stage to optimize the global search performance, improving the system cooperation ability and improving the lack of single AUV capability. Finally, simulation experiments indicate the evaluate of IABC-NGA algorithm. They show that IABC-NGA can obtain more effective assignment results under multiple constraints.","PeriodicalId":198531,"journal":{"name":"Global Oceans 2020: Singapore – U.S. Gulf Coast","volume":"34 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116401580","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}
Siyuan Cang, Xueli Sheng, Hang Dong, Long-xiang Guo, Jingwei Yin
{"title":"Underwater Acoustic Echo Time-Frequency Feature Extraction and Reconstruction using Second-order Synchrosqueezing Transform","authors":"Siyuan Cang, Xueli Sheng, Hang Dong, Long-xiang Guo, Jingwei Yin","doi":"10.1109/IEEECONF38699.2020.9389015","DOIUrl":"https://doi.org/10.1109/IEEECONF38699.2020.9389015","url":null,"abstract":"Time-frequency analysis-based signal feature extraction has been well studied in the active sonar area in recent years. When the non-stationary signal is analyzed, traditional approach, such as short-time Fourier transform, cannot provide an excellent time or frequency resolution simultaneously on account of the uncertainty principle. Furthermore, acoustic echo is strongly influenced by multi-path channel and complex environment of shallow water. In this paper, an enhanced time-frequency analysis method named second-order synchrosqueezing transform is investigated here to extract underwater acoustic echo characteristics and reconstruct them. Second-order synchrosqueezing transform is able to extract and clearly depict components with time-varying spectrum, and furthermore it allows individual reconstruction of these components. Synchrosqueezing is performed in three steps. First, the continuous wavelet transform of the received signal is calculated. Second, an initial estimate of the demodulated frequency is calculated based on the first step. Finally, continuous wavelet transform is squeezed via reassignment. In this way, synchrosqueezing representation is obtained, which process is invertible. Furthermore, second-order operation aims to adapt synchrosqueezing transform to the superposition of perturbed linear chirps.","PeriodicalId":198531,"journal":{"name":"Global Oceans 2020: Singapore – U.S. Gulf Coast","volume":"201 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122361823","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}
Dimitrios Stergianelis, Matthew McConnell, Ze Ji, Ya Huang, Hui Yu
{"title":"A simulation study of simple local path planning and control for unmanned surface vehicles","authors":"Dimitrios Stergianelis, Matthew McConnell, Ze Ji, Ya Huang, Hui Yu","doi":"10.1109/IEEECONF38699.2020.9389484","DOIUrl":"https://doi.org/10.1109/IEEECONF38699.2020.9389484","url":null,"abstract":"The present study provides a fundamental but necessary review and implementation of collision avoidance and path planning techniques based on 2D plane kinematics of a small test vessel. The developed path planning logics first consider a single obstacle avoidance scenario, and then move on to tackle multiple obstacle avoidance based on more realistic obstacle and environment representations using multiple overlapping circles of various sizes and positions. The manuscript elaborates the development process step by step with all codes provided in order to introduce the path planning and simulation workflow as a laid tutorial for learners and researchers entering the field.","PeriodicalId":198531,"journal":{"name":"Global Oceans 2020: Singapore – U.S. Gulf Coast","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122716347","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}
Liang Ruan, Shumin Chen, Jie Zhou, Daheng Zeng, Yuan-xin Xu
{"title":"A Factor Graph Method for AUV Navigation in the Mobile Docking Progress","authors":"Liang Ruan, Shumin Chen, Jie Zhou, Daheng Zeng, Yuan-xin Xu","doi":"10.1109/IEEECONF38699.2020.9389322","DOIUrl":"https://doi.org/10.1109/IEEECONF38699.2020.9389322","url":null,"abstract":"Navigation and localization in underwater environments for autonomous underwater vehicle (AUV) are particularly essential in the autonomous docking process. Because of unavailability of global positioning system (GPS) in the underwater, AUV need to estimate its attitude and position utilized proprioceptive sensors and external sensors on board. Past estimation algorithms are mostly based on filter approaches, such as Extended Kalman Filter (EKF), and few optimization-based methods. In this paper, we proposed a novel AUV navigation algorithm based on factor graph in the mobile docking process. We consider AUV kinetics model and measurements as factors adding to the factor graph, mobile object as a mobile landmark and add its motion model as a factor to the factor graph, then optimize the factor graph. Then we proposed a batch optimization approach which plays a trade between computing and accuracy. The simulation and pool experimental results both show the feasibility and accuracy of the algorithm.","PeriodicalId":198531,"journal":{"name":"Global Oceans 2020: Singapore – U.S. Gulf Coast","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114599979","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}
{"title":"Evaluation of Path Planning Algorithms of Autonomous Surface Vehicles Based on Safety and Collision Risk Assessment","authors":"Anete Vagale, Robin T. Bye, Ottar L. Osen","doi":"10.1109/IEEECONF38699.2020.9389481","DOIUrl":"https://doi.org/10.1109/IEEECONF38699.2020.9389481","url":null,"abstract":"Improved safety while navigating on waters and reduction of collision risk is a vital part of the guidance, navigation and control system of an autonomous surface vehicle. Another problem is, how to compare the performance of existing path planning and collision avoidance algorithms in a unified way. To tackle these problems, a novel evaluation simulator platform is proposed in this paper for simulation-based testing of algorithms. The platform is designed for generating different scenarios based on the system's inputs, such as static and dynamic obstacles, environmental disturbances, vessel's dynamic model, and environment, and to evaluate algorithm performance based on path fitness, risk assessment and, in the future, good seamanship practice. Formation of safety maps is used for assessment of each performance measure. Additionally, a root sum square method combines these measures into a total algorithm performance rating. Finally, multi-objective optimisation is applied to evaluate the algorithms based on their performance ratings.","PeriodicalId":198531,"journal":{"name":"Global Oceans 2020: Singapore – U.S. Gulf Coast","volume":"268 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122086089","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}
Edoardo Topini, Alberto Topini, Matteo Franchi, Alessandro Bucci, Nicola Secciani, A. Ridolfi, B. Allotta
{"title":"LSTM-based Dead Reckoning Navigation for Autonomous Underwater Vehicles","authors":"Edoardo Topini, Alberto Topini, Matteo Franchi, Alessandro Bucci, Nicola Secciani, A. Ridolfi, B. Allotta","doi":"10.1109/IEEECONF38699.2020.9389379","DOIUrl":"https://doi.org/10.1109/IEEECONF38699.2020.9389379","url":null,"abstract":"Autonomous Underwater Vehicles (AUVs) do need to be equipped with highly-accurate and robust navigation systems in order to perform challenging operations and complex missions. Nevertheless, since the Global Positioning System (GPS) is not a feasible, functional solution in the underwater scenario, the localization task is fulfilled by employing filtering techniques - as Extended Kalman Filter (EKF) and Unscented Kalman Filter (UKF) - or Dead Reckoning (DR) strategies. Even though the specific algorithm architecture may vary, such methodologies usually rely on the direct linear speed measurements provided by specialized and expensive sensors, such as the Doppler Velocity Log (DVL). Thus, DVL failures or fallacies alongside DVL-denied environments may arise as unexpected causes for a severe malfunction of the whole navigation system. Motivated by the aforementioned considerations and the outstanding performance of Long Short-Term Memory (LSTM) neural networks in time-series regression problems, an LSTM-based DR approach has been developed to estimate the surge and sway body-fixed frame velocities, without canonically employing the DVL sensor.","PeriodicalId":198531,"journal":{"name":"Global Oceans 2020: Singapore – U.S. Gulf Coast","volume":"101 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129762520","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}
{"title":"Few-shot Fish Image Generation and Classification","authors":"Zonghui Guo, Liqiang Zhang, Yufeng Jiang, Wenjie Niu, Zhaorui Gu, Haiyong Zheng, Guoyu Wang, Bing Zheng","doi":"10.1109/IEEECONF38699.2020.9389005","DOIUrl":"https://doi.org/10.1109/IEEECONF38699.2020.9389005","url":null,"abstract":"Scientific studies on fishes have considerable importance to the fishery industry and marine ecosystem. Fish images are typically collected by scuba divers or autonomous underwater vehicles and then annotated manually by marine biologists. In recent years, with the rise of deep learning, the field of image classification has been greatly improved. But, fish image classification which can be considered as fine-grained problem is more challenging than common image classification. Usually, fish images are low-quality and belong to small-scale data, while classic CNN models require a large quantity of high-quality data to gain excellent results. For few-shot fish images, it's hard to get rich diversity by rotation, slight transform, and upsampling. In this paper, our method could obtain few-shot classes which are translated from large-scale sample classes, while maintaining the intra-class similarity, diversity and inter-class identifiabilityso as to improve the classification effect of fish image,","PeriodicalId":198531,"journal":{"name":"Global Oceans 2020: Singapore – U.S. Gulf Coast","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129936424","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}