Autonomous Visual Perception for Unmanned Surface Vehicle Navigation in an Unknown Environment.

There are many unknown obstacles in the sea, so the autonomous navigation of unmanned surface vehicle needs to avoid them as soon as possible even under the condition of low control ability of the controller. To solve the problem, by combining the dynamic collision avoidance algorithm and tracking control, a dynamic collision avoidance control method in the unknown ocean environment is presented. In this article, in consideration of the unknown ocean environment and real-time dynamic obstacle avoidance problem, the collision avoidance controller using a velocity resolution method and backstepping tracking controller based on unmanned surface vehicle maneuvering motion model is designed. Simulation results show that the method is effective and accurate and can provide the reference for the unmanned surface vehicle intelligent collision avoidance control technology.

The reliable obstacle detection is a challenging task in autonomous navigation of unmanned surface vehicles. In this paper, we present a novel real-time obstacles detection based on monocular vision which can effectively tell apart obstacles on the sea surface from complex background. The main innovation of this paper is to propose a water-boundary-line algorithm based on semantic segmentation and random sample consistency line fitting. And use a simple and effective saliency detection method based on background prior and foreground prior to detect obstacles under the water-boundary-line. Our method can efficiently and quickly obtain obstacle information from images captured by shipborne cameras, and it has the ability to process more than 33 frames/s.

Object tracking is an important basis for the autonomous navigation of unmanned surface vehicles. However, several problems still must be addressed for a wide applicating of object tracking in unmanned surface vehicles. First, if multiple objects of the same classification exist in the same field of view, then stable extraction of an object is difficult. Second, in an environment with a complex background and large changes in object shape, the tracking accuracy is low, and object tracking errors and tracking loss can easily occur. Third, much time is required to detect a high-resolution real-time video stream, not meeting the delay requirement of the photoelectric servo stable tracking. To resolve these problems, this paper proposes an object detection-tracking algorithm based on a radar-photoelectric system. The algorithm combines an object detection algorithm with an object tracking algorithm and involves the following steps. First, a first-frame object extraction algorithm is used to extract the tracking object from the first frame. Second, a region of interest (ROI)-prediction algorithm is used to predict ROIs and detect objects in these ROIs. This algorithm can effectively solve the above problems in marine tests. When multiple objects of the same classification exist in the same field of view, the algorithm can extract the radar-guided object stably. When faced with a complex background and a large change in object shape, the algorithm substantially improves the accuracy and robustness of object tracking. Compared with the conventional object detection algorithm, the time consumption of this algorithm is reduced by 25.8%.

Water segmentation is essential for autonomous navigation of Unmanned Surface Vehicles (USV) in inland waterways. However, most of the existing methods are suitable for maritime environments. Due to their poor real-time performance and a high false-positive rate of water surface reflection and water-sky interference in inland waterways, we propose a radarassisted real-time water segmentation network (RaRWS) to solve the above problems. We obtain pseudo-waterlines based on radar data by deleting points, fitting them multiple times, and generating a radar mask. While simplifying the encoder backbone network, we use the Attention Refinement Module (ARM) to fuse radar masks to improve detection and waterside accuracy. In addition, a Feature Fusion Module (FFM) is introduced to help the decoder fuse high- and low-level features and further fuse radar data. RaRWS is tested in both normal and bad weather. The results show that RaRWS can achieve higher performance compared with the current state-of-the-art methods (F <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</inf> is above 99%), while gaining real-time performance (41.3fps).

SummaryEffective and intelligent path planning algorithms designed for operation in a dynamic marine environment are essential for the safe operation of unmanned surface vehicles (USVs). Most of the current research deals with the ‘dynamic problem’ by basing solutions on the nonpractical assumption that each USV has a robust communication channel to obtain essential information such as position and velocity of marine vehicles. In this paper, a Kalman filter‐based predictive path planning algorithm is proposed. The algorithm has been designed to predict the trajectories of moving ships, and the USV's own position in real time and accordingly assesses collision risk. For path planning, a weighted fast marching square method is proposed and developed to search for the optimal path. The path can be optimised for mission requirements such as minimum distance to travel and the most safety path by adjusting weighting parameters. The proposed algorithm has been validated using a number of simulations that include practical environmental aspects. The results show that the algorithms can sufficiently deal with complex traffic environments and that the generated practical path is suited for both unmanned and manned vessels. Copyright © 2015 John Wiley &amp; Sons, Ltd.

The Unmanned Surface Vehicles (USVs) have the characteristics of small volume, fast speed, and its application field is often in unknown complex ocean environment. Therefore, it is necessary to solve the fast collision avoidance problem of USV in unknown ocean environment. To solve this problem, a fast collision avoidance method based on velocity resolution is proposed. The motion constraint of the USV in the method is mainly based on the Lanxin USV of Dalian Maritime University. Finally, the safety and rapidity of the method are verified in the complex static and dynamic environments, which can provide a reference for the research of collision avoidance technology about USV.

This paper addresses the problems related to the mapless navigation control of wheeled mobile robots based on deep learning technology. The traditional navigation control framework is based on a global map of the environment, and its navigation performance depends on the quality of the global map. In this paper, we proposes a mapless Light Detection and Ranging (LiDAR) navigation control method for wheeled mobile robots based on deep imitation learning. The proposed method is a data-driven control method that directly uses LiDAR sensors and relative target position for mobile robot navigation control. A deep convolutional neural network (CNN) model is proposed to predict motion control commands of the mobile robot without the requirement of the global map to achieve navigation control of the mobile robot in unknown environments. While collecting the training dataset, we manipulated the mobile robot to avoid obstacles through manual control and recorded the raw data of the LiDAR sensor, the relative target position, and the corresponding motion control commands. Next, we applied a data augmentation method on the recorded samples to increase the number of training samples in the dataset. In the network model design, the proposed CNN model consists of a LiDAR CNN module to extract LiDAR features and a motion prediction module to predict the motion behavior of the robot. In the model training phase, the proposed CNN model learns the mapping between the input sensor data and the desired motion behavior through end-to-end imitation learning. Experimental results show that the proposed mapless LiDAR navigation control method can safely navigate the mobile robot in four unseen environments with an average success rate of 75%. Therefore, the proposed mapless LiDAR navigation control system is effective for robot navigation control in an unknown environment without the global map.

Path planning and obstacle avoidance of Unmanned Surface Vehicle (USV) is one of the hottest research topics in modem national defense and ocean engineering. Considering the issue of obstacle-free path planning of USV, this paper focuses on a 3-DoF USV and develops an algorithm design. We adopt Gauss pseudo-spectral method to discretize control model and make use of a hybrid algorithm to optimize which combines the advantage of genetic algorithm and sequential quadratic programming algorithm. Simulation results show that this method can quickly explore a high-precision route in an unknown environment which meets the mobility requirement of USV without setting the initial value artificially.

In the autonomous city explorer (ACE) project a mobile robot is developed, which is capable of finding its way to a given destination in an unknown urban environment. An exemplary mission is to find the way from our institute to the Marienplatz, a public place in the center of Munich, without any prior knowledge or GPS information. Inspired by the behavior of humans in unknown environments, ACE must find its way by asking pedestrians. The route is about 1.5 kilometers far and includes heavily traveled roads and crowded public places. In order to navigate safely in an unknown urban environment, some challenges arise for the vision system. Robust human detection, tracking and the estimation of human body poses is essential for natural interaction with pedestrians. Furthermore, the robot needs to be able to detect sidewalk and crossroads. A visual odometry system is used to support the conventional navigation. Outdoor experiments were conducted twice successfully. After about 5 hours and interacting with 25 and 38 persons respectively, ACE arrived the Marienplatz. This paper describes both, an architecture of the vision system used for ACE and the algorithms used to deal with the described challenges.

The rapid development of machine learning in recent years has made it possible to solve the problem of surface ship monitoring. This paper takes intelligent unmanned surface vessel (USV) as the application object. It aims to solve the problem of marine target motion intention recognition based on deep reinforcement learning. Using python to build a highly realistic USV planning platform. Solving the confidentiality of field test data results in the system's simulation results unable to obtain effective feedback from examples. Aiming at the problem of low uniform sampling efficiency of DDPG algorithm, this paper proposes a USV motion control method based on prioritized experience playback DDPG(PER-DDPG) algorithm. This method assists USV to realize the movement intention recognition of the target in the unknown marine environment. This paper also constructs an accurate USV motion model, and carefully sets the USV's action and state space. By setting a new reward function to evaluate the output decision of PER-DDPG, thus strengthening the decision network to control the movement of USV. The experimental results show that the simulation has good realism and environmental adaptability. Compared with the DDPG algorithm, this algorithm has a higher learning efficiency and can realize the recognition of the target movement intention by the USV.

Layer-by-layer description of the core is performed to understand the regularities of the structure of the geological section, predict the development of reservoirs, clarify stratigraphic boundaries and obtain calculation parameters for assessing hydrocarbon reserves. In this case, the name of the breed is one of the key parameters determined in the layer-by-layer description. This paper presents a comparative analysis of two approaches to determining the breed using machine learning methods: based on graphical identifiers and convolutional neural networks. The original sample contained photographs of core samples from the Tyumenskaya suite fields (8 fields, 15 wells, more than 2 km of core) under daylight. For the analysis, 4 main classes of rocks (siltstones, mudstones, sandstones, coals) were selected. For these rocks, windows of 5 × 5 cm were formed and compressed to 299 × 299 pixels. The total sample exceeded 90,000 windows: 70% — training sample (60,359 windows) and 30% — test (31,140 windows). The training and test samples contain photographs of core samples from different fields. The comparison was made between convolutional neural networks (ResNet, ResNeXt, Inception, etc.) and a classifier (such as XGBoost) based on graphic identifiers of two types: color (average color, dominant colors) and texture (entropy, Euler’s number, contrast, dissimilarity, uniformity, energy, correlation). According to the results of the experiments, the model based on convolutional neural networks turned out to be more sensitive to implicit features and made it possible to reduce the error in the weighted average f1-measure with respect to the ensemble of weak classifiers by 12.5% on the test sample even without optimization of hyperparameters. Thus, we can conclude that the model based on convolutional neural networks is more sensitive to implicit features that are difficult to extract using known graphic identifiers. On the other hand, the approach based on graphic identifiers and an ensemble of weak classifiers can be used without specialized computing power (video cards).

The detection of coreferent pairs within a text is one of the basic tasks in the area of natural language processing (NLP). The state‑ of‑ the‑ art methods of coreference resolution are based on machine learning algorithms. The key idea of the methods is to detect certain regularities between the semantic or grammatical features of text entities. In the paper, the comparative analysis of current methods of coreference resolution in English and Ukrainian texts has been performed. The key disadvantage of many methods consists in the interpretation of coreference resolution as a classification problem. The result of coreferent pairs detection is the set of groups in which elements refer to a common entity. Therefore it is advisable to consider the coreference resolution as a clusterization task. The method of coreference resolution using the set of filtering sieves and a convolutional neural network has been suggested. The set of filtering sieves to find candidates for coreferent pairs formation has been implemented. The training process of a multichannel convolutional neural network on a marked Ukrainian corpus has been performed. The usage of a multichannel structure allows analyzing of the different components of text units: semantic, lexical, and grammatical features of words and sentences. Furthermore, it is possible to process input data with unfixed size (words or sentences of a text) using a convolutional layer. The output result of the method is the set of clusters. In order to form clusters, it is necessary to take into account the previous steps of the model’s workflow. Nevertheless, such an approach contradicts the traditional methodology of machine learning. Thus, the training process of the network has been performed using the SEARN algorithm that allows the solving of tasks with unfixed output structures using a classifier model. An experimental examination of the method on the corpus of Ukrainian news has been performed. In order to estimate the accuracy of the method the corresponding common metrics for clusterization tasks have been calculated. The results obtained can indicate that the suggested method can be used to find coreferent pairs within Ukrainian texts. The method can be also easily adapted and applied to other natural languages.

In machine learning problems, we usually assume that the validation accuracy is a good estimation of prediction accuracy for datasets without ground truth. In reality, this assumption may not hold. Therefore, we propose an approach to estimate the prediction accuracy of a target model on unlabeled datasets. The proposed approach uses multiple target homogeneous models to assign each unlabeled sample a confidence value, based on the number of models agreeing on the predicted label. With the confidence values, the prediction accuracy of the target model on the datasets can be estimated. In the experiments, the target model is a convolutional neural network (CNN) model, and the homogeneous models only differ in initial weights. The experiments are conducted with datasets from a wide variety of music genres. The estimation performance of the proposed approach is compared with the reversed testing qualities (RTQ) and the ensemble average qualities (EAQ) approaches. The RTQ approach was proposed to estimate the prediction accuracy of trained models, and the EAQ approach was originally designed for estimating the predictive uncertainty of individual samples. We apply all three compared models to estimate prediction accuracy of datasets by using a linear model. The parameters of the linear model are either computed by using multiple labeled datasets or one labeled dataset. The experimental results show that when compared with the RTQ approach, the proposed approach has much lower estimation errors for some datasets. When compared with the EAQ approach, the proposed approach is more robust for datasets with large distribution shifts. Finally, we show an additional benefit of the proposed approach. In case that the estimated accuracy is unsatisfactory, we may re-train the target model with a new training set, which contains the original training samples plus new training samples with manual labeling from the unlabeled dataset. The experimental results confirm that it is more effective to select (and label) new samples from those with low confidence values than those randomly selected. Overall, the proposed approach is a promising approach for estimating prediction accuracy on unlabeled datasets.

Path planning of unmanned surface vessel in an unknown environment based on improved D*Lite algorithm

Sea-sky line detection (SSLD) is pivotal for applications such as unmanned surface vehicles (USVs) navigation and maritime target detection. However, existing algorithms are susceptible to interference from adverse weather, illumination change, and the presence of waves, leading to poor detection accuracy and robustness. To address these challenges, we propose a novel SSLD algorithm based on a deep semantic segmentation network. First, we integrate the strengths of convolutional neural networks (CNNs) and Transformers in a lightweight block named efficient vision transformer (E-ViT). This block enables efficient interaction and aggregation of local and global features with lower computational overhead. Building upon E-ViT, we develop an encoder module that significantly improves the accuracy of semantic segmentation while maintaining the network’s lightweight. Then, we design a robust postprocessing module, which leverages semantic information to effectively remove interferences and filter out candidate points for the sea-sky line, thereby achieving high-precision SSLD. Finally, we construct a well-labeled maritime scene dataset with diverse complex attributes to validate the proposed algorithm. Experimental results demonstrate that our method outperforms several state-of-the-art algorithms in terms of both accuracy and robustness in complex maritime scenarios.