Search

The development of ship technology and information technology has been driving the continuous improvement of ship intelligence, with safety being an inevitable requirement in the shipping industry. A machine vision-based ship collision warning method is proposed for high monitoring system cost and limited information acquisition in safety design of autonomous ship navigation. The method combines machine learning with image algorithms. Firstly, the backbone of YOLOv7 detector is replaced by EfficientFormerV2 network to achieve model lightweight while ensuring detection accuracy. Public datasets SeaShips , Flow and self-made ship pictures are combined, and the network is trained on this dataset. StrongSORT is used for target tracking. Secondly, a data fusion algorithm is introduced to determine the target point at the bow-bottom of the ship based on the time-varying attitude of the camera and the time-series features of the bounding boxes. Ship navigation trajectory estimation is performed using image algorithms. Finally, a collision evaluation model is established to calculate the collision risk index. Experimental results demonstrate that the improved YOLOv7 network maintains similar mAP.5 and Recall compared to the original model, while reducing the parameters by 31.2 % and GFLOPs by 58.4 %. The accuracy of target ship trajectory estimation is high, with MAE values below 1.5 % and RMSE values below 2 % in experiments. In ship collision warning experiments, the proposed method accurately identifies navigating vessels, estimates the trajectories, and provides timely warnings for imminent collision accidents. Compared to traditional ship collision warning methods, this paper offers a more intelligent and lightweight solution., Competing Interests: The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (© 2024 The Authors.)

The development of high-performance, low-cost unmanned aerial vehicles paired with rapid progress in vision-based perception systems herald a new era of autonomous flight systems with mission-ready capabilities. One of the key features of an autonomous UAV is a robust mid-air collision avoidance strategy. This paper proposes a vision-based in-flight collision avoidance system based on background subtraction using an embedded computing system for unmanned aerial vehicles (UAVs). The pipeline of proposed in-flight collision avoidance system is as follows: (i) subtract dynamic background subtraction to remove it and to detect moving objects, (ii) denoise using morphology and binarization methods, (iii) cluster the moving objects and remove noise blobs, using Euclidean clustering, (iv) distinguish independent objects and track the movement using the Kalman filter, and (v) avoid collision, using the proposed decision-making techniques. This work focuses on the design and the demonstration of a vision-based fast-moving object detection and tracking system with decision-making capabilities to perform evasive maneuvers to replace a high-vision system such as event camera. The novelty of our method lies in the motion-compensating moving object detection framework, which accomplishes the task with background subtraction via a two-dimensional transformation approximation. Clustering and tracking algorithms process detection data to track independent objects, and stereo-camera-based distance estimation is conducted to estimate the three-dimensional trajectory, which is then used during decision-making procedures. The examination of the system is conducted with a test quadrotor UAV, and appropriate algorithm parameters for various requirements are deduced.

Inertial localisation is an important technique as it enables ego-motion estimation in conditions where external observers are unavailable. However, low-cost inertial sensors are inherently corrupted by bias and noise, which lead to unbound errors, making straight integration for position intractable. Traditional mathematical approaches are reliant on prior system knowledge, geometric theories and are constrained by predefined dynamics. Recent advances in deep learning, which benefit from ever-increasing volumes of data and computational power, allow for data-driven solutions that offer more comprehensive understanding. Existing deep inertial odometry solutions rely on estimating the latent states, such as velocity, or are dependent on fixed-sensor positions and periodic motion patterns. In this work, we propose taking the traditional state estimation recursive methodology and applying it in the deep learning domain. Our approach, which incorporates the true position priors in the training process, is trained on inertial measurements and ground truth displacement data, allowing recursion and learning both motion characteristics and systemic error bias and drift. We present two end-to-end frameworks for pose invariant deep inertial odometry that utilises self-attention to capture both spatial features and long-range dependencies in inertial data. We evaluate our approaches against a custom 2-layer Gated Recurrent Unit, trained in the same manner on the same data, and tested each approach on a number of different users, devices and activities. Each network had a sequence length weighted relative trajectory error mean ≤0.4594 m, highlighting the effectiveness of our learning process used in the development of the models.

Multi-robot motion and observation generally have nonlinear characteristics; in response to the problem that the existing extended Kalman filter (EKF) algorithm used in robot position estimation only considers first-order expansion and ignores the higher-order information, this paper proposes a multi-robot formation trajectory based on the high-order Kalman filter method. The joint estimation method uses Taylor expansion of the state equation and observation equation and introduces remainder variables on this basis, which effectively improves the estimation accuracy. In addition, the truncation error and rounding error of the filtering algorithm before and after the introduction of remainder variables, respectively, are compared. Our analysis shows that the rounding error is much smaller than the truncation error, and the nonlinear estimation performance is greatly improved.

Driver situation awareness is critical for safety. In this paper, we propose a fast, accurate method for obtaining real-time situation awareness using a single type of sensor: monocular cameras. The system tracks the host vehicle's trajectory using sparse optical flow and tracks vehicles in the surrounding environment using convolutional neural networks. Optical flow is used to measure the linear and angular velocity of the host vehicle. The convolutional neural networks are used to measure target vehicles' positions relative to the host vehicle using image-based detections. Finally, the system fuses host and target vehicle trajectories in the world coordinate system using the velocity of the host vehicle and the target vehicles' relative positions with the aid of an Extended Kalman Filter (EKF). We implement and test our model quantitatively in simulation and qualitatively on real-world test video. The results show that the algorithm is superior to state-of-the-art sequential state estimation methods such as visual SLAM in performing accurate global localization and trajectory estimation for host and target vehicles.

Hazardous accidents often happen in construction sites and bring fatal consequences, and therefore safety management has been a certain dilemma to construction managers for long time. Although computer vision technology has been used on construction sites to identify construction workers and track their movement trajectories for safety management, the detection effect is often influenced by limited coverage of single cameras and occlusion. A multi-angle fusion method applying SURF feature algorithm is proposed to coalesce the information processed by improved GMM (Gaussian Mixed Model) and HOG + SVM (Histogram of Oriented Gradient and Support Vector Machines), identifying the obscured workers and achieving a better detection effect with larger coverage. Workers are tracked in real-time, with their movement trajectory estimated by utilizing Kalman filters and safety status analyzed to offer a prior warning signal. Experimental studies are conducted for validation of the proposed framework for workers' detection and trajectories estimation, whose result indicates that the framework is able to detect workers and predict their movement trajectories for safety forewarning.

This study presents a methodology for estimating passenger's spatio-temporal trajectory with personalization and timeliness by using incomplete Wi-Fi probe data in urban rail transit network. Unlike the automatic fare collection data that only records passenger's entries and exits, the Wi-Fi probe data can capture more detailed passenger movements, such as riding a train or waiting on a platform. However, the estimation of spatio-temporal trajectories remains as a challenging task because a few unfavorable situations could result into deficient data. To address this problem, we first describe the Wi-Fi probe data and summarize their common defects. Then, the n -gram method is developed to infer missing spatio-temporal location information. Next, an estimation algorithm is designed to generate feasible spatio-temporal trajectories for each individual passenger by integrating multiple data sources, i.e., urban rail transit network topology, Wi-Fi probe data, train schedules, etc. This proposed method is tested on both simulated data in blind experiments and real-world data from a complex urban rail transit network. The results of case study show that 93% of passengers' unique physical routes can be estimated. Then, for 80% of passengers, the number of feasible spatio-temporal trajectories can be reduced to one or two. Potential applications of the trajectory estimation approach are also identified., Competing Interests: The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (© 2020 Elsevier B.V. All rights reserved.)

Although geo-tagged mobility data (e.g., cell phone data and social media data) can be potentially used to estimate individual space-time travel trajectories, they often have low sample rates that only tell travelers' whereabouts at the sparse sample times while leaving the remaining activities to be estimated with interpolation. This study proposes a set of time geography-based measures to quantify the accuracy of the trajectory estimation in a robust manner. A series of measures including activity bandwidth and normalized activity bandwidth are proposed to quantify the possible absolute and relative error ranges between the estimated and the ground truth trajectories that cannot be observed. These measures can be used to evaluate the suitability of the estimated individual trajectories from sparsely sampled geo-tagged mobility data for travel mobility analysis. We suggest cutoff values of these measures to separate useful data with low estimation errors and noisy data with high estimation errors. We conduct theoretical analysis to show that these error measures decrease with sample rates and people's activity ranges. We also propose a lookup table-based interpolation method to expedite the computational time. The proposed measures have been applied to 2013 geo-tagged tweet data in New York City and 2014 cell-phone data in Shenzhen, China. The results illustrate that the proposed measures can provide estimation error ranges for exceptionally large datasets in much shorter times than the benchmark method without using lookup tables. These results also reveal managerial results into the quality of these data for human mobility studies, including their distribution patterns.

MR images are affected by system delays and gradient field imperfections which induce discrepancies between prescribed and actual k-space trajectories. This could be even more critical for non-Cartesian data acquisitions where even a small deviation from the assumed k-space trajectory results in severe image degradation and artifacts. Knowledge of the actual k-space trajectories is therefore crucial and can be incorporated in the reconstruction of high quality non-Cartesian images. A novel MR method for the calibration of actual gradient waveforms was developed using a combination of phase encoding increments and subsequent detection of the exact time point at which the corresponding trajectory is crossing the k-space origin. The measured sets of points were fitted to a parametrical model to calculate the complete actual acquisition trajectory. Measurements performed on phantoms and volunteers, positioned both in- and off-isocenter of the magnet, clearly demonstrate the improvement in reconstructed ultrashort echo time (UTE) images, when information from calibration of k-space sampling trajectories is employed in the MR image reconstruction procedure. The unique feature of the proposed method is its robustness and simple experimental setup, making it suitable for quick acquisition trajectory calibration procedures e.g. for non-Cartesian radial fast imaging., (Copyright © 2016 Elsevier Inc. All rights reserved.)

Purpose: In radial imaging, projections may become "miscentered" due to gradient errors such as delays and eddy currents. These errors may result in image artifacts and can disrupt the reliability of direct current (DC) navigation. The proposed parallel imaging-based technique retrospectively estimates trajectory error from miscentered radial data without extra acquisitions, hardware, or sequence modification., Theory and Methods: After phase correction, self-calibrated GRAPPA operator gridding (GROG) weights are iteratively applied to shift-miscentered projections toward the center of k-space. A search algorithm identifies the shift that aligns the peak k-space signals by maximizing the sum-of-squares DC signal estimate of each projection. The algorithm returns a trajectory estimate and a corrected radial k-space signal., Results: Data from a spherical phantom, the head, and the heart demonstrate that image reconstruction with the estimated trajectory restores image quality and reduces artifacts such as streaks and signal voids. The DC signal level is increased and variability is reduced., Conclusion: Retrospective phase correction and iterative application of GROG can be used to successfully estimate the trajectory error in two-dimensional radial acquisitions for improved image reconstruction without requiring extra data acquisition or sequence modification., (© 2015 Wiley Periodicals, Inc.)

An estimation method for a two-dimensional walking trajectory during free walking, such as forward walking, side stepping and backward walking, was investigated using a tiptoe-mounted inertial sensor. The horizontal trajectory of the toe-tip is obtained by double integration of toe-tip acceleration during the moving phase in which the sensor is rotated before foot-off or after foot-contact, in addition to the swing phase. Special functions that determine the optimum moving phase as the integral duration in every one step are developed statistically using the gait cycle and the resultant angular velocity of dorsi/planter flexion, pronation/supination and inversion/eversion so that the difference between the estimated trajectory and actual one gives a minimum value during free walking with several cadences. To develop the functions, twenty healthy volunteers participated in free walking experiments in which subjects performed forward walking, side stepping to the right, side stepping to the left, and backward walking at 39 m down a straight corridor with several predetermined cadences. To confirm the effect of the developed functions, five healthy subjects participated in the free walking experiment in which each subject performed free walking with different velocities of normal, fast, and slow based on their own assessment in a square course with 7 m side. The experimentally obtained results of free walking with a combination of forward walking, backward walking, and side stepping indicate that the proposed method produces walking trajectory with high precision compared with the constant threshold method which determines swing phase using the size of the angular velocity., (Copyright © 2015 Elsevier Ltd. All rights reserved.)

This study proposes a trajectory estimation scheme for tactical ballistic missiles (TBMs). Target information acquired from the ground-based radar system is investigated by incorporating input estimation (IE) and extended Kalman filtering techniques. In addition to estimate the missile's position and velocity, our special focus is put on the estimation of the TBMs evasive acceleration and ballistic coefficient. In the demonstrative example, radar measurement errors are served as specifications while characterizing the acquirable zone of the ground-based radar system. Effect of the proposed design is fully verified by examining the estimation performance., (© 2013 ISA Published by ISA All rights reserved.)

Large meteors (or superbolides [Ceplecha Z, et al. (1999) Meteoroids 1998:37-54]), although rare in recorded history, give sobering testimony to civilization's inherent vulnerability. A not-so-subtle reminder came on the morning of February 15, 2013, when a large meteoroid hurtled into the Earth's atmosphere, forming a superbolide near the city of Chelyabinsnk, Russia, ∼1,500 km east of Moscow, Russia [Ivanova MA, et al. (2013) Abstracts of the 76th Annual Meeting of the Meteoritical Society, 5366]. The object exploded in the stratosphere, and the ensuing shock wave blasted the city of Chelyabinsk, damaging structures and injuring hundreds. Details of trajectory are important for determining its specific source, the likelihood of future events, and potential mitigation measures. Earth-viewing environmental satellites can assist in these assessments. Here we examine satellite observations of the Chelyabinsk superbolide debris trail, collected within minutes of its entry. Estimates of trajectory are derived from differential views of the significantly parallax-displaced [e.g., Hasler AF (1981) Bull Am Meteor Soc 52:194-212] debris trail. The 282.7 ± 2.3° azimuth of trajectory, 18.5 ± 3.8° slope to the horizontal, and 17.7 ± 0.5 km/s velocity derived from these satellites agree well with parameters inferred from the wealth of surface-based photographs and amateur videos. More importantly, the results demonstrate the general ability of Earth-viewing satellites to provide valuable insight on trajectory reconstruction in the more likely scenario of sparse or nonexistent surface observations.