Target Motion Research Articles

Adaptive pseudo‐spectrum based track‐before‐detect for targets with uncertain existence

AbstractIn the existing velocity filtering based track‐before‐detect (VF‐TBD), target existence is generally assumed to be constant, and multiple successive frames in a sliding window are jointly processed. However, when the target existence changes abruptly in a processing batch, the incorporation of noise‐only frames may corrupt target energy integration. This results in degraded detection and estimation performances with serious decision delays. A pseudo‐spectrum based VF‐TBD is proposed to effectively detect and track targets with existence uncertainty. An adaptive integration strategy is developed to realise effective energy accumulation with only informative frames by matching target appearance and disappearance times in addition to the velocity. The change of target existence can be discovered in time, and the decision delay is reduced. An estimation strategy based on the maximum output SNR is presented to supply accurate estimations of not only target motion parameters (including the position and velocity) but also time parameters (including appearance and disappearance times). The target output envelope, SNR gain and receiver operating characteristic curve of targets with uncertain existence are analysed theoretically. Numerical simulations exhibit the validity of the proposed method.The cover image is based on the Original Article Adaptive pseudo‐spectrum based track‐before‐detect for targets with uncertain existence by Peiyuan Li et al., https://doi.org/10.1049/rsn2.12502.

New robust state‐estimation frameworks for angles‐only underwater target tracking problems

SummaryTracking a moving target using only angle measurements is a challenging and complex problem. This is because the measurements received from a sensor on a moving observer provide very little information about the target's movements. Additionally, there are uncertainties regarding the target's initial range and speed, which can further complicate the estimation process. The angle measurements themselves carry a lot of uncertainties as they can be affected by significant outliers, which do not fit the standard Gaussian distribution that estimation algorithms typically assume. Furthermore, the observer's position can also be uncertain due to various environmental factors. To address these challenges and to propose a robust, unified and accurate solution, two new estimation frameworks are proposed. They involve making use of a range and speed parametrization approach to compensate for the initial uncertainty in the target's range and speed. They also involve using a maximum correntropy (MC) criterion and a numerically stable centered error entropy (CEE) criterion to deal with measurement outliers and the observer motion uncertainties. The performance of the proposed frameworks is evaluated by comparing them to the traditional unscented Kalman filter (UKF) and its parametrized versions using different performance metrics. The simulations conducted showed that the developed algorithms perform better in terms of estimation accuracy than conventional methods. Specifically, the second framework based on the CEE criterion outperformed MC‐based estimators.

Magnetic Anomaly Detection Based on a Compound Tri-Stable Stochastic Resonance System.

In the case of strong background noise, a tri-stable stochastic resonance model has higher noise utilization than a bi-stable stochastic resonance (BSR) model for weak signal detection. However, the problem of severe system parameter coupling in a conventional tri-stable stochastic resonance model leads to difficulty in potential function regulation. In this paper, a new compound tri-stable stochastic resonance (CTSR) model is proposed to address this problem by combining a Gaussian Potential model and the mixed bi-stable model. The weak magnetic anomaly signal detection system consists of the CTSR system and judgment system based on statistical analysis. The system parameters are adjusted by using a quantum genetic algorithm (QGA) to optimize the output signal-to-noise ratio (SNR). The experimental results show that the CTSR system performs better than the traditional tri-stable stochastic resonance (TTSR) system and BSR system. When the input SNR is -8 dB, the detection probability of the CTSR system approaches 80%. Moreover, this detection system not only detects the magnetic anomaly signal but also retains information on the relative motion (heading) of the ferromagnetic target and the magnetic detection device.

Research on multi-target data association of the infrared fish-eye warning system

The multi-target data association method is studied in order to realize multi-target tracking in infrared fish-eye warning system. The Neural Joint Probabilistic Data Association (NJPDA) algorithm is analyzed. It is found that the NJPDA algorithm only considers the distance information between the measurement and the target in the data association process, and its tracking accuracy needs to be further improved. Therefore, a new method fused with direction information is proposed based on the NJPDA algorithm. The proposed algorithm defines the concept of direction difference, introduces the direction information of target motion, and modifies the likelihood function by Gaussian weighting method, so as to fuse the direction information of target motion into the calculation of data interconnection probability. Experimental results demonstrate that the tracking success rate of the proposed algorithm is nearly 10 % higher than that of JPDA and NJPDA algorithms and its consuming time meets the real-time requirement of the infrared fish-eye warning system.

Interception of vertically approaching objects: temporal recruitment of the internal model of gravity and contribution of optical information.

Introduction: Recent views posit that precise control of the interceptive timing can be achieved by combining on-line processing of visual information with predictions based on prior experience. Indeed, for interception of free-falling objects under gravity's effects, experimental evidence shows that time-to-contact predictions can be derived from an internal gravity representation in the vestibular cortex. However, whether the internal gravity model is fully engaged at the target motion outset or reinforced by visual motion processing at later stages of motion is not yet clear. Moreover, there is no conclusive evidence about the relative contribution of internalized gravity and optical information in determining the time-to-contact estimates. Methods: We sought to gain insight on this issue by asking 32 participants to intercept free falling objects approaching directly from above in virtual reality. Object motion had durations comprised between 800 and 1100ms and it could be either congruent with gravity (1g accelerated motion) or not (constant velocity or -1g decelerated motion). We analyzed accuracy and precision of the interceptive responses, and fitted them to Bayesian regression models, which included predictors related to the recruitment of a priori gravity information at different times during the target motion, as well as based on available optical information. Results: Consistent with the use of internalized gravity information, interception accuracy and precision were significantly higher with 1g motion. Moreover, Bayesian regression indicated that interceptive responses were predicted very closely by assuming engagement of the gravity prior 450ms after the motion onset, and that adding a predictor related to on-line processing of optical information improved only slightly the model predictive power. Discussion: Thus, engagement of a priori gravity information depended critically on the processing of the first 450 ms of visual motion information, exerting a predominant influence on the interceptive timing, compared to continuously available optical information. Finally, these results may support a parallel processing scheme for the control of interceptive timing.

The effects of object size on spatial orientation: an eye movement study.

The processing of visual information in the human brain is divided into two streams, namely, the dorsal and ventral streams, object identification is related to the ventral stream and motion processing is related to the dorsal stream. Object identification is interconnected with motion processing, object size was found to affect the information processing of motion characteristics in uniform linear motion. However, whether the object size affects the spatial orientation is still unknown. Thirty-eight college students were recruited to participate in an experiment based on the spatial visualization dynamic test. Eyelink 1,000 Plus was used to collect eye movement data. The final direction difference (the difference between the final moving direction of the target and the final direction of the moving target pointing to the destination point), rotation angle (the rotation angle of the knob from the start of the target movement to the moment of key pressing) and eye movement indices under conditions of different object sizes and motion velocities were compared. The final direction difference and rotation angle under the condition of a 2.29°-diameter moving target and a 0.76°-diameter destination point were significantly smaller than those under the other conditions (a 0.76°-diameter moving target and a 0.76°-diameter destination point; a 0.76°-diameter moving target and a 2.29°-diameter destination point). The average pupil size under the condition of a 2.29°-diameter moving target and a 0.76°-diameter destination point was significantly larger than the average pupil size under other conditions (a 0.76°-diameter moving target and a 0.76°-diameter destination point; a 0.76°-diameter moving target and a 2.29°-diameter destination point). A relatively large moving target can resist the landmark attraction effect in spatial orientation, and the influence of object size on spatial orientation may originate from differences in cognitive resource consumption. The present study enriches the interaction theory of the processing of object characteristics and motion characteristics and provides new ideas for the application of eye movement technology in the examination of spatial orientation ability.

A target tracking method based on adaptive occlusion judgment and model updating strategy

Target tracking is an important research in the field of computer vision. Despite the rapid development of technology, difficulties still remain in balancing the overall performance for target occlusion, motion blur, etc. To address the above issue, we propose an improved kernel correlation filter tracking algorithm with adaptive occlusion judgement and model updating strategy (called Aojmus) to achieve robust target tracking. Firstly, the algorithm fuses color-naming (CN) and histogram of gradients (HOG) features as a feature extraction scheme and introduces a scale filter to estimate the target scale, which reduces tracking error caused by the variations of target features and scales. Secondly, the Aojmus introduces four evaluation indicators and a double thresholding mechanism to determine whether the target is occluded and the degree of occlusion respectively. The four evaluation results are weighted and fused to a final value. Finally, the updating strategy of the model is adaptively adjusted based on the weighted fusion value and the result of the scale estimation. Experimental evaluations on the OTB-2015 dataset are conducted to compare the performance of the Aojmus algorithm with four other comparable algorithms in terms of tracking precision, success rate, and speed. The experimental results show that the proposed Aojmus algorithm outperforms all the algorithms compared in terms of tracking precision. The Aojmus also exhibits excellent performance on attributes such as target occlusion and motion blur in terms of success rate. In addition, the processing speed reaches 74.85 fps, which also demonstrates good real-time performance.

A Poisson multi-Bernoulli mixture filter for tracking multiple resolvable group targets

This paper presents a novel Poisson multi-Bernoulli mixture (PMBM) filter for tracking multiple resolvable group targets (MRGT) based on graph theory. Firstly, the number of groups and the relationships between members within the group are modelled by the adjacency matrix. Then, considering that a single dynamic evolution model is insufficient to guarantee stable tracking performance for group targets, the virtual leader-follower model (VLFM) is introduced to predict the evolution trend of unknown and potentially detected targets, respectively. Furthermore, we prove the conjugation of the proposed algorithm with the probability generating functionals (PGF) and give a detailed implementation of the Gaussian mixture (GM). Based on the coexistence scenario of splitting, merging and non-linear motion of the group targets, the simulation results show the effectiveness of the proposed algorithm in comparison with the existing methods.

Intent-informed state estimation for tracking guided targets

This paper proposes a state estimation and prediction for tracking guided targets using intent information. A conditionally Markov process is used to describe the destination-oriented target motion, and the collision intent is incorporated through the zero-effort-miss guidance information. The expected arrival time necessary for the conditionally Markov model is determined through the collision geometry and destination motion. Finally, the Kalman filter technique is used to estimate and predict the target state. Numerical simulations demonstrate that the proposed approach can improve state estimation accuracy in both static and dynamic destination cases.

Analysis of rotational Doppler shift with multi-ring vortex beams.

Vortex beams (VBs) with orbital angular momentum have shown great potential in the detection of transverse rotational motion of spatial targets which is undetectable in the classical radar scheme. However, most of the reported rotational Doppler measurements based on VBs can only be realized under ideal experimental conditions. The long-range detection is still a challenge. The detection distance based on rotational Doppler effect (RDE) is mainly limited by the scattered signal's signal-to-noise ratio (SNR). In this work, we investigated the influence of multi-ring vortex beams (MVBs) on the rotational Doppler frequency spectrum of scattered light from an object based on RDE and proposed a method of SNR enhancement of RDE signal. Firstly, different types of MVBs composed of a set of single-ring VBs with the same topological charge and different radii are designed, including multi-ring Laguerre Gaussian beam (MLGB), multi-ring perfect vortex beams (MPVB), and high-order Laguerre Gaussian beam (HLGB). Then, the influence of the number of rings and radial radius interval on the intensity profiles of MVBs and rotational Doppler frequency spectra under aligned and misaligned conditions is studied in detail. And the reasons why different types of MVBs lead to different SNR enhancement effectiveness with the increase of rings are also analyzed theoretically. Finally, proof-of-concept experiments were conducted to verify the effectiveness of the SNR enhancement method for RDE signals. The results showed that the amplitudes of the Doppler spectra generated by the MLGB and MPVB are improved substantially with the increase of rings, but the enhancement effect caused by the former is superior to the latter. The gain of HLGB on the RDE signal is the lowest. This study provides a useful reference for the optimization of rotational Doppler detection systems and may be of great application value in telemetry, long-range communication and optical imaging.

Evaluation of Drag-and-Drop Task in Virtual Environment: Effects of Target Size and Movement Distance on Performances and Workload

This study investigated the effects of the target size and movement distance on user performance and workload in a virtual reality (VR) environment. In a repeated-measures laboratory study, 36 participants (18 male and 18 female) performed the drag-and-drop task as a standard human–computer interaction (HCI) task with different target sizes (1, 1.5, 2, 2.5, and 3 cm) and movement distances (5, 9, 13, 17, and 20 cm). Task completion time (TCT), error rate, and movement time (MT) were measured as performance indices, whereas physical load and effort were assessed as workload indices. The results demonstrated that the target size and movement distance significantly affected all performance measures and workload indices. Large target sizes produced better performance and lower workloads; however, large movement distances decreased performance and increased workload. However, sex had no significant effect on the performance or workload during the drag-and-drop tasks. The best target sizes were 2.5 and 3 cm, and the worst size was 1 cm. The best movement distances were 5 and 9 cm, and the worst distance was 20 cm. The results of this study can provide useful reference information for developing VR technology based on human factors and demonstrate that additional basic research is required to reflect the distinctive features of VR in the future.

A deep learning model based on transformer structure for radar tracking of maneuvering targets

The motion complexity of maneuvering target causes the estimation uncertainty of target motion model, resulting in state estimation error. Especially for strong maneuvering target, the drastic change of target motion models makes the tracking methods hard to adapt and provide accurate state estimation promptly. To solve the state estimation problem of strong maneuvering targets, we propose a new transformer maneuvering target tracking model based on deep learning, named TrMTT model. The TrMTT model uses a new residual mapping between the observation trajectory and the real trajectory to estimate the target states, and is composed of the encoder and decoder branches while the two have the same input of observation trajectory. The encoder extracts the self-attention information for the input at each layer while the decoder implements cross-attention extraction and fusion between features in different layers, thus providing more correlation information between states for learning the transition law of rapidly changing states. Moreover, we propose an input module before the encoder–decoder structure to code the state features of the observation trajectory, and apply two kinds of normalization layers in the input module and the encoder–decoder structure, to project the input into a feature space which facilitates extracting the correlation information between states. Simulation results show that the proposed TrMTT model is superior in performance for maneuvering target tracking compared with other existing approaches.

Multi-agent target search strategy optimization: Hierarchical reinforcement learning with multi-criteria negative feedback

Recently, using unmanned platforms to perform target search tasks has received extensive attention and research. The complexity of the search scenario and the unpredictability of the target movement pose significant challenges for unmanned platforms to perform the search task. Developing efficient search strategies is crucial for their success. In this study, we model the problem as a Partially Observable Markov Decision Process (POMDP) and propose a Hierarchical Deep Q Network with Multi-criteria Negative Feedback method named MNF-HDQN to solve the problem efficiently. The MNF-HDQN incorporates point and area evaluations instead of the original sparsity calculation to enhance the cooperation competence of unmanned platforms in searching for a target in various search tasks. Additionally, the integration of convex polygon theory into reward shaping and the design of a new two-stage search strategy encouragement function further improve the performance of the proposed method. We conduct detailed experiments to verify the superiority of the MNF-HDQN. And experimental results show that our method provides a high successful search rate in a shorter timestep compared with state-of-the-art baselines. This advantage is more evident when the search scenario is more complex or the target movement is more unpredictable.

Target motion misjudgments reflect a misperception of the background; revealed using continuous psychophysics.

Determining the velocities of target objects as we navigate complex environments is made more difficult by the fact that our own motion adds systematic motion signals to the visual scene. The flow-parsing hypothesis asserts that the background motion is subtracted from visual scenes in such cases as a way for the visual system to determine target motions relative to the scene. Here, we address the question of why backgrounds are only partially subtracted in lab settings. At the same time, we probe a much-neglected aspect of scene perception in flow-parsing studies, that is, the perception of the background itself. Here, we present results from three experienced psychophysical participants and one inexperienced participant who took part in three continuous psychophysics experiments. We show that, when the background optic flow pattern is composed of local elements whose motions are congruent with the global optic flow pattern, the incompleteness of the background subtraction can be entirely accounted for by a misperception of the background. When the local velocities comprising the background are randomly dispersed around the average global velocity, an additional factor is needed to explain the subtraction incompleteness. We show that a model where background perception is a result of the brain attempting to infer scene motion due to self-motion can account for these results.

Tracking a moving visual target in the rhesus monkey: influence of the occurrence frequency of the target path.

Following previous studies documenting the ability to generate anticipatory responses, we tested whether the repeated motion of a visual target along the same path affected its oculomotor tracking. In six rhesus monkeys, we evaluated how the frequency of a target path influenced the onset, accuracy, and velocity of eye movements. Three hundred milliseconds after its extinction, a central target reappeared and immediately moved toward the periphery in four possible (oblique) directions and at a constant speed (20°/s or 40°/s). During each daily session, the frequency of one motion direction was either uncertain (25% of trials) or certain (100% of trials). Our results show no reduction of saccade latency between the two sessions. No express saccades were observed in either session. A slow eye movement started after target onset (presaccadic glissade) and its velocity was larger during the "certain" sessions only with the 40°/s target. No anticipatory eye movement was observed. Longer intersaccadic intervals were found during the "certain" sessions but the postsaccadic pursuit velocity exhibited no change. No correlation was found between the accuracy and precision of saccades (interceptive or catch-up) and the postsaccadic pursuit velocity. Repeatedly tracking a target that moves always along the same path does not favor the generation of anticipatory eye movements, saccadic or slow. Their occurrence is not spontaneous but seems to require preliminary training. Finally, for both sessions, the lack of correlation between the saccade-related and pursuit-related kinematic parameters is consistent with separate control of saccadic and slow eye movements.NEW & NOTEWORTHY Following previous studies documenting anticipatory movements, we investigated how the frequency of occurrence of a target path influenced the generation of tracking eye movements. When present, the effects were small. The limited performance that we found suggests that anticipatory responses require preliminary training, in which case, they should not be considered as a behavioral marker of the primates' ability to extrapolate but the outcome of learning and remembering past experience.

Optimizing SIFT algorithm parameters for better matching UAV and satellite images

Image registration has been increasingly employed in various applications such as target identification, 3D mapping, and motion tracking. The main idea of Image registration is aligning two or more images of the same scene captured from different viewpoints, at different times. Scale-invariant feature transform, SIFT, is considered one of the most robust algorithms used in image registration for extracting and matching features under different conditions. Using SIFT algorithm default parameters in Matching UAV and satellite Images provides unreliable results due to the nature of aerial images because the dynamic range is quite low. The number of extracted features depends on the image content and the selected parameters. In this paper we tuned SIFT parameters to get the best performance with aerial images, to increase the number of features (SM) and the correct match rate (CMR) which increases the efficiency of the process of registration. The algorithm is validated by matching a large number of aerial images taken by mini-UAV with satellite images for the same region.

Visual–inertial object tracking: Incorporating camera pose into motion models

Visual object tracking for autonomy of aerial robots could become challenging especially in the presence of target or camera fast motions and long-term occlusions. This paper presents a visual-inertial tracking paradigm by incorporating the camera kinematics states into the visual object tracking pipelines. We gathered a dataset of image sequences with the addition of camera’s position and orientation measurements as well as the object’s position measurement. For the cases of long-term object occlusion, we provide ground-truth boxes derived from mapping the measured object position onto the image frame. A search zone proposal method is developed based on the estimation of object future position represented in the inertial frame and projected back onto the image frame using the camera states. This search zone, which is robust to fast camera/target motions, is fused into the original search zone settings of the base tracker. Also proposed is a measure indicating the confidence of a tracking structure in keeping track of a correct target and reporting the tracking failure in-time. Accordingly, the model updating mechanism of base tracker is modulated to avoid recovering of wrong objects as the target. The proposed modifications are benchmarked on nine visual object tracking algorithms including five state-of-art deep learning structures , namely DiMP, PrDiMP, KYS, ToMP, and MixFormer. Most of the trackers are remarkably improved by the modifications with up to 8% increase in precision. Modified PrDiMP tracker yields the best precision of 68.4%, more than all considered original (and modified) trackers. Source code and dataset are made available online.1

Sample-balanced and IoU-guided anchor-free visual tracking

Siamese network-based visual tracking algorithms have achieved excellent performance in recent years, but challenges such as fast target motion, shape and scale variations have made the tracking extremely difficult. The regression of anchor-free tracking has low computational complexity, strong real-time performance, and is suitable for visual tracking. Based on the anchor-free siamese tracking framework, this paper firstly introduces balance factors and modulation coefficients into the cross-entropy loss function to solve the classification inaccuracy caused by the imbalance between positive and negative samples as well as the imbalance between hard and easy samples during the training process, so that the model focuses more on the positive samples and the hard samples that make the major contribution to the training. Secondly, the intersection over union (IoU) loss function of the regression branch is improved, not only focusing on the IoU between the predicted box and the ground truth box, but also considering the aspect ratios of the two boxes and the minimum bounding box area that accommodate the two, which guides the generation of more accurate regression offsets. The overall loss of classification and regression is iteratively minimized and improves the accuracy and robustness of visual tracking. Experiments on four public datasets, OTB2015, VOT2016, UAV123 and GOT-10k, show that the proposed algorithm achieves the state-of-the-art performance.

Underwater Doppler-bearing maneuvering target motion analysis based on joint estimated adaptive unscented Kalman filter

Noncooperative maneuvering target motion analysis is one of the challenging tasks in the field of underwater target localization and tracking for passive sonar. Underwater noncooperative targets often perform various maneuvers, and the targets are commonly modeled as a combination of constant-velocity models and coordinate-turn models with unknown turning rates. Traditional algorithms for Doppler-bearing target motion analysis are incapable of processing noncooperative maneuvering targets because the algorithms rely on a priori information of the turning rate and the center frequency. To address these shortcomings, this paper proposes the joint estimated adaptive unscented Kalman filter (JE-AUKF) algorithm. The JE-AUKF places the center frequency and turning rate into the state vector and constructs a time-varying state model that self-adapts to a maneuvering target. The JE-AUKF also introduces a time-varying fading factor into the process noise covariance matrix to improve the tracking performance. Simulations and sea trials are conducted to compare the performance of the JE-AUKF with the iterative unscented Kalman filter, the interacting multiple model-unscented Kalman filter, the interacting multiple model-iterative unscented Kalman filter, and the interacting multiple model-joint estimated unscented Kalman filter. The result shows that the JE-AUKF achieves better tracking performance for noncooperative maneuvering targets.

P056 - Ultra-high single-dose ablative radiation therapy with real-time target motion management for unfavorable localized prostate cancer

P056 - Ultra-high single-dose ablative radiation therapy with real-time target motion management for unfavorable localized prostate cancer