Most conventional target tracking algorithms assume that a target can generate at most one measurement per scan. However, there are tracking problems where this assumption is not valid. For example, ...multiple detections from a target in a scan can arise due to multipath propagation effects as in the over-the-horizon radar (OTHR). A conventional multitarget tracking algorithm will fail in these scenarios, since it cannot handle multiple target-originated measurements per scan. The Joint Probabilistic Data Association Filter (JPDAF) uses multiple measurements from a single target per scan through a weighted measurement-to-track association. However, its fundamental assumption is still one-to-one. In order to rectify this shortcoming, this paper proposes a new algorithm, called the Multiple-Detection Joint Probabilistic Data Association Filter (MD-JPDAF) for multitarget tracking, which is capable of handling multiple detections from targets per scan in the presence of clutter and missed detection. The multiple-detection pattern, which can account for many-to-one measurement set-to-track association rather than one-to-one measurement-to-track association, is used to generate multiple detection association events. The proposed algorithm exploits all the available information from measurements by combinatorial association of events that are formed to handle the possibility of multiple measurements per scan originating from a target. The MD-JPDAF is applied to a multitarget tracking scenario with an OTHR, where multiple detections occur due to different propagation paths as a result of scattering from different ionospheric layers. Experimental results show that multiple-detection pattern based probabilistic data association improves the state estimation accuracy. Furthermore, the tracking performance of the proposed filter is compared against the Posterior Cramér-Rao Lower Bound (PCRLB), which is explicitly derived for the multiple-detection scenario with a single target.
With the significant development of practicability in deep learning and the ultra-high-speed information transmission rate of 5G communication technology will overcome the barrier of data ...transmission on the Internet of Vehicles, automated driving is becoming a pivotal technology affecting the future industry. Sensors are the key to the perception of the outside world in the automated driving system and whose cooperation performance directly determines the safety of automated driving vehicles. In this survey, we mainly discuss the different strategies of multi-sensor fusion in automated driving in recent years. The performance of conventional sensors and the necessity of multi-sensor fusion are analyzed, including radar, LiDAR, camera, ultrasonic, GPS, IMU, and V2X. According to the differences in the latest studies, we divide the fusion strategies into four categories and point out some shortcomings. Sensor fusion is mainly applied for multi-target tracking and environment reconstruction. We discuss the method of establishing a motion model and data association in multi-target tracking. At the end of the paper, we analyzed the deficiencies in the current studies and put forward some suggestions for further improvement in the future. Through this investigation, we hope to analyze the current situation of multi-sensor fusion in the automated driving process and provide more efficient and reliable fusion strategies.
In this paper we propose an automatic trajectory data reconciliation to correct common errors in vision-based vehicle trajectory data. Given “raw” vehicle detection and tracking information from ...automatic video processing algorithms, we propose a pipeline including (a) an online data association algorithm to match fragments that describe the same object (vehicle), which is formulated as a min-cost network circulation problem of a graph, and (b) a one-step trajectory rectification procedure formulated as a quadratic program to enhance raw detection data. The pipeline leverages vehicle dynamics and physical constraints to associate tracked objects when they become fragmented, remove measurement noises and outliers and impute missing data due to fragmentations. We assess the capability of the proposed two-step pipeline to reconstruct three benchmarking datasets: (1) a microsimulation dataset that is artificially downgraded to replicate the errors from prior image processing step, (2) a 15-min NGSIM data that is manually perturbed, and (3) tracking data consists of 3 scenes from collections of video data recorded from 16–17 cameras on a section of the I-24 MOTION system, and compare with the corresponding manually-labeled ground truth vehicle bounding boxes. All of the experiments show that the reconciled trajectories improve the accuracy on all the tested input data for a wide range of measures. Lastly, we show the design of a software architecture that is currently deployed on the full-scale I-24 MOTION system consisting of 276 cameras that covers 4.2 miles of I-24. We demonstrate the scalability of the proposed reconciliation pipeline to process high-volume data on a daily basis.
•An efficient online data association algorithm, online negative cycle canceling, is developed.•A one-step convex program is formulated to rectify trajectory data.•A software architecture is presented to process large-scale, streaming data from I-24 MOTION.•Benchmarking datasets for the proposed algorithms are released.
Multi-object tracking can be achieved by detecting objects in individual frames and then linking detections across frames. Such an approach can be made very robust to the occasional detection ...failure: If an object is not detected in a frame but is in previous and following ones, a correct trajectory will nevertheless be produced. By contrast, a false-positive detection in a few frames will be ignored. However, when dealing with a multiple target problem, the linking step results in a difficult optimization problem in the space of all possible families of trajectories. This is usually dealt with by sampling or greedy search based on variants of Dynamic Programming which can easily miss the global optimum. In this paper, we show that reformulating that step as a constrained flow optimization results in a convex problem. We take advantage of its particular structure to solve it using the k-shortest paths algorithm, which is very fast. This new approach is far simpler formally and algorithmically than existing techniques and lets us demonstrate excellent performance in two very different contexts.
In this article, we demonstrate how a variation of the joint integrated probabilistic data association with interacting multiple models and a visibility state can be derived as a special case of the ...Poisson multi-Bernoulli mixture filter with a hybrid state representation and standard approximations. The proposed method is tested on two radar data sets that were recorded during maritime collision-avoidance experiments.
•Tracks and detections are modelled as random vectors in which uncertainties are taken into account.•The similarity between track and detection is evaluated based on a modified Kullback-Leibler ...divergence.•The level of track uncertainty is incorporated in the proposed cost function to guide the data association process.•Code has been made available at https://github.com/hejiawei2023/UG3DMOT. For the benefit of the community.
In the existing literature, most 3D multi-object tracking algorithms based on the tracking-by-detection framework employed deterministic tracks and detections for similarity calculation in the data association stage. Namely, the inherent uncertainties existing in tracks and detections are overlooked. In this work, we discard the commonly used deterministic tracks and deterministic detections for data association, instead, we propose to model tracks and detections as random vectors in which uncertainties are taken into account. Then, based on a modified Kullback-Leibler divergence, the similarity between two multidimensional distributions, i.e. track and detection, is evaluated for data association purposes. Lastly, the level of track uncertainty is incorporated in our cost function design to guide the data association process. Comparative experiments have been conducted on two typical datasets, KITTI and nuScenes, and the results indicated that our proposed method outperformed the compared state-of-the-art 3D tracking algorithms. For the benefit of the community, our code has been made available at https://github.com/hejiawei2023/UG3DMOT.
Visual SLAM (simultaneous localization and mapping) refers to the problem of using images, as the only source of external information, in order to establish the position of a robot, a vehicle, or a ...moving camera in an environment, and at the same time, construct a representation of the explored zone. SLAM is an essential task for the autonomy of a robot. Nowadays, the problem of SLAM is considered solved when range sensors such as lasers or sonar are used to built 2D maps of small static environments. However SLAM for dynamic, complex and large scale environments, using vision as the sole external sensor, is an active area of research. The computer vision techniques employed in visual SLAM, such as detection, description and matching of salient features, image recognition and retrieval, among others, are still susceptible of improvement. The objective of this article is to provide new researchers in the field of visual SLAM a brief and comprehensible review of the state-of-the-art.
A sequential probability ratio test (SPRT) is introduced as a track management (TM) mechanism for the joint probabilistic data association (JPDA) multitarget tracking algorithm. The SPRT is based on ...an approximate target likelihood ratio that is derived from the JPDA event and association probabilities. SPRT-based TM functions are defined and compared to the joint integrated probabilistic data association (JIPDA) algorithm, including a discussion of parameter tuning for both approaches. Simulation experiments are reported for multiple closely spaced constant-velocity targets over a range of values for detection probability and clutter density. The SPRT-based algorithm achieves performance similar to JIPDA for this data, but with fewer TM parameters and a more straightforward parameter-tuning process.
Multi-object tracking (MOT) involves the prediction of object identities and their corresponding bounding boxes within video or image sequences. While numerous models have been proposed for MOT, ...there is still a lack of discrimination of object features and severe ID switches during the tracking stage. This paper presents a novel fusion detection and re-identification (ReID) embedding with hybrid attention for multi-object tracking to address this issue. It incorporates two major cores: a hybrid attention module (HAM) and an embedding association module (EAM). Firstly, the HAM comprises spatial-aware attention, scale-aware attention, and task-aware attention, aiming to obtain more informative features. By integrating these mechanisms, the proposed model can effectively handle variations in object scales and spatial relationships to promote discrimination and balance two tasks (detection and ReID). Secondly, we introduce an embedding association module to address the unreliable similarity matching during the tracking. Specifically, the EAM not only considers the appearance similarity but also ponders on geometric attributes to improve the ability to track in the presence of object occlusions and brief disappearances. Extensive experiments are conducted on the public MOT Challenge datasets, demonstrating that our method performs better than other advanced methods.
•We present a fusion detection and ReID embedding mechanism for MOT. Extensive experiments illustrate our model achieves competitive performance.•We design a HAM module to extract discriminative features for better performance in subsequent detection and ReID tasks.•We introduce an EAM module to handle the unreliable similarity matching during tracking.
The accurate extraction of micro-Doppler (m-D) curves is crucial for target recognition for space targets. However, the strong overlap of m-D curves and the noise disturbance of multi-component ...non-stationary signals in the time-frequency domain make it a great challenge to separate these curves using conventional algorithms. To address this problem, a curve extraction algorithm for multicomponent m-D signals based on the reassociation Viterbi algorithm (RVA) is proposed in this paper. First, the Viterbi algorithm is combined with an adaptive curve removal method for curve detection in a multicomponent m-D signal. Then, a two-step association method based on distance and velocity is presented to achieve correct curve separation. In addition, a variable threshold algorithm is introduced to extract both strong and weak components. Finally, the improved validity and robustness of the proposed algorithm have been verified by experiments compared to conventional methods.