Target Track Enhancement Based on Asynchronous Radar and Camera Fusion in Intelligent Driving System

In urban rail transit systems, the intelligent driving system is gradually replacing manual driving for its high safety, punctuality, and stopping accuracy. With the development of big data analytics, the data-driven intelligent driving system becomes a research focus. Traditional datadriven intelligent driving systems suffer from inadequate data. Due to lacking effective incentives and trust, data from different urban rail operators cannot be shared directly. In this paper, we propose a framework that uses blockchain technology to realize sharing and collaborative training of intelligent driving models between operators. In this framework, we use blockchain-based distributed federated reinforcement learning methods to complete intelligent driving calculations. We use smart contracts to implement the management of the entire federal reinforcement learning. Operators use local historical data to participate in intelligent driving training based on Q-network by exchanging encrypted model parameters, and optimize the safety distance, energy consumption, and punctuality of urban rail transit systems. Simulation results show that our proposed distributed federated reinforcement learning method can significantly improve the intelligent driving system performance.

Target positioning and tracking in wireless sensor networks (WSNs) are necessary for real ap-plications. This paper uses an artificial fish swarm algorithm (AFSA) with receive signal strength indicator (RSSI) for indoor target positioning and tracking. The performance of random and fixed artificial fish deployment topologies in AFSA for target positioning and tracking is investigated when the number of artificial fish is 12, 24, 52, 72, and 100, respectively. The simulation results show that the difference in average error between the random and fixed deployments is 5%, and the average positioning time is 2% for target position. The target tracking performance of fixed deployment is better than that of random one when the number of artificial fish is small.

This paper reviews the literature from the biological and cognitive sciences in sensory integration and derives principles for use in constructing intelligent sensor fusion systems. In particular, it presents psychophysical and neurophysical studies on how sensor fusion is accomplished and cognitive models of associated activities, including optimization of sensing configurations, improvement of sensing quality, and filtering of noise. The sensor fusion effects architecture for robot navigation is also presented as one example of how these insights from the biological and computer science can be applied to robotic sensor fusion. Experimental results demonstrates the utility of the biological and cognitive insights, especially that of fusion modes. Other representative architectures for robotic sensor fusion are contrasted with the biological and cognitive principles.

Many visual quality metrics are used to measure the quality of lossy compressed images and videos, and are integrated in the rate-distortion optimization of hybrid video codecs. However, most of the metrics focus on the average objective quality in a picture. In certain applications, like medical image processing, the maximum absolute error should be more weighted. In this paper, the rate-distortion optimization of HEVC is extended by integrating this error metric. Thus, rate, average error, and maximum absolute error are jointly optimized. Furthermore, a weighting factor α is included into the calculation of the optimization for balancing the ratio between average and maximum absolute error. For HEVC intra with $\alpha =0.25$ an average maximum absolute error reduction of −25.63 can be achieved, while the bitrate increases slightly by 0.59%. Furthermore, the visual quality of the medical volumes improves and the data fidelity increases, i.e. less block artifacts appear and less structure disappear.

We present a method to acquire 3D position measurements for decentralized target tracking with an asynchronous camera network. Cameras with known poses have fields of view with overlapping projections on the ground and 3D volumes above a reference ground plane. The purpose is to track targets in 3D space without constraining motion to a reference ground plane. Cameras exchange line-of-sight vectors and respective time tags asynchronously. From stereoscopy, we obtain the fused 3D measurement at the local frame capture instant. We use local decentralized Kalman information filtering and particle filtering for target state estimation to test our approach with only local estimation. Monte Carlo simulation includes communication losses due to frame processing delays. We measure performance with the average root mean square error of 3D position estimates projected on the image planes of the cameras. We then compare only local estimation to exchanging additional asynchronous communications using the Batch Asynchronous Filter and the Sequential Asynchronous Particle Filter for further fusion of information pairs' estimates and fused 3D position measurements, respectively. Similar performance occurs in spite of the additional communication load relative to our local estimation approach, which exchanges just line-of-sight vectors.

In this talk we present our works on decision support analytic for autonomous systems. Decision support analytic process multiple sensory information collected by an autonomous system such as lidar, camera, RGBD, acoustic to perform signal detection, target tracking, object recognition. As multiple sensors are involved, our system uses sensor registration, data association and fusion to combine sensory information. The next layer of the proposed decision support system orients the processed sensory information at feature and classification levels to perform situation assessment and treat evaluation. Based on the assessment, the decision support system will recommend decision. If the uncertainty is high, actions including resource allocation, planning will be used to extract or reassess the sensory information to get a recommended decision with lower uncertainty. This talk will also presents the applications of the proposed decision support analytic in four industrial projects including 1) goal-driven net-enabled distributed sensing for maritime surveillance, 2) autonomous navigation and perception of humanoid service robots, 3) distance learning for oil and gas drilling and 4) cognitive vehicles.

In this paper, we present a method for updating robotic state belief through contact with uncertain surfaces and apply this update to a Kalman filter for more accurate state estimation. Examining how guard surface uncertainty affects the time spent in each mode, we derive a novel guard saltation matrix- which maps perturbations prior to hybrid events to perturbations after - accounting for additional variation in the resulting state. Additionally, we propose the use of parame-terized reset functions - capturing how unknown parameters change how states are mapped from one mode to the next - the Jacobian of which accounts for additional uncertainty in the resulting state. The accuracy of these mappings is shown by simulating sampled distributions through uncertain transition events and comparing the resulting covariances. Finally, we integrate these additional terms into the “uncertainty aware Salted Kalman Filter”, uaSKF, and show a peak reduction in average estimation error by 24–60% on a variety of test conditions and systems.

Device-Free Localization (DFL) is a promising technique for applications in living environments and disaster sites, as it does not require users to carry any radio devices. Multipath-assisted Radio Tomographic Imaging (multipath-RTI) is one effective method for enabling DFL reducing the number of anchor nodes. This method maps a person’s location in a target area by using numerous wireless links within an radio sensor network. However, improving the localization accuracy in DFL poses a significant challenge, particularly in optimizing the regularization parameters required to solve the inverse problem. As usual, regularization parameters are heuristically selected. In this study, we propose a method for identifying the optimal regularization parameters through simulations using Bayesian optimization, thereby improving localization accuracy. The experimental results show that the proposed method can effectively improve localization accuracy—achieving a 10% reduction in average error—which highlights its applicability in practical DFL systems.

Accurate endoscopic camera pose estimation is crucial for real-time AR navigation systems. While current methods primarily use depth and optical flow, they often ignore structural inconsistencies between images. Leveraging the RAFT framework, we process sequential stereo RGB pairs to extract optical flow and depth features for pose estimation. To address structural inconsistencies, we refine the weights for both 2D and 3D residuals by computing SSIM indices for the left and right views, as well as pre- and post-optical flow transformations. The SSIM metric is also used in the loss function. Experiments on the StereoMIS dataset demonstrate our method's improved pose estimation accuracy compared to rigid SLAM methods, showing a lower accumulated trajectory error (ATE-RMSE: 18.5mm). Additionally, ablation experiments achieved an 11.49% reduction in average error. The pose estimation accuracy has been improved by incorporating SSIM. The code is available at: https://github.com/lianrq/pose-estimation-by-SSIM-Integration.

Environmental perception is a key technology for Advanced Driver Assistance Systems (ADAS) and Autonomous Vehicles, and thus a crucial prerequisite for traffic situation assessment and reasonable decision making. Sensor fusion can make full use of the characteristics of multi-sensors to achieve complementary advantages to improve the target recognition precision in various weather conditions. In this paper, a robust target detection algorithm based on millimeter wave (MMW) radar and camera fusion is proposed. Firstly, the images captured by the camera with low visibility in foggy weather are defogged. Then, the effective target is filtered by MMW radar, and mapped to the camera image to obtain the corresponding regions of interest (ROIs). Finally, the method of weighted is used to combine camera vision network recognition results with the radar target estimation results, and obtain the final result of the ROIs. The simulation results indicates that the detection precision through fusion of MMW radar and camera significantly outperforms that of single sensor.

<para xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> In this paper, we address the problem of target tracking in a collaborative acoustic sensor network. To cope with the inherent characteristics and constraints of wireless sensor networks, we present a novel target-tracking algorithm with power-aware concerns. The underlying tracking methodology is described as a multiple-sensor tracking/fusion technique based on particle filtering. As discussed in the most recent literature, particle filtering is defined as an emerging Monte Carlo state estimation technique with proven superior performance in many target-tracking applications. More specifically, in our proposed method, each activated sensor transmits the received acoustic intensity and the direction of arrival (DOA) of the target to the sensor fusion center (a dedicated computing and storage platform, such as a microserver). The fusion center uses each received DOA to generate a set of estimations based on the state partition technique, as described later in this paper. In addition, a set of sensor weights is calculated based on the acoustic intensity received by each activated sensor. Next, the weighted sum of the estimates is used to generate the proposal distribution in the particle filter for sensor fusion. This technique renders a more accurate proposal distribution and, hence, yields more precise and robust estimations of the target using fewer samples than those of the traditional bootstrap filter. In addition, since the majority of the signal processing efficiently resides on the fusion center, the computation load at the sensor nodes is limited, which is desirable for power-aware systems. Last, the performance of the new tracking algorithm in various tracking scenarios is thoroughly studied and compared with standard tracking methods. As shown in the theory and demonstrated by our experimental results, the state-partition-based centralized particle filter reliably outperforms the traditional method in all experiments. </para>

This paper aims to increase the Unmanned Aerial Vehicle's (UAV) capacity for target tracking. First, a control model based on fuzzy logic is created, which modifies the UAV's flight attitude in response to the target's motion status and changes in the surrounding environment. Then, an edge computing-based target tracking framework is created. By deploying edge devices around the UAV, the calculation of target recognition and position prediction is transferred from the central processing unit to the edge nodes. Finally, the latest Vision Transformer model is adopted for target recognition, the image is divided into uniform blocks, and then the attention mechanism is used to capture the relationship between different blocks to realize real-time image analysis. To anticipate the position, the particle filter algorithm is used with historical data and sensor inputs to produce a high-precision estimate of the target position. The experimental results in different scenes show that the average target capture time of the algorithm based on fuzzy logic control is shortened by 20% compared with the traditional proportional-integral-derivative (PID) method, from 5.2 s of the traditional PID to 4.2 s. The average tracking error is reduced by 15%, from 0.8 m of traditional PID to 0.68 m. Meanwhile, in the case of environmental change and target motion change, this algorithm shows better robustness, and the fluctuation range of tracking error is only half of that of traditional PID. This shows that the fuzzy logic control theory is successfully applied to the UAV target tracking field, which proves the effectiveness of this method in improving the target tracking performance.

In the process of artificial intelligence and intelligent algorithms driving computer technology to move forward in the direction of high precision and strong adaptability, aviation target tracking technology faces the dual challenges of accuracy and stability in complex interference environments. To solve the above problems, an aviation anti-interference target tracking technology based on the fusion of secondary radar information is proposed. The dynamic sliding window is used to detect pulse signals, and the tracking stability is guaranteed by the track maintenance system. The short-range tracking and scanning track data are fused to build a distributed multi-radar fusion system for state estimation, and the three-level Kalman filter is used to solve the time-space alignment problem, and finally the aviation anti-interference target tracking is realized. Experiments show that the research method significantly reduces the absolute error to an average of 12.45, which is more than 56% lower than the comparison method, and greatly improves the tracking success rate to 78% under G8 high-intensity interference, which is 41 percentage points higher than the comparison method. Experiments show that the fusion of secondary radar information effectively enhances the accuracy and stability of target tracking, providing a solid technical support for aviation safety in complex electromagnetic environments.

Target association in sensor data fusion often assumes that both the target tracks and the measurements are described with Gaussian random variables. For some sensor reports, such as passive acoustics, this assumption creates a poor approximation. For the association of measurement to target track, significant errors can occur as a uniform distribution is warped such that the centroid is weighted significantly more than the edges when a single Gaussian is used as the approximation. Using the chi-squared metric to associate a new measurement to an existing target track under this condition can increase the likelihood of association or reduce it significantly. In this paper, a fuzzy-logic approach to data association is enhanced. The original approach showed that it could emulate the chi-squared metric comparing two Gaussian random variables. Here, the approach is enhanced to handle the cases of two uniformly distributed random variables and the case where a uniform measurement is compared to a Gaussian track

Multistatic active target tracking in GPS-denied scenarios is complicated by the fact that the emitter and receiver locations are unknown and must be estimated jointly with the target track. Maximum a posteriori algorithms for solving this joint estimation problem are complicated by the nonlinearities in the likelihood function of the bistatic range measurement. A new integral representation of this likelihood function is presented for small measurement error variances. Remarkably, target state appears linearly in this integral. This paper presents a new approach to the basic problem of target state estimation for known sensor locations. The optimal estimator derived from the integral representation is an iteratively re-weighted linear Kalman smoother. Joint estimators for target and emitter-receiver field tracking will be reported elsewhere.