With the increasing complexity of equipment systems, the requirements for testability design are becoming more demanding. This paper proposes using an integer-encoded fault dictionary as the testability model, integra...
详细信息
ISBN:
(数字)9798331544577
ISBN:
(纸本)9798331544584
With the increasing complexity of equipment systems, the requirements for testability design are becoming more demanding. This paper proposes using an integer-encoded fault dictionary as the testability model, integrating fault propagation and test detectability for sensor allocation. The proposed approach is carried out in three main steps. First, an integer-encoded fault dictionary model is developed to represent the relationship between faults and tests. Second, the model’s ability to diagnose faults is evaluated by defining the performance index, which considers fault propagation and the detectability of tests. Third, the Discrete Cuckoo Search (DCS) algorithm is applied to optimize the sensor allocation based on the defined performance index. Experimental results demonstrate that this approach can achieve more accurate testability models under imperfect conditions, thereby enhancing fault diagnosis performance.
The distributed nonconvex optimization problem of minimizing a global cost function formed by a sum of n local cost functions by using local information exchange is *** problem is an important component of many machin...
详细信息
The distributed nonconvex optimization problem of minimizing a global cost function formed by a sum of n local cost functions by using local information exchange is *** problem is an important component of many machine learning techniques with data parallelism,such as deep learning and federated *** propose a distributed primal-dual stochastic gradient descent(SGD)algorithm,suitable for arbitrarily connected communication networks and any smooth(possibly nonconvex)cost *** show that the proposed algorithm achieves the linear speedup convergence rate O(1/(√nT))for general nonconvex cost functions and the linear speedup convergence rate O(1/(nT)) when the global cost function satisfies the Polyak-Lojasiewicz(P-L)condition,where T is the total number of *** also show that the output of the proposed algorithm with constant parameters linearly converges to a neighborhood of a global *** demonstrate through numerical experiments the efficiency of our algorithm in comparison with the baseline centralized SGD and recently proposed distributed SGD algorithms.
3D object detection is crucial for autopilot and augmented reality applications. However, accurate feature extraction is the key to achieve high-precision detection, but traditional methods face the challenges of info...
详细信息
ISBN:
(数字)9798350366600
ISBN:
(纸本)9798350366617
3D object detection is crucial for autopilot and augmented reality applications. However, accurate feature extraction is the key to achieve high-precision detection, but traditional methods face the challenges of information loss and computational inefficiency. To address this, we propose an enhanced feature extraction method. We first divide the 3D space into voxel grids of varying scales. Then, using 3D sparse convolution, we extract features which are compressed along the Z axis to form BEV (Bird’s Eye View) feature images. An efficient multi-scale attention mechanism is then employed to enhance these BEV features. The refined feature map is subsequently located and classified through a CT-stacking module and RPN (Region Proposal Network). Voxel features from the proposal and 3D backbone are passed through a voxel RoI pool to directly extract RoI features from the 3D features for further refinement. Finally, a cascaded attention network refines the proposal and strengthens confidence prediction. Experiments on test set validate our method effectiveness, showing improved accuracy over traditional voxel methods. Overall, this enhanced feature extraction technique offers a promising solution for 3D target detection, particularly in scenarios demanding high real-time performance and precision.
The aim of this study is to explore optimization control methods for suspension systems during the flight of quadcopters. The traditional control method of suspension system has some problems such as oscillation and i...
详细信息
ISBN:
(数字)9798350366600
ISBN:
(纸本)9798350366617
The aim of this study is to explore optimization control methods for suspension systems during the flight of quadcopters. The traditional control method of suspension system has some problems such as oscillation and instability, so an optimization scheme based on Gaussian pseudo-general algorithm is proposed in this paper. Through numerical simulation and actual flight test, the effectiveness of the optimization algorithm in improving system stability and reducing oscillation amplitude is verified. At the same time, this study conducted a comparative analysis of the suspension system of the quadrotor vehicle and the free fall body, and the optimization algorithm is found to be feasible and can well improve the control accuracy and stability of the system, and provide strong support for the safety of the quadrotor vehicle in various application scenarios.
Path planning algorithms are current research hotspots. Heuristic algorithms that can solve dynamic environment problems are gradually becoming the mainstream research direction. The D∗ algorithm, as one of the new he...
详细信息
Aiming at velocity/height tracking control and attitude stabilization for the longitudinal flight of cruise vehicle, the research of intelligent flight control method is carried out by combining the backstepping contr...
Aiming at velocity/height tracking control and attitude stabilization for the longitudinal flight of cruise vehicle, the research of intelligent flight control method is carried out by combining the backstepping control theory and the deep reinforcement learning method. Firstly, the longitudinal flight control system of the vehicle is decomposed into velocity and height subsystems, and the control law is designed based on backstepping control theory. Then, aiming at the problems that the selection of the multiple control gains will directly affect the control performance of the closed-loop system and the adjustment process of control gains is cumbersome, the twin delayed deep deterministic policy gradient algorithm (TD3) will be adopted to train the cruise vehicle, so that it can determine the control gains online according to the current flight states. The numerical simulation results show that, the proposed intelligent flight control method can guarantee the high-precision tracking with respect to height and velocity commands, and also realize the attitude stability of cruise vehicle.
It is essential to achieve real-time fault detection of the industrial process to reduce the occurrence of accidents during the industrial process. However, there are some problems in the actual monitoring process, su...
详细信息
Most data-driven soft sensor methods can model nonlinear time-varying characteristics of biochemical processes. However, the intrinsic relationship between variables, which is helpful for understanding model behavior,...
详细信息
The development of an accurate data-driven fault diagnosis model faces several obstacles. One significant challenge is how to combine the model with the process mechanism rationally and improve the model interpretabil...
详细信息
This paper addresses the problem of multi-robot cooperative localization in non-Gaussian noise environments. Traditionally, the Kalman filter（KF） based algorithms are used as solutions to cooperative localization（C...
详细信息
ISBN:
(数字)9789887581581
ISBN:
(纸本)9798350366907
This paper addresses the problem of multi-robot cooperative localization in non-Gaussian noise environments. Traditionally, the Kalman filter（KF） based algorithms are used as solutions to cooperative localization（CL） problems. However,the KF algorithms cannot obtain accurate state estimation when noises are non-Gaussian. To deal with this issue, an H filter based cooperative localization algorithm is proposed. Unlike the KF algorithm, the H filter algorithm does not require the assumption of Gaussian noise, thus achieving better localization accuracy for non-Gaussian scenarios. The effectiveness of the proposed algorithm is demonstrated through simulation examples.
暂无评论