Abstract:Marginal Fisher analysis (MFA) exploits the margin criterion to compact the intraclass data and separate the interclass data, and it is very useful to analyze the high-dimensional data. However, MFA just considers the structure relationships of neighbor points, and it cannot effectively represent the intrinsic structure of hyperspectral imagery (HSI) that possesses many homogenous areas. In this paper, we propose a new dimensionality reduction (DR) method, termed local geometric structure Fisher analysis (LGSFA), for HSI classification. Firstly, LGSFA uses the intraclass neighbor points of each point to compute its reconstruction point. Then, an intrinsic graph and a penalty graph are constructed to reveal the intraclass and interclass properties of hyperspectral data. Finally, the neighbor points and corresponding intraclass reconstruction points are used to enhance the intraclass-manifold compactness and the interclass-manifold separability.LGSFA can effectively reveal the intrinsic manifold structure and obtain the discriminating features of HSI data for classification. Experiments on the Salinas, Indian Pines, and Urban data sets show that the proposed LGSFA algorithm achieves the best classification results than other state-of-the-art methods.
A dynamic model that includes friction and tooth profile error excitation for herringbone gears is proposed for the dynamic analysis of variable speed processes. In this model, the position of the contact line and relative sliding velocity are determined by the angular displacement of the gear pair. The translational and angular displacements are chosen as generalized coordinates to construct the dynamic model. The friction is calculated using a variable friction coefficient. The tooth profile error excitation is assumed to depend on the position along the contact line and to vary with the angular displacement of the driving gear. Thus, the proposed model can be used in the dynamic analysis of the variable speed process of a herringbone gear transmission system. An example acceleration process is numerically simulated using the model proposed in this paper. The dynamics responses are compared with those from the model utilizing a constant friction coefficient and without friction in cases where the profile error excitations are included and ignored.
In the sheet metal assembly process, welding operations join two or more sheet metal parts together. Since sheet metals are subject to dimensional variation resulting from manufacturing randomness, a gap may be generated at each weld pair prior to welding. These gaps are forced to close during the welding operation and accordingly undesirable structural deformation results. Optimizing the welding pattern (the number and locations of weld pairs) in the assembly process was proven to improve significantly the quality of the final assembly. This paper presents a genetic-algorithm-based optimization method to search automatically for the optimal weld pattern so that assembly deformation is minimized. The application result for a real industrial part demonstrated that the proposed algorithm effectively achieved the objective.
The drum driving system is one of the weakest parts of the long-wall shearer, and some methods are also needed to monitor and control the long-wall shearer to adapt to the important trend of unmanned operation in future mining systems. Therefore, it is essential to conduct an electromechanical dynamic analysis for the drum driving system of the long-wall shearer. First, a torsional dynamic model of planetary gears is proposed which is convenient to be connected to the electric motor model for electromechanical dynamic analysis. Next, an electromechanical dynamic model for the drum driving system is constructed including the electric motor, the gear transmission system, and the drum. Then, the electromechanical dynamic characteristics are simulated when the shock loads are acted on the drum driving system. Finally, some advices are proposed for improving the reliability, monitoring the operating state, and choosing the control signals of the long-wall shearer based on the simulation.
Optimizing the locator positions and clamping schemes of a ®xture was proven to improve the dimensional and form accuracy of a workpiece signi®cantly. A number of approaches have been developed to optimize the designs of sheet-metal assembly ®xtures and machining ®xtures. However, in these previous works, the optimal selection of the positions of locators and clamps were based on a stationary set of locator and clamp conditions; i.e. the numbers of the locators and clamps were ®xed during optimization. This paper proposes a genetic algorithm (GA)-based optimization method to select automatically the optimal numbers of locators and clamps as well as their optimal positions in sheet-metal assembly ®xtures, such that the workpiece deformation due to the gravity e ect and resulting variation due to part dimensional variation are simultaneously minimized. The application result of a real industrial part demonstrated that the proposed algorithm e ectively achieves the objective.
Surface quality is a major factor affecting the performance of a component. The machined surface quality is strongly influenced by the external loads during the fixturing and machining processes. In machining process development, it is highly desirable to predict the quality of a machined surface. For this purpose, an integrated finite element analysis (FEA) model of the entire fixture-workpiece system is developed to investigate the influence of clamping preload and machining force on the surface quality of the machined workpiece. The effects of fixture and machine table compliance (from experimental data), and the workpiece and its locators/clamps contact interaction, and forced vibration, on the machined surface quality are taken into account. This simulation model provides a better understanding of the causes of surface error and a more realistic prediction of the machined surface quality. The deck face of a V-type engine block subjected to fixture clamping and a face milling operation is given as an example. A comparison between the simulation result and experimental data shows a reasonable agreement.
This paper considers the underdetermined blind separation of multiple input multiple output (MIMO) radar signals that are insufficiently sparse in both time and frequency domains under noisy conditions, while traditional algorithms are usually applied in the ideal sparse environment. An effective separation method based on single source point (SSP) identification and time-frequency smoothed l 0 norm (TF-SL0) is proposed. Firstly, a preprocessing step of the moving average filter and a novel argument-based time-frequency SSPs detection are employed to improve the signal-to-noise ratio and signal sparsity of the observed signals, respectively. Then, the mixing matrix is obtained by using clustering algorithms. Secondly, to obtain the optimal solution of underdetermined sparse component analysis, the smoothed l 0 norm (SL0) is introduced to preliminarily achieve signal separation in the time-frequency domain. Finally, time-frequency ridge estimation is proposed to jointly enhance the reconstruction accuracy of the MIMO radar signals, and the time domain waveforms are recovered by the model of the signals. Simulations illustrate the validity of the method and show that the proposed method outperforms the traditional methods in source separation, especially in the non-cooperative electromagnetic case where the prior information is unknown.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.