To address the challenges of low accuracy and poor robustness in industrial crayfish weight estimation caused by variable postures, this paper proposes a lightweight method that integrates pose awareness. First, a multi-task perception model, Crayfish-YOLO, is developed based on the YOLOv8s-Seg framework. By reconstructing the backbone with MobileNetV3 and integrating Coordinate Attention (CA), CARAFE upsampling, and the Wise Intersection over Union (Wise-IoU) loss function, the model is significantly compressed while enhancing its ability to output high-fidelity pixel-level masks and pose categories. Second, a pose-adaptive weight estimation strategy is proposed, which leverages perceived pose information to dynamically invoke the optimal regression model from a pre-constructed heterogeneous model library. Using seven core geometric features extracted from the segmentation masks, the system achieves precise weight estimation. Experimental results on a self-built dataset show that Crayfish-YOLO reduces parameters by 75.2% compared to YOLOv8s-Seg, while core segmentation accuracy (mAP(50 similar to 95) (Seg)) improves by 1.1%. The integrated end-to-end system achieves a Mean Absolute Error (MAE) of 2.1 g and a mean coefficient of determination (R-2) of 0.92, significantly outperforming comparative algorithms. This research provides an efficient visual perception and estimation solution for the automated grading of crayfish and similar non-rigid aquatic products.
Ripple currents on the direct current (DC) bus in variable frequency drive (VFD) systems originate from motor load current fluctuations and the high-frequency switching of power devices. The resulting Joule heating within the DC-link capacitors is a primary driver of lifespan degradation. To address the lack of systematic models for multi-phase H-bridge inverters and the over-design caused by empirical methods, this paper proposes a novel analytical method that incorporates the 2k pi/N phase difference of parallel units for precise ripple current quantification. First, a dynamic DC-link capacitor model is established based on a single-phase H-bridge inverter, and the expressions for the instantaneous, average, and root mean square (RMS) input currents are derived. Furthermore, by introducing the 2k pi/N phase difference (where k = 0, 1, & mldr;, N - 1) among N parallel H-bridge units, a universal analytical expression for the RMS input current and its harmonic spectrum in a multi-phase system is obtained. The analysis reveals that ripple current harmonics concentrate at 2m & times; fsw (where m is a positive integer and fsw is switching frequency) and their sidebands (2m & times; fsw +/- fo, fo is output fundamental frequency), and the coupling influence of modulation index and power factor angle on ripple amplitude is quantitatively characterized. A 12 & times; 160 kW twelve-phase H-bridge inverter is taken as a case study, and MATLAB (v2023b) simulations and hardware experiments demonstrate that the theoretical calculations are in close agreement with the simulated and measured results, with the errors of input current harmonic amplitudes all below 5%. Compared with traditional empirical design, the proposed method reduces the capacitor volume and cost by approximately 15-20% while ensuring system reliability. This method is directly extensible to other multi-phase inverter topologies, providing a theoretical foundation for the accurate selection of DC-link capacitors.
This article delves into the exploration of state-space methods applied to the modeling and estimation of systems with time-varying parameters. While typically existing approaches rely on the assumption that the parameters satisfy the Markov evolution and require the knowledge of the transfer matrix, this article develops an explicit autoregressive (AR) model for time-varying parameters in which the invariant matrix represents the dynamic changes in the parameters. Unlike the previous work, the state-space model is constructed by stacking the invariant matrix and time-varying parameters into the unknown state vector. Then, the joint state estimation (JSE) algorithm is deduced based on the Kalman filtering principle, aiming to reduce the dependence on the prior knowledge of the invariant matrix. Through the numerical simulation and Monte Carlo test, it is indicated that the developed algorithm maintains reliability under various random white noises. In addition, the practical estimation results with the real-time series also verify the validity.
This paper considers recursive parameter identification for autoregressive output-error autoregressive moving average (AR-OE-ARMA) systems from the perspective of computational efficiency. By means of the hierarchical identification principle, we propose an auxiliary model hierarchical generalized extended stochastic gradient algorithm (AM-HGESG), an auxiliary model hierarchical multi-innovation generalized extended stochastic gradient (AM-HMI-GESG) algorithm, an auxiliary model hierarchical generalized extended recursive gradient algorithm (AM-HGERG), an auxiliary model hierarchical multi-innovation generalized extended recursive gradient (AM-HMI-GERG) algorithm, an auxiliary model hierarchical generalized extended least squares algorithm (AM-HGELS), and an auxiliary model hierarchical multi-innovation generalized extended least squares (AM-HMI-GELS)algorithm by using the multi-innovation identification theory. The proposed hierarchical identification methods can be extended to other linear and nonlinear multivariable stochastic systems with colored noises.
This article investigates the identification issue of multivariable ARX systems with colored noise. To address the bias caused by colored noise, a data filtering method is applied to whiten the original multivariable system, which filters the input-output data without altering their inherent dynamics and yields a filtered identification model. Considering the computational complexity and burden in multivariable system identification, a three-stage filtered stochastic gradient algorithm is proposed based on the filtered identification model with a hierarchical strategy. In addition, the historical innovations are utilized to further improve estimation accuracy and convergence performance, resulting in a three-stage filtered multi-innovation stochastic gradient algorithm. The numerical examples verify the effectiveness of the proposed algorithms in identifying multivariable ARX systems.
