The paper presents memristor crossbar architectures for implementing layers in deep neural networks, including the fully connected layer, the convolutional layer, and the pooling layer. The crossbars achieve positive and negative weight values and approximately realize various nonlinear activation functions. Then the layers constructed by the crossbars are adopted to build the memristor-based multi-layer neural network (MMNN) and the memristor-based convolutional neural network (MCNN). Two kinds of in-situ weight update schemes, which are the fixed-voltage update and the approximately linear update, respectively, are used to train the networks. Consider variations resulted from the inherent characteristics of memristors and the errors of programming voltages, the robustness of MMNN and MCNN to these variations is analyzed. The simulation results on standard datasets show that deep neural networks (DNNs) built by the memristor crossbars work satisfactorily in pattern recognition tasks and have certain robustness to memristor variations.

The purpose of this paper is to accurately establish a model of a special vehicle driven by in-wheel motors (IWMs) and investigate its dynamic characteristics. This study proposes a novel integrated vehicle modeling strategy, focusing on the IWM and other key subsystems. Based on this strategy, the vehicle model is developed and simulated in ADAMS. The natural frequencies of the vehicle and transmission characteristics of key components are compared with MATLAB simulation results to validate the accuracy of the model. In the time-domain analysis, the vehicle system’s time-domain characteristics are obtained by using random road spectra of different road grades as excitations. Our simulation results demonstrate that vibration input can be reduced by between 97% and 99% after multi-stage vibration reduction. This work will provide relevant parameters for the design of special vehicles driven by IWMs.

The performance of ultra-precision positioning stages is particularly critical in the field of ultra-precision manufacturing. This paper presents the design of an eight-actuator redundant six-degree-of-freedom micropositioning stage with magnetic suspension gravity compensation (6-DOF RDMS). All drives are electromagnetic direct-drive systems based on voice coil motors (VCMs). The vertical motors incorporate a Halbach array and a magnetic suspension gravity compensation structure. A mathematical model and a finite element model are established for the designed positioning stage. The models are utilized for magnetic field and magnetic force analysis, thermal analysis, and analysis of the dynamic force-induced deformation of stator to ensure operational stability. A decoupling control system for the six-degree-of-freedom micropositioning stage is developed, mapping the six degrees of freedom to the control loops of the eight actuators to achieve high-precision positioning. Finally, a prototype was constructed, validating the effectiveness and feasibility of the structural design and the decoupling control system.