Search

Graph learning models are critical tools for researchers to explore graph-structured data. To train a capable graph learning model, a conventional method uses sufficient training data to train a graph model on a single device. However, it is prohibitive to do so in real-world scenarios due to privacy concerns. Federated learning provides a feasible solution to address such limitations via introducing various privacy-preserving mechanisms, such as differential privacy on graph edges. Nevertheless, differential privacy in federated graph learning secures the classified information maintained in graphs. It degrades the performances of the graph learning models. In this paper, we investigate how to implement differential privacy on graph edges and observe the performances decreasing in the experiments. We also note that the differential privacy on graph edges introduces noises to perturb graph proximity, which is one of the graph augmentations in graph contrastive learning. Inspired by that, we propose to leverage the advantages of graph contrastive learning to alleviate the performance dropping caused by differential privacy. Extensive experiments are conducted with several representative graph models and widely-used datasets, showing that contrastive learning indeed alleviates the models' performance dropping caused by differential privacy.

The performance of electric vehicles is affected by the shift quality of multi-gear transmission. The realization of dual-target tracking control requires the transmission control unit (TCU) to accurately measure and process the input signals of the gear-shifting control system and precisely control the drive motor torque and the position of shift motors. An electric-vehicle-dedicated TCU was designed to meet the above design requirements. Its function modules included a single-chip control circuit, shift position signal sampling circuit, signal conditioning circuit of the rotational speed and angle, controller area network communication circuit, and shift motor drive circuit. A hardware-in-the-loop simulation test system showed that the TCU design scheme met measurement accuracy requirements and coordinated the actions of the shift actuator and motor control unit to achieve fast and smooth shifting before the road test. The power interruption time of the shifting process was within 350 ms. The reliability of the TCU design was further verified in a 150,000-km vehicle road test.

Hydrogels are known to be soft and wet smart materials that respond to external stimuli. Numerous hydrogel actuators have been developed that perform volume and shape changes, and reversible motions. Heterogeneity, periodicity, and integration of hydrogels with different properties are needed to realize biomimetic motions. In order to achieve specific actuations with predictable pathways, it is critical to fabricate hydrogel structures in a well-controlled manner. This review article summarizes recent progress in representative methods to fabricate hydrogel actuators by different methods, ranging from sequential synthesis, macroscopic supramolecular assembling, field-induced alignment, photolithography, ionoprinting, 3D printing, and gradient structuring. The advantages and limits of these methods are compared and analyzed. Finally, a brief perspective and conclusion are presented to point out some important issues for further studies.

Additional file 2: Figs. S1-S9. Supplementary figures and their corresponding figure legends.

In order to calculate the iron loss resistance of the permanent magnet synchronous motor (PMSM) in real time, this paper presents an identification method of iron loss resistance based on model reference adaptive control (MRAC), and designs the adaptive mechansim by Popov's hyperstability theory. According to the equivalent phase d-q circuit model of PMSM considering the iron loss resistance, the simplified PMSM model in the original Simulink motor library is modified. The MRAC-based identification method is simulated on the modified PMSM system under steady and dynamic conditions. The simulation results show that the estimated value of iron loss resistance can effectively converge to the real value. By adjusting the PI parameters and adding a low-pass filter, the robustness and dynamic characteristics of the identification system are improved.

For urban rail transit station simulation evacuation, domestic and foreign scholars for panic, conformity, reason, gender ratio research has been more mature. However, in the actual evacuation process, the line of sight will be blocked due to the block of buildings, no scholars have studied the influence of the obstruction of sight on the evacuation. In this paper, using Anylogic software, an experimental model is established to compare and analyze the influence of line of sight obstruction on evacuation time, taking a subway station in Southwest China as an example, this paper introduces the factor of line of sight obstruction, when setting up evacuation route, it integrates the evacuation distance and exit congestion degree, establishes the dynamic selection model of pedestrian evacuation path, obtains the evacuation bottleneck of the station, and puts forward the evacuation bottleneck of the station Optimize the scheme and shorten the evacuation time of the station.