No abstract

An automotive suspension determines both the driving stability and comfort of the vehicle occupants. This paper establishes two kinds of two degrees of freedom for the automobile suspension vibration model and uses the PID controller to establish an automobile suspension adaptive open-loop and closed-loop control system. Respectively, by step interference, white noise and sinusoidal interference for the input, studying the vibration characteristics of the vibration model in the vertical direction. By numerical simulation, we obtain the suspension of the vertical displacement and acceleration-time graphs. The simulation results show that the vibration characteristics of the first model are more in accordance with the actual situation of the car, and the closed-loop control is better than the open-loop control. The adaptive closed-loop control system can reduce the output displacement of automobile suspension to around 1% of the interference road input displacement. The output acceleration value is small, and the acceleration changes smoothly. The results verify the rationality and validity of the automobile suspension model and adaptive control system, which provides a theoretical foundation for the design and optimization of the automobile suspension system.

No abstract

The fault diagnosis method based on ResNet50 convolutional neural network and migration learning is proposed as a way of improving the gearbox fault diagnosis. Firstly, exporting the normal and faulty data of the gearbox. Then converting the exported data into one-dimensional images to and generate the corresponding training and test sets, train in the model to get accuracy of the test set and training set. After fine adjustment, it is used for gearbox fault diagnosis, compared with the VGG16, ResNet101, and GoogleNet model, the accuracy of ResNet50 is above 86.6%. It has a good prospect of application, and its effect is obviously better than that of other models.

Cross-domain recommender systems are known to provide solutions to the cold start and data sparsity problems in recommender systems. This can be achieved by leveraging sufficient ratings and users' profiles in one domain to enhance accurate recommendations in another domain. However, domains with sufficient ratings are not willing to share their users' ratings with other recommender systems or domains due to users' privacy and legal concern. Hence this shows a need for a privacy-preserving mechanism that encourages secure knowledge transfer between different domains. This study proposes a privacy-preserving cross-domain recommender system based on matrix factorization. Specifically, the study formally described the privacy requirements of a cross-domain recommender system, which are different from a single domain recommender system. It designs a new framework for a privacy-preserving cross-domain recommender system and then utilized the somewhat homomorphic encryption (SWHE) scheme to ensure users' privacy. The SWHE scheme was used to encrypt users' ratings in different domains, shared latent factor approach was implemented between the domains and extracted knowledge was securely transferred from the source domain to the target domain. We prove that users' privacy is secured throughout the stages involved in the proposed protocol. Experiments on both synthetic and real datasets demonstrate the efficiency of our protocol.