Data privacy has become a growing concern with advances in machine learning. Federated learning (FL) is a type of machine learning invented by Google in 2016. In FL, the main aim is to train a high-accuracy global model by aggregating the local models uploaded by participants, and all data in the process are kept locally. However, compromises to security in the cloud server or among participants render this process insufficiently secure. To solve the problem, this article presents an identity-based multireceiver homomorphic proxy re-encryption (IMHPRE) scheme that utilizes homomorphism operations and re-encryption to provide improved encrypted-data processing and access control. When this scheme is employed, participants can directly use public identities for encryption. The IMHPRE scheme is also secure against the chosen-plaintext attacks. Comparison results indicated that the IMHPRE outperforms its counterparts because it allows a cloud server to perform model aggregation on re-encrypted models for multiple receivers.

In this study, an intelligent method is used at the optimal parameters of a time delay based proportional integral (PI) controller for brushless direct current (BLDC) motor speed control system problem. The intelligent method uses the particle swarm optimization (PSO) method to search the optimal proportional gain and the integral gain of the PI control parameter space when the delay time varies. Then, a radial basis function neural network (RBFNN) method is used to obtain an optimal parametric fitting curve from the results of the stochastic inertia weight in PSO (SIWPSO) method. The 3D stability boundaries is shown in graphical form in the control parameter space. The optimal PI control parameters for the speed control are illustrated in a BLDC motor with a digital signal processor (DSP) based control platform. Details of the simulated results and DSP-based experimental results are also presented in this study.

Automotive technology has been recently challenged with the issue of ensuring and improving road safety. Several academic institutions and automobile manufacturers are making efforts to develop technology for automotive safety. This study proposes an object recognition system based on the adaptive boosting algorithm that integrates a laser range finder and a camera. The laser range finder is used to measure the distances of objects in front of the vehicle and then, their central position and length can be calculated. The camera is used to capture sequences of images. The positions of objects in the image are extracted using the corner detection method and optical flow. Coordinate transformation based on curve fitting is used to integrate the distance information from the laser range finder with the information contained in the image. The adaptive boosting algorithm is adopted to recognize objects in front of the vehicle as pedestrians or cars. Finally, the system will trigger an alarm once an obstacle is detected in a region of interest. The experimental results show that the accuracy rate is improved after the combination of the sensors.

As urbanization increases, urban traffic has become increasingly dense and dangerous. Motivated by the problem of traffic congestion, this work develops a safety monitoring systems for use at intelligent intersections that is based on a laser range finder and a camera. An artificial neural network (ANN) is used to execute an algorithm that classifies data from the laser range finder (concerning, for example, vehicles and bikes). To compensate for data loss and obstruction of the line-of-sight of the laser range finder, template matching is applied to classify the objects in the images of the camera. Also, the system transmits the image from the camera, compressed by vector quantization, as well as status in real time to remote control sites through a wireless network and a 3G mobile network for drivers and a control center. Finally, the laser range finder and the camera are integrated in a manner consistent with Bayesian theorem to improve the reliability of classification.

With the rapid growth in the volume, complexity, and heterogeneity of information in the World Wide Web (WWW), the role of user-data interaction paradigms is becoming increasingly critical to the success of web-based information retrieval and assimilation. Currently, the paucity of mature paradigms in this problem area contrasts sharply with the advances in design of search techniques that allow indexing large volumes of information and efficiently executing keyword-based search. State of the art research in mediating user-data interactions over large information repositories such as the WWW has seen the proposition of techniques such as page categorization, page summarization, content-based page clustering, as well as algorithms for recognizing semantic correlations between web-pages having heterogeneous content and supporting experiential and unified interactions across them. In this context, a key challenge involves studying and analyzing user behaviors and usage patterns in such information interaction frameworks. In this paper, we present a study that investigates this issue by analyzing both quantitatively and qualitatively how users interact, query, explore, and assimilate information in such an environment. We also investigate how efficacious such interaction paradigms are as compared to the standard approach of presenting results by ordering them in terms of page rank. Results from our investigation provide important insights into user behavior in heterogeneous information organization and interaction frameworks and will be valuable in further development of information presentation, querying, and interaction techniques.

As the complexity of the information available on the Web increases, the role of user-data interaction paradigms is becoming increasingly critical for the success of Web information retrieval. Recent years have witnessed significant advances in techniques for indexing and querying Web data. However, in the same period, limited advancements have been made in developing paradigms and researching algorithmic issues associated with the design of interfaces for Web-search. In this paper, we propose a novel paradigm for enabling multiple-perspective query and interaction in Web search. Underlying the proposed metaphor are information and pattern analysis techniques that help determine semantic correlations between Web pages, identify and extract information critical for intuitive understanding and hypothesis generation, and support effective and multiple-perspective interactions between users and the data. We provide a comprehensive study on the effectiveness and efficiency of the proposed approach in query-retrieval scenarios involving complex information goals. Our investigations point to the importance of developing novel ways to mediate interactions during Web-search and are useful in the development of the next generation of real-world solutions for Web information retrieval

This paper describes a novel approach to the presentation of results from Web-search by utilizing the semantic correlations amongst the retrieved results as well as their spatio-temporal characteristics. The fundamental goals of this research are threefold; first, it seeks to facilitate easier access to information than what is provided by the current paradigm of listing the retrieved hits in order of relevance. Second, and more importantly, it attempts to recognize, retain, and present context within the retrieved data. Such support is essential for assimilation of the information obtained as the consequence of a query. Finally, it seeks to support both direct search behavior as well as more exploratory search strategies. In this paper, we describe the algorithmic underpinnings, based on latent semantic analysis, of our approach as well as the strategy for spatio-temporal information display and querying. Examples involving different types of queries and comparative experiments illustrate the idea and verify its efficacy