Ubiquitous Computing and Embedded Architectures

  • Research Area: Ubiquitous Computing and Embedded Architectures 

    The research of ubiquitous computing is aimed at developing a semantic network service platform for aerospace information repositories, and accesses. The research of embedded architectures focuses on the support of computational and communication need of a suite of designated aerospace applications. The tasks of the proposed work are described in the following subsections.


  • Research Activity: Distributed Tuple-space Server for Ubiquitous Computing

  • Figure 1. shows the proposed semantic network service platform. Users are able to generate, integrate, publish, and access information of different formats related to the aerospace community. The technologies relevant to the use of this service platform include: telescope imagery; decentralized control signals for precision pointing of the telescope; aerospace database, sensor network; and visualization for flight dynamics and control, among others. The service of the proposed network service platform is location free by using a distributed tuple-space programming model. With customized interface design, the users may connect to the network service platform utilizing different devices and through different means. The individual participant can search information through the network service platforms. A task of keyword match is performed based on the tuple-space technology. However, the power of the service platform lies in the “metadata” generated through the efforts of individuals on information analysis, integration, and synthesis to facilitate the understanding, characteristics, and management usage of data, and hence, give the “semantics” of the information.


  • Research Activity: Aerospace Information Organizer

  • The contents of the tuple space, i.e., the metadata, are published by the participants. Thus, the success of the server platform will rely on the efficiency of generating such metadata. By simply storing bookmarks of separate web pages, the users need to digest information scattered among the referenced materials, and then, generate another material. The drawbacks are two-fold: 1) the information, even stored in hypertext format, is organized deductively according to the authors’ perspective. The readers must store the entire link in which, perhaps, most of the information is not needed, and 2) the contexts of the material are stored in human-but not machine-recognizable format. These drawbacks have been identified by researchers from MIT (URL: simile.mit.edu/). Such drawbacks make the automation of semantic-based information organization very difficult. In the proposed research, an aerospace information organizer will be implemented. The organizer will be context-aware and will allow the users to perform the following key operations: 1) Searching the entire distributed tuple space for existing metadata; 2) Scrapping of the raw data in different formats, and extracting the selected information; 3) Generating the associated metadata based on a standard format; and 4) Publishing the metadata to the distributed tuple space.

  • Research Activity: Embedded Architectures, Data Processing, and Visualization

  • The proposed research activity in this area focuses on the development of embedded sub-systems for data acquisition, data analysis, and visualization. Such subsystems facilitate the research of the designated applications including decentralized control, combustion/propulsion control, and thermal analysis. On-site data processing requirements, memory buffer constraints, and I/O bandwidth will be investigated to guide the configuration of the embedded sub-systems. Interfaces will be defined for data format conversions, user/machine interactions, and visualization. Since embedded applications typically need customized and special-purpose hardware/software designs, several difficulties arise due to the lack of industrial standards and the essential barrier to merging of techniques. Meanwhile, the techniques employed to support real-time embedded systems span over a wide spectrum of the disciplines of computer areas. To tackle such difficulties, the proposed research focuses on: computational power analysis, performance analysis for cost effectiveness of embedded architectures, experimental study of the performance against the computational models, and interfaces for supporting mobility and ubiquity.