A data-compression system is presented using a neural-network-based architecture that will accept a whole scene of data, compress it for storage or transmission, retrieve it, and reconstruct the original scene. The architecture of the network consists of a feedforward four-layer network. The network is trained by using a nested training algorithm. The authors (Proc. Int. Joint Conf. on Neural Networks, Washington, DC, Jan. 1990) had presented limited results on the performance of this system. A more detailed performance evaluation of this system is presently given using multiple gray-level inputs and variations of training methods

The authors defines multiple-access method suitable for a centralized multiple satellite network, with onboard processing in the central node. This multiple-access method (CDMA/Aloha/BFN) is a combination of code-division multiple-access (CDMA), random multiple access (RMA) or Aloha, and a set of electronic beam forming networks (BFN). CDMA/Aloha/BFN multiple access maintains the advantages of CDMA and RMA techniques and avoids their disadvantages by using some processing on board the central satellite stations. The major characteristics of this method are: no need for network synchronization resistance to jamming, efficient use of the available system power and bandwidth, low equipment complexity, low mutual interference, ability to operate in systems supporting highly dynamic users, and good delay-throughput performance. The theory of operation of the CDMA/Aloha/BFN multiple-access technique is described and its performance is analyzed.>

This correspondence deals with the application of a bit slice microprocessor to a satellite born packet switch. A system architecture for accomplishing this task is proposed, and the performance of the packet switch is evaluated by obtaining an upper bound on the system throughput.

The design and performance evaluation of a single microprocessor based packet switch for use on-board a satellite are presented. The primary design criterion is to maximize the system throughput. A number of protocols and their impact on the switch design are discussed. System and processor hardware architectures are described. The flow charts of the software required to accomplish the essential functions are presented and modification needed for use in a multisatellite network are also indicated. Maximum throughput attainable is calculated. Desirable characteristics of microprocessors for this application are pointed out. A queue theoretic model of the packet switch has been developed. Analytical relationships of the average waiting times and the queue sizes at the queues and the overall average response time with the various design parameters of the switch have been obtained. A number of graphs showing the effect of variation in the number of input and output lines, data rates, destination fucntions, speed of output lines and some other design parameters on the performance of the switch are presented and discussed.

A class of distribution-free detectors forM-ary pulse amplitude modulated digital signals in the presence of certain signal dependent interference and additive noise of unknown distribution is formulated, evaluated in terms of probability of error and compared to a matched filter detector in Gaussian noise. The noise distribution is unknown except that it satisfies certain broad assumptions and that its value at the median is known. A decision method based on the value of a test statistic is developed. An analytical expression for the average probability of error and a method of calculating the optimum decision thresholds that minimize this probability of error are presented. The probability of error is independent of the form of the noise distribution but depends on the value of the noise density function at the median. An example of a ternary signal set in white Gaussian noise is worked out for illustration. The optimum decision thresholds are calculated and the resulting average probability of error is plotted as a function of signal-to-noise ratio. Average probability of error of a matched filter detector is also evaluated under similar conditions and a comparison is made with that of the proposed detector.