Distributed decision fusion using empirical estimation

The problem of optimal data fusion in multiple detection systems is studied in the case where training examples are available, but no a priori information is available about the probability distributions of errors committed by the individual detectors. Earlier solutions to this problem require some knowledge of the error distributions of the detectors, for example, either in a parametric form or in a closed analytical form. Here we show that, given a sufficiently large training sample, an optimal fusion rule can be implemented with an arbitrary level of confidence. We first consider the classical cases of Bayesian rule and Neyman-Pearson test for a system of independent detectors. Then we show a general result that any test function with a suitable Lipschitz property can be implemented with arbitrary precision, based on a training sample whose size is a function of the Lipschitz constant, number of parameters, and empirical measures. The general case subsumes the cases of nonindependent and correlated detectors.     

Finite-dimensional filters with nonlinear drift. III: Duncan-Mortensen-Zakai equation with arbitrary initial condition for the linear filtering system and the Benes filtering system

We consider the Duncan-Mortensen-Zakai (DMZ) equation for the Kalman-Bucy filtering system and Benes filtering system. We show that this equation can be solved explicitly with an arbitrary initial condition by solving a system of ordinary differential equations and a Kolmogorov-type equation, Let n be the dimension of state space. We show that we need only n sufficient statistics in order to solve the DMZ equation.     

Finite-dimensional filters with nonlinear drift. V: solution to Kolmogorov equation arising from linear filtering with non-Gaussian initial condition

Despite its usefulness, the Kalman-Bucy filter is not perfect. One of its weaknesses is that it needs a Gaussian assumption on the initial data. Recently Yau and Yau introduced a new direct method to solve the estimation problem for linear filtering with non-Gaussian initial data. They factored the problem into two parts: (1) the on-line solution of a finite system of ordinary differential equations (ODEs), and (2) the off-line calculation of the Kolmogorov equation. Here we derive an explicit closed-form solution of the Kolmogorov equation. We also give some properties and conduct a numerical study of the solution.     

FUZZY PETRI NET FOR FAULT DIAGNOSIS

ＦＵＺＺＹＰＥＴＲＩＮＥＴＦＯＲＦＡＵＬＴＤＩＡＧＮＯＳＩＳＷｕｙａｎｆａｎｇ；ＷｅｉＺｈｏｎｇｘｉｎ．（ＤｅｐａｒｔｍｅｎｔｏｆＭｅｃｈａｎｉｃａｌＥｎｇｉｎｅｅｒｉｎｇ，ＮａｎｊｉｎｇＵｎｉｖｅｒｓｉｔｙｏｆＡｅｒｏｎａｕｔｉｃｓａｎｄＡｓｔｒｏ...     

Inside and outside of supernovae

A newly formed neutron star in a supernova finds itself in a dense environment, in which the gravitational energy of accreting matter can be lost to neutrinos. For the conditions in SN 1987A, 0.1M may have fallen back onto the central neutron star on a timescale of hours after the explosion, after which the accretion rate is expected to drop sharply. Radiation is trapped in the flow until the mass accretion rate drops to 2×10^–4 M yr^–1 at which point radiation can begin to escape from the shocked envelope at an Eddington limit luminosity. Between this neutrino limit and the Eddington limit, 3×10^–8 M yr^–1, there are no steady, spherical solutions for neutron star accretion. SN 1987A should have reached the neutrino limit within a year of the explosion; the current lack of an Eddington luminosity can be attributed to black hole formation or to a clearing of the neutron star envelope. There is no evidence for newly formed neutron stars in supernovae. Radio supernovae, which were initially interpreted as pulsar activity, probably involve circumstellar interaction; SN 1993J shows especially good evidence for outer shock phenomena.     

The effect of realistic surface properties on low temperature space observatories

We highlight the effect on space-telescope temperatures of thedirectionality of the radiative properties of materials, by showing results from a Monte-Carlo simulation of telescope cooling. The need for further measurements of directional properties is stressed.     

IMF connection for energetic protons observed at Ulysses via mid-latitude solar wind rarefaction regions

As Ulysses moved inward and southward from mid-1992 to early 1994 we noticed the occasional occurrence of inter-events, lasting about 10 days and falling between the recurrent events, observed at proton energies of 0.48–97 MeV, associated with Corotating Interaction Regions (CIR). These inter-events were present for several sequences of two or more solar rotations at intensity levels around 1% of those of the neighbouring main events. When we compared the Ulysses events with those measured on IMP-8 at 1 AU we saw that the inter-events appeared at Ulysses after the extended emission (>10 days) of large fluxes of solar protons of the same energy that lasted at least one solar rotation at 1 AU. The inter-events fell completely within the rarefaction regions (dv/dt<0) of the recurrent solar wind streams. The interplanetary magnetic field (IMF) lines in the rarefactions map back to the narrow range of longitudes at the Sun which mark the eastern edge of the source region of the high speed stream. Thus the inter-events are propagating at mid-latitudes to Ulysses along field lines free from stream-stream interactions. They are seen in the 0.39–1.28 MeV/nucleon He, which exhibit a faster decay, but almost never in the 38–53 keV electrons. We show that the inter-events are unlikely to be accelerated by reverse shocks associated with the CIRs and that they are more likely to be accelerated by sequences of solar events and transported along the IMF in the rarefactions of the solar wind streams.     

Pioneer and Voyager observations of large-scale spatial and temporal variations in the solar wind

The Pioneer 10, Pioneer 11, and Voyager 2 spacecraft were launched in 1972, 1974, and 1977, respectively. While these three spacecraft are all at compartively low heliographic latitudes compared with Ulysses, their observation span almost two solar cycles, a range of heliocentric distances from 1 to 57 AU, and provide a unique insight into the long-term variability of the global structure of the solar wind. We examine the spatial and temporal variation of average solar wind parameters and fluxes. Our obsevations suggest that the global structure of the outer heliosphere during the declining phase of the solar cycle at heliographic latitudes up to 17.5°N was charaterized by two competing phenomena: 1) a large-scale increase of solar wind density, temperature, mass flux, dynamic pressure, kinetic energy flux, and thermal enery flux with heliographic latitude, similar to the large-scale latitudinal gradient of velocity seen in IPS observations, 2) a small-scale decrease in velocity and temperature, and increase in density near the heliospheric current sheet, which is associated with a band of low speed, low temperature, and high density solar wind similar to that observed in the inner heliosphere.     

Reconnection on open field lines ahead of coronal mass ejections

Plasma and magnetic field signatures from 29 November 1990 indicate that the Ulysses spacecraft passed through a series of interplanetary structures that were most likely formed by magnetic reconnection on open field lines ahead of a coronal mass ejection (CME). This reconnection changed the magnetic topology of the upstream region by converting normal open interplanetary magnetic field into a pair of regions: one magnetically disconnected from the Sun and the other, a tongue, connected back to the Sun at both ends. This process provides a new method for producing both heat flux dropouts and counterstreaming suprathermal electron signatures in interplanetary space. In this paper we expand upon the 29 November case study and argue that reconnection ahead of CMEs should be less common at high heliolatitudes.     

Three-dimensional cosmic-ray simulations: Heliographic latitude and current-sheet tilt

We present the results from a study of the variations of the cosmic-ray intensity with time, heliographic latitude, and longitude, and for varying interplanetary conditions, using our three-dimensional, time-dependent computer code for cosmic-ray transport in the heliosphere. Our code also produces a solar-wind and interplanetary magnetic field (IMF) configuration which is compared with observations. Because of the fully threedimensional nature of the model calculations, we are able to model time variations which would be expected to be observed along Ulysses's trajectory as it moves to high latitudes. In particular we can model the approximately 13-and 26-day solar-rotation induced variations in cosmic rays, solar wind and IMF, as a function of increasing heliographic latitude, as one moves poleward of the interplanetary current sheet. Our preliminary model results seem to be in general form quite similar to published data, but depend on the physical parameters used such as cosmic-ray diffusion coefficients, boundary conditions, and the nature of the solar wind and IMF and current sheet.