Convergence of filters with applications to the Kalman-Bucy case

For each N, and each fixed time T, a signal X^N and a `noisy' observation Y^N are defined by a pair of stochastic difference equations. Under certain conditions (X^N, Y^N) converges in distribution to (X, Y, where dX(t)= f(t, X(t))dt+dV( t), dY(t)=g(t, X( t))dt+dW(t). Conditions are found under which convergence in distribution of the conditional expectations E{F(X^N)|Y^N} to E{F(X)|Y} follows, for every bounded continuous function F. The case in which the conditional expectations still converge but the limit is not E{ F(X)|Y} is also studied. In the situation where f and g are linear functions of X, an examination of this limit leads to a Kalman-Bucy-type estimate of X ^N which is asymptotically optimal and is an improvement on the usual Kalman-Bucy estimate     

The Zak transform and some counterexamples in time-frequencyanalysis

It is shown how the Zak transform can be used to find nontrivial examples of functions f, g∈L²(R) with f×g≡0≡F×G, where F, G are the Fourier transforms of f, g, respectively. This is then used to exhibit a nontrivial pair of functions h, k∈L²(R), h≠k, such that |h|=|k|, |H |=|K|. A similar construction is used to find an abundance of nontrivial pairs of functions h, k∈L² (R), h≠k, with |A_h |=|A_k| or with |W_h|=|W _k| where A_h, A_k and W_h, W_k are the ambiguity functions and Wigner distributions of h, k, respectively. One of the examples of a pair of h, k∈L²(R), h≠k , with |A_h|=|A_k| is F.A. Grunbaum's (1981) example. In addition, nontrivial examples of functions g and signals f₁≠f₂ such that f₁ and f₂ have the same spectrogram when using g as window have been found     

Balancing sets of vectors

For n>0, d⩾0, n≡d (mod 2), let K(n, d) denote the minimal cardinality of a family V of ±1 vectors of dimension n, such that for any ±1 vector w of dimension n there is a v∈V such that |v- w|⩽d, where v-w is the usual scalar product of v and w. A generalization of a simple construction due to D.E. Knuth (1986) shows that K(n , d)⩽[n/(d+1)]. A linear algebra proof is given here that this construction is optimal, so that K(n, d)-[n/(d+1)] for all n≡d (mod 2). This construction and its extensions have applications to communication theory, especially to the construction of signal sets for optical data links     

Optimal probability-of-error thresholds and performance for twoversions of the sign detector

Expressions are obtained for specifying the optimal error probability (minimum P_e) thresholds λ₀₁ and λ₀₂ for the traditional and modified sign detectors, respectively. These thresholds are shown to depend on the parameters p, P₁, and M where: M is the number of observations z_i used in the test statistic; P₁=P(H₁ ) is the prior probability for hypothesis H₁ that signal s₁ is present and 1-P₁ =P(H₀) corresponds to the hypothesis H₀ that signal s₀ is present; and p=Pr{z_i⩾0|H₁} with s₀=0 for the traditional sign detector and p=Pr{z_i⩾λ|H₁ }=Pr{z_i<λ|H₀} with λ =(s₀+s₁)/2 for the modified sign detector. The expressions for λ₀₁ and λ₀₂, are given explicitly, and shown to be independent of P₁ for sufficiently large M. Optimal P_e versus M performance curves, corresponding to both versions of the sign detector, are obtained for a representative range of values for p and P₁     

Limiting efficiencies of burst-correcting array codes

The author evaluates the limiting efficiencies e(-S ) of burst-correcting array codes A(n₁,n₂, -s) for all negative readouts -s as n₂ tends to infinity and n₁ is properly chosen to maximize the efficiency. Specializing the result to the products of the first i primes donated by s_i (1⩽i<∞), which are optimal choices for readouts, gives the expression e(-s_i)=(2p_i+1 -2)/(2p_i+1-1) where p_i+1 is the next prime. Previously, it was known only that e(-2)⩾4/5 and e(-1)⩾2/3. This result reveals the existence of burst-correcting array codes with efficiencies arbitrarily close to 1 and with rates also arbitrarily close to 1     

Correlation between 1/f noise and hFElong-term instability in silicon bipolar devices

Accelerated life tests with high-temperature storage and electric aging for n⁺-p-n silicon planar transistors were carried out. Current gain h_FE increases monotonously with time during the tests, and the h_FE drift is correlated with initial measured 1/f noise in the transistors, i.e. the drift amount significantly increases with the increase of noise level. The correlation coefficient of relative drift Δh_FE /h_FE and 1/f noise spectral density S_iB(f) is far larger than that of Δ h_FE/h_FE and initial DC parameters of the transistors. A quantitative theory model for the h _FE drift has been developed and explains the h _FE drift behavior in the tests, which suggests that the h _FE drift and 1/f noise can be attributed to the same physical origin, and both are caused by the modulation of the oxide traps near the Si-SiO₂ interface to Si surface recombination. 1/f noise measurement, therefore, may be used as a fast and nondestructive means to predict the long-term instability in bipolar transistors     

On the Hamming distance properties of group codes

Under certain mild conditions, the minimum Hamming distance D of an (N, K, D) group code C over a non-abelian group G is bounded by D⩽N -2K+2 if K⩽N/2, and is equal to 1 if K>N/2. Consequently, there exists no (N, K, N-K+1) group code C over an non-abelian group G if 1<K<N. Moreover, any normal code C with a non-abelian output space has minimum Hamming distance equal to D=1. These results follow from the fact that non-abelian groups have nontrivial commutator subgroups. Finally, if C is an (N, K, D) group code over an abelian group G that is not elementary abelian, then there exists an (N, K, D) group code over a smaller elementary abelian group G'. Thus, a group code over a general group G cannot have better parameters than a conventional linear code over a field of the same size as G     

Evidence for continuous visible chemical lasing from the fast nearresonant energy transfer pumping of atomic sodium

Energy transfer from selectively formed metastable states of SiO is used to pump sodium-atom laser amplifiers at λ≈569 nm (4d²D-3p²P), λ≈616 nm (5s²S-3p² P), and λ≈819 nm (3d²D -3p²P). The a³Σ⁺ and b³Π states of SiO are generated in high yield from the Si+N₂O→SiO+N₂ reaction. The energy stored in the triplet states is transferred in a highly efficient collisional process to pump sodium atoms to their lowest excited 3d² D, 4d²D, and 5s²S states. Adopting a sequence in which high concentrations of silicon and sodium atoms are mixed and oxidized, a continuous amplification (gain condition) is monitored which suggests the creation of a population inversion among the receptor sodium-atom energy levels and forms the basis for full cavity oscillation on the Na4 d²D-3p²P transition at 569 nm     

Calculation of the binomial survivor function [reliabilityapplications]

A method is presented for calculating the binomial SF (cumulative binomial distribution), binfc(k;p,n), especially for a large n, beyond the range of existing tables, where conventional computer programs fail because of underflow and overflow, and Gaussian or Poisson approximations yield insufficient accuracy for the purpose at hand. This method is used to calculate and sum the individual binomial terms while using multiplication factors to avoid underflow; the factors are then divided out of the partial sum whenever it has the potential to overflow. A computer program uses this technique to calculate the binomial SF for arbitrary inputs of k, p, and n. Two other algorithms are presented to determine the value of p needed to yield a specified SF for given values of k and n and calculate the value where p=SF for a given k and n. Reliability applications of each algorithm/program are given, e.g. the value of p needed to achieve a stated k-out-of-n :G system reliability and the value of p for which k -out-of-n:G system reliability equals p     

A model for fast switching transients in power systems: the nearzone concept

The author presents a simple time-domain model which makes it possible to predict the order of magnitude of the highest di/ dt values generated by closing switches in electrical power systems. The model is based on traveling-wave analysis. It is demonstrated that two different approaches must be applied, according to whether (a) the closing time, T_s, of the switch is faster than twice the traveling time to the first reflection point or (b) T_s is much slower. Under condition (b) the well-known quasistationary approach di/dt_max=U₀/L can be used, where U₀ is the switched voltage and L is the self-inductance of the line between the stray capacitances located to the left and the right of the switching device. Under condition (a) a new formula must be applied: di/dt _max≈2 U₀/ZT_s, where Z is the line impedance of the line in which the switching device is installed and T_s is the time during which the voltage across the switch collapses from U₀ to zero. Experimental results are given from both fast and slow closing switches     

Asymptotic performance of M-ary orthogonal modulation ingeneralized fading channels

The asymptotic (M→∞) probability of symbol error P_e,_m for M-ary orthogonal modulation in a Nakagami-m fading channel is given by the incomplete gamma function P(m, mx) where x=In 2/(E_b/N₀) and E_b is the average energy per bit. For large signal-to-noise ratio this leads to a channel where the probability of symbol error varies as the inverse mth power of E_b/N₀. These channels exist for all m⩾1/2. The special case of m=1 corresponds to Rayleigh fading, an inverse linear channel     

On the competitive optimality of Huffman codes

Let X be a discrete random variable drawn according to a probability mass function p(x), and suppose p(x), is dyadic, i.e., log(1/p(x)) is an integer for each x. It is shown that the binary code length assignment l(x)=log(1/p(x)) dominates any other uniquely decodable assignment l'(x) in expected length in the sense that El(X)<El'(X), indicating optimality in long run performance (which is well known), and competitively dominates l'(x), in the sense that Pr{ l (X)<l'(X)}>Pr{l ( X)>l'(X)}, which indicates l is also optimal in the short run. In general, if p is not dyadic then l=[log 1/p] dominates l'+1 in expected length and competitivity dominates l'+1, where l' is any other uniquely decodable code     

Index system and separability of constant weight Gray codes

A number system is developed for the conversion of natural numbers to the codewords of the Gray code G(n,k) of length n and weight k, and vice versa. The focus is on the subcode G(n,k) of G(n) consisting of those words of G(n) with precisely k 1-bits, 0<k<n. This code is called the constant weight Gray code of length n and weight k. As an application sharp lower and upper bounds are derived for the value of |i-j|, where i and j are indices of codewords g_i and g_j of G(n,k) such that they differ in precisely 2 m bits     

Error-correcting codes for list decoding

In the list-of-L decoding of a block code the receiver of a noisy sequence lists L possible transmitted messages, and is in error only if the correct message is not on the list. Consideration is given to (n,e,L) codes, which correct all sets of e or fewer errors in a block of n bits under list-of-L decoding. New geometric relations between the number of errors corrected under list-of-1 decoding and the (larger) number corrected under list-of-L decoding of the same code lead to new lower bounds on the maximum rate of (n,e,L) codes. They show that a jammer who can change a fixed fraction p<1/2 of the bits in an n-bit linear block code cannot prevent reliable communication at a positive rate using list-of- L decoding for sufficiently large n and an L⩽n. The new bounds are stronger for small n , but weaker for fixed e/n in the limit of large n and L than known random coding bounds     

On the influence of coding on the mean time to failure fordegrading memories with defects

The application of a combined test-error-correcting procedure is studied to improve the mean time to failure (MTTF) for degrading memory systems with defects. The degradation is characterized by the probability p that within a unit of time a memory cell changes from the operational state to the permanent defect state. Bounds are given on the MTTF and it is shown that, for memories with N words of k information bits, coding gives an improvement in MTTF proportional to (k/n) N(d_min^-2)/(d_min ^-1), where d_min and (k/n) are the minimum distance and the efficiency of the code used, respectively. Thus the time gain for a simple minimum-distance-3 is proportional to N^-1. A memory word test is combined with a simple defect-matching code. This yields reliable operation with one defect in a word of length k+2 at a code efficiency k/(k+2)     

Secret key agreement by public discussion from common information

The problem of generating a shared secret key S by two parties knowing dependent random variables X and Y, respectively, but not sharing a secret key initially, is considered. An enemy who knows the random variable Z, jointly distributed with X and Y according to some probability distribution P_XYZ, can also receive all messages exchanged by the two parties over a public channel. The goal of a protocol is that the enemy obtains at most a negligible amount of information about S. Upper bounds on H(S) as a function of P_XYZ are presented. Lower bounds on the rate H (S)/N (as N→∞) are derived for the case in which X=[X₁, . . ., X _N], Y=[Y₁, . . ., Y_N] and Z=[Z₁, . . ., Z_N] result from N independent executions of a random experiment generating X_i, Y_i and Z_i for i=1, . . ., N. It is shown that such a secret key agreement is possible for a scenario in which all three parties receive the output of a binary symmetric source over independent binary symmetric channels, even when the enemy's channel is superior to the other two channels     

Analysis of a modified Hebbian rule

A modified Hebbian rule using the matrix G=sgn(X ^TX) to induce a certain mapping is discussed. This mapping g is specified as soon as one has chosen the m×n matrix X over U to construct G by use of the above expression. The analysis of g relies on simple counting arguments and on the use of Stirling's approximation to obtain asymptotic results     

Electric fields and vector potentials of thin cylindrical antennas

The vector potential and elastic field generated by the current in a center-driven or parasitic dipole antenna that extends from z=-h to z=h are investigated for each of the several components of the current. These include sin k (h-|z|), sin k|z|-sin kh , cos kz-cos kh, and cos kz/2-cos kh /2. Of special interest are the interactions among the variously spaced elements in parallel, nonstaggered arrays. These depend on the mutual vector potentials. It is shown that at a radial distance ρ~ h and in the range -h⩽z⩽h, the vector potentials due to all four components become alike and have an approximately plane-wave form. Simple approximate formulas for the electric fields and vector potentials generated by each of the four distributions are derived and compared with the exact results. The application of the new formulas to large arrays is discussed     

On the number of memories that can be perfectly stored in a neuralnet with Hebb weights

Let {w_ij} be the weights of the connections of a neural network with n nodes, calculated from m data vectors v¹, ···, v^m in {1,-1}ⁿ, according to the Hebb rule. The author proves that if m is not too large relative to n and the v^k are random, then the w_ij constitute, with high probability, a perfect representation of the v^k in the sense that the v ^k are completely determined by the w_ij up to their sign. The conditions under which this is established turn out to be less restrictive than those under which it has been shown that the v^k can actually be recovered by letting the network evolve until equilibrium is attained. In the specific case where the entries of the v^k are independent and equal to 1 or -1 with probability 1/2, the condition on m is that m should not exceed n/0.7 log n