This paper introduces an overview of throughput and delay analyses of IEEE 802.11 networks, mainly from authors' researches. In our opinion, the difficulty of wireless multi-hop network analysis comes from the network-dynamics multiplicity between uniformity and locality in spatiotemporal fields. From the background, the concept of “bottom-up analysis” was advocated, which is Network-layer modeling over MAC-layer one with respect to each node. Generalities of analytical expressions increase by applying the bottom-up analysis concept. Additionally, it is possible to express end-to-end delays by including the queuing theory in the bottom-up analysis concept. The validity of the obtained model is confirmed by comparisons with simulation results.
Signal filtering is necessary for wireless communication. However it causes the signal amplitude to fluctuate and affects the performance of stochastic resonance (SR) receivers. In this study, we evaluate the bit error rate (BER) performance of filtered binary phase-shift keying(BPSK) on an SR receiver. The results show that filtering improves the BER performance of the SR receiver because the amplitude fluctuation contributes to improving the SR effect. We also evaluate the effect of the roll-off factor, which determines the bandwidth of the filter and the amplitude fluctuation. The results demonstrate the applicability of the SR receiver to bandlimited BPSK signals.
Vector Stream Cipher (VSC) is a stream cipher which consists of permutation polynomial over a ring of modulo 2w. The algorithm for generating key stream is very simple and the encryption is very fast. Some theoretical attacks for VSC have been reported so far since the invention of VSC in 2004. Then, the authors proposed some improvements and developed “Vector Stream Cipher 2.0 (VSC 2.0)” to be immune against the theoretical attacks. In this paper, we propose further improvement of VSC 2.0 to publish as a new chaos cipher “Vector Stream Cipher 2.1 (VSC2.1)”. VSC 2.1 is faster and more secure than VSC 2.0. Our result suggests that permutation polynomials over a ring of modulo 2w are useful for cryptography.
Post-processing scheme using linear feedback shift registers (LFSRs) for generating chaos-based random bit sequences with infinite period for use in Monte-Carlo methods is discussed. We theoretically analyze auto-correlation functions of choatic binary sequences generated by the Bernoulli chaotic map and a class of binary funcitons. The theoretical analyses can be applied to the LFSR-based post-processing for generating random bit sequences with prescribed auto-correlations because LFSRs can be regarded as a finite bit approximation of the Bernoulli map. Some results of numerical experiments are also given.
In this paper, we propose a Bayesian branch-prediction circuit, consisting of an instruction-feature extractor and a naïve Bayes classifier (NBC), as a machine learning approach for branch prediction. A branch predictor predicts the outcome of a branch instruction by analyzing the pattern of the previous branch outcome. In other words, branch prediction can be viewed as a type of pattern recognition problem, and such problems are often solved using neural networks. A perceptron branch predictor has already been proposed as one example of a neural branch prediction architecture, which predicts the next branch outcome by using past branch history to form feature vectors. The proposed circuit is constructed by replacing the arithmetic unit (neurons) in conventional neural branch predictors with an NBC. By introducing an approximate Bayesian computation and its parallel architectures, the NBC circuit completes branch prediction within two clock cycles per instruction. This constitutes a suitable replacement for conventional branch predictors in modern pipelined reduced instruction set computing microprocessors.
The Stirling engine, originally development in 1816 by R. Stirling, has been considered as one of the most efficient systems that convert energy resource to electric power using a thermodynamic cycle ideally close to the Carnot cycle. Use of the Stirling engine for real-world problems, however, has been limited because of its relatively low output power. Towards its more practical applicability, simultaneous operation of many individual Stirling engines is indispensable to increase the output power. This paper presents an experimental study of entraining two Stirling engines to an external pacemaker. Our aim is to achieve synchronized oscillations of the Stirling engines without lowering their oscillation frequencies, because both synchrony and high frequencies are important factors to enhance the total output power. Compared to our previous study of directly coupled Stirling engines, it is shown that the output power is significantly improved in the present framework.
Quadratic Assignment Problem (QAP) is one of the combinatorial optimization problems which are classified into Nondeterministic Polynomial time solvable (NP)-hard problems. Therefore, it is important to develop algorithms for finding good approximate solutions in short time. In this paper, we proposed an algorithm for approximately solving QAP by using chaotic neurodynamics. The proposed algorithm has three characteristics. First, compared with the conventional method, the number of neurons was substantially reduced. Second, the effect of external inputs to neurons was changed. Third, a new parameter tuning method was introduced. As a result, our algorithm can find good solutions compared with the conventional method using chaotic neurodynamics.