首页>
外国专利>
ELECTRONIC CIRCUIT ARRANGEMENT FOR MEASURING STATISTICAL PARAMETERS OF STOCHASTICAL PROCESSES
ELECTRONIC CIRCUIT ARRANGEMENT FOR MEASURING STATISTICAL PARAMETERS OF STOCHASTICAL PROCESSES
展开▼
机译:测量随机过程中统计参数的电子电路布置
展开▼
页面导航
摘要
著录项
相似文献
摘要
1,207,394. Selective signalling. NORMA FABRIK ELEKTRISCHER MESSGERATE G.m.b.H. 6 Oct., 1967 [6 Oct., 1966], No. 45810/67. Heading G4H. [Also in Division G1] A device is described for measuring various statistical parameters of a stochastic process represented by an electrical waveform. A stochastical process is a process that is subject to random variations, due for example to machine tolerances-such a process may be the continuous production of wire which should be of a nominal diameter, but of which the actually diameter varies in a random manner about the nominal diameter, and is illustrated in Fig. 2a. In the basic embodiment, Fig. 1, the stochastic wave 1 is repeatedly sampled 2 by clock pulses 3 and each sample is compared 4a, 4b with two reference potentials U sa , U sb so as to mark firstly either none, one or both of two leads X, Y, and then, via a logic circuit 6 (Fig. 3, not shown), one of three leads A 1 , A 2 , A 3 , in accordance with the range in which the sample lies. Thus a sequence of pulses occurs on each of the leads A 1 , A 2 , A 3 , each pulse indicating that the sample then being taken lies in the associated range. The Specification considers how to derive from these signals measures of various statistical parameters, and the simplest of these are the three " total probability " measures P 1 , P 2 , P 3 , each of which gives the probability that a sample chosen at random will lie in the associated range. Thus, if for example, P 2 has a value of 0À75, then the probability of a sample selected at random lying in range 2 is 0À75. Clearly, since a sample must lie in one of the three ranges, P 1 +P 2 +P 3 = 1. The value of P 2 , for example, is obtained by applying the pulses A 2 to a trigger which is reset regularly by clock pulses, and then by integrating in either an analogue or a digital manner, the resulting uniformly shaped pulses, so as to obtain a measure of their average frequency. The more complex parameters include joint probabilities, such as P 13 for example, which is the probability that two consecutive samples will lie in ranges 1 and 3 in that order. This probability is evaluated (Fig. 5, not shown) in a similar manner by gating one-bit delayed A 1 pulses with A 3 pulses, and shaping and integrating the resulting pulse train to obtain a measure of its average frequency. The Specification also refers to the autocorrelation function and the power spectral density. If a stochastic wave consists of random variations superimposed upon a known signal, for example a signal, Fig. 9, having three different levels, then in order to analyse the random fluctuations either the known signal with no fluctuations can be subtracted from the stochastic wave, or the threshold levels in the system described above can be adjusted in accordance with the average signal level taken over an appropriate period (Fig. 8, not shown).
展开▼