Many of the super-resolution DOA (Direction-of-arrival) estimation algorithms presume the number of signal sources. Since the sample snapshot is limited, however, the available algorithms on source number estimation may fail at lower SNR (signal-to-noise).

The transmitled signal used is linear frequency modulation (LFM) signal and the received signal within observation time is represented by the summation of multiple sinusoid signals in both range and Doppler dimension. The time-delay and Doppler information are used jointly to estimate the sample covariance matrix,and the final resolution algorithm is a high-resolution source number estimation method-the new modified Gerschgorin disk estimation (NMGDE) method.

The detailed solution to the problem of number estimation was explained,and Monter-Carlo trials demonstrated that the performance of this method was better than the usual algorithms.

Anther width had a relatively lower heritability of 50% and the number of genes involved in the trait was estimated to be up to 13,while for other 9 traits the heritabilities and gene numbers were estimated to be 68%-93% and 7-10,respectively.

A new source number estimation method based on the beam eigenvalue

Most source number estimation methods based on the eigenvalues are decomposed by covariance matrix in MUSIC algorithm.

Both of the theory analysis and the simulation results show that the BEM method can estimate the source number for correlated signals and can be more effective at lower signal to noise ratios than the normal source number estimation methods.

A general expression for gene number estimation which encompasses the conventional formula was derived.

It provides a basis for gene number estimation from the data of recurrent selection experiments that are not of sufficient duration to measure total response to selection.

During the course of studies on the population dynamics of the European corn borerin Peking from 1961 to 1964, data were secured on the egg populations in a total of 19fields of spring-sown corn at whorl-stage. Based upon these data, the following pointswere analyzed: (1) The distribution pattern of the egg masses in corn fields. (2)Relations between the percentage of plants with one or more egg masses and the numberof total egg masses accumulated during the entire whorl-stage. (3) Relations betweenthe number...

During the course of studies on the population dynamics of the European corn borerin Peking from 1961 to 1964, data were secured on the egg populations in a total of 19fields of spring-sown corn at whorl-stage. Based upon these data, the following pointswere analyzed: (1) The distribution pattern of the egg masses in corn fields. (2)Relations between the percentage of plants with one or more egg masses and the numberof total egg masses accumulated during the entire whorl-stage. (3) Relations betweenthe number of egg masses at the peak stage, i.e., the highest number of egg masses foundon any one day, and the total egg masses accumulated during the entire whorl-stage. The analyses of these field data show: 1. Since the observed frequency of plantsreceiving various numbers of egg masses during entire whorl-stage fitted the expectedfrequency according to the Poisson series closely, the distribution of the egg masses in cornfields is fully random. Therefore, the percentage of plants with one or more egg massesis proportional to the total number of egg masses accumulated. 2. The relations between the percentage of plants receiving one or more egg masses(P) and the number of egg masses per 100 plants (X) during entire whorl-stage of 15corn fields was shown in Figure 1. This relationship can be generally expressed by P=1--e~(-ax~b)where a and b are constant. Putting m=ax~b, the values of a and b can be calculated bytransforming above equation to log m=log a + b log XBy estimating the parameter a and b, the following equation was obtained P=1--e~(-0.00985x~(0.9984))Table 3 gives the calculated values of x to different percentages of P. 3. Data on the total number of egg masses per 100 plants for the entire whorl-stage(Y) and the highest number of egg masses per 100 plants found on any one day (X)in 17 corn fields are presented in Table 4. A X~2 test for goodness-of-fit reveals alinearship between the two variates, indicating the peak of egg deposition is directly pro-portional to the total number of egg masses deposited during the entire whorl-stage. Theregression equation is Y=11.47 + 2.64XThe straight line that corresponds to this equation is plotted in Figure 2. By using theabove equation, the value of Y would be determined for any given value of X. There-fore, the use of the latter record as an index of the former appears to be biologicallysound.

A procedure of simultaneous calculations for correcting regression coefficients due to deletion and/or addition of observations is developed, which includes the Plackett result as a special case. Considerations with respect to deleting observations in relation to the missing values problem in the factorial experiment lead to a different, yet feasible, approach to the correction problem. That is, with the deleted observations being first estimated, the correct regression coefficients turn up naturally.As for...

A procedure of simultaneous calculations for correcting regression coefficients due to deletion and/or addition of observations is developed, which includes the Plackett result as a special case. Considerations with respect to deleting observations in relation to the missing values problem in the factorial experiment lead to a different, yet feasible, approach to the correction problem. That is, with the deleted observations being first estimated, the correct regression coefficients turn up naturally.As for estimation of missing values, the principle of least squares may be worked into a tabular form equivalent in the wide sense to the method of cross-contrasts as was proposed in a previous paper by one of the authors.

This paper deals with the applioations of multiple-knot cardinal δ-spline function interpolation to the problem of fitting in a great deal data, as a continuation to past(Ⅰ) and(Ⅱ)of the same title by the same author.The author has established the least squares fit by multiple-knot δ-spline, pointed out that coefficient matrix of the system of its normal equation has the characteristics of band and majorant and so on,and estimated its condition number for degree of m.c.δ-spline k≤3.The given intrisical property...

This paper deals with the applioations of multiple-knot cardinal δ-spline function interpolation to the problem of fitting in a great deal data, as a continuation to past(Ⅰ) and(Ⅱ)of the same title by the same author.The author has established the least squares fit by multiple-knot δ-spline, pointed out that coefficient matrix of the system of its normal equation has the characteristics of band and majorant and so on,and estimated its condition number for degree of m.c.δ-spline k≤3.The given intrisical property and its explicit expression would bring facilities for its applications, especially for computer-aided geometricdesign.