All issues
Volume 3, Issue 6
Displaying 1-9 of 9 articles from this issue
- |<
- <
- 1
- >
- >|
-
[in Japanese]1964 Volume 3 Issue 6 Pages 409-414
Published: June 10, 1964
Released on J-STAGE: October 21, 2010
JOURNAL FREE ACCESSDownload PDF (3349K) -
Etsujiro SHIMEMURA, Yukio YOSHIDA, Kazuhiko KAWAMURA, Yuichiro KOMATSU1964 Volume 3 Issue 6 Pages 415-420
Published: June 10, 1964
Released on J-STAGE: October 21, 2010
JOURNAL FREE ACCESSIn this thesis the time-optimal control in linear time-invariant system with time-delay is investigated by applying the method of maximum principle. It is shown under some assumptions that anoptimal control is a piecewise constant function of time. Theorems on the existence and uniquenessof the optimal control are established. A brief discussion on an essential characteristic of the optimal feedback controller is also given and it is shown that the optimal controller should have a delayelement which gives the same amount of delay to the detected system variables as that of thecontrolled object.View full abstractDownload PDF (859K) -
Koji NEZU1964 Volume 3 Issue 6 Pages 421-425
Published: June 10, 1964
Released on J-STAGE: October 21, 2010
JOURNAL FREE ACCESSIn stroboscopic measurement of rotational speed, the observed pattern is stationary at everyrotational speed which is multiple of primary speed. Since suitable multiple measuring method hasnot been known, there has been serious troubles in measuring. On the other hand, suitable measuring device is useful not only for removing such troubles but for extending measurable speedrange of the stroboscope.
In this paper two practical methods, method 1 and method 2 are discussed. Operating principlesof these methods are quite different from each other; the former makes use of mathematical natureof series of flush frequency when single stationary pattern appears, but the latter makes use of therelation between the phase difference of two flushes with equal frequency, and the angular differenceof double stationary patterns.
Biased multivibrators modified for the above two methods are presented, and satisfactory result hasbeen obtained. The electronic circuit for method 2 is a little more complicated than that for method 1, but it was proved by experiments that the former is much more convenient for mesuring operation.View full abstractDownload PDF (702K) -
Kamekichi SHIBA, Tatsumi TOZAWA1964 Volume 3 Issue 6 Pages 426-428
Published: June 10, 1964
Released on J-STAGE: October 21, 2010
JOURNAL FREE ACCESSThe wet-and-dry-plate dew point hygrometer is composed of a wet-and-dry-plate hygrometerand a cooling apparatus.
Being cooled properly, the air to be measured flows from the dry-part to the wet-part of micaplate set in the wet-and-dry-plate hygrometer. The temperature difference t-tw of both parts ismeasured by the thermopile, and the air temperature t (dry-part temperature) is measured bythermister thermometer.
When the air is cooled down with constant pressure, the dew point tD is always the same so faras air temperature is above the dew point. Therefore, the temperature difference t-tw is functionof the air temperature t and its dew point tD. As t-tw is small and the dew point tD is constant, the temperature difference t-tw is proportional to t-tD, in other words, it varies linearly with thetemperature t. This fact is proved by experiment as expected theoretically.
The dew point tD is determined by extrapolating the curve which shows the relation betweent and (t-tw).View full abstractDownload PDF (448K) -
Tadashi YAMASHITA1964 Volume 3 Issue 6 Pages 429-439
Published: June 10, 1964
Released on J-STAGE: May 24, 2011
JOURNAL FREE ACCESSRecently the artificial hands have been investigated from various viewpoints and some hardwares have been presented. Compared these hands with natural hands, they are extremely lackingin degrees of freedom of the motion, and they can not do dexterous works. The author has made amodel of mechanical fingers having many degrees of freedom in order to investigate the possibilityof making highly efficient mechanical fingers.
The model made by the author has three fingers, which correspond to the thumb, index fingerand middle finger respectively. Each finger has three degrees of freedom, consequently, there arenine degrees of freedom, and these nine joints can be moved independently by compressed air. Eachjoint has sense organ which detects the angular displacement of a joint.
Manipulators are worked by human operators, but this model can be worked automatically according to predetermined programs without aids of men. As an example the rotation of a graspedobject by sequentially changing grasping fingers is performed, such a complicated work can not be, done by other artificial hands which have been made till now.
These artificial fingers possessing advanced functions may be used as powerful control organsof materials handling, which is contained almost every industrial process. However, actual applications of mechanical fingers to industrial processes have many difficult problems, the largest problemmay be the pattern recognition.View full abstractDownload PDF (6503K) -
Masakazu MATSUBARA1964 Volume 3 Issue 6 Pages 440-446
Published: June 10, 1964
Released on J-STAGE: May 24, 2011
JOURNAL FREE ACCESSThe equivalent dead time D of a meromorphic function G(s) in |s| < ∞ is defined as an integral_??_dt, where u (t) is the unit step function, i (t) the indicial response of G (s), andK the final value of i (t). It is shown that G (s) can be written in the form K(1+a1s+a2s2+ ……)/(1+b1s+b2s2+……) and D is equal to b1-a1 if D exists. It is also shown that D corresponds to thecenter of gravity of the impulse response of G (s).
Then, the author proposes to use D as a new measures of delay. Using D, the delay is explicitly defined in a sense. Relations between the pole-zero configuration of G (s) and the delay areinvestigated.
It is shown that the equivalent dead time of a unity feedback system is equal to that of theopen loop multiplied by the steady-state positional error for input u (t) in the case of a type0 system, to the steady-state velocity error for input tu (t) in the case of a type 1 system, and to 0 in the cases of type 2, 3, … systems. Consequently, if the type of system is other than 0, we canobtain the steady-state velocity error by calculating D from the result of the indicial responseexperiment.
To approximate a dead time by a delay with distortion or vice versa, it is proposed to make theequivalent dead time of the approximate transfer function equal to that of the true transfer function.As an example, it is shown that such a method gives much better approximation than the usualone in the case where the transfer function of a process is approximated by a lst-order delay and adead time.View full abstractDownload PDF (1012K) -
Kazuo TSUCHIYA1964 Volume 3 Issue 6 Pages 447-452
Published: June 10, 1964
Released on J-STAGE: October 21, 2010
JOURNAL FREE ACCESSThere is more or less dead-time in the human transfer-function in manual operation. It isconsidered that dead-time is consisted to be in 3 parts, i. e., (a) time for recognizing the stimmulation, (b) time to calculate controlling the muscles, and (c) time to transmit the control signalthrough the nurves.
Active current was measured to obtain the mechanism of stimmulouse-recognition. The actioncurrent was induced from the surface of arm applying the indirect-induce electrodes. It wasamplified by a. c. amplifier (gain 100 dB, frequency range 1-1000c/s) and input stimulation toelectromagnetic oscillograph and the actual response were recorded simmultaneously.
The action current which is induced by indirect electrode is composition of controlling signalto the muscle and of feedback signal from the muscle. The controlling signal precedes about50-80ms to the actual movement of hand.
Both amplitudes of controlling signal and the feedback signal are proportionate to the logarithmof the stimmulation amplitude. Therefore they are the function of bits number of input-information, but there is some difference on the level of between two.
The movement of hand can be predicted by the controlling signal in the active current. Itseems that the controlling signal has two states, as the start signal and reset signal, and as it iscombination of feedback signal, it seems as if there are three states in the wave-envelope. Thatbecomes much clear when alignment amplifier of 200-400 c/s of high Q is provided. These signalsmay utilize for distance remote control devices for many industried purposes.View full abstractDownload PDF (1010K) -
[in Japanese], [in Japanese]1964 Volume 3 Issue 6 Pages 453-456
Published: June 10, 1964
Released on J-STAGE: October 21, 2010
JOURNAL FREE ACCESSDownload PDF (891K) -
[in Japanese], [in Japanese], [in Japanese], [in Japanese]1964 Volume 3 Issue 6 Pages 457-466
Published: June 10, 1964
Released on J-STAGE: October 21, 2010
JOURNAL FREE ACCESSDownload PDF (2824K)
- |<
- <
- 1
- >
- >|