JPH0348163A - Very small constant integration tracking type measuring device - Google Patents

Abstract

PURPOSE:To prevent that a measuring system gets into disorder with respect to strong pulse-like noise appearing at low frequency by allowing an estimated value to follow the most probable value of physical quantity when the value calculated by subtracting the estimated value from a measured value becomes positive or negative. CONSTITUTION:Now, primary measured rows are set to A1, A2...Ai and the estimated values thereof are set to B1, B2...Bi and DELTA1=A1-B1 is calculated with respect to the first measured value A1 and the first estimated value B1. Next, the code of DELTAis discriminated to perform the next calculation. That is, at the time of DELTA1>0, B2=B1+B is formed and, at the time of DELTA1=0, B2=B1+0 is formed and, at the time of DELTA1<0, B2-B is formed. Herein, DELTAB is a properly determined constant. Next, the same calculation is repeated with respect to the measured value A2 of the second time. That is, at the time of DELTA2=A2-B2 and DELTA2>0, B3=B2+DELTAB is formed and, at the time of DELTA2=0, B3=B2+0 is formed and, at the time of DELTA2<0, B3=B2-DELTAB is formed. By this method,when the estimated value B is corrected at every measured value and the next estimated value B is corrected to obtain the next estimated value, the value of B shown in a steady measuring state approaches the mean value of measured values Ai.

Description

[Detailed Description of the Invention] A plurality of physical quantities Ii! When measured, the measured values are not necessarily consistent down to the smallest digit, but rather generally show considerable variation. 1 including these errors
If there are many single price points, it may be okay to use that value as is, but normally it is inconvenient and some kind of smoothing process is used to improve accuracy and profit. In this case, the conventional common method was a moving average. However, this moving average does not work well with an offline method in which primary measured values are recorded on a suitable recording medium and then numerically processed later. The online method is often inconvenient when you want to know the merged value in real time.In theory, with a moving average, the corresponding time of the obtained value is long before the time of the last measured value. Because it will be.

The present invention relates to a measuring device that calculates the most probable value of a measured quantity in real time, without using the moving average plus 6 as in the conventional method.

Now, the -th measurement value sequence is ^1 + A2 +...A.

The estimated value is 81+82+...B.

and Δs=A for the first measured value Al+estimated value B1
Calculate +-B+ (1) 0th order Determine the sign of this Δ and perform the next calculation. Namely.

Here, ΔB is an appropriately determined constant. Next, repeat the same calculation for the 28th m measured value A2.
That is, Δ* = Ae-82 (3) If the estimated value B is corrected for each measured value and the next estimated value is obtained, the value of B shown in the steady state of measurement will be the measured value A.
I(approaching the average of 1+r+-n*sl.

Here, ΔB is set to 17211 degrees, which is the required minimum measurement resolution.

3. Example A method for measuring ocean currents using a sonic current meter installed on a ship is to use a gyro compass to measure the ship's heading and water speed by measuring the reflected waves of ultrasonic pulses emitted into the sea from the bottom of the ship and reflected from each layer of the water. The Doppler frequency and ground speed are measured using the Doppler frequency of seafloor reflected waves or a radio navigation device, and are determined as the difference between the ground and water speeds. Namely.

The speed over water measured by the sonic log is v - the ship's heading indicated by the gyro compass. zH is the speed over the ground measured by navigation radio waves. v0 is the ship's heading direction measured by radio waves.
.

Then, the water velocity component in the north-south and east-west directions v88. ν1 is the north-south and east-west ground velocity component V. From this, Nw VOE is the absolute value V of the ocean current. However, the ocean current values determined in this way contain considerable errors and are not accurate.

Conventionally, the above calculations were performed on data obtained by accumulating past data in memory and performing a moving average on two individual pieces of data, but by applying the principles of this invention, such complicated calculations can be avoided. The necessary accuracy can be ensured.

Equation (7) is rewritten as follows.V.Novol- is determined so that it becomes 0'', but at this time, the estimated value is substituted as VCN+VCE.

That is, for the data of 1st vR12, calculate the north-south and east-west direction components of the ocean current VCNsVCIE.

This sign is calculated in the same way for the second data 4, and the sign of feature 1, (1), (3), etc. is determined to be positive or negative, and the estimated value is set so that the frequency of positive and negative expressions is 50/50. Since the value is corrected and tracked using a minute constant integration method, the magnitude of the variation in measured values is not directly related to the stabilization of the measurement system, and the measurement system becomes unstable due to strong pulse noise that appears at low frequencies. I'm never there.

2. Accuracy can be increased by setting ΔB small.

In this way, calculations are performed one after another for the data from the third time onwards (.

Of course, the ocean current is determined by the VC cap tVcl:l (8) ~
Find it by executing equation (10). If ΔB is set smaller than the minimum measurement resolution, VCNIell:El' after a certain period of time is limited to the average of the flow determined by equation (7). approach, most finally ±ΔB
The variation in the results is unavoidable.

5. Precautions for use 1. This method can be used to measure various physical quantities, but
When the rate of change of the measured quantity is large, care must be taken regarding the tracking speed.

2. The tracking speed is determined by the magnitude of the infinitesimal constant ΔB and the signal S
/N is involved.

that's all

Claims (1)

[Claims] "The value Δ that is calculated by subtracting the estimated value from the measured value of a certain physical quantity
An infinitesimal constant integration tracking method measuring device that adds an infinitesimal constant to the estimated value when it is positive, and subtracts it when it is negative, so that the estimated value tracks the most probable value of that physical quantity.