SU1218397A1 - Device for determining convolution of two functions - Google Patents

Abstract

The invention relates to the field of computing and can be used in the processing of random processes by gliding weight smoothing in real time. The purpose of the invention is to expand the functionality by performing a convolution of the input signals with an arbitrary two-level weight function. The device achieves the ability to invert codes from the output of the unitary signal generator in accordance with the values of the control signals from the register outputs, in which the code of the two-level weighting function is permanently stored, the application of which is new in the device. 1 il. (L 00 00 with

Description

The invention relates to computing and can be used in the processing of random processes by sliding weight smoothing in real time.

The purpose of the invention is to expand the functionality of the device by performing a convolution of the input signals with an arbitrary two-level function.

With the passage of unitary signals through the computing medium, they are convolved and sliding averaged by calculating the modular function according to the expression i

u |.-jb-i - i Si 1,

U1

where X j is the input unitary code value,

S-, are the binary values of the weight function.

In the particular case of all values, the device performs a calculation of the moving average of a random process.

The drawing shows a block diagram of the device.

The device contains a shaper 1 unitary signals, a homogeneous computing environment 2, a multichannel converter 3 unitary signals into binary code, RS-flip-flops 4, adders 5, register 6, the second element AND-NOT 7, the first element AND-NOT 8, the elements OR 9, as well as outputs, devices.

The device works as follows.

Before the operation of the device, the code of the two-level weight function is stored and permanently stored in register 6. Processing of measurement results by their digital convolution with a two-level, weighting function and sliding averaging takes place in one cycle of operation of the device. At the beginning of each cycle of operation, a short impulse is formed at the first output of block 1, along the rise front of which the information stored in the computing environment 2 is shifted. At the same time, the adders 5 of the computing environment of the i-th row are entered into the S value of the two-level weight function. direct or inverted values

2183972

the states of the triggers of 4 elements of the computing environment (i-l) -ro row.

Thus, after performing shift operations in the computational

5 environment 2 at the outputs of the device, being the outputs of the multichannel converter 3 unitary signals into binary code and the outputs of the elements of the latter (,.) - r (

 row of computing environment 2, a code is generated for the sliding averages of the modular function. The operation of the input elements .7-9 is described by a logical equation

 (X, -i J + S;)

Claims (1)

Invention Formula 20 An apparatus for determining the convolution of two functions, comprising a shaper of unitary signals, a homogeneous computing environment, a multichannel converter of unitary  signals in binary code, each (i, j} -H element (, 3, ..., 2, 3, ..., K) of a homogeneous computing environment contains an adder and RS-trigger, outputs of adders 1-x 30 () elements of a homogeneous computing environment are connected to the corresponding inputs of the multichannel converter of unitary signals into a binary code, the control inputs 35 triggers are combined and connected to the output of a formator of unitary signals, the first input of the adder (i, j) -ro of the medium element is connected to the output of the sum of the adder (i-i, j) -ro 40 elements of the homogeneous computing environment, the second input of the adder is connected to the forward output of the trigger (i, j) -ro of the element of the homogeneous computing environment, the third input The 45 adder (i, j) -ro of the homogeneous computing environment element is connected to the transfer output of the adder (ijJ-O-ro of the homogeneous computing environment element, and the outputs of the 1st The 50 (upon) elements of the computing environment, starting with the K-th bit, and the outputs of the multichannel converter of the unitary signals into binary code are the outputs 55 devices, characterized in that, with the aim of expanding. neither the functionality due to the convolution of the input signals with an arbitrary two-level weight function, the code register of the two-level weight function is additionally entered into it, and in each (1, o) -th element of the homogeneous computing environment the first and second elements of NAND and OR are entered, and ff-input of the trigger (i , j) -r element of the homogeneous computing environment is connected to the output of the first AND-NOT element, the first input of which is connected to the output of the second AND-NOT element, and the second input to the output of the OR element, the first input of which is combined with the same input 2183974 the second element AND-NOT of each 1st element (for j const) of the computing environment and connected to the corresponding output of the code register of the two-level weight function, the second input of the OR element is combined with the same input of the second AND-NOT element of each (i, j) -ro element homogeneous computing environment and connected for each (i,) - ro element to the corresponding output of the group of the generator of unitary signals, and for each (i, j + l) -ro element to the trigger output (i, j) -ro of the element of uniform computing environment. . (ABOUT f5 ffftt-i fOn VNIIPI Order 1133/57 Circulation 673 Subscription Branch Patent, Uzhgorod, st. Project, 4