RU2610146C1 - Radio-wave doppler detector - Google Patents

Abstract

FIELD: electronics.

SUBSTANCE: invention relates to signals detection and classification equipment, recieved from objects in radio wave (super high frequency UHF range) device. Method of increasing resistance to false triggering in case of animals appearance consists in following: SPDRWD (single-position Doppler radio wave detector) is capable to recognize human and animal by information, contained in UHF signal reflected from object. During reflection signal is modulated and carries information on object. From received signal Doppler component is selected, which is amplified, filtered and supplied to detector. When detecting signal exceeding threshold level, features selection "circuit" is launched. Selection of criteria is obtaining information from signal contained in it during reflection from object. Information is transmitted to decision making "circuit", which according to classification model loaded into it (binary solutions trees set weighted voting) decides on issuing of alarm signal.

EFFECT: technical result is increase in single-position Doppler radio wave detector (SPDRWD) noise-immunity in case of animals appearance in detection zone.

1 cl, 10 dwg

Description

The invention relates to technical means for detecting and classifying signals received from objects in a microwave radio frequency (microwave) device and can be used to protect areas of the area with high intensity of the appearance of animals in the detection zone.

Known device KSM - OLD [1], consisting of a microwave module and a signal processing circuit, which is a Doppler one-way detection tool with adjustable dimensions of the detection zone (ZO). The detector emits a microwave signal; when the intruder moves in the AO, the microwave module receives a signal reflected from the object, modulated in frequency depending on the speed of the object. The received signal is detected to isolate the Doppler component and compared with a threshold value. When the signal exceeds the threshold value, a decision is made to issue an alarm.

Known device "Agate" [2], which is a radio wave detection tool. According to the principle of operation, the detector is a Doppler radar with selection of moving targets in range based on the frequency modulation of the emitted signal. The detector emits a frequency modulated microwave signal. When the intruder moves in the AO, a reflected signal appears, which is received by the detector. Based on the difference between the frequencies of the emitted and reflected signal, range selection is performed. Intruder movement is selected based on the analysis of the frequency spectrum of the Doppler component of the reflected signal.

In both devices, the decision about the violation is made by the threshold device according to the energy criterion, provided that the Doppler frequency shift of the signal lies in a certain range, and the duration of its exposure is not less than a given value.

Thus, the principle of operation of detection means - prototypes is based on registration by the receiver of the microwave module of changes in the "interference pattern" of the electromagnetic field by an object moving in the detection zone.

When deciding on the presence of an intruder, the algorithm of sequential evaluation of three informative features is used (Fig. 1):

- the frequency of the signal, depending on the radial velocity of the object (carried out by the filter);

- the duration of the signal, depending on the time the intersection of the detection zone by the object (carried out by the integrator);

- the signal amplitude, when using the amplifier-corrector and frequency modulation of the signal, depending on the size of the object (carried out by a threshold device).

Due to the fact that many animals do not differ from a person in speed and size, this method of deciding on the presence of an object does not have the fundamental ability to distinguish between a person and an animal. This disadvantage is inherent in all radio wave Doppler detection tools and leads to a high level of false alarms when operating these tools in places with their frequent occurrence (for example, dogs).

Thus, the disadvantage of these devices is the low noise immunity when the animal appears in the detection zone, if its size and speed of movement are comparable with the size and speed of movement of a person.

The aim of the invention is to increase the noise immunity of the radio wave Doppler detection means when an animal comparable in size to a person appears in the detection zone.

To achieve this goal, the problem is solved - the creation of a device with the ability to distinguish a person from an animal and allowing its effective use in conditions of high probability of the appearance of animals in the AO.

An increase in the noise immunity of the detection means when an animal appears in the detection zone, the parameters of which are comparable with those of a person, is achieved by introducing into the structure of the detection means a scheme for generating a vector of signs that have more pronounced differences between humans and animals, and a decision making scheme (classifier).

When an object moves in the detection zone, the signal reflected from it is modulated in all parameters due to the geometric features of the structure of the object and the method of movement typical of this class of objects. Crucial in improving noise immunity are new informative features, the principle of construction of which is based on the extraction of information from a signal about the method of movement of an object characteristic of a person or animal.

The model of signal generation (MSO) of the centimeter-wave radio-frequency Doppler SO Doppler is based on the radar equation and the description of the Doppler effect, taking into account that the received signal is detected by a phase detector, and the detection object is a complex target and is a combination of elementary reflectors (elements):

where E _d (t) is the voltage at the output of the phase detector (microwave module), V;

K _d - proportionality coefficient, depending on the characteristics of the receiver;

K (α (t)) is the coefficient of proportionality, depending on the angle α between the axis of the antenna pattern (BOTTOM) and the direction to the j-element;

U ₀ is the level of the probing signal, V;

λ is the wavelength of the probe signal, m;

D _j (t) is the distance from the microwave module to the j-th element, m;

V _j (t) is the radial velocity of the j-th element, m / s;

Δt _j (t) - the time during which the wave front passes from the microwave module to the j-th element and vice versa, sec;

σ _j (t) is the effective scattering surface (EPR) of the j-th element, [m ² ];

j is the number of the element of the object.

It can be seen from the MCO that only the radial velocity of the jth element, the EPR of the jth element, the number of elements n and their ratio (EPR and speeds) can affect the recognition of objects.

The proposed device is a radio wave Doppler detector, consisting of a microwave module 1 (a sensor detecting a signal), a signal preprocessing circuit 2, a feature vector generating circuit 3, a decision circuit 4, and an actuator 5 (FIG. 2).

The signal preprocessing circuit includes: a range determining unit 6, an analog-to-digital converter 7, an amplifier 8, a filter 9, an integrator 10, a threshold device 11.

The structure of the formation of the vector of signs includes blocks 12 allocation of signs (by the number of signs) and block 13 of the formation of the vector of signs.

The structure of the decision-making scheme includes blocks 14 for making decisions and multiplying the results by the significance of the block (by the number of decision trees) and block 15 of weighted voting.

The invention is illustrated by drawings, where:

- FIG. 1 is a diagram of a decision algorithm in prototype tools;

- FIG. 2 is a structural diagram of a radio wave Doppler detector;

- FIG. 3 is a diagram of an object passing through a sensitivity zone of a microwave module;

- FIG. 4 is a view of a signal generated at the output of a microwave module when an object passes through a detection zone;

- FIG. 5 is a graph of a spectrum of a Doppler signal;

- FIG. 6 is a drawing representing a person walking and running;

- FIG. 7 is a drawing showing the rotation of the pelvis of a person when walking;

- FIG. 8 is a drawing representing the movement of a common center of mass of a person when walking;

- FIG. 9 - drawing of the options for horse movements, where a) - step, b), c) - lynx, d) - gallop;

спектра доплеровского сигнала при проходе человека через ЗО под углом 10° к оси ДН антенн СВЧ модуля.- FIG. 10 is a graph of the frequency variation of the main maximum

the spectrum of the Doppler signal when a person passes through the AO at an angle of 10 ° to the axis of the bottom of the antenna of the microwave module.

The proposed device operates as follows.

The microwave module 1 emits a signal into space, receives a signal reflected from the object. From the received signal, it selects the Doppler component, which is fed to the signal preprocessing circuit 2 (amplifier input 8) and the range determining unit 6.

The signal preprocessing circuit 2 is intended for preliminary amplification of the signal in the amplifier 8, filtering the signal in the filter 9, integrating the signal in the integrator 10 and detecting that the signal exceeds the predetermined level by the threshold device 11, generating a control trigger signal for the feature vector generating circuit 3. The output of the amplifier 8 is connected to the input of the filter 9, the filter 9 is connected to the input of the integrator 10 and the input of the ADC 7. The integrator 10 is connected to the input of the threshold device 11.

The signal received by the signal preprocessing circuit 2 is amplified, filtered in a predetermined frequency range and fed to an integrator for detuning from short-term emissions, then to threshold device 11. Triggering of threshold device 11 is perceived as object detection. When this signal from the output of the threshold device 11 starts the circuit 3 of the formation of the vector of signs and circuit 4 decision. From the output of the filter 9 of the pre-processing circuit 2, the signal enters the second channel of the ADC 7.

To determine the range to the object, the range determination unit 6 supplies a modulating signal to the microwave module 1 and receives a low-frequency Doppler signal from it. By the ratio of the frequencies of the modulating and received signals, the range determining unit 6 provides a signal to the ADC 7 whose level is proportional to the range to the detected object.

The ADC 7 has 2 channels and converts the range signal and the Doppler signal into digital form and transmits them to the feature vector generation circuit 3.

Scheme 3 of the formation of signs is a program that calculates informative signs based on input data. Input data for the circuit are:

- filtered and amplified Doppler signal in digital form;

- data on the distance to the detected object in digital form;

- start signal of circuit 3 (from a threshold device).

Scheme 3 of the formation of the vector of signs includes several blocks. Each block selects (calculates) its characteristic.

Signs are estimates of the specific parameters of the Doppler signal that characterize the properties of the detected object. Symptoms are determined from the following considerations.

A. Man and animal differ in size (in part). The signal amplitude characterizes the reflectivity of the detection object (effective scattering area - EPR), therefore, with approximately the same composition of animals and humans, the EPR characterizes its size. The ESR of the object (sign P1) was adopted as a sign for assessing this difference. To calculate the characteristic, additional information is needed - the distance to the object:

The ESR of the object is calculated by the formula:

where D is the distance to the object, m;

U _max - maximum signal at the detector output, V;

K (α) is a coefficient depending on the characteristics of the microwave module and the direction between the axis of the radiation pattern (LH) of the antennas and the direction to the object.

Since at the intersection of the object the object crosses the axis of the beam (Fig. 3), at this moment the maximum signal is present at the output of the module (Fig. 4). Therefore, when using the EPR calculation formula U _max, one can take K (α) = K (0), thereby eliminating the dependence on the angle between the axis of the antenna array and the direction to the object.

B. Man and animal differ in speed of movement (partially). According to [3], the position of the maximum of the frequency spectrum of the signal characterizes the radial velocity averaged over the time of observation of the largest reflecting part of the object (body). As a sign for assessing this difference, the maximum frequency spectrum of the Doppler signal was adopted.

The sign is calculated by the formula:

where: f _glam is the main maximum of the spectrum of the Doppler signal (Fig. 4).

B. Man and animal differ in the spread of speeds of moving elements. Human walking is a series of rhythmic and alternating movements of the limbs and trunk (Fig. 6), which lead to the forward movement of the center of gravity. When walking, the general center of mass (JMC) of the human body performs rhythmic up and down movements when moving forward. The shoulder girdle and pelvis rotate in a step cycle in the opposite direction (Fig. 7). When the mass is transferred from one leg to the other, the pelvis and torso move toward the supporting leg. When moving forward, the common center of mass not only performs rhythmic up and down movements (Fig. 8), but also sways from side to side. While the pelvis rotates around the supporting leg (to eject the forward leg forward), the shoulder girdle also rotates in the opposite direction. The arm on the supporting side at this time swings, it compensates for the mass of the rotating body of the portable leg on the opposite side. When a person runs, the general picture of the movement of body parts changes slightly (Fig. 6). Thus, for a uniform translational movement, a person makes additional balancing movements with his hands. Since the speed of limbs in this case differs from the speed of movement of the body, the frequencies of the signals reflected from them are different from the frequency of the main maximum, which leads to the expansion of the frequency spectrum of the signal from a person.

Using a horse as an example, a typical four-legged way of movement is considered. When the horse walks in a step, the right front leg is first moved forward, followed by the left hind leg. Next is the left front and, finally, the right rear (Fig. 9). This completes the cycle. At the same time, animals for walking do not make additional balancing movements. When animals are running, the changes are much more pronounced, therefore, the difference between humans and animals during running can be reduced.

As a sign for assessing this difference, the ratio of the frequency of the maximum of the Doppler signal spectrum to its width (signal broadband) is taken:

The sign is calculated by the formula:

- главный максимум амплитудно-частотного спектра доплеровского сигнала (фиг. 3);Where

- the main maximum of the amplitude-frequency spectrum of the Doppler signal (Fig. 3);

- максимальная частота спектра сигнала по уровню S_min;

- the maximum frequency of the signal spectrum in terms of S _min ;

- минимальная частота спектра сигнала по уровню S_max.

- the minimum frequency of the signal spectrum in terms of S _max .

D. A person and an animal differ in the ratio of the sizes of body parts moving at a speed greater than the speed of the body (the main reflective element) and body parts moving at a speed less than the speed of the body. Since for a uniform translational movement a person makes additional balancing movements with his hands, it turns out that in certain phases of a step in a person, two limbs move faster than the body, and the other two limbs slower (the situation does not change when running). Since the animal does not need to make balancing movements for walking, when moving with a step, only one limb moves faster than the body, and the remaining limbs are used as a support. As a sign for assessing this difference, the symmetry of the frequency spectrum of the Doppler signal is accepted:

The sign is calculated by the formula:

- площадь S1 (фиг. 5) под кривой амплитудно-частотного спектра доплеровского сигнала на участке от

до

;Where:

- area S1 (Fig. 5) under the curve of the amplitude-frequency spectrum of the Doppler signal in the area from

before

;

- площадь S2 (фиг. 5) под кривой амплитудно-частотного спектра доплеровского сигнала на участке от

до

.

- area S2 (Fig. 5) under the curve of the amplitude-frequency spectrum of the Doppler signal in the area from

before

.

D. Man and animal differ in rhythmic acceleration of the body. At each step, a person leans forward, his common center of mass (JMC) extends beyond the support area and begins to “fall”, while his body accelerates (Fig. 8). The leg is extended to prevent the fall. After it touches the ground, body weight is transferred to it. The JMC is inside the support area. The knee bends, cushioning the fall, and straightens, returning the body to its original height, while the body slows down. Thus, the body of a person when walking is subject to rhythmic accelerations and decelerations.

When the horse walks in a step, the right front leg is first moved forward, followed by the left hind leg. Next is the left front and, finally, the right rear. This completes the cycle in order to repeat from the right front foot, etc. The projection of the center of gravity always falls inside the triangle, or between two fulcrum, formed by limbs located on the ground. This allows the animal to be in a stable state at any time, and its body when walking is less prone to rhythmic accelerations and decelerations, since the need to balance the instability of the animal’s body position is less than that of humans.

Increasing the speed, the animal switches to a lynx: each leg comes off the ground before the limb that precedes it in the cycle described above touches the ground. There is a moment when an animal rests on the ground with only two limbs. This increases the speed of movement, but reduces stability, while rhythmic acceleration increases. In horses, cats, dogs, and many other mammals with such a quick gait in the air, two diagonally located limbs are simultaneously located in the air: the left front and the right rear or vice versa. With this movement, the body of the animal is subject to rhythmic accelerations and decelerations comparable to humans.

When galloping, the animal tears the limbs off the ground in a different order than walking in steps: right front - left front - right rear - left rear or, conversely, left front - right front - left rear - right rear. At the same time, for some time the animal is entirely in the air, without touching any of the limbs on the ground. With this movement, the rhythmic acceleration and deceleration of the body of the animal can exceed human performance.

Thus, the difference in rhythmic acceleration can manifest itself to a greater extent when walking an object.

, что характеризует радиальное ускорение основной отражающей части объекта между ближайшими локальными минимумами и максимумами:As a sign for assessing this difference, the rate of rise of the global maximum of the amplitude-frequency spectrum averaged over the observation time was taken

, which characterizes the radial acceleration of the main reflecting part of the object between the nearest local minima and maxima:

The sign is calculated by the formula:

(фиг. 10).where: α _n is the angle of inclination of the line connecting the n-th maximum and the n-th minimum of the behavior graph

(Fig. 10).

The set of all the features formed by this scheme is a feature vector M and is a column matrix in which each cell is a numerical estimate of one of the features:

The calculated vector of signs M is fed to the decision-making scheme. A decision-making scheme is a program that implements a classification model (decision rule). The classification model consists of a set of j decision trees constructed by various algorithms, for example [4, 5] and a weighted voting block (Fig. 2). The feature vector M arrives simultaneously at all decision trees, each tree independently makes a decision h about the class of the detected object (h can take values -1 or +1) and has its own weight α. Further, the results of the work of all the trees are sent to the weighted voting block [5]. Voting takes place according to rule (2):

where sign is the sign of the result;

k is the number of the decision tree.

As a result of a weighted vote on a set of trees, a decision is made on the class of the detected object: “plus” is the intruder, “minus” is the interference. [4, 5]

If it is decided that the detected object belongs to the class of intruders, then the corresponding signal is sent to the executive device, and it gives an alarm to the user in the form that is necessary for him.

The proposed detector has increased noise immunity and can improve the quality of technical means of protection. Improving the noise immunity is achieved by making changes to the signal processing circuit — using the recognition subsystem in the detection tool, which includes a circuit for generating a feature vector and a decision circuit.

Information sources

1. Single-position KSM-OLD detection tool, operation manual BAZHK.425919.018 RE.

2. Volumetric radio wave security detectors ИО407-15 "АГАТ-СП5У" and ИО-407-15 / 1 "АГАТ-СП5 / 1" Operation manual USDP.425144.003 RE. Closed Joint-Stock Company Firm UMIRS.

3. Marshalov T.A. Radar recognition of the equipped intruder and animals / Evdokimov N.O., Kostenko EA // Problems and prospects of development, improvement of TSOG. Interuniversity collection. Part 4. - 2003. No. 12. - Kaliningrad: KVI FPS of Russia, p. 44-48.

4. Chubukova I.A. Data Mining: a tutorial. - 2nd ed., Rev. - M: Internet - University of Information Technology; BINOMIAL. Laboratory of Knowledge, 2010 .-- 383 p.

5. Vorontsov K.V. Lectures on logical classification algorithms. 12/21/2007 http: www.machinelearning.ru

Claims (1)

A radio wave Doppler detector containing a microwave module, which is a sensor, a recording signal, a range determination unit, an analog-to-digital converter (ADC), a signal preprocessing circuit, a feature vector generation circuit, a decision making circuit, and an executive device, while a microwave the module is connected to the range determination unit and the input of the signal preprocessing circuit, the output of the signal preprocessing circuit is connected to the input of the formation circuit feature vector, the output of the feature vector circuit is connected to the input of the decision circuit, the output of the decision circuit is connected to the input of the actuator, in addition, the outputs of the range determination unit are connected to the input of the microwave module and the input of the ADC, the outputs of the ADC are connected to the input of the circuit of the vector of features characterized in that the pattern of generating a feature vector generates a column matrix in which each row is a numerical assessment of one of the features of the detected object - P1, P2, P3, P4, P5, while P1 - e effective scattering area (EPR) of the detected object, P2 is the maximum amplitude-frequency spectrum of the Doppler signal, P3 is the ratio of the frequency of the maximum amplitude-frequency spectrum of the Doppler signal to its width, P4 is the symmetry of the amplitude-frequency spectrum of the Doppler signal, P5 is averaged over the observation time the rise rate of the maximum amplitude-frequency spectrum, the generated vector of signs M is fed to a decision-making scheme in which a decision on the presence of an object is made based on the generated cl an assimilation model, moreover, if the detected object belongs to the class of intruders, the executive device gives an alarm to the user.

Images