RU2455615C1 - Method for non-coherent accumulation of optical location signals - Google Patents

Abstract

FIELD: physics.

SUBSTANCE: method comprises a series of N probing cycles. In each probing cycle, a probing light pulse S₀(t,t₀) is sent and the time is quantised into units with duration T. In each time unit, the reflected signal S(t, t_D) is received and its value S_jm is determined. Values S_jm are accumulated in each j-th range cell by forming their sums

Based on the array of sums {S_j}, the delay τ of the received signal relative the probing pulse T=t_D-t₀ is determined and the range to the target D=cτ/2 is determined. Duration t_i of the probing pulse is established. The probing pulse undergoes preliminary digitisation by determining its sampling values S_0j with sampling period equal to the clock period T and recording the array {S_0j} of those sampling values. At the end of accumulation, the array {S_j} is shifted stepwise relative {S_0j}, on each step p=1, 2, … P_max, while verifying the degree of their coincidence based on a set criterion, e.g., a correlation coefficient

The index number of the step P on which the degree of coincidence of arrays {S_j} and {S_0j} optimally correspond to the accepted criterion is determined and the range D to the target is determined using the formula D=cPT/2. The accumulation threshold is established C_N=Qσ_s+M_s.

EFFECT: achieving maximum possibility of rapid measurement of range with maximum operational range and minimum power consumption.

3 cl, 3 dwg

Description

The invention relates to laser technology, namely to laser ranging.

Known methods of laser sensing of remote objects to obtain information about their range and other characteristics. A known method for determining the distance to the target, including sending a laser pulse S ₀ (t, t ₀ ) to the target, recording the sending moment t ₀ , receiving the radiation reflected by the target S (t, t _D ), registering the receiving moment t _D and determining the time interval τ = t _D -t ₀ , which is used to judge the distance D to the target [1] according to the formula D = cτ / 2, where c is the speed of light. The disadvantage of this method is the need to use powerful emitters and highly sensitive receivers to provide a wide range of range measurements.

The energy potential of the range meter can be increased by repeatedly repeating measurements (accumulation), in particular by the method of incoherent accumulation. With this method, the detection characteristics (the probability of an erroneous measurement or signal skipping) can be improved due to the statistical processing of the results of multiple sounding based on their combination [2].

и по массиву сумм {S_j} судят о задержке τ принятого сигнала относительно зондирующего импульса τ=t_R-t₀, по которой определяют дальность до цели D=сτ/2, где с - скорость света, j=1, 2,… J - порядковый номер ячейки дальности, начиная от момента t₀, J - количество ячеек дальности; m=1, 2,… N - порядковый номер зондирования [3].The closest in technical essence to the proposed method is a method of incoherent accumulation of radar signals, including a series of N sounding cycles, in each sounding cycle, a probe light pulse S ₀ (t, t ₀ ) is sent, where t is the current time, t ₀ is the moment radiation of the probe pulse, quantize the time into separate samples of duration T, thereby forming a range cell, in each of the time samples, receive the reflected signal S (t, t _D ), where t _D is the moment of reception of the reflected pulse, op assign its value S _jm , accumulate the values of S _jm in each j-th range cell by forming their sums

and by the array of sums {S _j } judge the delay τ of the received signal relative to the probe pulse τ = t _R -t ₀ , which determines the distance to the target D = cτ / 2, where c is the speed of light, j = 1, 2, ... J is the ordinal number of the range cell, starting from the moment t ₀ , J is the number of range cells; m = 1, 2, ... N is the sounding sequence number [3].

With a large maximum range to the target, that is, with a weak signal reflected by the target, this method requires a significant amount of accumulation N and, accordingly, a large measurement time. This may be undesirable and even unacceptable with high requirements for the efficiency of measurements, for example, if the target moves or is available for observation for a short time.

The objective of the invention is to increase the efficiency of measuring ranges with maximum range and minimum energy costs.

, и по массиву сумм {S_j} судят о задержке τ принятого сигнала относительно зондирующего импульса τ=t_R-t₀, по которой определяют дальность до цели R=сτ/2, где с - скорость света, j=1, 2,… J - порядковый номер ячейки дальности, начиная от момента t₀, J - количество ячеек дальности, m=1, 2,… N - порядковый номер зондирования, устанавливают длительность t_и зондирующего импульса в пределах от 2Т до ΔT_D, где ΔТ_D - заданная разрешающая способность по дальности, предварительно производят оцифровку зондирующего импульса путем определения его выборочных значений S_0j с периодом выборок, равным тактовому периоду Т, где j=1, 2,… К - порядковый номер выборки начиная от момента t₀, К=t_и/T - количество выборок, и регистрации массива {S_0j} этих выборочных значений, а по завершении накопления пошагово сдвигают массив {S_j} относительно {S_0j}, на каждом шаге р=1, 2,… Р_mах проверяя степень их совпадения по заранее установленному критерию, например, по коэффициенту корреляции

, где Р_mах - максимальное число шагов, соответствующее диапазону измерения дальности, определяют порядковый номер шага Р, на котором степень совпадения массивов {S_j} и {S_0j} оптимально соответствует принятому критерию, и определяют дальность D до цели по формуле D=сРТ/2, причем в процессе накопления определяют оценки среднеквадратического отклонения σs выборочных значений S_jm и их среднего значения M_s, устанавливают порог накопления C_N=Qσ_S+M_S, где Q - коэффициент превышения порога над шумом, определяемый из условия обеспечения максимальной вероятности правильного обнаружения в каждой ячейке дальности при заданной вероятности ложной тревоги во всем диапазоне измеряемых дальностей, и прекращают процесс накопления при количестве зондирований N, когда хотя бы в одной из ячеек дальности накопленная сумма S_j превысит пороговую величину C_N.This problem is solved due to the fact that in the known method of incoherent accumulation of radar signals, which includes a series of N sounding cycles, a probe light pulse S ₀ (t, t ₀ ) is produced in each sounding cycle, where t is the current time, t ₀ - the moment of emission of the probe pulse are quantized into separate discrete unit time period t, thereby forming a range cell in each of the discrete-time reception is performed reflected signal S (t, t _D), where t _D - the moment of receiving the reflected pulse, determine its values s S _jm, accumulate values S _jm in each j-th cell range by forming their sum

, and by the array of sums {S _j } judge the delay τ of the received signal relative to the probe pulse τ = t _R -t ₀ , which determine the distance to the target R = сτ / 2, where c is the speed of light, j = 1, 2, ... J is the ordinal number of the range cell, starting from the moment t ₀ , J is the number of range cells, m = 1, 2, ... N is the sensing ordinal number, the duration t _{and the} probe pulse are set from 2T to ΔT _D , where ΔT _D - preset range resolution, pre-digitize the probe pulse by determining its sample values S _0j with a sampling period equal to the clock period T, where j = 1, 2, ... K is the serial number of the sample starting at time t ₀ , K = t _and / T is the number of samples, and the array {S _0j } of these sample values is registered , and upon completion of accumulation step by step shift the array {S _j } relative to {S _0j }, at each step p = 1, 2, ... P _max checking the degree of their coincidence according to a pre-established criterion, for example, according to the correlation coefficient

, where Р _max - the maximum number of steps corresponding to the range of range measurement, determine the sequence number of step Р, at which the degree of coincidence of the arrays {S _j } and {S _0j } optimally corresponds to the accepted criterion, and determine the distance D to the target using the formula D = сРТ / 2, moreover, during the accumulation process, estimates of the standard deviation σs of the selected values of S _jm and their average value of M _s are determined, the accumulation threshold is set to C _N = Qσ _S + M _S , where Q is the coefficient of exceeding the threshold over noise, determined from the condition of ensuring maximum the occurrence of correct detection in each range cell at a given probability of false alarm in the entire range of measured ranges, and the accumulation process is terminated with the number of soundings N when the accumulated sum S _j exceeds at least one of the range cells the threshold value C _N.

;

, причем количество выборочных значений в этой группе предварительно устанавливают из условия обеспечения заданной точности оценок, например, с помощью уравненияTo estimate the standard deviation σ _{S of the} sample values of S _jm and their average value of M _S, one can select in the array {S _jm }, for example, in the array {S _j1 } a group of statistically independent sample values in the amount of M, and then determine the average value estimates M _S and standard deviation σ _S according to the formulas

;

moreover, the number of sample values in this group is preliminarily set from the condition of ensuring a given accuracy of estimates, for example, using the equation

Where

q is the permissible error of the estimate;

γ is the reliability of the estimate;

- плотность распределения случайной величины %;

- distribution density of random variable%;

;

;

- оценка величины обусловленного шумом среднеквадратического разброса выборок в массиве {S_j};

- estimate the magnitude of the noise-induced mean-square spread of samples in the array {S _j };

σ> 0 is a continuous parameter having dimension σ _S.

The probability of correct detection of a signal in the entire range of measured ranges can be increased if, at the beginning of the measured range of ranges, the coefficient Q is set higher than the average value, and as the serial number j increases, it is reduced so that the probability of a false alarm in the entire measured range of ranges does not exceeded the specified limit, and the initial excess of the coefficient Q and its values corresponding to the current range are set from the condition for compliance with the required signal detection probabilities in each range cell.

Figure 1 presents a timing diagram of the sensing process and its relationship to the clock frequency. Figure 2 illustrates the formation of an array {S _j } by accumulating arrays {S _jm }. Figure 3 shows the formation of the correlation function R (p) of the arrays {S _j } and {S _0j }.

;

, где J₁ и J₂ - границы части массива, выделенной для определения M_S и σ_S1, а затем определяют величину среднеквадратического разброса

, причем количество выборочных значений в этой группе М=(J₂-J₁+1) предварительно устанавливают из условия обеспечения заданных доверительных интервалов разброса этих оценок. Например, если задана допустимая погрешность определения доверительного интервала q с надежностью оценки γ, то по известной методике [4, стр.221] можно определить необходимое для этого количество выборок М. Для этого необходимо решить уравнениеAt time t ₀ , a probe pulse S ₀ (t, t ₀ ) 1 is sent in the direction of the target, corresponding to its sample values stored in the array {S _0j } 2, and then the pulse S (t, t _D ) 3 reflected by the target is received. The radiation moment of the probe pulse 1 is tied to the clock sequence 4 by assigning to the pulse the clock sequence that coincides with the moment t ₀ , serial number j = 1. Clock pulses are generated using a highly stable source with a frequency of F _T = 1 / T, where T is the repetition period of clock pulses. Array 2 is preliminarily obtained by trial sending a probe pulse and digitizing it with a clock frequency F _T. Array 2 is stored in the system’s memory and updated in the process of preparing for measurements each time it is required by the conditions of metrological calibration. With each m-th sounding, the received signal 3 is digitized by determining its sample values S _jm at the moments of generation of each j-th clock pulse; the numbering of samples S _jm is from the clock pulse with number j = 1. The obtained sample values are recorded in arrays {S _jm } stored in the system memory (Figure 2). From the sampled values S _{j1 of the} array {S _j1 }, the statistical characteristics of the received mixture of the signal with noise are determined. To do this, select the part of the array {S _j1 }, limited by the numbers of the range cells J ₁ and J ₂ , and determine the estimates of the average value of M _S and the standard deviation σ _S1 by the formulas

;

, where J ₁ and J ₂ are the boundaries of the part of the array allocated to determine M _S and σ _S1 , and then determine the magnitude of the mean square spread

moreover, the number of sample values in this group M = (J ₂ -J ₁ +1) is preliminarily determined from the condition of ensuring given confidence intervals for the spread of these estimates. For example, if a permissible error in determining the confidence interval q with a reliable estimate of γ is given, then by the well-known method [4, p. 211], the number of samples M necessary for this can be determined. For this, it is necessary to solve the equation

Where

γ is the reliability of the estimate;

- плотность распределения случайной величины %;

- distribution density of random variable%;

;

;

- оценка величины обусловленного шумом среднеквадратического разброса выборок в массиве {S_j};

- estimate the magnitude of the noise-induced mean-square spread of samples in the array {S _j };

σ is a continuous parameter having dimension σ _S.

In practice, to determine M, the error q and the reliability of the estimate γ are preliminarily set, after which M is determined from pre-calculated dependencies, for example, from [4, p. 464, Appendix 4]. So, with a permissible twenty-percent error (q = 0.2) and reliability γ = 0.95, the size of the sample array is M = 50. If the total number of range cells in the array {S _j1 } significantly exceeds this value, then, obviously, there is no need to use this entire array and it is enough to limit it to a group of M samples. It is advisable to choose this group from the final part of the array {S _j1 }, where the magnitude of the signal reflected by the target is small and cannot distort the estimation results. After determining the statistical characteristics of M _S and σ _S , the value of C _N is determined for the accumulation threshold 6 C _N = Qσ _S + M _S , where Q is the coefficient of exceeding the threshold over noise, determined in advance from the condition for the admissible probability of a false measurement. For example, if for each measurement the probability of false alarm is set to F = 0.01, then with a uniform distribution of this parameter along the path in each range cell, the probability of false alarm should not exceed F _j = F / J, where J is the number of range cells. So, with J = 1000, the probability F _j = 10 ^-5 . The coefficient Q is determined from the condition 0.5-F (Q) = F _j , where

To determine Q from a given value of F _j, we can use the approximate relation

, с высокой точностью справедливым для Q>2.

, with high accuracy, valid for Q> 2.

For the specified value F _j = 10 ^{-5, the} value of Q ~ 4.4.

In accordance with the proposal, the coefficient Q is changed depending on the range (from the number of the range cell) so that in the far zone it decreases, providing a higher probability of signal detection. Moreover, the total probability of false alarm in the entire range does not worsen.

Example.

The number of range cells J = 1000. These cells are divided into two subranges - the near (I) in the amount of 800 cells and the far (II) in the amount of 200 cells. Probabilities of false alarm in the subbands F _I = 0.002, F _II = 0.008. F _I + F _II = 0.01, as in the previously considered case with a uniform probability distribution F over range cells. In this example, in the far subband II, the probability F _{j (II)} = F _II / 200 = 0.008 / 200 = 4 · 10 ^-5 . In subband I, respectively, F _{j (II)} = F _I / 800 = 0.002 / 800 = 2.5 · 10 ^-6 . The coefficients Q for these probabilities are respectively equal:

Q _(II) = 4; Q _(I) = 4.6.

This means that in the far part of the range of measured ranges, the sensitivity improves in Q / Q _(II) = 4.4 / 4 = 1.1 times, which leads to a corresponding increase in the range of the rangefinder. In the near part of the range, the sensitivity deteriorates by Q _(I) / Q = 4.6 / 4.4 = 1.05 times, which does not affect the quality of measurements, since this deterioration is compensated by a more significant increase in the received signal: it is known [1] that for small-sized targets, its amplitude is inversely proportional to the fourth power of range, therefore, the ratio of signal amplitudes corresponding to the 800th and 1000th range cells is (1000/800) ⁴ ~ 2.5 times, which significantly exceeds the loss due to an increase in Q in this area.

. Затем определяют шаг р=Р, на котором функция R(p) максимальна, и вычисляют дальность D до цели по формуле D-сРТ/2, где с - скорость света.At the end of the process of accumulation and allocation of the array {S _j } it is compared with the array {S _0j } according to a predetermined rule, for example, by determining the correlation coefficients R (p) of the arrays {S _j } and {S _0j } with a stepwise shift of the second array relative to the first at p = 1,2, ... P _max steps [5]. The result is a correlation function

. Then determine the step p = P, at which the function R (p) is maximum, and calculate the distance D to the target using the formula D-cPT / 2, where c is the speed of light.

To reduce the measurement time, it is recommended to search for the sequence number of step P, at which the value of R (P) is optimal, by phased approximation by the method of dividing the segment in half, namely, at each stage of the search, estimates of R (p _n ) and R (p _in ) are determined lower p _n and upper p _{within the} boundaries of the search zone and at the next stage, the search zone is set between p ~ (p _in + p _n ) / 2 and that of the extreme values p of the search zone, the estimate of which r (p) at this approximation step is most optimal , the search in the first stage p _n = p (S _C) -t _aND / T, and p _a = p _{and n} + 2t / T, where p (S _C) - the number of the range cell in which the accumulated sum S _j = S _C exceeded the set threshold C _N.

The proposed method of incoherent accumulation of radar signals provides a range measurement with a minimum accumulation volume, with a minimum amount of statistical data used to determine the optimal accumulation mode, as well as for a minimum number of steps. The decrease in the accumulation volume N is also significantly facilitated by the increase in the energy potential of the device, due to the increase in the duration of the probe pulse within the limits allowed by the given resolution: this increases the energy of the probe signal and the noise in the receiving channel due to the corresponding narrowing of its passband. As a result, the maximum possible range measurement efficiency is achieved at the maximum range and minimum energy costs.

Information sources

1. V.A. Volokhatyuk, V.M. Kochetkov, P.P. Krasovsky "Issues of optical location". M., Publishing House "Soviet Radio", 1971, p.176.

2. Ya. D. Shirman, VN Manzhos “Theory and technique of processing radar information against the background of interference”. M., Publishing House "Radio and Communications", 1981, S. 70, 81.

3. RF patent No. 2359226 according to the application 2007137271/28 from 10.10.2007. "The method of incoherent accumulation of radar signals." - The prototype.

4. V.E. Gmurman "Probability Theory and Mathematical Statistics", Moscow, Publishing House "Higher School", 1977

5. Patent US No. 5805468.

Claims (3)

1. A method of incoherent accumulation of radar signals, including a series of N sounding cycles, in each sounding cycle, a probe light pulse S ₀ (t, t ₀ ) is sent, where t is the current time, t ₀ is the moment of radiation of the sounding pulse, time is quantized by individual samples of duration T, thereby forming a range cell, in each of the time samples receive the reflected signal S (t, t _D ), where t _D is the moment of reception of the reflected pulse, determine its value S _jm , accumulate the values of S _jm in each j far cell aws by forming their sums

, and from the array of sums {S _j } judge the delay τ of the received signal relative to the probe pulse τ = t _D -t ₀ , which determine the distance to the target D = cτ / 2, where c is the speed of light, j = 1, 2, ... J is the ordinal number of the range cell, starting from the moment t ₀ ; J is the number of range cells; m = 1, 2, ... N is the sensing sequence number, characterized in that the duration t _{and the} probe pulse are set in the range from 2T to ΔT _D , where ΔT _D is the preset range resolution, the probe pulse is first digitized by determining its sample values of S _0j with a sampling period equal to the clock period T, where j = 1, 2, ... K is the serial number of the sample, starting from the moment t ₀ ; K = t _and / T is the number of samples and registration of the array {S _0j } of these sample values, and upon completion of accumulation, the array {S _j } is shifted step by step relative to {S _0j }, at each step p = 1, 2, ... P _max checking the degree of their coincidence according to a predetermined criterion, for example, by correlation coefficient

where P _max is the maximum number of steps corresponding to the range of range measurement, determine the sequence number of step P, at which the degree of coincidence of the arrays {S _j } and {S _0j } optimally matches the accepted criterion, and determine the distance D to the target using the formula D = cPT / 2, and during the accumulation process, estimates of the standard deviation σ _{S of the} selected values of S _jm and their average value M _S are determined, the accumulation threshold C _N = Qσ _S + M _{S is set} , where Q is the coefficient of exceeding the threshold over noise, determined from the condition of ensuring the maximum probability the correct detection in each range cell at a given probability of false alarm in the entire range of measured ranges, and stop the accumulation process with the number of soundings N, when the accumulated sum S _j exceeds at least one of the range cells the threshold value С _N. 2. The method according to claim 1, characterized in that in one of the arrays {S _jm }, for example, in the array {S _j1 }, a group of statistically independent sampled values is selected in the amount of M, and then estimates of the mean value of M _S and standard deviation are determined σ _S by the formulas

;

moreover, the number M of sample values in this group is preliminarily set from the condition of ensuring a given accuracy of estimates, for example, using the equation

Where
q is the permissible error of the estimate;
γ is the reliability of the estimate;

is the distribution density of the random variable χ.

- estimate the magnitude of the noise-induced mean-square spread of samples in the array {S _j };
σ> 0 is a continuous parameter having dimension σ _S. 3. The method according to claim 1, characterized in that at the beginning of the measured range of ranges the coefficient Q is set in excess of the average value, and as the serial number j increases, it is reduced so that the probability of a false alarm in the entire measured range of ranges does not exceed the specified the limit, and the initial excess of the coefficient Q and its values corresponding to the current range are set from the condition for observing the required probability of detecting a signal in each of the range cells.

Images