A Photon Quantum Adaptive Ranging Method in Harsh Environments

Abstract

The present invention proposes a method for optical quantum adaptive ranging in harsh environments. First, the method uses pump light to generate idler photons and signal photons through the process of parametric down-conversion. Then, they are separated by a dual-wavelength polarization beam splitter and a beam expander. The idler light is directly detected by a local single-photon detector, and the signal light is sent to the target to be measured and reflected back to the local area for detection by another single-photon detector. Secondly, the errors caused by the environmental noise of the ranging system and the inherent errors of the optical path are measured. Thirdly, the arrival times of the idler photons and signal photons are grouped, and their relative quantum efficiency is calculated, and the group length is dynamically adjusted using the relative quantum efficiency. Then, a bidirectional long short-term memory network is used to judge the reliability of each group of data. Finally, the groups with higher reliability scores are selected, and the target distance is estimated using the coincidence counting algorithm.

Description

Technical Field

[0001] The present invention relates to the field of optical ranging, and particularly to a method for improving the high-precision ranging of optical quanta in harsh environments through reliable packet data selection. Background Art

[0002] Ranging technology, as a basic technology for realizing the positioning and navigation of unknown targets, is closely related to people's lives and national defense construction. With the continuous development of technology and society, people have higher and higher requirements for the accuracy and resolution of target detection, especially in the fields of precision instrument processing, manufacturing and assembly of large components, aerospace exploration, wind power generation, etc.

[0003] Due to the small divergence angle, concentrated energy, and strong anti-interference ability of laser propagation in the atmosphere, laser ranging technology is widely used in long-distance and high-precision scenarios such as earth gravity detection and space distance measurement. Traditional laser ranging technology mainly uses the excellent coherence and directivity of laser for high-precision measurement, mainly including optical pulse ranging, phase modulation ranging, and optical interference ranging. Optical pulse ranging maintains good performance during long-distance transmission, and the measurement method is simple, but the accuracy is not high; the accuracy of phase modulation ranging technology is limited by the laser repetition frequency, and there is a cycle ambiguity error; although the accuracy of optical interference ranging is high, the conditions for the establishment of interference are relatively harsh, and it has no ability to distinguish background light and signal light, so the application scenario is very limited.

[0004] In recent years, with the rapid development of lasers, single photon detectors (Single Photon Detector, SPD), and time digital conversion technology (Time Digital Convert, TDC), the arrival time of a single photon can achieve a measurement accuracy of picoseconds, which has greatly promoted the rapid development of precision ranging technology based on single photon measurement. Single photon ranging technology is developed on the basis of photon counting technology and has been widely used in fields such as low-light-level imaging, satellite altimetry, and remote sensing imaging. After detecting low-intensity, high-repetition-rate single photon optical pulses through high-precision SPD, this technology mainly uses the time-correlated single photon counting method to analyze the relevant statistical characteristics of single photon pulses to achieve the detection of targets. However, the realization of high precision and high resolution of this technology depends on the accumulation of photons on a very long time scale (for example, several seconds), which is not conducive to improving the stability and real-time performance of the ranging system. In this case, entangled photon ranging technology plays an increasingly important role in related fields.

[0005] The ranging technology based on quantum entanglement mainly realizes ranging by using the structures of Michelson, HOM (Hong-Ou-Mandel), and HBT (Hanbury Brown Twiss) interferometers. There is an integer ambiguity problem in the correlation measurement of entangled twin photons using a Michelson interferometer; the HOM interferometer can achieve attosecond time resolution, but this method depends to a certain extent on an optical delay line and is vulnerable to phase noise interference; the ranging method based on the HBT interferometer uses the correlation measurement of optical field intensity fluctuations to replace the measurement of the optical field intensity distribution. In essence, it measures the second-order correlation function of entangled photon pairs and can achieve picosecond time resolution on a 3-km optical fiber. The entanglement ranging method based on the HBT interferometer is insensitive to phase noise, and there is a unique correspondence between the maximum value of its second-order correlation function and the optical path difference. It has been widely used in quantum relay schemes for long-distance transmission. The above schemes mainly explore the theoretical limit of entanglement ranging in a laboratory environment, and there are still many challenges in practical entanglement ranging. The preparation of the entanglement source, the distribution of entangled photons, device performance, and many other factors limit the ranging accuracy of the entanglement ranging system. Summary of the Invention

[0006] The object of the present invention is to provide an optical quantum adaptive ranging method for harsh environments. This method separates entangled photons and noise photons by designing a data grouping algorithm and a Bi-directional Long Short-Term Memory (Bi-LSTM) network, which can effectively handle data anomalies in harsh environments and has good ranging accuracy and robustness in different test environments.

[0007] The technical solution adopted by the present invention is as follows: an optical quantum adaptive ranging method based on relative quantum efficiency and a Bi-LSTM network, specifically including the following steps:

[0008] Step 1: Use a continuous narrow-linewidth semiconductor laser (SL) with a bandwidth of 160 MHz to generate pump light with a wavelength of 405 nm.

[0009] Step 2: Make the 405-nm pump light sequentially pass through a half-wave plate (HWP), a quarter-wave plate (QWP), a polarizing beam splitter (PBS), and a dual-wavelength reflector (DR) to complete the filtering and horizontal polarization of the pump light.

[0010] Step 3: Let the filtered and horizontally polarized pump light enter the Sagnac loop, and perform Spontaneous Parametric Down Conversion (SPDC) through a Periodically Poled KTP (PPKTP) crystal. During the SPDC process, a pump photon with a wavelength of 405 nm splits into two photons with a wavelength of 810 nm after passing through the PPKTP crystal. These two photons are in a polarization entangled state after passing through the Sagnac interference ring structure;

[0011] Step 4: Let the entangled photon pair pass through a Dual-wavelength Polarizing Beam Splitter (DPBS), and separate it into an idler photon and a signal photon using the polarization state;

[0012] Step 5: Let the idler photon pass through the DR, be coupled into the optical fiber through an Optical Collimator (OC), and finally be detected by SPD1. The signal photon is emitted after passing through the PBS, QWP, Beam Expander (BE) and mirror, and is finally detected by SPD2 after being reflected by the target;

[0013] Step 6: Save the arrival time sequences of the signal photon and the idler photon locally through the TDC technology, denoted as CH1 and CH2 respectively;

[0014] Step 7: Measure the error caused by noise in the environment where the ranging system is located and the inherent error of the optical path;

[0015] Step 8: After grouping the data of CH1 and CH2, calculate the relative quantum efficiency of each group, and dynamically adjust the length of each group using the relative quantum efficiency;

[0016] Step 9: Plot the coincidence counting curve of each group of data, input it into the Bi-LSTM network for reliability judgment, and return the reliability score c;

[0017] Step 10: Screen the groups with c≥0.8. First, count the peak offset of the coincidence counting function of these groups, then subtract the error caused by noise in the environment and the calculated inherent error of the optical path and perform filtering processing, and finally estimate the target distance.

[0018] The steps in Step 7 include the following steps:

[0019] Step 7(1): Place the mirror directly in front of the BE, and record the arrival time sequences CH1 1 and CH2 1 ;

[0020] Step Seven (2), place the mirror directly in front of BE at a distance d 0 and record the arrival time sequences CH1 2 and CH2 2 of the idle photons and the signal photons respectively.

[0021] The eighth step includes the following steps:

[0022] Step Eight (1), set the single - test time to T c , the time length T g of a cluster, the number of packets p contained in a single cluster, and the threshold value α;

[0023] Step Eight (2), divide the arrival time sequences CH1 and CH2 of the idle photons and the signal photons into q = T c / T g clusters, each cluster containing p packets, which can be expressed as:

[0024]

[0025]

[0026] Step Eight (3), calculate the relative quantum efficiency η g in each packet:

[0027]

[0028] where n g (g = 1, 2, … p) is the number of photon pairs in the g - th packet, and N k (k = 1, 2, … q) is the number of photons in each cluster;

[0029] Step Eight (4), store the relative quantum efficiencies of each packet in the p×q matrices E and F:

[0030]

[0031]

[0032] Step Eight (5), calculate the sum of variances v of matrices E and F. When v ≤ α, retain the packets; otherwise, let T g = T g + 100, and return to Step Eight (2).

[0033] The ninth step includes the following steps:

[0034] Step Nine (1), assume that the time resolution of the SPD is T s , the length of the packet data is T, E = {t 1,1 t1,2 …t 1,i …t 1,m} and F = {t 2,1 t 2,2 …t 2,j …t 2,n}, set the iteration step size in the coincidence counting process to s, the coincidence gate width to δ, and denote the lengths of E and F as m and n respectively;

[0035] Step Nine (2), initialize the coincidence count value to cn = 0; initialize the variable kk = -1, the number of iterations where represents rounding down;

[0036] Step Nine (3), let kk = kk + 1, i = 0, if kk ≤ k, τ kk = s·kk, execute Step Nine (4), otherwise execute Step Nine (6);

[0037] Step Nine (4), let j = 0, i = i + 1; if i ≤ m, execute Step Nine (5); if i > m, let the coincidence count value count(kk) = cn, and execute Step Nine (3);

[0038] Step Nine (5), let j = j + 1; if j ≤ n, judge whether |t 2,j -τ kk -t 1,i | ≤ δ and |t 2,j -t 2,j-1 | < T s hold simultaneously. If they hold simultaneously, let cn = cn + 1 and then enter Step Nine (4), otherwise enter Step Nine (5); if j > n, execute Step Nine (4);

[0039] Step Nine (6), draw the curve of count and input it into the Bi-LSTM network to calculate the reliability score c.

[0040] The steps in Step Ten include the following steps:

[0041] Step Ten (1), calculate the peak offset τ 2 of the coincidence calculation function count for CH1 2 and CH2 1 and τ 2 ;

[0042] Step Ten (2), use the coincidence counting algorithm to calculate the peak offset τ r (r = 1, 2,... R) of the count of R reliable grouped data, and perform filtering processing;

[0043] Step Ten (3), solve the distance d of the target:

[0044]

[0045] where c = 3×10 8 m / s represents the speed of light. Description of the Drawings

[0046] Figure 1 is the optical path diagram of the optical quantum ranging of the present invention

[0047] Figure 2 is the curve diagram of the coincidence counting function count of the present invention;

[0048] Figure 3 is the Bi-LSTM network structure diagram of the present invention;

[0049] Figure 4 is the ranging error diagram of the present invention Specific Embodiment

[0050] The present invention will be further described in detail below with reference to the accompanying drawings:

[0051] Step 1: Use a continuous narrow linewidth semiconductor laser (SL) with a bandwidth of 160 MHz to generate pump light with a wavelength of 405 nm;

[0052] Step 2: Make the 405 nm pump light pass through a half-wave plate (HWP), a quarter-wave plate (QWP), a polarizing beam splitter (PBS), and a dual-wavelength reflector (DR) in sequence to complete the filtering and horizontal polarization of the pump light;

[0053] Step 3: Make the filtered and horizontally polarized pump light enter the Sagnac loop and undergo spontaneous parametric down conversion (SPDC) through a periodically poled potassium titanyl phosphate (PPKTP) crystal. In the SPDC process, a pump photon with a wavelength of 405 nm is split into two photons with a wavelength of 810 nm after passing through the PPKTP crystal. These two photons are in a polarization-entangled state after passing through the Sagnac interference ring structure. They are unpredictable and indistinguishable in time and polarization state, and satisfy conservation in energy and momentum, that is, the frequency satisfies w p = w i + w s , and the wave vector satisfies

[0054] Step 4: Pass the entangled photon pair through a Dual-wavelength Polarizing Beam Splitter (DPBS), and separate it into an idler photon and a signal photon by using the polarization state;

[0055] Step 5: Let the idler photon pass through the DR, then couple it into an optical fiber through an Optical Collimator (OC), and finally detect it by SPD1. The signal photon is emitted after passing through the PBS, QWP, Beam Expander (BE) and a mirror, and is finally detected by SPD2 after being reflected by the target. Among them, the probability of the SPD detecting a photon can be expressed as:

[0056]

[0057] where t is the time when the photon is detected, d is the distance between the SPD and the light source, and are the positive-frequency term and the negative-frequency term of the electric field , satisfying Since the electric field has been quantized, so and can be expressed in terms of the number of photons per second, represents the expectation. Since the signal photon and the idler photon are a pair of photons entangled in frequency, according to quantum field theory, the probabilities of detecting the signal photon and the idler photon at distances d 1 and d 2 from the light source are proportional to the second-order correlation function of the optical field intensity fluctuations:

[0058]

[0059] where t 1 , t 2 are the times when the signal photon and the idler photon are detected respectively. After normalizing the second-order correlation function, it can be expressed as:

[0060]

[0061] Step 6: Save the arrival time sequences of the signal photon and the idler photon locally through the TDC technology, denoted as CH1 and CH2 respectively;

[0062] Step 7: Measure the error caused by noise in the environment where the ranging system is located and the inherent error of the optical path;

[0063] Step 8: After grouping the data of CH1 and CH2, calculate the relative quantum efficiency of each group, and dynamically adjust the length of each group by using the relative quantum efficiency;

[0064] Step 9. In practice, a coincidence counting algorithm is often used to screen entangled photon pairs, instead of the relevant measurements on the photon field. By moving CH2 with different time delays τ, the coincidence counting function count(τ) at different time delays can be obtained. When the coincidence gate width δ is much smaller than the coherence time τ of the optical field c , count(τ) and the second-order correlation function g (2) (d 1 , d 2 , t 1 , t 2 ) satisfy the relationship:

[0065]

[0066] where T c is the data sampling time, R 1 and R 2 are the photon counting rates of SPD1 and SPD2 respectively; plot the coincidence counting curves of each group of data and input them into the Bi-LSTM network for reliability judgment, and return the reliability score c;

[0067] Step 10. Screen the groups with c≥0.8. First, count the peak offset of the coincidence counting function of these groups, then subtract the error caused by noise in the environment and the inherent error of the optical path, calculate and perform filtering processing, and finally estimate the target distance.

[0068] The seventh step includes the following steps:

[0069] Step 7(1). Place the mirror directly in front of BE, and record the arrival time sequences CH1 1 and CH2 1 of the idler photons and the signal photons respectively;

[0070] Step 7(2). Place the mirror at a distance d 0 in front of BE, and record the arrival time sequences CH1 2 and CH2 2 of the idler photons and the signal photons respectively.

[0071] The eighth step includes the following steps:

[0072] Step 8(1). Set the single test time as T c , the time length T g of a cluster, the number p of packets contained in a single cluster, and the threshold value α;

[0073] Step 8(2). Divide the arrival time sequences CH1 and CH2 of the idler photons and the signal photons into q = T c / T g clusters, each cluster contains p packets, which can be expressed as:

[0074]

[0075]

[0076] Step Eight (3), calculate the relative quantum efficiency η in each group g :

[0077]

[0078] where n g (g = 1, 2, … p) is the number of photon pairs in the g-th group, and N k (k = 1, 2, … q) is the number of photons in each cluster;

[0079] Step Eight (4), store the relative quantum efficiencies of each group in the p×q matrices E and F:

[0080]

[0081]

[0082] Step Eight (5), calculate the sum of variances v of matrices E and F. When v ≤ α, retain the groups; otherwise, let T g = T g + 100, and return to Step Eight (2).

[0083] The following steps are included in Step Nine:

[0084] Step Nine (1), assume the time resolution of the SPD is T s , the length of the grouped data is T, E = {t 1,1 t 1,2 …t 1,i …t 1,m} and F = {t 2,1 t 2,2 …t 2,j …t 2,n}, set the iteration step size in the coincidence counting process as s, the coincidence gate width as δ, and denote the lengths of E and F as m and n respectively;

[0085] Step Nine (2), initialize the coincidence count value as cn = 0; initialize the variable kk = -1, and the number of iterations where represents rounding down;

[0086] Step Nine (3), let kk = kk + 1, i = 0. If kk ≤ k, τ kk = s·kk, execute Step Nine (4); otherwise, execute Step Nine (6);

[0087] Step Nine (4): Let j = 0 and i = i + 1; if i ≤ m, execute Step Nine (5); if i > m, let the compliance count value count(kk) = cn, and execute Step Nine (3);

[0088] Step Nine (5): Let j = j + 1; if j ≤ n, judge whether |t 2,j -τ kk -t 1,i | ≤ δ and |t 2,j -t 2,j-1 | < T s hold simultaneously. If they hold simultaneously, let cn = cn + 1 and then enter Step Nine (4); otherwise, enter Step Nine (5); if j > n, execute Step Nine (4);

[0089] Step Nine (6): Plot the curve of count and input it into the Bi-LSTM network to calculate the reliability score c.

[0090] The following steps are included in Step Ten:

[0091] Step Ten (1): Calculate the peak offset τ 2 of CH1 2 and CH2 1 that conform to the calculation function count, and τ 2 ;

[0092] Step Ten (2): Use the compliance counting algorithm to calculate the peak offset τ r (r = 1, 2,..., R) of count for R reliable grouped data, and perform filtering processing;

[0093] Step Ten (3): Solve the distance d of the target:

[0094]

[0095] where c = 3×10 8 m / s represents the speed of light.

Claims

1. A method for optical quantum adaptive ranging in harsh environments, characterized in that it includes the following steps: Step 1: Use a continuous narrow-linewidth semiconductor laser SL with a bandwidth of 160 MHz to generate pump light with a wavelength of 405 nm; Step 2: Let the 405-nm pump light pass through a half-wave plate HWP, a quarter-wave plate QWP, a polarization beam splitter PBS, and a dual-wavelength mirror DR in sequence to complete the filtering and horizontal polarization of the pump light; Step 3: Let the filtered and horizontally polarized pump light enter the Sagnac loop and undergo spontaneous parametric down-conversion SPDC through a periodically poled potassium titanyl phosphate PPKTP crystal; during the SPDC process, a 405-nm pump photon splits into two 810-nm photons after passing through the PPKTP crystal, and these two photons are in a polarization-entangled state after passing through the Sagnac interference ring structure; Step 4: Let the entangled photon pair pass through a dual-wavelength polarization beam splitter DPBS and separate it into an idler photon and a signal photon using the polarization state; Step 5: Let the idler photon pass through DR, be coupled into the optical fiber through a collimator OC, and finally be detected by SPD1; the signal photon is emitted after passing through PBS, QWP, a beam expander BE, and a mirror, and is finally detected by SPD2 after being reflected by the target; Step 6: Save the arrival time sequences of the signal photon and the idler photon locally through TDC technology, denoted as CH1 and CH2 respectively; Step 7: Measure the error caused by noise and the inherent error of the optical path in the environment where the ranging system is located; Step 8: After setting the cluster length, the number of groups, and the threshold value, group the CH1 and CH2 data and calculate the relative quantum efficiency of each group; if the sum of the variances of the relative quantum efficiency matrix is less than the threshold value, it is judged that the grouping is valid; Otherwise, increase the cluster length and re-group, repeat this process until the variance is less than the threshold value; Step 9: Plot the coincidence counting curve of each grouped data and input it into the Bi-LSTM network for reliability judgment, and return the reliability score c; Step 10: Screen the groups with c≥0.8, first count the peak offset of the coincidence counting function of these groups, then subtract the error caused by noise and the inherent error of the optical path in the environment after calculation and perform filtering processing, and finally estimate the target distance.

2. A method for optical quantum adaptive ranging in harsh environments according to claim 1, characterized in that: the following steps are included in step 7: Step Seven (1): Place the mirror directly in front of BE, and record the arrival time series CH1 of the idle photons and the signal photons respectively 1 and CH2 1 ; Step Seven (2): Place the mirror directly in front of BE at a distance d 0 and record the arrival time series CH1 2 of the idle photons and the signal photons respectively 2 .

3. A method for optical quantum adaptive ranging in harsh environments according to claim 1, characterized in that: the following steps are included in step 8: Step VIII (1), set the single test time as T c , the time length T of a cluster g , the number p of groups contained in a single cluster and the threshold α; Step VIII (2), divide the arrival time sequences CH1 and CH2 of the idle photons and the signal photons into q = T c / T g clusters, each cluster containing p packets, which can be expressed as: Step VIII (3), calculating the relative quantum efficiency η in each group g :[[]]END]] where n g (g = 1, 2, … p) is the number of photon pairs in the g-th group, and N k (k = 1, 2, … q) is the number of photons in each cluster; Step 8(4): Store the relative quantum efficiency of each group in matrices E and F of p×q; Step VIII (5), calculate the sum of variances v of matrices E and F. When v ≤ α, retain the grouping; otherwise, let T g = T g + 100, and return to Step VIII (2).

4. A method for optical quantum adaptive ranging in harsh environments according to claim 1, characterized in that: the following steps are included in step 9: Step Nine (1). Assume that the time resolution of the SPD is T s , the length of the grouped data is T, E = {t 1,1 t 1,2 … t 1,i …t 1,m} and F = {t 2,1 t 2,2 … t 2,j … t 2,n}, set the iteration step size in the coincidence counting process as s, the coincidence gate width as δ, and the lengths of E and F are denoted as m and n respectively; Step Nine (2), initialize the compliance count value to cn = 0; initialize the variable kk = -1, the number of iterations wherein represents rounding down; Step Nine (3), let \(kk = kk + 1\), \(i = 0\). If \(kk\leq k\) and \(\tau\) kk \(= s\cdot kk\), execute Step Nine (4); otherwise, execute Step Nine (6). Step 9(4): Let j = 0, i = i + 1; if i≤m, execute step 9(5); if i>m, let the coincidence count value count(kk)=cn, and execute step 9(3); Step Nine (Five), let j = j + 1; if j ≤ n, determine whether |t 2,j -τ kk -t 1,i | ≤ δ and |t 2,j -t 2,j-1 | < T s hold simultaneously. If they hold simultaneously, let cn = cn + 1 and then enter Step Nine (Four); otherwise, enter Step Nine (Five). If j > n, execute Step Nine (Four); Step 9(6): Plot the curve of count and input it into the Bi-LSTM network to calculate the reliability score c.

5. A method for optical quantum adaptive ranging in a harsh environment according to claim 1, characterized in that: the following steps are included in step ten: Step Ten (1), calculate CH1 2 and CH2 2 to conform to the peak offset τ of the calculation function count 1 and τ 2 ; Step Ten (Two), calculate the peak offset τ of the count of R reliable packet data using the coincidence counting algorithm r (r = 1, 2, … R), and perform filtering processing; step ten (iii), let c represent the speed of light, and the distance d of the target can be expressed as: Among them, the magnitude of the speed of light is 3×10 8 m / s.

Images