Laser ranging device based on FPGA and single photon avalanche diode

Abstract

The application provides a laser ranging device based on FPGA and single photon avalanche diode. The ranging device is characterized by comprising a field programmable logic gate array calculation core (1), a quenching circuit (2), a single photon avalanche diode (3), a pulse laser driving circuit (4), a pulse laser (5) and an optical lens system (6), wherein the field programmable logic gate array calculation core (1) comprises time-to-digital conversion logic (11) and calculation logic (12), the time-to-digital conversion logic comprises correction algorithm logic (111) and carry chain logic (112), the device adopts the single photon avalanche diode and dark count correction algorithm to solve the problems of low precision, low linearity and short measuring distance of the existing ranging device. The application can be used for laser detection distance and can be widely used in the fields of automobile electronics, medical treatment, industrial robots, national defense and aerospace.

Description

(I)TECHNICAL FIELD

[0001] The application relates to a laser ranging device based on FPGA and single photon avalanche diode, which can be used in the fields of laser radar, industrial ranging and aerospace ranging and belongs to the technical field of photoelectric detection. (II)BACKGROUND

[0002] The single photon detection technology adopted by the device is a photoelectric detection technology, and the principle is to count single photons by using photoelectric effect to realize detection of extremely weak signals. It is the core of extremely weak light measurement such as laser ranging, DNA sequencing, quantum key distribution, laser radar and medical imaging. With the progress of semiconductor process technology and the improvement of integrated circuit design level, it is possible to design single photon avalanche diodes using standard integrated circuit technology and integrate single photon avalanche diodes and related integrated circuits such as avalanche event sensing circuits, quenching circuits and back-end signal processing circuits on the same chip. Compared with traditional single photon detection devices such as photomultiplier tubes (PMT) or microchannel plates (MCP), such products operate at a lower bias voltage, have a smaller volume, lower power consumption, are not sensitive to electromagnetic noise, have high integration, low cost and good repeatability, and have a broad application prospect.

[0003] In a single photon avalanche diode-based photon detection / imaging system, the system detects and counts the avalanche events excited by the hole-electron pairs generated inside the single photon avalanche diode. The avalanche breakdown events in the single photon avalanche diode can not only be excited by photon absorption, but also by other factors. The avalanche events excited in the absence of light are irrelevant to photon absorption, and the counting of these avalanche events is called dark counting. The dark counting rate is an important parameter of the single photon avalanche diode. High dark counting rate will reduce the signal-to-noise ratio of the single photon system, increase the system error rate, increase the integration (exposure) time required by the imaging system and reduce the detection sensitivity. Dark counting has random characteristics and is approximately uniformly distributed in the entire time range. The device using SPAD for ranging uses direct time-of-fly (DTOF) technology. This technology needs to measure the time difference between the signal sent by the laser and the received signal, and convert it into a digital signal using time-to-digital conversion (TDC) for subsequent algorithm processing. The current mainstream solutions include 1) using an external integrated TDC chip, and 2) using FPGA internal delay chain resources. The use of FPGA internal delay chain resources requires correction of the delay chain, and average correction counting or code density correction technology is usually used.

[0004] In the scheme using external integrated TDC chip, the measurement frame rate is usually limited to about 100 frames due to the communication limitation with the external chip, and an additional dedicated chip is needed, so the system complexity and cost are high. In the scheme using FPGA internal delay chain resource, a higher measurement frame rate can be achieved. In the average correction scheme, the linearity is poor, and linear measurement cannot be achieved. In the traditional code density correction scheme, if the internally generated random signal is used as the correction source, the correction scheme will be started when the chip is powered on for the first time, and the linearity error will still occur when the chip temperature changes. If the correction is performed after the temperature changes, an additional temperature sensor and temperature monitoring logic need to be added. The fixed time correction scheme cannot measure during the correction process.

[0005] The present application proposes a laser ranging device based on FPGA and single photon avalanche diode, which uses FPGA internal delay chain resource to realize TDC, improves the code density correction scheme, and is different from the traditional internal random signal generation scheme. The dark count of SPAD in the ranging interval time is used as the random signal source for code density correction. The probability histogram is counted during ranging, and the correction parameters are updated after 10000 times of counting. The improved code density correction scheme for this device is called dark count correction scheme. The device can continuously correct the linearity without affecting real-time measurement, and greatly enhances the linearity and stability of the system without increasing the system complexity. (III) SUMMARY

[0006] The present application aims to provide a low-noise single-photon detection system, which comprises a field programmable logic gate array computing core (1), a quenching circuit (2), a single photon avalanche diode (3), a pulsed laser driving circuit (4), a pulsed laser (5), and an optical lens system (6). The field programmable logic gate array computing core (1) comprises time-to-digital conversion logic (11) and calculation logic (12). The time-to-digital conversion logic comprises correction algorithm logic (111) and carry chain logic (112).

[0007] The purpose of the application is achieved: the ranging device is composed of a field programmable logic gate array computing core (1), a quenching circuit (2), a single photon avalanche diode (3), a pulsed laser driving circuit (4), a pulsed laser (5), an optical lens system (6), wherein the field programmable logic gate array computing core (1) contains time-to-digital conversion logic (11), computing logic (12), wherein the time-to-digital conversion logic contains correction algorithm logic (111), carry chain logic (112); the system generates a signal to make the pulsed laser driving circuit (4) drive the pulsed laser (5) to generate pulsed laser, the pulsed laser is emitted by the optical lens system (6) to produce high-power collimated pulsed laser, the laser is reflected by the target object and passes through the optical lens system (6) to cause the single photon avalanche diode (3) to avalanche breakdown, and the quenching circuit (2) is used to quench the single photon avalanche diode and convert the avalanche breakdown signal into a pulse square wave and transmit it to the field programmable logic gate array computing core (1). The time-to-digital conversion logic (11) of the field programmable logic gate array computing core (1) directly converts the signal transmitted to the laser pulse driving circuit (4) to digital quantity through the quenching circuit (2), and the computing logic (12) converts the time into digital statistics and calculates the distance. The time-to-digital conversion logic mainly records the time difference by the carry chain logic (112), and the correction algorithm logic (111) corrects the linearity of the output digital quantity time difference, wherein the correction algorithm logic needs to be driven by a random signal, and the random signal source contains two, the first is the random signal generated by the programmable logic gate array computing core (1) itself, and the second is the dark count of the single photon avalanche diode (3) after being processed by the quenching circuit (2); both kinds of random signals have randomness; after the above process and algorithm, the device finally outputs distance information with high accuracy and good linearity.

[0008] The device uses a single photon avalanche diode and a quenching circuit combination as a laser pulse receiving part, when a reverse bias voltage higher than the avalanche voltage is applied across the SPAD, the avalanche breakdown effect is used, a high internal current gain is obtained, and a chain reaction is triggered. At this time, a larger current is generated, quenched by the rear-end quenching circuit, and an avalanche breakdown signal is output. However, in the semiconductor, electron-hole pairs can be generated by heat, even without any photons, that is, in the dark condition, the avalanche will also be triggered, and this avalanche is called dark noise. This avalanche has strong randomness and can be approximately understood as uniformly distributed in the entire time range. The output digital signal is output after the quenching circuit, and the behavior of the signal is called dark count. Dark count will have a certain impact on the ranging performance, but dark count cannot be completely eliminated. A 905nm waveband infrared laser is used as the generating device, and a pulsed laser driving circuit is used to drive the laser to generate a high-power laser pulse signal with a width of less than 5ns.

[0009] The time-to-digital converter (TDC) function is implemented using the internal carry chain resources of the FPGA. The FPGA uses the Xilinx A7 series, and the carry chain is implemented by cascading internal CARRY4. One CARRY4 outputs four taps (CO0, CO1, CO2, CO3) corresponding to different delays. In the design, 50 delay units are used to generate 200 taps, and there are 200 time units in total, with an average delay of about 33 ps. The entire TDC conversion logic includes coarse counting and fine counting. The coarse counting clock period roughly records the number of clock periods t1 in the time difference, and the fine counting is used to reflect the fixed time difference between the clock edge and the real edge. This time is reflected by the number of carry chain time units. The final calculation method is:

[0010] T = kt1 + iτ - jτ

[0011] In the formula, T represents the time difference, k is the number of coarse counting times, i is the number of carry chain time units of the start signal, j is the number of carry chain time units of the end signal, and τ is the average delay of the carry chain time unit.

[0012] However, due to the internal structure characteristics of the delay chain, there is a time deviation between different time units, and it will change with the change of chip temperature. Generally, for scenes with low linearity requirements, the average correction method is used, and for scenes with high linearity requirements, the code density method is generally used for linearity correction. The ranging device has high linearity requirements, and the linearity correction scheme based on the improved code density method is used. Code density method needs to input a group of random signals, which randomly fall in a certain position in the delay chain after inputting the delay chain. A large number of random results are counted as a histogram. A large number of random jump pulses are input into the delay chain, and the rising edge of the random pulse is sampled. The data obtained by sampling is the code density data. Since the jump signal is a random signal, it is considered that the probability of falling into any phase point within the reference clock (0, T) is the same, that is, the time interval t of any pulse signal and the rising edge of the reference clock is also completely random, where t is in the range of (0, T). The probability of random time interval t < t i is:

[0013]

[0014] The probability of the pulse signal falling into the ith time unit is When the test times are sufficient, the delay time τ i of the ith time unit can be obtained as:

[0015]

[0016] In the formula, T represents a reference clock period, h(i) represents the height of the i-th time unit in the histogram, and N represents the total number.

[0017] The delay chain correction needs to use a random signal, and if the chip temperature changes after correction, the linearity will also change, at which time the accuracy can only be corrected again. The application proposes that the first delay chain correction uses an internal random signal generator of the FPGA for correction, and after the first correction is completed, the position of the SPAD dark count in the delay chain is continuously counted into the histogram during the ranging gap, and the number of dark counts is counted. When the dark count reaches a certain amount N, the histogram is calculated to obtain a new linearity correction relationship, which is used to ensure that the calibrated TDC linearity meets the expected value.

[0018] The TDC after real-time correction has good linearity, and the statistical result is filtered by using an average value algorithm, and the optical flight time is converted into distance information Dis

[0019]

[0020] In the formula, Δt i represents the flight time of the laser obtained each time, K represents the cumulative number of times, and λ represents the speed of light. Taking the average value of multiple results can ensure data stability. After calculation by the algorithm, the absolute distance between the measurement target and the ranging device is obtained. (IV) DESCRIPTION OF DRAWINGS

[0021] Figure 1 is a ranging device composed of a field programmable logic gate array computing core (1), a quenching circuit (2), a single photon avalanche diode (3), a pulsed laser driving circuit (4), a pulsed laser (5), and an optical lens system (6). The field programmable logic gate array computing core (1) includes time-to-digital conversion logic (11) and calculation logic (12), wherein the time-to-digital conversion logic includes correction algorithm logic (111) and carry chain logic (112).

[0022] Figure 2 is a histogram of the distribution of the dark count of the SPAD in the time domain under normal working conditions. As can be seen from the histogram, the dark count of the SPAD in the time domain presents randomness, which meets the requirement of the random signal in the linearity correction algorithm. This signal can be recorded and used for correction algorithm statistics, which is used for linearity correction of the TDC implemented in the FPGA.

[0023] Figure 3is the linear correction flowchart in the device, the device starts first to carry out the first linear correction, the random signal is generated by the FPGA in the first linear correction, when the first correction parameter is completed, the ranging can be carried out, this process belongs to the part of the initialization, when the first linear correction is completed, the normal ranging is carried out, each ranging process is to carry out laser emission, record the SPAD receiving signal, after 100 times of signal accumulation, the algorithm calculation is carried out, the laser emission is not carried out in the algorithm calculation process, at this time, the SPAD generates the probability avalanche breakdown, and there is strong randomness, the quenching signal in this time period is recorded for statistics; the quenching signal between the ranging intervals is counted and counted, when 10000 times are reached, the new linear correction data is calculated, the data is updated after the calculation is completed, and the subsequent ranging result is corrected according to the updated data.

[0024] Figure 4 is the distribution statistical histogram of the signal on the carry chain after the avalanche breakdown signal generated by the SPAD in the ranging interval is processed by the quenching circuit and input to the FPGA, the linear correction parameter of the carry chain is further calculated according to the histogram, the random signal falls in a certain carry chain, if the delay of the carry chain is large, the probability of the random signal falling in the carry chain is higher, when the data quantity is enough, the histogram reflects that the position with larger carry chain delay has higher intensity, and the position with smaller carry chain delay has lower intensity.

[0025] Figure 5 is a comparison line chart of the influence of the device with three different correction schemes of only first correction, temperature reference correction and dark count automatic correction on the ranging result; the only first correction scheme can ensure the ranging accuracy within a short time after starting, but when the environmental temperature changes, the ranging accuracy accumulates errors, the more the environmental temperature changes, the larger the error, and the faster the environmental temperature changes, the larger the error slope; the temperature reference correction scheme combines the temperature sensor in the system, and when the temperature deviation reaches a certain value, a correction is triggered, but the measurement needs to be paused during the correction, and then the measurement is started after the correction is completed; in the dark count automatic correction scheme, the SPAD dark count in the ranging interval is counted, and when a certain amount is accumulated, a correction is carried out; it can be seen from the line chart that the dark count automatic correction scheme has the best repeated accuracy.

[0026] Figure 6 is the re-measurement under different distances of the devices with different schemes after 10 hours of continuous work, and the effect diagram of the influence of different linear calibration schemes on time and temperature superposition. The horizontal axis represents different distances, and the vertical axis represents the distance output of the device. The ideal curve is a standard line, it can be seen that the only first correction curve is greatly affected after 10 hours of continuous work, the temperature reference correction scheme is also affected to a certain extent, and the dark count automatic correction scheme is almost not affected. (Five) Specific embodiments

[0027] Example 1:

[0028] Combination Figure 4 , Figure 5 This embodiment is described below. Figure 4 In the embodiment of the laser ranging device based on FPGA and single-photon avalanche diode, the histogram generated by dark counting during the dark counting automatic correction process is used. Figure 5 This is a performance comparison chart of this ranging device with other devices that have the same function but different principles. The ranging device consists of a field-programmable gate array (FPGA) computing core (1), a quenching circuit (2), a single-photon avalanche diode (3), a pulsed laser driving circuit (4), a pulsed laser (5), and an optical lens system (6). The FPGA computing core (1) contains time-to-digital conversion logic (11) and calculation logic (12). The time-to-digital conversion logic contains correction algorithm logic (111) and carry chain logic (112). The FPGA (1) in the system generates a signal that drives the pulsed laser driving circuit (4) to drive the pulsed laser (5) to generate a pulsed laser. The pulsed laser is emitted as a high-power collimated pulsed laser through the optical lens system (6). After being reflected by the target object, the laser passes through the optical lens system (6) and causes the single-photon avalanche diode (3) to avalanche and break down. The quenching circuit (2) quenches the single-photon avalanche diode and converts the avalanche breakdown signal into a pulsed square wave, which is then transmitted to the FPGA computing core (1). The field-programmable gate array (FPGA) computing core (1) time-to-digital conversion logic (11) directly converts the signal transmitted to the laser pulse drive circuit (4) to the received quenching circuit (2) into a digital quantity. The calculation logic (12) converts the time into a digital quantity and converts it into distance. The time-to-digital conversion logic mainly uses carry chain logic (112) to record the time difference, and the correction algorithm logic (111) corrects the linearity of the output digital quantity time difference. The correction algorithm logic requires random signal drive. There are two random signal sources. The first is the random signal generated inside the FPGA computing core (1), and the second is the signal recorded after the single-photon avalanche diode (3) dark count is processed by the quenching circuit (2). Both random signals have randomness.

[0029] After placing the device at a distance of 50.35 cm from the target object, the dark counting correction method proposed in this invention was compared with other correction methods. The device was then turned on, and according to… Figure 3 The flowchart first shows the initial linear correction, followed by distance measurement. The measured distance D is 50.35 meters. The measurement is maintained, and the dark counting correction histogram during device operation is derived. Figure 4 The histogram shows the first bin height as 47 and the second bin height as 52, allowing us to calculate the delay of the first time unit in the carry chain. The second unit delay τ2= 26 ps; the correction delay of each delay unit can be calculated according to the formula. In one measurement, the coarse count K = 5, the fine count i = 150, and the fine count j = 49; the time interval T = 60470 ps is calculated, which includes the calibration system delay 57110 ps, and the final TOF = 3359 ps, and the converted distance is 50.35 cm, which is consistent with the actual distance.

[0030] The entire ranging process lasts for 180 minutes, and the distance value output by the device is continuously recorded. In order to reflect the excellent performance of the dark count correction algorithm adopted by the device, the measurement means is kept unchanged, and the device with different algorithm schemes is used. Two experiments of the device with only the first correction scheme and the device with temperature reference correction scheme are increased. The final experimental data is shown in Figure 5 The experimental data is arranged, and the calculation formula is used The root mean square error under different conditions is calculated, wherein the root mean square error RMSE1 of the output of the dark count automatic correction algorithm is 0.0052, which is less than the design expectation 0.006; the root mean square error RMSE2 of the output of the device with only the first correction algorithm is 0.0372, which is poor and is greatly affected by the environmental temperature; the root mean square error RMSE3 of the output of the temperature reference correction algorithm is 0.0071, which is greater than RMSE1, and an additional temperature sensor needs to be added to the device, which has higher cost and complexity. By comparing the ranging output RMSE of the three algorithms, the ranging device composed of a single photon avalanche diode and a dark count correction algorithm has high measurement accuracy and low root mean square error. At the same time, due to the high sensitivity of the single photon avalanche diode, the device has great innovation and improvement in ranging accuracy, linearity, maximum distance measured, and cost compared with traditional ranging devices.

[0031] Embodiment two:

[0032] In combination with Figure 6 This embodiment, Figure 6 is a comparison chart of the output data of the device and the devices with the same function but different principles under the same conditions. In order to consider the influence of different temperatures and working time on the device, after 10 hours of continuous work, the ranging module is re-measured at different distances, and the final output results of the devices with different algorithms are compared; the results are shown in Figure 6 The horizontal axis represents different distances, and the vertical axis represents the distance output by the device. The ideal curve is a standard line, and it can be seen that the curve of the device with only the first correction is greatly affected after 10 hours of continuous work, the curve of the device with temperature reference correction is also affected to a certain extent, and the curve of the device with dark count automatic correction scheme is hardly affected. The closeness to the ideal curve can be obtained by calculating the Pearson correlation coefficient r of the ideal curve, and the calculation formula is:

[0033]

[0034] In the formula and The average value of the two groups of data is represented.

[0035] It is calculated that the Pearson correlation coefficient r1 of the first correction scheme device is 0.987, the Pearson correlation coefficient r2 of the temperature reference correction scheme in the yellow paper is 0.991, and the Pearson correlation coefficient r3 in the device is 0.997; From the data, it can be seen that the automatic correction scheme of dark count has great advantages in maintaining linearity under long-time working conditions.

Claims

1. A laser ranging device based on FPGA and single photon avalanche diode; characterized in that: The ranging device is composed of a field programmable logic gate array computing core (1), a quenching circuit (2), a single photon avalanche diode (3), a pulsed laser driving circuit (4), a pulsed laser (5), and an optical lens system (6), wherein the field programmable logic gate array computing core (1) contains time-to-digital conversion logic (11) and computing logic (12), wherein the time-to-digital conversion logic contains correction algorithm logic (111) and carry chain logic (112); in the system, the field programmable logic gate array computing core (1) generates a signal to drive the pulsed laser driving circuit (4) to generate pulsed laser for the pulsed laser (5), the pulsed laser is emitted as high-power collimated pulsed laser through the optical lens system (6), the laser is reflected by a target object and then passes through the optical lens system (6) to cause avalanche breakdown of the single photon avalanche diode (3), and the quenching circuit (2) quenches the single photon avalanche diode and converts the avalanche breakdown signal into a pulse square wave and transmits it to the field programmable logic gate array computing core (1); the time-to-digital conversion logic (11) of the field programmable logic gate array computing core (1) directly converts the signal transmitted to the pulsed laser driving circuit (4) to digital quantity after receiving the quenching circuit (2), and the computing logic (12) converts the time into digital statistics and calculates the distance; wherein the time-to-digital conversion logic mainly records the time difference by the carry chain logic (112), and the correction algorithm logic (111) corrects the linearity of the output digital quantity time difference, wherein the correction algorithm logic needs to be driven by a random signal, and the random signal source contains two, the first is a random signal generated by the field programmable logic gate array computing core (1) itself, and the second is a signal recorded after the dark count of the single photon avalanche diode (3) is processed by the quenching circuit (2), both of which have randomness.

2. The FPGA and single-photon avalanche diode based laser ranging device of claim 1, wherein: The Single Photon Avalanche Diode (SPAD) and quenching circuit (QC) combination is used in the device as a laser pulse receiving part. When a reverse bias voltage higher than the avalanche voltage is applied across the SPAD, avalanche breakdown effect can be used to obtain high internal current gain, thereby triggering a chain reaction. At this time, a larger current is generated, quenched by the rear-end quenching circuit, and an avalanche breakdown signal is output. However, in the semiconductor, electron-hole pairs can be generated by heat, even without any photons, that is, in the dark condition, the avalanche will also be triggered. This avalanche is called dark noise, which has strong randomness and can be understood as uniformly distributed in the entire time range. The digital signal output through the quenching circuit is called dark count, which will have a certain impact on the ranging performance, but the dark count cannot be completely eliminated. The 905nm band infrared laser is used as the generating device, and the pulse laser driving circuit is used to drive the laser to generate a high-power laser pulse signal with a width of less than 5ns. The Direct Time offlight (DTOF) scheme is used to calculate the measured distance.

3. The FPGA and single-photon avalanche diode based laser ranging device of claim 1, wherein: The time-to-digital converter (TDC) function is implemented using the internal carry chain resources of the FPGA. The FPGA uses the Xilinx A7 series, and the carry chain is implemented by cascading internal CARRY4. A CARRY4 outputs four taps (CO0, CO1, CO2, CO3) corresponding to different delays. In the design, 50 delay units are used to generate 200 taps, and there are 200 time units in total, with an average delay of about 25 ps. However, due to the internal structure characteristics of the delay chain, there is a time deviation between different time units, and it will change with the change of the chip temperature. For scenarios with low linearity requirements, an average correction method is used, and for scenarios with high linearity requirements, a code density method is used for linearity correction. The code density method requires inputting a set of random signals, which randomly fall into a certain position in the delay chain after inputting the delay chain. A large number of random results are counted as a histogram. A large number of random jump pulses are input into the delay chain, and the rising edge of the random pulse is sampled. The data obtained by sampling is the code density data. Since the jump signal is a random signal, it is assumed that the probability of falling into any phase point within the reference clock (0, T) is the same, that is, the time interval t between any pulse signal and the rising edge of the reference clock is completely random, where the range of t is (0, T). The probability of random time interval is: ， The probability of the pulse signal falling in the i-th time unit is When the number of tests is large, the delay time of the i-th time unit can be obtained as : ， In the equation, T represents a reference clock cycle, represents the height of the i-th time unit in the histogram after statistics, and N represents the total number.

4. The FPGA and single-photon avalanche diode based laser ranging device of claim 1, wherein: The device uses a dark count linear correction scheme for the delay chain correction part. The delay chain correction requires the use of a random signal, and after correction, if the chip temperature changes, the linearity will also change. At this time, in order to maintain accuracy, the correction must be performed again. The first delay chain correction uses the FPGA internal random signal generator for correction. After the first correction is completed, the position of the SPAD dark count in the delay chain is continuously counted into a histogram during the ranging gap, and the number of dark counts is counted. When the dark count reaches a large amount, the histogram is calculated to obtain a new linearity correction relationship. Good linearity correction can be used to ensure that the root mean squared error (Root Mean Squared Error, RMSE) of the calibration TDC linearity and the ideal value is less than 0.

006. The RMSE calculation formula is: ， where n represents the number of data, represents the current data, represents the average of all data, which reflects the repeatability of the distance measuring device.

Images