Obstacle avoidance radar constant false alarm processing method

Abstract

The invention relates to an obstacle avoidance radar constant false alarm processing method, which adopts a modular architecture of step-by-step processing in the range direction and the beam direction and comprises the following steps of: firstly, carrying out adaptive unit average (CFAR) processing on range dimension data, and dynamically eliminating abnormal values to accurately estimate a background level; connecting the beam dimension data end to end to form a closed loop chain, and processing by adopting a minimum (SO-CFAR) strategy to effectively suppress edge clutter interference; and finally, multiplying two-dimensional processing results, and judging the target through a dynamic threshold. According to the method, the detection performance and the contour keeping capability in a strong clutter and multi-target scene are effectively improved, the data processing amount and the transmission overhead are remarkably reduced, and the method is suitable for obstacle avoidance radar systems in the fields of maritime affairs, automatic driving and the like.

Description

Technical Field

[0001] This invention relates to the field of radar detection technology, and in particular to a method for handling constant false alarm rates in obstacle avoidance radar. Background Technology

[0002] Radar detects target signal information through radio transmission. The parameters of these target signals are obtained through the reflection of electromagnetic waves on the target signal. Obstacle avoidance radar is a type of radar system used to perceive the surrounding environment in real time, detect obstacles, and assist in obstacle avoidance decisions. With the rapid development of autonomous driving, drones, and robotics, "autonomous obstacle avoidance" has become a core requirement. Traditional solutions relying on manual labor or single sensors are insufficient to meet these needs, directly driving research into obstacle avoidance radar.

[0003] In obstacle avoidance radars (such as vehicle-mounted radars, UAV radars, and industrial detection radars), constant false alarm rate (CFAR) is one of the core signal processing technologies. Its core objective is to maintain a constant false alarm probability (probability of incorrect detection) of the radar in complex and variable environments (such as clutter, noise, and multi-target interference) by adaptively adjusting the detection threshold, while ensuring effective detection of real targets.

[0004] Common methods for handling constant false alarms (CFA) include cell-averaged CFA, ordered statistical CFA, and selective large / small CFA. However, each of these methods has its specific advantages, disadvantages, and limitations. In real-world applications, considering the complexity of radar operating environments such as thunderstorms, fog, and waves, interference can affect radar operation. This interference mainly includes static and dynamic clutter. Large areas of clutter can overwhelm the target's echo signal, making target observation extremely difficult. Especially in the presence of strong clutter, the target signal is easily obscured, making it impossible to extract. Obstacle avoidance radar not only needs to detect moving or stationary targets and calculate the target's DCPA (closest encounter distance) and TCPA (time to nearest encounter point), but in some fields, such as maritime obstacle avoidance radar, it also needs to detect the outlines of obstacles (ships, reefs, floating objects) to guide the optoelectronic system for target confirmation. However, common constant false alarm rate (CFAR) systems typically use fixed protection and reference units. When the protection unit is too large, target detection accuracy decreases; conversely, when the protection unit is too small, the target detection probability decreases. When the reference unit is set too large, it is prone to including interfering targets or clutter edges, leading to performance degradation. When the reference unit is set too small, background estimation errors are large, and false alarms are unstable. This is especially problematic with strong reflectors such as large ships, shore-based structures, and islands / reefs, where there is a risk of missed or failed detections.

[0005] Therefore, there is an urgent need for a method to overcome the inability of conventional obstacle avoidance radar to perform constant false alarm rate (CFAR) processing on targets quickly and effectively in complex noise environments, especially in maritime obstacle avoidance radar where it is necessary to detect the outline of obstacles. This method solves the problem of adaptive CFAR processing in complex environments for existing obstacle avoidance radar. Summary of the Invention

[0006] To address the aforementioned technical problems, this invention provides a method for handling constant false alarms (CFAM) in obstacle avoidance radar. The method comprises three modular steps: range-oriented CFAM processing, beam-oriented CFAM processing, and combined processing. Specifically, it includes the following steps:

[0007] Step S1: Range-oriented constant false alarm rate (CFAR) processing; according to beam sequence, for each range-oriented detection unit, after removing the protection unit, select the reference unit, calculate the mean value E of the reference unit, and dynamically eliminate units with values ​​greater than the mean value N. C The outlier is multiplied by a factor to ensure that the number of remaining reference units is within the valid range, and the mean E of the remaining reference units after removal is calculated. C Finally, the distance-to-false alarm ratio of the detection unit is obtained;

[0008] Step S2: Beam-direction constant false alarm rate (CFAR) processing; after all beams have undergone range processing, the beam-direction data are linked end-to-end in range order to form a closed-loop data link; for each beam-direction detection unit, after removing the protection unit, a reference unit is selected, and the minimum value X within the reference unit is found. min The beamwidth constant false alarm value K of the detection unit is obtained. Bi ;

[0009] Step S3: Combination and discrimination processing; The distance corresponding to each detection point is converted to the constant false alarm rate ratio K. Di The ratio of beamwidth to constant false alarm rate K Bi Multiply to obtain the final constant false alarm rate (CFAR) result Ki for that point; compare Ki with the dynamic threshold Di to determine whether the point is a target. If Ki is greater than Di, it is determined to be a target; otherwise, it is determined to be clutter.

[0010] In one embodiment of the present invention, the process of dynamically removing outliers in step S1 includes: if the number of remaining points after removal does not meet a preset range, adjusting the removal threshold N in fixed steps. C The value is adjusted until the remaining points meet the requirements or the maximum number of adjustments is reached.

[0011] In one embodiment of the present invention, the rejection threshold N C The dynamic adjustment rule is as follows: if the number of remaining reference units after removing outliers exceeds 3 / 4 times the preset range, then N is increased. C Step value; if the remaining quantity is less than 1 / 4 of the preset range, then adjust N down. CStep value; after the number of adjustments exceeds the preset threshold, the state after the last adjustment is maintained.

[0012] In one embodiment of the present invention, in the distance processing of step S2, for the special case where reference units on both sides cannot be obtained at the beginning and end of the data segment, only one side of valid data is taken as the reference unit for calculation.

[0013] In one embodiment of the present invention, the formula for calculating the reference cell mean (E) in step S1 is as follows:

[0014]

[0015] Among them, R D The distance is the reference cell length; for data within a segment in special cases, the calculation formula is:

[0016]

[0017] In the formula X n This is the amplitude of the nth reference unit point.

[0018] In one embodiment of the present invention, the distance-to-direction constant false alarm ratio K Di The calculation formula is:

[0019]

[0020] Where X i For the amplitude of the detection unit, E c This is the average value of the remaining reference unit amplitudes.

[0021] In one embodiment of the present invention, the beam-directed constant false alarm rate ratio K Bi The calculation formula is:

[0022]

[0023] Where X min This is the minimum amplitude value within the reference unit.

[0024] In one embodiment of the present invention, the formula for calculating the final constant false alarm rate result Ki is:

[0025] K i =K Di ×K Bi .

[0026] In one embodiment of the present invention, the preset threshold Di is a dynamically adjustable value, with an initial value of 10, which can be adjusted in real time according to the processing results.

[0027] In one embodiment of the present invention, the data packet for all beams contains 48 beams, each beam contains 1024 range points, and the preset parameters include: range-direction protection unit length P. D The reference element length is 4, and the reference element length is R. D The beam-to-protection unit length P is 20. B The reference element length is 2, and the reference element length is R. B The threshold is 4; the elimination threshold N is also 4. C The initial value is 2.

[0028] Compared with the prior art, the above-mentioned technical solution of the present invention has the following advantages: The obstacle avoidance radar constant false alarm rate (CFAR) processing method of the present invention effectively solves the adaptive CFAR processing problem of obstacle avoidance radar in the face of complex noise environment, especially marine obstacle avoidance radar, which needs to detect the outline of obstacles by processing the range and beam directions in stages, and effectively improves the ability to extract different targets; at the same time, it is simple and fast to implement, saving the resource overhead of radar processing unit; and while ensuring that target information is not lost, it greatly reduces the amount of target data after processing, and reduces the data transmission pressure from radar processing module to radar terminal. Attached Figure Description

[0029] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings.

[0030] Figure 1 This is a design block diagram of the obstacle avoidance radar CFAR processing method of the present invention.

[0031] Figure 2-1 This is a design block diagram of the distance-oriented constant false alarm processing module described in this invention.

[0032] Figure 2-2 This is a design block diagram of the beam-directed constant false alarm processing module described in this invention.

[0033] Figure 2-3 This is a design block diagram of the combined processing module described in this invention.

[0034] Figure 3 This is a waveform diagram before CFAR processing as described in this invention.

[0035] Figure 4 This is the waveform diagram after CFAR processing as described in this invention. Detailed Implementation

[0036] like Figure 1As shown, this embodiment provides a method for handling constant false alarm rate (CFAR) of obstacle avoidance radar. The method consists of three processing units: a range-oriented CFAR processing module, a beam-oriented CFAR processing module, and a combined processing module group. Taking a phased array radar as an example, this patented method includes 48 beams, each beam containing 1024 range points, and sets the range-oriented protection unit length P. D The reference element length is 4, and the reference element length is R. D The beam-to-protection unit length P is 20. B The reference element length is 2, and the reference element length is R. B The threshold is 4; the elimination threshold N is also 4. C The initial value is 2. It also includes the following specific steps:

[0037] Step 1: Request a 48×1024 data address to store multi-beam data packets;

[0038] Step 2: Following the beam sequence, sequentially target the range-direction detection points X. i (Detection unit) Remove P on both sides D Data from each protection unit, then take R from both the front and rear sides. D One data point is used as a reference unit;

[0039] Step 3: Calculate the mean value E of the extracted reference cell;

[0040] Step 4: Remove the three values ​​N that are greater than the mean of the three values. C The data is multiplied by 10 to ensure that the number of remaining points after removal satisfies 1 / 4 of 2×R. D To 3 / 4 of 2×R D If the condition is not met, adjust N. C The value is used to ensure that the number of remaining points after removing large values ​​meets the condition;

[0041] Step 5: Calculate the mean E of the remaining reference cell data after removing large values. C ;

[0042] Step 6: Calculate the detection unit X i With E C The ratio K Di And record it;

[0043] Step 7: Repeat steps 2 to 6 until all (48×1024) distance points have been calculated;

[0044] Step 8: Connect the 48 data points in the beam direction end to end in order of distance to form 1024 closed-loop data links;

[0045] Step 9: According to the beam sequence, perform beam testing on point X. i (Detection unit) Remove P on both sides B Take one data point, then take R from both the front and back sides.B One data point is used as a reference unit;

[0046] Step 10: Find the minimum value Min within the reference cell;

[0047] Step 11: Calculate the detection unit X i The ratio K to the minimum value Min Bi And record it;

[0048] Step 12: Repeat steps 8 to 11 until all (48×1024) beam points have been calculated;

[0049] Step 13: Calculate all points K sequentially. Di ×K Bi The value is used to obtain the constant false alarm rate (CFAR) result Ki for that point;

[0050] Step 14: Determine whether the point Ki is greater than the threshold Di by using threshold discrimination. If it is greater, it is determined to be a target; otherwise, it is determined to be clutter.

[0051] Step 15: Repeat steps 13-14 until all (48×1024) thresholds are determined.

[0052] In one embodiment of the present invention, when taking the reference unit values ​​on the front and back sides of the distance detection unit, there is a data segment without a front or back side data segment at the beginning of the data. This segment of data is specially processed, that is, when the front end data is less than the length of the reference unit, only the back side data is taken as the reference unit data for processing; similarly, when the back end data is less than the length of the reference unit, only the front end data is taken as the reference unit data for processing.

[0053] In one embodiment of the present invention, the mean value E of the reference cell in step 3 is calculated as follows:

[0054]

[0055] If the data is within a special first or last segment, then:

[0056]

[0057] In the formula, E is the mean of all values ​​within the reference cell, and X... n For the magnitude of the nth reference cell point, due to the distance to the reference cell length R D The total number of data points is 40, and since the data is taken from both sides, the total number of data points is 20. In special segments, only one side of the data is taken, so the total number of data points is 20.

[0058] In one embodiment of the present invention, the threshold N is removed in step 4. C The value is 2, meaning that values ​​greater than 2×E within the sampled reference cell are discarded. The number of remaining reference cell points after discarding satisfies:

[0059]

[0060] That is, the remaining number of points N 剩余 If the data is within a special segment at the beginning and end of a range of 10 to 30 numbers, then the remaining number of points is N. 剩余 Between 5 and 15 numbers.

[0061] If the conditions are not met, adjust the rejection threshold N. C The value can be adjusted by stepping 0.5, meaning that if the remaining points exceed 3 / 4 times, N will be increased. C The value is 2.5. If the remaining points are less than 1 / 4, N is reduced. C The value is 1.5. After adjustment, the number of values ​​removed is checked. If the condition is met, the process stops and proceeds to the next step. If the condition is not met, the adjustment continues until the condition is met. A threshold for the number of adjustments can also be set here. If the condition is not met after adjusting beyond the set value, the adjustment stops, the state after the last adjustment is maintained, and the next step is executed. Generally, the adjustment threshold is set to 5 times.

[0062] In one embodiment of the present invention, such as Figure 2-1 The diagram illustrates the distance-oriented constant false alarm rate (CFAR) processing module. In step 6, K... Di This is the distance-to-far value of the i-th point, and its formula is:

[0063]

[0064] In the formula X i Let E be the amplitude of the i-th detection unit. c X is the mean of the numbers within the remaining reference cells. n For the magnitude of the nth reference unit point, N 剩余 This represents the number of remaining reference cell points after removing large values ​​through the removal threshold.

[0065] In one embodiment of the present invention, in step 8, the beam-direction data is connected end to end to form a closed-loop data chain. The reason is that the beam-direction data is generally short. When performing constant false alarm rate (CFAR) processing on the beam-direction, after setting reference units and protection units, the positions of the front and rear reference units can be completely sampled. Based on the beam-direction broadening characteristics of the target echo, it is considered to connect the beam-direction data end to end so that all points can be sampled with the same number of reference units and protection units.

[0066] In one embodiment of the present invention, the minimum value of the reference unit in step 10 is:

[0067] X min =Min(X1,X2,X3...X8)

[0068] In the example, the beam is directed towards the reference element R. B The value is 4, therefore the total number of reference units is 8, X min This is the minimum value of the amplitude of these 8 reference units.

[0069] In one embodiment of the present invention, such as Figure 2-2 The diagram illustrates the beamforming constant false alarm rate (CFAR) processing module. In step 11, K... Bi That is, the beamwidth cfar value at point i, and its formula is:

[0070]

[0071] In one embodiment of the present invention, such as Figure 2-3 The diagram illustrates the combined processing module, specifically step 13 where K... i This is the final CFAR value at point i, and its formula is:

[0072] K i =K Di ×K Bi

[0073] In one embodiment of the present invention, the criterion for determining whether the i-th detection point is a target in step 14 is: the point K i If the value is greater than the threshold Di, then the point is considered the target, and the target amplitude is K. i If the value is not greater than 0, then the point is a clutter point, and the amplitude of that point can be set to zero. The threshold Di can be dynamically adjusted according to the processing results, and is generally set to 10.

[0074] In summary, after processing the range-direction constant false alarm rate (CFAR), beam-direction constant false alarm rate (CFAR), and combining these processes, the CFAR processing for this detection point is completed.

[0075] To verify the feasibility of this design method, the following verification scenario was designed using an existing hardware platform:

[0076] 1) Target A simulates a shore-based target at sea. This target has a significant widening in the range direction, but not a significant widening in the beam direction.

[0077] 2) Target B simulates a large ship at sea. This target has a significant beamwidth in the beam direction, but not a significant beamwidth in the range direction.

[0078] 3) Target C simulates a floating object at sea, which does not show significant widening in either the range or beam direction;

[0079] 4) The radar uses the CFAR processing method of this design scheme to view the data waveform of the CFAR processing result.

[0080] Waveforms before and after processing are shown below Figure 3 , Figure 4As shown.

[0081] pass Figure 3 and Figure 4 The comparison shows that, after processing using this method, the radar can effectively detect conventional maritime targets and retain their distribution in the airspace. The radar backend can then use this distribution to infer the size and characteristics of obstacles, allowing ships to adopt different obstacle avoidance strategies. The processed target amplitude can be compressed to below 15 (binary 1111), which can be represented by 4 bits. This is the data compression ratio, which reduces the target information (amplitude) from 32 bits to 4 bits. Therefore, the amount of data transmitted to the terminal is greatly reduced.

[0082] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A method for handling constant false alarm rates (CFAR) in obstacle avoidance radar, characterized in that, The method consists of three modular steps in sequence: range-oriented constant false alarm rate (CFAR) processing, beam-oriented CFAR processing, and combined processing. Specifically, it includes the following steps: Step S1: Range-oriented constant false alarm rate (CFAR) processing; according to beam sequence, for each range-oriented detection unit, after removing the protection unit, select the reference unit, calculate the mean value E of the reference unit, and dynamically eliminate units with values ​​greater than the mean value N. C The outlier is multiplied by a factor to ensure that the number of remaining reference units is within the valid range, and the mean E of the remaining reference units after removal is calculated. C Finally, the distance-to-false alarm ratio of the detection unit is obtained; Step S2: Beam-direction constant false alarm rate (CFAR) processing; after all beams have undergone range processing, the beam-direction data are linked end-to-end in range order to form a closed-loop data link; for each beam-direction detection unit, after removing the protection unit, a reference unit is selected, and the minimum value X within the reference unit is found. min The beamwidth constant false alarm value K of the detection unit is obtained. Bi ; Step S3: Combination and discrimination processing; The distance corresponding to each detection point is converted to the constant false alarm rate ratio K. Di The ratio of beamwidth to constant false alarm rate K Bi Multiply to obtain the final constant false alarm rate (CFAR) result Ki for that point; compare Ki with the dynamic threshold Di to determine whether the point is a target. If Ki is greater than Di, it is determined to be a target; otherwise, it is determined to be clutter.

2. The obstacle avoidance radar constant false alarm rate processing method according to claim 1, characterized in that: The process of dynamically removing outliers in step S1 includes: if the number of remaining points after removal does not meet the preset range, adjusting the removal threshold N in fixed steps. C The value is adjusted until the remaining points meet the requirements or the maximum number of adjustments is reached.

3. The obstacle avoidance radar constant false alarm rate processing method according to claim 2, characterized in that: The elimination threshold N C The dynamic adjustment rule is as follows: if the number of remaining reference units after removing outliers exceeds 3 / 4 times the preset range, then N is increased. C Step value; if the remaining quantity is less than 1 / 4 of the preset range, then adjust N down. C Step value; after the number of adjustments exceeds the preset threshold, the state after the last adjustment is maintained.

4. The obstacle avoidance radar constant false alarm rate processing method according to claim 1, characterized in that: In the distance processing of step S2, in the special case where reference units on both sides cannot be obtained at the beginning and end of the data segment, only the valid data on one side is taken as the reference unit for calculation.

5. The obstacle avoidance radar constant false alarm rate processing method according to claim 1, characterized in that: The formula for calculating the mean value E of the reference unit in step S1 is as follows: Among them, R D The distance is the reference cell length; for data within a segment in special cases, the calculation formula is: In the formula X n This is the amplitude of the nth reference unit point.

6. The obstacle avoidance radar constant false alarm rate processing method according to claim 1, characterized in that: The distance-to-constant false alarm ratio K Di The calculation formula is: Where X i For the amplitude of the detection unit, E c This is the average value of the remaining reference unit amplitudes.

7. The obstacle avoidance radar constant false alarm rate processing method according to claim 1, characterized in that: The beam direction constant false alarm ratio K Bi The calculation formula is: Where X min This is the minimum amplitude value within the reference unit.

8. The obstacle avoidance radar constant false alarm rate processing method according to claim 1, characterized in that: The final constant false alarm rate result K i The calculation formula is: K i =K Di ×K Bi 。 9. The obstacle avoidance radar constant false alarm rate processing method according to claim 1, characterized in that: The preset threshold Di is a dynamically adjustable value, with an initial value of 10, which can be adjusted in real time according to the processing results.

10. The obstacle avoidance radar constant false alarm rate processing method according to claim 1, characterized in that: The data packet for all beams contains 48 beams, each beam contains 1024 range points, and the preset parameters include: range-oriented protection unit length P. D The reference element length is 4, and the reference element length is R. D The beam-to-protection unit length P is 20. B The reference element length is 2, and the reference element length is R. B The threshold is 4; the elimination threshold N is also 4. C The initial value is 2.

Images