CN-115527129-B - Unmanned aerial vehicle ground dynamic target recognition and tracking scene reliability test case generation method

Abstract

The invention provides an unmanned aerial vehicle-oriented ground dynamic target recognition and tracking scene reliability test case generation method. The method comprises the steps of (1) analyzing and determining static reliability influence factors in a ground dynamic target identification and tracking scene of the unmanned aerial vehicle, (2) analyzing and determining shielding influence caused by building influence factors in the ground dynamic target identification and tracking scene of the unmanned aerial vehicle, (4) collecting data, performing fitting goodness test on a distribution function, selecting optimal distribution of ground dynamic target identification and tracking influence factor parameters of the unmanned aerial vehicle, and (5) performing importance sampling to generate an unmanned aerial vehicle target identification and tracking test environment and determine test cases.

Inventors

• YANG DEZHEN
• LIU YEYANG
• REN YI
• SUN BO
• FENG QIANG
• WANG ZILI

Assignees

• 北京航空航天大学

Dates

Publication Date

20260505

Application Date

20220919

Claims (1)

1. 1. A method for generating a test case of ground dynamic target identification and tracking scene reliability of an unmanned aerial vehicle is characterized by comprising the following steps: analyzing and determining static reliability influence factors in a ground dynamic target identification and tracking scene of the unmanned aerial vehicle; step 1, analyzing and determining weather influencing factors; (1) The weather has w weather types such as sunny days, rains, fog, snow and the like, which are respectively marked as { alpha 1 ,α 2 ,α 3 ,…,α w }; (2) There are 4 weather classes for each weather type, general, severe, particularly severe, denoted { β 1 ,β 2 ,β 3 ,β 4 }; (3) The wind power has 12 grades which are respectively marked as { c 1 ,c 2 ,c 3 ,…,c 12 }; (4) The illumination intensity of each day is different, and the illumination intensity is d and d is more than 0, and the unit is luxlx; Step2, analyzing and determining road influence factors; Collecting road environments of different cities, numbering each road in the urban road environments, recording as { R 1 ,R 2 ,…,R n }, wherein n represents the total number of the roads in the city, taking a certain crossroad as an origin, taking the forward direction as the x-axis forward direction, taking the forward north direction as the y-axis forward direction, taking the direction vertical to the urban ground as the z-axis forward direction, establishing a three-dimensional coordinate system in the urban road environments, recording the position of each road, numbering expressions on the left side and the right side of each road respectively, and recording as { R l1 ,R l2 ,…,R ln },{R r1 ,R r2 ,…,R rn }; In a three-dimensional coordinate system, respectively numbering the buildings on the left side and the right side of each road, namely { b 1 ,b 2 ,…,b m }, recording the coordinate positions of two vertexes of each building on the z=0 plane, which are close to one side of the road, namely b a :{(x a ,y a ,0),(x' a ,y' a , 0) } (a epsilon (1, 2, the..m)), and recording the heights of each building, namely { h 1 ,h 2 ,…,h m }; Analyzing dynamic reliability influence factors in the ground dynamic target recognition and tracking scene of the unmanned aerial vehicle; step 1, analyzing and determining influencing factors of ground dynamic targets and other interference targets; Ground dynamic targets and other influencing factors that interfere with the target, including type, speed, width, length, and color of the target: (1) Constructing a set of types of ground dynamic targets and other interference targets, and recording the set as { T 0 ,T 1 ,T 2 ,…,T n }, wherein T 0 represents the type of the ground dynamic targets, and T i (i > 0) represents the other interference target type with the number of i; (2) After the types of the ground dynamic targets and other interference targets are determined, the widths and the lengths of the ground dynamic targets are also determined and are respectively marked as { W 0 ,W 1 ,W 2 ,…,W n } and { L 0 ,L 1 ,L 2 ,…,L n }, wherein W 0 represents the width of the ground dynamic targets, W i (i > 0) represents the widths of other interference targets with the number i, L 0 represents the lengths of the ground dynamic targets, and L i (i > 0) represents the lengths of other interference targets with the number i; (3) Constructing a speed set of a ground dynamic target and other interference targets, and recording the speed set as { v 0 (t),v 1 (t),v 2 (t),…,v n (t) }, wherein v 0 (t) represents the speed of the ground dynamic target at the time t, v i (t) (i > 0) represents the speed of the other interference targets with the number i at the time t, and simultaneously recording the central position of the ground dynamic target at the time t in a three-dimensional coordinate system, and recording the central position as (x o (t),y o (t), 0); (4) Constructing a color set of the ground dynamic target and other interference targets, denoted as { c 1 ,c 2 ,c 3 ,…,c n }, wherein c 0 represents the color of the ground dynamic target, and c i (i > 0) represents the color of the other interference targets numbered i; step2, analyzing influence factors of the unmanned aerial vehicle; the speed of the unmanned aerial vehicle is expressed in a three-dimensional coordinate system, v x (t) is used for expressing the speed of the unmanned aerial vehicle in the x-axis direction at the time t, v y (t) is used for expressing the speed of the unmanned aerial vehicle in the y-axis direction at the time t, v z (t) is used for expressing the speed of the unmanned aerial vehicle in the z-axis direction at the time t, and the coordinate position (x u ,y u ,h u ) when the unmanned aerial vehicle starts is recorded, so that the reliability of the unmanned aerial vehicle target recognition and tracking can be further analyzed by presetting different initial speeds and positions of the unmanned aerial vehicle; Analyzing and determining shielding influence caused by building influence factors in a scene for identifying and tracking the ground dynamic target by the unmanned aerial vehicle; step1, judging whether a building can bring shielding; at time t, the position coordinates of the unmanned aerial vehicle are as follows The ground dynamic target position coordinates are (x o (t),y o (t), 0) and are positioned on a road with the number p, expressions on the left side and the right side of the road are R lp :A p x+B p y+C lp ＝0,z＝0(-l＜x＜l'),R rp :A p x+B p y+C rp ＝0,z＝0(-l＜x＜l'), respectively, wherein A p and B p are coefficient constants for jointly determining the direction of the road, the coefficient constants are obtained by collecting a plurality of sample points on the same side of the road and performing straight line fitting calculation, C lp 、C rp is a constant term of offset of the road on the left side and the right side relative to an origin of coordinates, under the condition that A p and B p are known, points on the left side and the right side of the road are collected and are respectively substituted into the expressions to be solved, x represents the horizontal coordinate of the road in a coordinate system, and y represents the vertical coordinate of the road in the coordinate system; calculating the distance from the position of the unmanned aerial vehicle to the two sides of the road, when When the building is used, the building can bring shielding to the ground dynamic target, otherwise, the building has no shielding influence; Step2, calculating the shielding range of the building; after determining that the building can bring shielding influence, judging which side of the road the unmanned aerial vehicle is on, and when When the method is used, the left building can bring shielding influence to the ground dynamic target, at the moment, the nearest building on the left road and the ground dynamic target is b a , the height is h a , and at the moment, the shielding range w brought by the building is: similarly, when When the method is used, the right building can bring shielding influence to the ground dynamic target, at the moment, the nearest building on the right road and the ground dynamic target is b s , the height is h s , and at the moment, the shielding range w brought by the building is: Step 3, calculating the shielding range of the building to the ground dynamic target; (1) When unmanned aerial vehicle is in the road left side, the building of road left side brings the shielding for ground dynamic target vehicle, and the distance of ground dynamic target vehicle to the road left side this moment is: from this, can obtain the building on road left side and bring to the ground dynamic target vehicle shelter from scope w 0 is: (2) When unmanned aerial vehicle is on the road right side, the building on road right side brings the shielding for ground dynamic target vehicle, and the distance of ground dynamic target vehicle to road right side this moment is: Therefore, the shielding range w 0 of the building on the right side of the road for the ground dynamic target vehicle can be obtained as follows: collecting data, performing goodness-of-fit test on the distribution function, and selecting optimal distribution of the unmanned aerial vehicle on the ground dynamic target identification and tracking influence factor parameters; Step 1, collecting data of an unmanned aerial vehicle in executing target recognition and tracking tasks, and processing and analyzing the data to obtain distribution types possibly obeyed by each parameter; (1) Based on the ground dynamic target identification and tracking scene of the unmanned aerial vehicle, collecting data { X 1 ,X 2 ,X 3 …,X n } of reliability influence factors of the unmanned aerial vehicle in the flight scene; (2) Screening the collected data of the unmanned aerial vehicle in the real scene, removing noise data, taking the minimum value and the maximum value [ X min ,X max ] of the unmanned aerial vehicle reliability influence parameter data as the value reference range of the parameter, sorting the data from small to large, dividing the data into t groups by using an empirical formula t=1+3.3lgn (n is the number of data, and rounding up t), and determining group distances according to a formula Vx= (X max -X min )/t, and calculating the frequency in each group (X 1 ,x 2 ,…,x t ): (3) According to the obtained data, drawing a distribution map of each reliability parameter of the unmanned aerial vehicle, fitting the distribution to obtain a distribution function, and comparing the distribution function with a theoretical distribution function to obtain possible distribution of the parameters; In the step 1, available data in a ground dynamic target identification and tracking scene of the unmanned aerial vehicle is { x 1 ,x 2 ,x 3 ,…,x t }, and in a possible distribution model, measured values are as follows: Where x i ' is the value of the possible position corresponding to x i in the distribution model, a critical value can be obtained for a given significance level α=0.05 When (when) Is less than or equal to the critical value The influencing parameter may be considered to obey the distribution; Step 3, checking continuous distribution parameters in the reliability influence parameters of the unmanned aerial vehicle by using a gray correlation method based on a gray theory, and selecting an optimal distribution type; The method comprises the steps of recording available data in a ground dynamic target identification and tracking scene of the unmanned aerial vehicle as a reference sequence x (0) ：{x 1 ,x 2 ,x 3 ,…,x t }, obtaining a comparison sequence x (1) through calculation of frequencies of positions corresponding to the reference sequence in a distribution type 1, obtaining the comparison sequence x (2) ,x (3) ,…,x (j) similarly, and calculating association coefficients of distribution types of reliability influence parameters of the unmanned aerial vehicle according to a gray theory analysis method: Wherein, xi (i)k represents the association coefficient of the ith comparison sequence at the kth point relative to the reference sequence, which is used for representing the approaching degree of the two sequences at the moment, ρ is a resolution coefficient, which is a value between 0 and 1 and is used for adjusting the resolution capability of the association coefficient xi (i)k and improving the calculation stability, thus the association degree can be calculated as follows: finally, selecting a comparison sequence with the maximum association coefficient to obtain the optimal distribution of the unmanned aerial vehicle reliability influence parameters; step five, importance degree sampling is carried out, an unmanned aerial vehicle target recognition and tracking test environment is generated, and a test case is determined; (1) Obtaining probability density functions f 1 (x),f 2 (x),…,f n (x) of all reliability influence parameters of the unmanned aerial vehicle through the fourth step, and determining important sampling probability density functions of all influence parameters of the unmanned aerial vehicle by using a cross entropy optimizing method Weighting of (2) Important sampling probability density function of each independent reliability influence parameter of unmanned aerial vehicle Respectively sampling n groups of parameter sets { alpha, beta, c, d, }, and generating a test environment for identifying and tracking the ground dynamic target by the n groups of unmanned aerial vehicles; (3) And placing the unmanned aerial vehicle in a test scene for testing, setting test time, and recording the time for successfully identifying and tracking the ground dynamic target of the unmanned aerial vehicle in the test process.

Description

Unmanned aerial vehicle ground dynamic target recognition and tracking scene reliability test case generation method Technical Field The invention provides an unmanned aerial vehicle-oriented ground dynamic target recognition and tracking scene reliability test case generation method. Aiming at a typical task of target vehicle identification and tracking of an unmanned aerial vehicle, in a road network with a certain width, the method finds out the influence of a building on the identification and tracking of the unmanned aerial vehicle by analyzing static and dynamic reliability influence factors in a task environment, determines key parameters and distribution types thereof, and finally performs importance sampling, so that a reliability important scene test scene of the unmanned aerial vehicle on the ground dynamic target identification and tracking is generated, and support is provided for the reliability evaluation of the unmanned aerial vehicle on the ground dynamic target identification and tracking. The invention belongs to the field of reliability engineering. Background Because unmanned aerial vehicle possesses advantages such as with low costs, mobility is strong, convenient to use, security height, people use it in various important task activities such as target identification and tracking, aerial photography, public security control. Meanwhile, the reliability problem of the unmanned aerial vehicle is more and more prominent, and a designer is required to accurately evaluate the reliability level of the unmanned aerial vehicle in a typical task scene in the development process of the unmanned aerial vehicle, so that a basis is provided for the reliability design optimization of the unmanned aerial vehicle. In the development stage, the data of unmanned aerial vehicle reliability assessment mainly originate from tests, and how to ensure test coverage directly relates to the accuracy of the assessment result. When the unmanned aerial vehicle executes the target vehicle identification and tracking task in the urban road, static influence factors such as weather, roads, buildings and the like are often interfered by dynamic factors such as other vehicles, the movement of the target vehicle and the initial height speed of the unmanned aerial vehicle, but the reliability cost for identifying and tracking the target vehicle of the unmanned aerial vehicle is high, the efficiency is low and the repeatability is poor by using a simulation method to test the target vehicle of the unmanned aerial vehicle. In view of this, the present invention proposes a test case generating method based on importance. Static influence factors and dynamic influence factors of the unmanned aerial vehicle on a target vehicle identification and tracking scene are analyzed, a proper distribution model is constructed for the influence factors, finally importance degree sampling is carried out to generate a reliability test case, and input is provided for accurate evaluation of reliability of unmanned aerial vehicle ground dynamic target identification and tracking. Disclosure of Invention The invention provides a method for generating a test case of ground dynamic target recognition and tracking scene reliability of an unmanned aerial vehicle. The invention mainly comprises the following steps: analyzing and determining static reliability influence factors in a ground dynamic target recognition and tracking scene of the unmanned aerial vehicle. Static reliability influencing factors for unmanned aerial vehicle target recognition and tracking include weather, roads and buildings. Step 1, analyzing and determining weather influencing factors Weather influencing factors include weather type, weather level, wind power level and illumination intensity, and the selection of weather type and weather level influences the selection of illumination intensity. (1) The weather has w weather types such as sunny days, rains, fog, snow and the like, which are respectively marked as { alpha 1,α2,α3,…,αw }; (2) There are 4 weather classes for each weather type, general, severe, particularly severe, denoted { β 1,β2,β3,β4 }; (3) The wind power has 12 grades which are respectively marked as { c 1,c2,c3,…,c12 }; (4) The illumination intensity per day is different, and the illumination intensity is d and d is more than 0, and the unit is luxlx. Step 2, analyzing and determining road influencing factors Road influencing factors include road strip number, length, width. Road environments of different cities are collected, and each road is numbered in the urban road environment and is marked as { R 1,R2,…,Rn }, wherein n represents the total number of roads in the city. And by taking any crossroad as an origin, taking the forward direction as the positive x-axis direction, taking the positive north direction as the positive y-axis direction, taking the direction vertical to the urban ground as the positive z-axis direction, establishing a three-dimen