Method for calibrating safe release interval of logistics unmanned aerial vehicle based on speed error

Abstract

The present application relates to the technical field of unmanned aerial vehicle flight safety, in particular to a logistics unmanned aerial vehicle safety release interval calibration method based on speed error, comprising the following steps: S1 safety release key influence factor analysis; S2 safety target level and release interval determination; S3 speed error model determination; S4 collision template determination; S5 same route collision determination; S6 cross route collision determination. The logistics unmanned aerial vehicle safety release interval calibration method based on speed error, for the same route operation and cross route operation two unmanned aerial vehicle operation release management scenes, considering the unmanned aerial vehicle climbing and cruising stages, based on speed instability, an unmanned aerial vehicle collision risk assessment model is established, a release interval calibration method is provided, and the probability of unmanned aerial vehicle flight accident is reduced.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) flight safety technology, specifically to a method for calibrating the safe release interval of logistics UAVs based on speed error. Background Technology

[0002] When the volume of logistics drone operations is high enough, release intervals become a crucial issue that must be considered within the organization. From a safety perspective, the main concerns can be divided into two aspects: first, when multiple drones fly along the same route, will a later drone overtake the one in front and cause a collision; second, if there are intersecting routes at the same altitude, will drones collide at the intersection point? The former requires controlling the release interval at the takeoff point, while the latter requires coordinated control of the release intervals between the two takeoff points.

[0003] The core of these two types of interval management lies in the timing control of the UAV's longitudinal trajectory. If the trajectories of all UAVs can be precisely controlled according to the predetermined times, then these interval management issues become very simple. However, due to the influence of many factors such as the UAV control system and its mass, its operating speed is difficult to control precisely, resulting in a certain gap between the actual position and the predetermined position of the UAV. This leads to the uncertainty of the longitudinal trajectory, complicating the release interval management, and the instability of speed becomes a key factor in analyzing release safety.

[0004] Therefore, this study takes the speed instability of UAVs as the starting point, analyzes the safety issues of release management in two scenarios: same route and cross route, and proposes a calibration method for safe release interval of logistics UAVs based on speed error. Summary of the Invention

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0006] This invention provides a method for calibrating the safe release interval of logistics drones based on speed error, comprising the following steps:

[0007] S1: Analysis of key influencing factors for safe release. The flight process of the UAV mainly goes through four stages: climb stage, cruise acceleration stage, cruise constant speed stage and deceleration and landing stage.

[0008] In the case of running on the same route, the landing phase avoids occupying the landing point through existing command control and visual detection. However, during the climb, cruise acceleration, and cruise constant speed phases, there is a problem of the following aircraft catching up with the preceding aircraft and colliding. The main influencing factors include the vertical speed instability of the UAV during the climb phase, the horizontal speed instability during the cruise acceleration and cruise constant speed phases, the takeoff release interval, and the size of the UAV. By conducting a chase-and-collision study, we can analyze the relationship between the release interval and the collision risk under specific UAV speed and size attributes.

[0009] In the cross-route problem, potential collisions between UAVs mainly occur at the intersection of their routes. These collisions are primarily related to factors such as the instability of the UAV's operating speed, the distance between the takeoff points and the intersection, the size of the UAV, and the release interval between the two takeoff points. Taking a 90° cross-route as an example, this study analyzes the relationship between the distance between the two takeoff points and the intersection, the release interval, and the collision risk under specific UAV speed and size attributes.

[0010] S2: Determine the safety target level and clearance interval, and set the acceptable minimum clearance collision probability level; based on the operating area, the drone and the value of the transported goods, formulate the safety target level for collision probability, and then determine the minimum safe clearance interval;

[0011] Based on the principle of equivalent safety level, the safety target level for unmanned aerial vehicles (UAVs) should be at least no lower than that for manned aircraft; referring to the civil aviation field, the safety target level is set at 10. -7 Incident / flight hours.

[0012] S3: Determine the speed error model;

[0013] a. The first stage is the climb stage. The drone's climb speed is unstable due to various factors such as the control system and climb program. The climb stage is relatively short. Assume that the average climb speed of a specific drone model in this stage follows a mean of μ. climb The variance is The normal distribution, i.e., its average climbing speed v climb satisfy:

[0014]

[0015] b. The second stage is the cruise acceleration stage, which is shorter in both time and distance. Assume the drone's acceleration during this stage is a. cruise_a Stable, this stage is considered as uniformly accelerated motion, serving as a transitional phase between the preceding and following stages.

[0016] c. The third stage is the constant speed cruise stage. Assuming the UAV is designed to cruise at a constant speed, the speed during this stage is assumed to follow a normal distribution, i.e., the speed v of the UAV during the constant speed cruise stage...cruise satisfy:

[0017]

[0018] In actual drone flight, speed instability should be continuously changing, and the speed change characteristics should satisfy:

[0019] |v cruise_m+1 -v cruise_m |≤τa max m = 1, 2, ... (3)

[0020] S4: Determine the collision template;

[0021] Referring to the original Reich model collision box, this collision risk model is mainly based on four concepts: flight path system, interval vector, adjacent layer, and collision template. Based on the physical structure characteristics of the rotor shaft of the rotary-wing UAV being of equal length, and the pitching caused by turbulence during flight, a sphere is selected as the collision template.

[0022] Assume the drone's height is h. drone Twice the length of the rotor shaft is l drone Based on the characteristics of the drone's body, the minimum circumscribed sphere of the drone can be simulated. The radius R of the minimum circumscribed sphere is:

[0023]

[0024] A collision occurs when the smallest circumscribed spheres of two drones come into contact or overlap. Assuming the smallest circumscribed sphere radii of drone 1 and drone 2 are R1 and R2 respectively, combining the smallest circumscribed spheres of the two drones yields a sphere with a radius of R1 + R2. This combined sphere is then placed on a drone; this sphere serves as the collision template.

[0025] S5: Collision detection along the same flight path;

[0026] The main form of collisions on the same flight path is that the following aircraft catches up with the preceding aircraft. UAVs have speed instability during both the climb and cruise phases. Therefore, collisions on the same flight path are divided into climb phase collisions and cruise phase collisions.

[0027] a. Collision analysis during the ascent phase;

[0028] Suppose that in a series of consecutive releases of two drones, drone 1 is released first, followed by drone 2, v climb1 v climb2 The average speeds of drone 1 and drone 2 during the climb phase are respectively If the unit is m / s, then:

[0029]

[0030] The judgment result C: Two drones on the same flight path collided during the climb phase. climb for:

[0031]

[0032] b. Cruise phase collision analysis;

[0033] Assuming that the two drones are released sequentially without collision during the climb phase, when drone 1 reaches the cruise phase, drone 2 is still in the climb phase and needs to continue climbing. The remaining time t required for drone 2 to climb to the top is... climb_r2 for:

[0034]

[0035] When drone 2 is at t climb_r2 When reaching the summit within the allotted time, it is necessary to determine whether drone 1 has reached the cruise target speed v within that time. target1 :

[0036]

[0037] If t cruise_a1 Greater than t climb_r2 This means that when drone 2 climbs to the top, drone 1 is still in the cruise acceleration phase, and conversely, drone 1 enters the cruise constant speed phase; then we need to discuss the two scenarios separately.

[0038] a. The first scenario is: when drone 2 reaches the summit, drone 1 is still in the cruise acceleration phase. The distance d1 that drone 1 has flown during this cruise phase is:

[0039]

[0040] When drone 1 reaches the cruise target speed v target1 At that time, we can obtain:

[0041]

[0042]

[0043] When drone 2 reaches the cruise target speed v target2 At that time, we can obtain:

[0044]

[0045]

[0046] b. The second scenario is: when drone 2 reaches the summit, drone 1 has already entered the cruise phase at a constant speed. At this time, the distance d1 that drone 1 has flown during the cruise phase is:

[0047]

[0048] When drone 2 reaches the cruise target speed v target2 At that time, the distances d1 and d2 of the two UAVs during the cruise phase are respectively:

[0049]

[0050]

[0051] Afterwards, both drones maintained a constant speed. The distances d1 and d2 traveled by the two drones during the cruise phase were as follows:

[0052]

[0053]

[0054] The speed v of UAV 2 during the constant speed cruise phase cruise_j satisfy

[0055] The assessment regarding the collision between two drones traveling on the same route during the cruise phase is as follows:

[0056]

[0057] During the cruise phase, if L-d2 > R1+R2 at the end of the cruise phase, where L is the total distance of the cruise phase, it indicates that no overtaking collision occurred during the cruise phase.

[0058] S6: Cross-route collision detection;

[0059] During the climb phase, there is no need to consider collision issues between the two aircraft; collisions only occur near the intersection point. Therefore, the flight conditions for crossing routes differ slightly from those for flying along the same route, and can be divided into two scenarios:

[0060] a. The first scenario is that UAV 1 is released first and reaches the cruise phase first. This scenario is basically the same as the flight of two aircraft on the same route, except that they belong to different routes. The specific steps are shown in formulas (7) to (18) in S5.

[0061] b. The second scenario is that drone 2 is released later but reaches the cruising phase first. At this time, drone 1 is still in the climb phase and needs to continue climbing. Assume that the remaining time required for drone 1 to climb to the top is t. climb_r1 We can obtain:

[0062]

[0063] When drone 1 enters the cruise phase, it is necessary to determine the phase in which drone 2 is located.

[0064] If t climb_r1 Less than t cruise_a2 At this time, UAV 2 is in the cruise acceleration phase, and the flight distance d2 of UAV 2 during the cruise phase is:

[0065]

[0066] If t climb_r1 Greater than or equal to t cruise_a2 At this time, UAV 2 is in the cruise phase with a constant speed. The distance d2 that UAV 2 flies during the cruise phase is:

[0067]

[0068] When drone 1 reaches the cruise target speed v targe At time t1, the distance d1 that UAV 1 flew during the cruise phase is:

[0069]

[0070] If t climb_r1 +t cruise_a1 <t cruise_a2 The cruising flight distance d2 of UAV 2 is:

[0071]

[0072] If t climb_r1 +t cruise_a1 ≥t cruise_a2 The cruising flight distance d2 of UAV 2 is:

[0073]

[0074] Since collision detection for intersecting flight paths only occurs near the intersection point, the intersection point needs to be analyzed separately. Assuming the horizontal distances from the takeoff points of UAV 1 and UAV 2 to the intersection point are D1 and D2 respectively, collision detection begins when UAV 1 is flying D1-(R1+R2) during its cruise phase and ends when it is flying D1+R1+R2. During the detection process, d1 and d2 still represent the distances the two aircraft fly during their cruise phases. The intersection length of the collision template circle and the flight path of UAV 2 is half the length of s1. Therefore:

[0075]

[0076]

[0077] As a preferred technical solution of the present invention, in equation (3), τ is the velocity change period, v cru i se_m v represents the velocity in the previous time period. cruise_m+1 For the velocity in the next time period, a max This represents the maximum change in acceleration.

[0078] As a preferred technical solution of the present invention, UAV 1 is used as the UAV with an external collision template. The center of the collision template is the center of mass of UAV 1. At this time, another UAV can be regarded as a point mass, referred to as UAV 2 in this article. When the point mass is located at the edge of the collision template or enters the collision template, it is considered that UAV 1 and UAV 2 collide.

[0079] As a preferred technical solution of the present invention, in equations (5) and (6), t climb_diff The time difference between the time when the following aircraft just touches the collision template and the time when the preceding aircraft reaches the top is the difference in meters (m); H is the flight path altitude, i.e., the altitude during the climb phase, in meters (m); R1+R2 is the radius of the collision template, in meters (m); and the interval between the two aircraft is T, in seconds (s).

[0080] As a preferred technical solution of the present invention, in equations (7) and (8): a cruise_a1 The acceleration of UAV 1 during the cruise acceleration phase is expressed in m / s². -2 ;t cruise_a1 The time taken by UAV 1 during the cruise acceleration phase is expressed in seconds.

[0081] As a preferred technical solution of the present invention, in equations (9), (10), (11), (12) and (13), a cruise_a2 The acceleration of UAV 2 during the cruise acceleration phase is expressed in m / s². -2 d1 and d2 represent the values ​​when drone 1 reaches v. target1 The distance between the two aircraft during the cruise phase, measured in meters (m); t cruise_a2 The time taken by UAV 2 during the cruise acceleration phase is measured in seconds; the speed of UAV 1 during the constant speed cruise phase is measured in seconds.

[0082] As a preferred technical solution of the present invention, in formula (21), a cruise_a2 The acceleration of UAV 2 during the cruise acceleration phase is expressed in m / s². -2 ;t cruise_a2 The time taken by UAV 2 during the cruise acceleration phase is expressed in seconds.

[0083] The beneficial effects of this invention are: This method for calibrating the safe release interval of logistics drones based on speed error, for two drone operation release management scenarios of same-line operation and cross-line operation, takes into account the drone's climb and cruise phases, establishes a drone collision risk assessment model based on speed instability, and provides a release interval calibration method, thereby reducing the probability of drone flight accidents. Attached Figure Description

[0084] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0085] Figure 1 This is a schematic diagram of the same route operation of the logistics drone safety release interval calibration method based on speed error according to the present invention;

[0086] Figure 2 This is a schematic diagram of the cross-route operation of the logistics drone safety release interval calibration method based on speed error according to the present invention;

[0087] Figure 3 This is a top view of the intersection point of the present invention's method for calibrating the safe release interval of logistics drones based on speed error;

[0088] Figure 4 This is a schematic diagram of the simulation results of the same flight route for the safety release interval calibration method for logistics drones based on speed error according to the present invention;

[0089] Figure 5 This is a schematic diagram of the cross-route simulation results (D1=D2) of the safety release interval calibration method for logistics drones based on speed error according to the present invention;

[0090] Figure 6 This is a schematic diagram of the cross-route simulation results (T=0) of the safety release interval calibration method for logistics drones based on speed error according to the present invention;

[0091] Figure 7 This is a schematic diagram of the T, D, and P relationship in the safety release interval calibration method for logistics drones based on speed error according to the present invention;

[0092] Figure 8 This is a schematic diagram of a sample cutoff point for the method of calibrating the safe release interval of logistics drones based on speed error according to the present invention;

[0093] Figure 9 This is a schematic diagram of the sample point fitting line of the present invention, which is a method for calibrating the safe release interval of logistics drones based on speed error.

[0094] Figure 10This is a schematic diagram of the T, D, and P fitting Gaussian surfaces for the safety release interval calibration method for logistics drones based on speed error in this invention. Detailed Implementation

[0095] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0096] Example: Figure 1-3 As shown, the method for calibrating the safe release interval of logistics drones based on speed error includes the following steps:

[0097] S1: Analysis of key influencing factors for safe release. The flight process of the UAV mainly goes through four stages: climb stage, cruise acceleration stage, cruise constant speed stage and deceleration and landing stage.

[0098] In the case of running on the same route, the landing phase avoids occupying the landing point through existing command control and visual detection. However, during the climb, cruise acceleration, and cruise constant speed phases, there is a problem of the following aircraft catching up with the preceding aircraft and colliding. The main influencing factors include the vertical speed instability of the UAV during the climb phase, the horizontal speed instability during the cruise acceleration and cruise constant speed phases, the takeoff release interval, and the size of the UAV. By conducting a chase-and-collision study, we can analyze the relationship between the release interval and the collision risk under specific UAV speed and size attributes.

[0099] In the cross-route problem, potential collisions between UAVs mainly occur at the intersection of their routes. These collisions are primarily related to factors such as the instability of the UAV's operating speed, the distance between the takeoff points and the intersection, the size of the UAV, and the release interval between the two takeoff points. Taking a 90° cross-route as an example, this study analyzes the relationship between the distance between the two takeoff points and the intersection, the release interval, and the collision risk under specific UAV speed and size attributes.

[0100] S2: Determine the safety target level and clearance interval, and set the acceptable minimum clearance collision probability level; based on the operating area, the drone and the value of the transported goods, formulate the safety target level for collision probability, and then determine the minimum safe clearance interval;

[0101] Based on the principle of equivalent safety level, the safety target level for unmanned aerial vehicles (UAVs) should be at least no lower than that for manned aircraft; referring to the civil aviation field, the safety target level is set at 10. -7 Incident / flight hours.

[0102] S3: Determine the speed error model;

[0103] a. The first stage is the climb stage. The drone's climb speed is unstable due to various factors such as the control system and climb program. The climb stage is relatively short. Assume that the average climb speed of a specific drone model in this stage follows a mean of μ. climb The variance is The normal distribution, i.e., its average climbing speed v climb satisfy:

[0104]

[0105] b. The second stage is the cruise acceleration stage, which is shorter in both time and distance. Assume the drone's acceleration during this stage is a. cruise_a Stable, this stage is considered as uniformly accelerated motion, serving as a transitional phase between the preceding and following stages.

[0106] c. The third stage is the constant speed cruise stage. Assuming the UAV is designed to cruise at a constant speed, the speed during this stage is assumed to follow a normal distribution, i.e., the speed v of the UAV during the constant speed cruise stage... cruise satisfy:

[0107]

[0108] In actual drone flight, speed instability should be continuously changing, and the speed change characteristics should satisfy:

[0109] |v cruise_m+1 -v cruise_m |≤τa max m = 1, 2, ... (3)

[0110] S4: Determine the collision template;

[0111] Referring to the original Reich model collision box, this collision risk model is mainly based on four concepts: flight path system, interval vector, adjacent layer, and collision template. Based on the physical structure characteristics of the rotor shaft of the rotary-wing UAV being of equal length, and the pitching caused by turbulence during flight, a sphere is selected as the collision template.

[0112] Assume the drone's height is h. drone Twice the length of the rotor shaft is l drone Based on the characteristics of the drone's body, the minimum circumscribed sphere of the drone can be simulated. The radius R of the minimum circumscribed sphere is:

[0113]

[0114] A collision occurs when the smallest circumscribed spheres of two drones come into contact or overlap. Assuming the smallest circumscribed sphere radii of drone 1 and drone 2 are R1 and R2 respectively, combining the smallest circumscribed spheres of the two drones yields a sphere with a radius of R1 + R2. This combined sphere is then placed on a drone; this sphere serves as the collision template.

[0115] S5: Collision detection along the same flight path;

[0116] The main form of collisions on the same flight path is that the following aircraft catches up with the preceding aircraft. UAVs have speed instability during both the climb and cruise phases. Therefore, collisions on the same flight path are divided into climb phase collisions and cruise phase collisions.

[0117] a. Collision analysis during the ascent phase;

[0118] Suppose that in a series of consecutive releases of two drones, drone 1 is released first, followed by drone 2, v climb1 v climb2 The average speeds of drone 1 and drone 2 during the climb phase are respectively If the unit is m / s, then:

[0119]

[0120] The judgment result C: Two drones on the same flight path collided during the climb phase. climb for:

[0121]

[0122] b. Cruise phase collision analysis;

[0123] Assuming that the two drones are released sequentially without collision during the climb phase, when drone 1 reaches the cruise phase, drone 2 is still in the climb phase and needs to continue climbing. The remaining time t required for drone 2 to climb to the top is... climb_r2 for:

[0124]

[0125] When drone 2 is at t climb_r2 When reaching the summit within the allotted time, it is necessary to determine whether drone 1 has reached the cruise target speed v within that time. target1 :

[0126]

[0127] If t cruise_a1 Greater than t climb_r2This means that when drone 2 climbs to the top, drone 1 is still in the cruise acceleration phase, and conversely, drone 1 enters the cruise constant speed phase; then we need to discuss the two scenarios separately.

[0128] a. The first scenario is: when drone 2 reaches the summit, drone 1 is still in the cruise acceleration phase. The distance d1 that drone 1 has flown during this cruise phase is:

[0129]

[0130] When drone 1 reaches the cruise target speed v target1 At that time, we can obtain:

[0131]

[0132]

[0133] When drone 2 reaches the cruise target speed v target2 At that time, we can obtain:

[0134]

[0135]

[0136] b. The second scenario is: when drone 2 reaches the summit, drone 1 has already entered the cruise phase at a constant speed. At this time, the distance d1 that drone 1 has flown during the cruise phase is:

[0137]

[0138] When drone 2 reaches the cruise target speed v target2 At that time, the distances d1 and d2 of the two UAVs during the cruise phase are respectively:

[0139]

[0140]

[0141] Afterwards, both drones maintained a constant speed. The distances d1 and d2 traveled by the two drones during the cruise phase were as follows:

[0142]

[0143]

[0144] The speed v of UAV 2 during the constant speed cruise phase cruise_j satisfy

[0145] The assessment regarding the collision between two drones traveling on the same route during the cruise phase is as follows:

[0146]

[0147] During the cruise phase, if L-d2 > R1+R2 at the end of the cruise phase, where L is the total distance of the cruise phase, it indicates that no overtaking collision occurred during the cruise phase.

[0148] S6: Cross-route collision detection;

[0149] During the climb phase, there is no need to consider collision issues between the two aircraft; collisions only occur near the intersection point. Therefore, the flight conditions for crossing routes differ slightly from those for flying along the same route, and can be divided into two scenarios:

[0150] a. The first scenario is that UAV 1 is released first and reaches the cruise phase first. This scenario is basically the same as the flight of two aircraft on the same route, except that they belong to different routes. The specific steps are shown in formulas (7) to (18) in S5.

[0151] b. The second scenario is that drone 2 is released later but reaches the cruising phase first. At this time, drone 1 is still in the climb phase and needs to continue climbing. Assume that the remaining time required for drone 1 to climb to the top is t. climb_r1 We can obtain:

[0152]

[0153] When drone 1 enters the cruise phase, it is necessary to determine the phase in which drone 2 is located.

[0154] If t climb_r1 Less than t cruise_a2 At this time, UAV 2 is in the cruise acceleration phase, and the flight distance d2 of UAV 2 during the cruise phase is:

[0155]

[0156] If t climb_r1 Greater than or equal to t cruise_a2 At this time, UAV 2 is in the cruise phase with a constant speed. The distance d2 that UAV 2 flies during the cruise phase is:

[0157]

[0158] When drone 1 reaches the cruise target speed v targe At time t1, the distance d1 that UAV 1 flew during the cruise phase is:

[0159]

[0160] If t climb_r1 +t cruise_a1 <t cruise_a2 The cruising flight distance d2 of UAV 2 is:

[0161]

[0162] If t climb_r1 +t cruise_a1 ≥t cruise_a2 The cruising flight distance d2 of UAV 2 is:

[0163]

[0164] Since collision detection for intersecting flight paths only occurs near the intersection point, the intersection point needs to be analyzed separately. Assuming the horizontal distances from the takeoff points of UAV 1 and UAV 2 to the intersection point are D1 and D2 respectively, collision detection begins when UAV 1 is flying D1-(R1+R2) during its cruise phase and ends when it is flying D1+R1+R2. During the detection process, d1 and d2 still represent the distances the two aircraft fly during their cruise phases. The intersection length of the collision template circle and the flight path of UAV 2 is half the length of s1. Therefore:

[0165]

[0166]

[0167] (3) In the formula, τ is the period of velocity change, v cruise_m v represents the velocity in the previous time period. cruise_m+1 For the velocity in the next time period, a max This represents the maximum change in acceleration.

[0168] With UAV 1 as the external collision template, the center of the collision template is the center of mass of UAV 1. At this time, another UAV can be regarded as a point mass, referred to as UAV 2 in this paper. When the point mass is located at the edge of the collision template or enters the collision template, it is considered that UAV 1 and UAV 2 collide.

[0169] In equations (5) and (6), t climb_diff The time difference between the time when the following aircraft just touches the collision template and the time when the preceding aircraft reaches the top is the difference in meters (m); H is the flight path altitude, i.e., the altitude during the climb phase, in meters (m); R1+R2 is the radius of the collision template, in meters (m); and the interval between the two aircraft is T, in seconds (s).

[0170] In equations (7) and (8): a cruise_a1 The acceleration of UAV 1 during the cruise acceleration phase is expressed in m / s². -2 ;t cruise_a1 The time taken by UAV 1 during the cruise acceleration phase is expressed in seconds.

[0171] In equations (9), (10), (11), (12), and (13), a cruise_a2The acceleration of UAV 2 during the cruise acceleration phase is expressed in m / s². -2 d1 and d2 represent the values ​​when drone 1 reaches v. target1 The distance between the two aircraft during the cruise phase, measured in meters (m); t cruise_a2 The time taken by UAV 2 during the cruise acceleration phase is measured in seconds; the speed of UAV 1 during the constant speed cruise phase is measured in seconds.

[0172] (21) In the formula, a cruise_a2 The acceleration of UAV 2 during the cruise acceleration phase is expressed in m / s². -2 ;t cruise_a2 The time taken by UAV 2 during the cruise acceleration phase is expressed in seconds.

[0173] This method for calibrating the safe release interval of logistics drones based on speed error determines the safe release interval, which is the release time interval selected according to a set acceptable collision probability level; it depends on the relationship between the release time interval and the collision probability. Based on the collision determination method presented in this paper, Monte Carlo simulation becomes an important method for determining the relationship between different release intervals and collision probabilities. The idea is to use the law of large numbers and statistical inference methods to repeatedly conduct experiments (or use sufficiently long simulation runs) to obtain the desired results. Combining theoretical judgment and this simulation method, the steps for calibrating the safe release interval are as follows:

[0174] 1) Safety level selection. Set the minimum acceptable probability of a collision before departure;

[0175] 2) UAV performance parameter analysis. Based on historical test and operational data of the UAV model used, determine the flight speed performance characteristics of the UAV at each stage;

[0176] 3) Collision Probability Simulation Calculation. Using the Monte Carlo simulation method, based on parameters such as altitude and distance in specific operational scenarios (including same-line operation and same-altitude intersecting line operation), as well as the UAV's operational performance characteristics, a release interval T is set (in intersecting lines, the distance from the takeoff point of both aircraft to the intersection point also needs to be set), and operational simulations are performed to determine whether the two aircraft will collide. Assuming that each scenario has N simulations, and the number of times the two aircraft collide in N simulations is C, then the collision probability P under the corresponding release interval T is P = C / N;

[0177] 4) Fitting the relationship between release interval and collision probability. For same-line operations, the probability of overtaking collision under multiple release interval parameters is obtained through simulation analysis, and then the curve of release interval and collision probability function is fitted; for cross-line operations, by setting different release intervals and distance parameters from the takeoff point to the cross point, the cross-collision probability under different conditions is obtained through simulation analysis, and then the relationship between release interval, distance and collision probability function is fitted.

[0178] 5) Selection of release intervals based on acceptable safety levels. Based on the fitted collision probability influence function, and considering the acceptable collision probability level, the release interval is determined. For intersecting routes, the specific distance parameter between the takeoff point and the intersection point needs to be considered to finally determine the safe release interval.

[0179] To verify the feasibility of the method, a case study analysis was conducted using a specific type of rotary-wing logistics drone operating in a certain city as the research object. The case study assumed a collision probability of 10. -7 To achieve the target safety level, historical flight data from 135 flight missions of this UAV model were collected, all operating under similar environmental conditions. Parameter analysis was performed on the operational characteristics of each phase, yielding performance parameters as shown in Table 1. Specifically, 135 average climb velocities were obtained for the climb phase, while for the cruise constant speed phase, due to limitations in data collection technology, only one velocity data point per second was obtained, totaling 56,655 velocity data points. The Kolmogorov-Sminov test (KS test) was used to verify the velocity data distribution characteristics of the two phases. The results showed significance levels of 0.20 and 0.053, respectively, indicating that the velocity characteristics of the two phases conform to a normal distribution, consistent with the assumptions made in this paper regarding velocity instability.

[0180] Table 1. Parameters related to UAVs and airways

[0181]

[0182] Based on the parameters in Table 1, Monte Carlo simulations were performed using MATLAB R2021b to analyze the collision probability characteristics of this aircraft type under different release intervals during operation on the same and cross routes.

[0183] 1) Results of the same route

[0184] Simulation analysis was conducted to determine the collision probability under different release intervals during operation on the same flight route. By fitting the trend of the results, the relationship between the collision probability and the release interval for two aircraft on the same route was obtained. For ease of observation, the ordinary coordinate system was converted to a logarithmic coordinate system, such as... Figure 4 As shown in the figure, the dashed line represents the release interval when the collision probability is 10⁻⁷. Figure 4 The data shows that as the clearance interval increases, the probability of collision between the two drones gradually decreases. When two drones are cleared at a time interval of 122 seconds, the collision probability between the two drones just meets the ELOS requirement.

[0185] 2) Results of cross-route

[0186] In intersecting flight paths, the probability of a collision at the intersection point is related to both the distance between the takeoff points of the two aircraft and the intersection point, as well as the clearance interval. First, we analyze the case where the two distances are equal, i.e., D1 = D2. Simulations yield the following relationship between the probability of a collision and the clearance interval: Figure 5 As shown. In this scenario, to meet ELOS requirements, the safe release interval between the two drones should be at least 86 seconds.

[0187] However, the horizontal distances from the takeoff points of the two aircraft to the intersection point are not the same; that is, there is a difference D between D1 and D2. After fitting the simulation results, it was found that if the two aircraft take off simultaneously (T=0), the relationship between the collision probability and D is as follows: Figure 6 As shown in the figure. Clearly, the distance between the takeoff points of the two drones and the intersection point affects the safe clearance interval between them. Therefore, the clearance interval T and the difference D are simulated simultaneously, and the relationship between T, D, and the collision probability P is obtained as follows: Figure 7 As shown.

[0188] like Figure 7 As shown, the surface formed by T, D, and P approximates a Gaussian cylinder structure. A Gaussian cylinder is a surface obtained by translating a Gaussian function curve on a plane along the normal direction of that plane. Although Figure 7 The surface contains Gaussian peaks caused by the discreteness of the simulation data, but there is an approximately linear relationship between T and D, making the collision probability of each point on the same straight line equal.

[0189] To verify this feature, the surface is intersected at regular intervals using equiprobability surfaces, yielding several points with equal probability to the cross-sections. For example, with collision probabilities of 0.001, 0.01, and 0.02, the intercept points obtained after intersecting the surface are as follows: Figure 8 As shown, they all exhibit approximately linear characteristics, and the straight lines obtained after fitting are as follows: Figure 9 As shown (R2 = 0.996).

[0190] Finally, using this method, the multiple lines with different probabilities are fitted into a Gaussian cylinder, as shown in the following result. Figure 10 As shown, under this result, taking the collision probability of 10⁻⁷ as the result and extracting the result on the Gaussian cylinder, we can obtain the relationship between T and D that just satisfies the ELOS requirement: T = (D ± 1199.97) / 14 (T ≥ 0). That is, for this UAV model, the safe release interval is a function of the distance difference between the takeoff point and the intersection point, and the release interval can be determined based on the distance difference D.

[0191] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for calibrating the safe release interval of logistics drones based on speed error, characterized in that, Includes the following steps: S1: Analysis of key influencing factors for safe release. The flight process of the UAV mainly goes through four stages: climb stage, cruise acceleration stage, cruise constant speed stage and deceleration and landing stage. In the case of running on the same route, the landing phase avoids occupying the landing point through existing command control and visual detection. However, during the climb, cruise acceleration, and cruise constant speed phases, there is a problem of the following aircraft catching up with the preceding aircraft and colliding. The main influencing factors include the vertical speed instability of the UAV during the climb phase, the horizontal speed instability during the cruise acceleration and cruise constant speed phases, the takeoff release interval, and the size of the UAV. By conducting a chase-and-collision study, we can analyze the relationship between the release interval and the collision risk under specific UAV speed and size attributes. In the cross-route problem, the main possible collisions of UAVs occur at the intersection of the routes of each UAV. The collisions here are mainly related to factors such as the instability of the UAV's operating speed, the distance between the take-off points of the two routes and the intersection point, the size of the UAV, and the release interval between the two take-off points. Taking a 90° cross-route as the object, we analyze the relationship between the distance between the two take-off points and the intersection point and the release interval and the collision risk under specific UAV speed and size attributes. S2: Determine the safety target level and clearance interval, and set the acceptable minimum clearance collision probability level; based on the operating area, the drone and the value of the transported goods, formulate the safety target level for collision probability, and then determine the minimum safe clearance interval; Based on the principle of equivalent safety level, the safety target level for unmanned aerial vehicles (UAVs) should be at least no lower than that for manned aircraft; referring to the civil aviation field, the safety target level is set at 10. -7 Incident / flight hours; S3: Determine the speed error model; a. The first stage is the climb stage, during which the drone's climb speed is unstable due to various factors such as the control system and the climb program; The climb phase is relatively short. Assuming that the average climb rate of a specific drone model during this phase follows a mean of [value missing], [value missing]. The variance is The normal distribution of its average climbing speed satisfy: （1） b. The second stage is the cruise acceleration stage, which is shorter in both time and distance. Assume the drone's acceleration during this stage is... Stable, this stage is considered as uniformly accelerated motion, serving as a transitional phase between the preceding and following stages; c. The third stage is the constant speed cruise stage. Assuming the UAV is designed to cruise at a constant speed, the speed during this stage is assumed to follow a normal distribution. satisfy: （2） In actual drone flight, speed instability should be continuously changing, and the speed change characteristics should satisfy: （3） In equation (3), For the period of velocity change, The speed of the previous period. For the speed of the next period, The maximum change in acceleration; S4: Determine the collision template; Referring to the original Reich model collision box, this collision risk model is mainly based on four concepts: flight path system, interval vector, adjacent layer, and collision template. Based on the physical structure characteristics of the rotor shaft of the rotary-wing UAV being of equal length, and the pitching caused by turbulence during flight, a sphere is selected as the collision template. Assuming the drone's body height is Twice the length of the rotor shaft is Based on the characteristics of the drone's body, the minimum circumscribed sphere of the drone can be simulated, and the radius of the minimum circumscribed sphere is... for: （4） A collision is considered to occur when the smallest circumscribed spheres of two drones come into contact or overlap; assuming the smallest circumscribed sphere radii of drone 1 and drone 2 are respectively... and By combining the smallest circumscribed spheres of the two drones, a sphere with radius [missing information] is obtained. + The sphere is assembled and placed on a drone; this sphere is the collision template. S5: Collision detection along the same flight path; The main form of collisions on the same flight path is that the following aircraft catches up with the preceding aircraft. UAVs have speed instability during both the climb and cruise phases. Therefore, collisions on the same flight path are divided into climb phase collisions and cruise phase collisions. a. Collision analysis during the ascent phase; Suppose that in a series of consecutive releases of two drones, drone 1 is released first, followed by drone 2. , The average speeds of drone 1 and drone 2 during the climb phase are respectively , The unit is m / s, then we have: （5） The assessment result of two drones on the same flight path colliding during the climb phase. for: （6） In equations (5) and (6), The time when the rear machine just touches the collision template while the front machine is climbing to the top is the difference between the time when the front machine reaches the top and the time when the front machine reaches the top, in meters; This refers to the altitude during the climb phase, measured in meters (m). + The radius of the collision template is in meters; the release interval between the two machines is... The unit is seconds (s). b. Cruise phase collision analysis; Assuming the two drones are released sequentially without collision during the climb phase, when drone 1 reaches the cruise phase, drone 2 is still in the climb phase and needs to continue climbing. The remaining time required for drone 2 to climb to the top is... for: （7） When drone 2 is When reaching the summit within the allotted time, it is necessary to determine whether drone 1 has reached the cruise target speed within that time. : （8） In equations (7) and (8): The acceleration of UAV 1 during the cruise acceleration phase is expressed in m / s². 2 ; The time taken by UAV 1 during the cruise acceleration phase is expressed in seconds. if Greater than This means that when drone 2 climbs to the top, drone 1 is still in the cruise acceleration phase, and conversely, drone 1 enters the cruise constant speed phase; then we need to discuss the two scenarios separately. a. The first scenario is: when drone 2 reaches the summit, drone 1 is still in the cruise acceleration phase. The distance drone 1 has flown during this cruise phase is... for: (9) When drone 1 reaches the cruise target speed At that time, we can obtain: (10) (11) When drone 2 reaches the cruise target speed At that time, we can obtain: (12) (13) In equations (9), (10), (11), (12), and (13), The acceleration of UAV 2 during the cruise acceleration phase is expressed in m / s². 2 ; and When drone 1 reaches The distance between the two aircraft during the cruise phase, expressed in meters (m). The time taken by UAV 2 during the cruise acceleration phase is expressed in seconds. Speed ​​of UAV 1 during the constant speed cruise phase ; b. The second scenario is: when drone 2 reaches the summit, drone 1 has already entered the cruise phase at a constant speed. The distance drone 1 has flown during this cruise phase is... for: （14） When drone 2 reaches the cruise target speed At that time, the distance the two drones flew during the cruise phase , They are respectively: （15） （16） Both drones then maintained a constant speed, and the distance they flew during the cruise phase was... , They are respectively: （17） （18） Speed ​​of UAV 2 during the constant speed cruise phase satisfy ; The assessment regarding the collision between two drones traveling on the same route during the cruise phase is as follows: （19） During the cruise phase, if the preceding aircraft finishes its cruise... ,in If the total distance during the cruise phase is 1, it indicates that no collision occurred during the cruise phase. S6: Cross-route collision detection; During the climb phase, there is no need to consider collision issues between the two aircraft; collisions only occur near the intersection point. Therefore, the flight conditions for crossing routes differ slightly from those for flying along the same route, and can be divided into two scenarios: a. The first scenario is that UAV 1 is released first and reaches the cruise phase first. This scenario is basically the same as the flight of two aircraft on the same route, except that they belong to different routes. The specific steps are shown in formulas (7) to (18) in S5. b. The second scenario is that drone 2 is released later but reaches the cruising phase first. At this time, drone 1 is still in the climb phase and needs to continue climbing. Assume that the remaining time required for drone 1 to climb to the top is... We can obtain: （20） When drone 1 enters the cruise phase, it is necessary to determine the phase in which drone 2 is located. like Less than At this time, UAV 2 is in the cruise acceleration phase, and the distance UAV 2 flies during the cruise phase is... for: （21） In equation (21), The acceleration of UAV 2 during the cruise acceleration phase is expressed in m / s². 2 ; The time taken by UAV 2 during the cruise acceleration phase is expressed in seconds. like Greater than or equal to At this time, UAV 2 is in the cruise constant speed phase, and the distance UAV 2 flies during the cruise phase is... for: （22） When drone 1 reaches the cruise target speed At that time, the distance that UAV 1 flew during the cruise phase for: （23） like The cruising flight distance of drone 2 for: （24） like The cruising flight distance of drone 2 for: （25） Since collision detection for intersecting flight paths only occurs near the intersection point, the intersection point needs to be analyzed separately. Assume the horizontal distances from the takeoff points of UAV 1 and UAV 2 to the intersection point are respectively... and Collision determination begins from the time drone 1 is in cruise phase. From the start of the flight End of judgment process; and This still represents the distance the two aircraft flew during the cruise phase; half the length of the intersection line between the collision template plane circle and the flight path of UAV 2 is... We can obtain: （26） （27）。 2. The method for calibrating the safe release interval of logistics drones based on speed error according to claim 1, characterized in that, With UAV 1 as the external collision template, the center of the collision template is the center of mass of UAV 1. At this time, another UAV can be regarded as a point mass, referred to as UAV 2 in this paper. When the point mass is located at the edge of the collision template or enters the collision template, it is considered that UAV 1 and UAV 2 collide.

Images