An optimization method of a UAV communication and perception integrated system

Abstract

This invention provides an optimization method for an integrated communication and sensing system for unmanned aerial vehicles (UAVs), comprising: constructing an integrated communication and sensing system model assisted by a UAV; wherein the system model includes several constraints; under the several constraints, jointly optimizing sensing scheduling, communication scheduling, UAV flight trajectory, UAV transmission power, and beamforming to construct a non-convex optimization problem that minimizes the average information age of the sensed information; repeatedly solving the non-convex optimization problem to continuously update the average information age until it is determined that the average information age update has no change, at which point the optimization ends. This invention can minimize the information age of the sensed information and improve the timeliness of the sensed information.

Description

Technical Field

[0001] This invention relates to the field of integrated communication and sensing, and in particular to an optimization method for an integrated communication and sensing system for unmanned aerial vehicles (UAVs). Background Technology

[0002] Currently, both academia and industry are beginning to study the vision, requirements, scenarios, key technologies, and system architecture of 6G networks. However, due to the massive number of connected devices and the rapid development of the wireless communication industry, spectrum resources are becoming increasingly scarce. Therefore, the auction price of wireless spectrum has risen sharply in recent years, and network providers are actively seeking reusable frequency bands, with radar bands considered one of the best candidates for shared communication bands. Meanwhile, many visions for next-generation wireless communication, such as smart cities and industry, require high-quality communication capabilities and high-precision sensing capabilities, especially in applications requiring location / environment awareness. For these two reasons, Integrated Sensing and Communications (ISAC) has recently received extensive research from academia.

[0003] Unlike traditional Radar-Communication Coexistence (RCC) technologies where radar and communication modules are designed with separate components and signals, ISAC allows the communication and radar modules to use the same hardware and spectrum resources, further improving the efficiency of spectrum, energy, and hardware utilization. Furthermore, through hardware and waveform design, the two wireless functions can complement each other, such as communication assisting sensing or sensing assisting communication. However, while next-generation wireless communication technologies, such as the millimeter-wave band, offer higher transmission bandwidth and provide higher range resolution for sensing, high-frequency signals have weak transmission performance and face more severe losses during transmission, including more severe free path loss, atmospheric absorption, and rain attenuation. Therefore, for millimeter-wave information transmission, multiple-input multiple-output (MIMO) technology is often used in conjunction with beamforming design of the transmitted signal. Even with MIMO support, due to limited transmission power and obstructions from surrounding obstacles, ground-based ISAC base stations can only provide service within a fixed range.

[0004] Thanks to the rapid and flexible deployment and high mobility of small drones, their research and application in the field of communications have attracted widespread attention from scholars both domestically and internationally. Drones can be used as aerial base stations, adjusting their altitude to avoid obstacles and increasing the possibility of line-of-sight communication, complementing ground networks and providing high-speed, low-latency supplementary services to hotspot areas. Therefore, drones hold promise as an excellent aerial ISAC platform, enhancing the services of ground-based ISAC base stations.

[0005] The existing technologies similar to the scheme in this application are mainly found in the following two journal articles: "Joint Maneuver and Beamforming Design for UAV-Enabled Integrated Sensing and Communication" and "Throughput Maximization for UAV-enabled Integrated Periodic Sensing and Communication". In the first article, the authors studied an UAV ISAC system, maximizing the overall communication throughput of the system while meeting the sensing requirements by optimizing the UAV's flight trajectory, communication beam, and sensing beam. In the second article, the authors proposed the "asymmetry" between sensing and communication, arguing that it is sufficient to sense objects of interest at a certain frequency, and proposed a periodic ISAC framework to optimize the UAV's trajectory and beamforming to maximize the overall communication throughput of the system.

[0006] However, both of the above studies neglected the feedback phase of perceived information, which is not conducive to its application in real-time decision-making. In order to achieve real-time decision-making, the UAV must transmit the perceived information back to the data center / perceiving user as soon as possible after completing the perception process. Summary of the Invention

[0007] To overcome the shortcomings of the prior art, this invention provides an optimization method for an integrated communication and sensing system for unmanned aerial vehicles (UAVs), which can minimize the information age of the sensed information and improve the timeliness of the sensed information.

[0008] An embodiment of the present invention provides an optimization method for an integrated communication and sensing system for unmanned aerial vehicles (UAVs), comprising the following steps:

[0009] Construct a communication and sensing integrated system model assisted by unmanned aerial vehicles (UAVs); wherein the system model includes several constraints.

[0010] Under the aforementioned constraints, a non-convex optimization problem is constructed that minimizes the average information age of the sensed information by jointly optimizing the sensing scheduling, communication scheduling, UAV flight trajectory, UAV transmission power, and beamforming.

[0011] The non-convex optimization problem is solved repeatedly to keep the average information age updated until it is determined that the average information age has no change, at which point the optimization ends.

[0012] Furthermore, the construction of the UAV-assisted integrated communication and sensing system model specifically includes:

[0013] Define user set k = V + U, where the first V users in the user set are sensing users, and the remaining U users are communication users. The set of sensing targets is defined as follows: Each element corresponds one-to-one with the first V users in k;

[0014] Define the number of antennas carried by the drone as N. t The flight cycle of the UAV is T, which is evenly divided into N time slots, where the length of each time slot is τ = T / N;

[0015] Let the time slot set be denoted as The drone's flight trajectory is then approximated as follows:

[0016]

[0017] in It is the three-dimensional position of the UAV in time slot n. h[n] is the horizontal position of the UAV, and h[n] is the flight altitude of the UAV in time slot n;

[0018] Define the maximum horizontal flight speed of the UAV during the entire mission cycle as The maximum vertical flight speed is And the drone needs to start from a given location The drone departs and returns to the same point at the end of its flight cycle, with its maximum and minimum flight altitudes recorded as h. max and h min The trajectory of the drone is subject to the following constraints:

[0019] q u [1] = q U ;

[0020]

[0021]

[0022]

[0023]

[0024]

[0025] Let p[n] be the transmit power of the UAV in time slot n. Then the transmit power meets the following peak and total value constraints:

[0026]

[0027]

[0028] Among them, Pmax The instantaneous maximum transmit power, This represents the total power that the drone can launch;

[0029] Model the probability of a direct link between the drone and communication user k:

[0030]

[0031] Where a and b are the parameters of the S-curve. Let $\frac{ ... The distance from the drone to the communication user k in time slot n;

[0032] Define a binary variable Used to represent the scheduling status of UAVs to communication users in time slot n, when When the drone communicates with user k, otherwise...

[0033] The following communication flight elevation angle constraints also exist for drones:

[0034]

[0035] in, It is the probability of acceptable line-of-sight. The corresponding angle of elevation;

[0036] If a drone is limited to communicating with a maximum of one user at a time, then:

[0037]

[0038]

[0039] According to the free-space loss model, the channel gain from the UAV to the communication user k is:

[0040]

[0041] Where β0 is the channel power gain when the reference distance is 1m;

[0042] The baseband equivalent channel from the drone to communication user k is represented as:

[0043]

[0044] Where, λ c The wavelength of the transmitted signal, Let be the guidance vector from the UAV to the communication user k. It is the elevation angle of the drone to the communication user k in time slot n;

[0045] Let the information sent by the drone to user k be s. k [n], s k It follows a standard normal distribution, and the information sent to each user is independent of each other, that is... The signal transmitted by the drone is Where w[n] is the precoding vector in time slot n, which meets the following constraints:

[0046]

[0047] Therefore, the throughput of user k in time slot n is:

[0048]

[0049] Where B is the signal bandwidth, σ 2 The power of augmented Gaussian white noise;

[0050] Let R be the throughput threshold for communication users. th Then the throughput constraint for communication users is:

[0051]

[0052] Define another binary variable This is used to represent the perception scheduling of a UAV to target point v in time slot n. The UAV only... When sensing the target point v, the sensing elevation angle constraint is expressed as:

[0053]

[0054] in, It is the distance from the drone to the target point v. It is the acceptable flight elevation angle threshold during perception.

[0055] If a drone is required to hover while performing sensing operations, then there are hovering constraints:

[0056]

[0057] The beam intensity reaching the sensing point is denoted as... in It is the guidance vector of the UAV to the target point v. If the elevation angle of the UAV to the target point v is the beam angle, then the beam strength must be greater than the given beam strength threshold Γ. th ,Right now:

[0058]

[0059] If the drone is limited to sensing only one target at a time, and each target must be sensed once during the entire flight cycle, then:

[0060]

[0061]

[0062]

[0063] Let R be the amount of information generated in a single perception. s The time when the drone completes transmitting the perceived information back to the sensing user is recorded as The throughput constraint for perceived users is then expressed as:

[0064]

[0065]

[0066] The scheduling variable meets the following causality constraints:

[0067]

[0068]

[0069]

[0070] Furthermore, the non-convex optimization problem of constructing the minimum average information age of the perceived information specifically includes:

[0071] By unifying constraints (12) and (19), we obtain the following constraints:

[0072]

[0073]

[0074] in, It is the throughput threshold for user k;

[0075] Let the time from the generation of perceived information to the completion of the feedback of perceived information be defined as minimizing the average information age of the perceived information. Then, the average information age of minimizing the perceived information is:

[0076]

[0077] in, The moment when the drone senses the target;

[0078] If minimizing the average information age of perceived information is taken as the optimization objective, then the following non-convex optimization problem exists:

[0079]

[0080] st(1)-(10),(13)-(18),(20)-(25).

[0081] Furthermore, solving the nonconvex optimization problem specifically includes:

[0082] The non-convex optimization problem is simplified by applying constraints to the closed-form solution of the beamforming to obtain a simplified problem.

[0083] The simplified problem is broken down into two sub-problems and solved separately.

[0084] Furthermore, the simplified problem obtained by simplifying the non-convex optimization problem based on the closed-form solution of the beamforming specifically includes:

[0085] Given a communication schedule A c Perception Scheduling A p Flight trajectory Q, transmission power P, mission completion time N F In this case, optimize the beamforming W to obtain all closed-form solutions of the beamforming and the lower bound of the throughput of communication user k in time slot n. R k [n];

[0086] Based on all the closed-form solutions and the lower bound of the throughput R k [n], constraints (15) and (24) can be expressed as:

[0087]

[0088]

[0089] The simplified problem (P1) after simplifying the constraints is:

[0090]

[0091] st(1)-(9),(13),(14),(16)-(18)

[0092] (20)-(23),(25),(36),(37).

[0093] Preferably, the optimized beamforming W yields a closed-form solution w for the beamforming. * and its corresponding lower bound R of throughput k [n] specifically includes:

[0094] When the drone communicates with communication user k only in time slot n To maximize the signal strength received by the user To optimize the objective, the first closed-form solution is obtained as follows:

[0095]

[0096] Substituting equation (28) into equation (11), we can obtain the throughput of the UAV when it is only communicating:

[0097]

[0098] in, The distance from the drone to the communication user;

[0099] When the drone senses target point v in time slot n To maximize beam strength To optimize the objective, the second closed-form solution is obtained as follows:

[0100]

[0101] And the throughput of the drone when it only senses the target point is:

[0102]

[0103] When the UAV senses target point v while communicating with user k in time slot n, Under constraints (10) and (15), the goal is to maximize the signal strength received by the user. To optimize the objective, consider the following questions:

[0104]

[0105] st(10),(15)

[0106] Solution (P2) yields the third closed-form solution w3 for the beamforming. * [n] and its lower bound on throughput for:

[0107]

[0108]

[0109] in,

[0110] After rearranging equations (29), (31), and (34), the lower bound of the throughput of the communication user k in time slot n can be obtained. Rk [n] is:

[0111]

[0112] Furthermore, the process of breaking down the simplified problem into two sub-problems specifically includes:

[0113] Based on the objective function of (P3), the simplified problem is decomposed into two sub-problems;

[0114] In subproblem one, by optimizing communication scheduling A c Flight trajectory Q, transmission power P, and sensing mission completion time N F To maximize total perception time

[0115] In subproblem two, by optimizing communication scheduling A c Flight trajectory Q and transmission power P are used to minimize the total sensing mission completion time.

[0116] Furthermore, the solution process for sub-problem one is as follows:

[0117] Initialize variables; wherein, the variables include communication scheduling, sensing scheduling, flight trajectory, transmission power, and sensing mission completion time;

[0118] The iteration begins, the initial solution of the CB algorithm is updated, and the CB algorithm is run once; wherein, updating the initial solution of the CB algorithm includes updating the communication schedule using a greedy algorithm and updating the flight trajectory and the transmission power using a continuous convex optimization technique;

[0119] In the next iteration, update the initial solution of the CB algorithm and run the CB algorithm again;

[0120] When the solutions for the total sensing time obtained in two consecutive iterations are equal, the iteration ends, and the solution for the total sensing time obtained in the last iteration is output as the maximum total sensing time.

[0121] Preferably, the CB algorithm is a linked list-based iterative algorithm, specifically comprising:

[0122] The iteration begins by selecting any one of the sensing targets as the optimization object, optimizing the sensing time with the goal of maximizing the sensing time of the optimization object, and then updating the sensing schedule, communication schedule, flight trajectory, transmission power, and sensing task completion time based on the obtained optimized sensing time using a linked list structure.

[0123] After one iteration is completed, any unoptimized perception target is selected as the optimization object for the next iteration;

[0124] Once all the perception targets have been optimized, the optimized perception times for all the perception targets are summed to obtain the total perception time.

[0125] Furthermore, the solution process for sub-problem two is as follows:

[0126] Initialize the loop variable and input the loop variable into the outer loop; wherein, the loop variable includes the first communication schedule, the first flight trajectory, the first transmission power, and the first sensing task completion time;

[0127] The outer loop begins, updating the first flight trajectory based on the first communication schedule and the first sensing task completion time to obtain the second flight trajectory. Subsequently, the first communication schedule, the first transmission power, the first sensing task completion time, and the second flight trajectory are input into the inner loop.

[0128] The inner loop begins, and the first communication scheduling and the first transmission power are optimized by the Lagrange duality method. At the same time, the second flight trajectory is optimized by the continuous convex optimization technique to obtain the second communication scheduling, the second transmission power and the third flight trajectory.

[0129] The completion time of the first sensing task is updated according to the second communication schedule, the second transmission power, and the third flight trajectory to obtain the completion time of the second sensing task. The second communication schedule, the second transmission power, the third flight trajectory, and the completion time of the second sensing task are then output from the inner loop.

[0130] If the completion times of the first and second sensing tasks are not consistent, the second communication scheduling, the second transmission power, the third flight trajectory, and the completion time of the second sensing task are re-inputted into the inner loop until the completion time of the sensing task input into the inner loop is equal to the completion time of the sensing task output into the inner loop. Then the inner loop is stopped, and the inner loop communication scheduling, inner loop transmission power, inner loop flight trajectory, and inner loop sensing task time obtained from the last inner loop are output.

[0131] If the inner loop sensing task time is compared with the sensing task completion time input to the outer loop, and the two are inconsistent, the inner loop communication scheduling, inner loop transmission power, inner loop flight trajectory, and inner loop sensing task time are re-inputted into the outer loop until the sensing task completion time input to the outer loop is equal to the sensing task completion time output to the outer loop, at which point the outer loop stops.

[0132] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0133] (1) In the ISAC UAV-assisted wireless communication system, two different needs are considered: those of sensing users and those of communication users. For communication users, the UAV needs to complete its throughput task throughout the flight phase. For sensing users, the UAV needs to sense the target points of interest and transmit the collected sensing information back to the corresponding sensing user. In this system, the present invention considers minimizing the information age of the sensing information, thereby improving the timeliness of the sensing information.

[0134] (2) This invention proposes a perception time reselection algorithm that packages variables according to time slots for movement, solving the problem of deep coupling between perception scheduling and UAV flight trajectory, thereby enabling rapid optimization of perception scheduling and updating of perception time. Furthermore, this invention also proposes an initial solution optimization problem for communication scheduling, UAV flight trajectory, and transmission power based on this reselection algorithm.

[0135] (3) This invention proposes to transform the problem of minimizing the completion time of the sensing task into maximizing the throughput of the sensing users, and further proposes optimization problems for communication scheduling, UAV flight trajectory, and transmission power. It solves the problem that the completion time is difficult to solve as the upper limit of the summation.

[0136] (4) In solving the problem of maximizing the perceived user throughput in (3), this invention proposes a method to solve mixed integer non-convex problems by applying the Lagrange multiplier method, which can quickly obtain suboptimal communication scheduling and power. Attached Figure Description

[0137] Figure 1 This is a flowchart illustrating an optimization method for an integrated communication and sensing system for unmanned aerial vehicles (UAVs) according to an embodiment of the present invention.

[0138] Figure 2 This is a schematic diagram of an ISAC system assisted by an unmanned aerial vehicle (UAV) according to an embodiment of the present invention.

[0139] Figure 3 This is a schematic diagram of a process for maximizing total sensing time according to an embodiment of the present invention.

[0140] Figure 4 This is a flowchart illustrating a method for minimizing the total completion time of a sensing task, as provided in an embodiment of the present invention.

[0141] Figure 5 This is a schematic diagram of the three-dimensional flight trajectory of a UAV when the throughput threshold of the communication user is 5000 Mbit, provided as an embodiment of the present invention.

[0142] Figure 6 This is a schematic diagram of a two-dimensional flight trajectory of a drone when the throughput threshold of the communication user is 0 Mbit, provided as an embodiment of the present invention.

[0143] Figure 7 This is a schematic diagram of a two-dimensional flight trajectory of a drone when the throughput threshold of the communication user is 3000 Mbit, provided as an embodiment of the present invention.

[0144] Figure 8 This is a schematic diagram of a two-dimensional flight trajectory of a drone when the throughput threshold of a communication user is 5000 Mbit, provided as an embodiment of the present invention.

[0145] Figure 9 This is a schematic diagram showing the change in average perceived information age as a function of the throughput threshold of a communication user, obtained by different schemes provided in an embodiment of the present invention. Detailed Implementation

[0146] The accompanying drawings are for illustrative purposes only and should not be construed as limiting the scope of this patent.

[0147] It will be understood by those skilled in the art that certain well-known structures and their descriptions may be omitted in the accompanying drawings.

[0148] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0149] Reference Figure 1 The following is a flowchart illustrating an optimization method for an integrated communication and sensing system for unmanned aerial vehicles (UAVs) according to an embodiment of the present invention, comprising the following steps:

[0150] S1: Construct a communication and sensing integrated system model assisted by unmanned aerial vehicles (UAVs); wherein, the system model includes several constraints;

[0151] S2: Under the aforementioned constraints, jointly optimize perception scheduling, communication scheduling, UAV flight trajectory, UAV transmission power, and beamforming to construct a non-convex optimization problem that minimizes the average information age of the perception information;

[0152] S3: Repeat the solution of the non-convex optimization problem to keep the average information age updated until it is determined that the average information age update has no change, then end the optimization.

[0153] For step S1, specifically, the construction of the UAV-assisted integrated communication and sensing system model includes:

[0154] Define user set K = V + U, where the first V users in the user set are sensing users, and the remaining U users are communication users. The set of sensing targets is defined as follows: Each element corresponds one-to-one with the first V users in K;

[0155] Define the number of antennas carried by the drone as N. t The flight cycle of the UAV is T, which is evenly divided into N time slots, where the length of each time slot is τ = T / N;

[0156] Let the time slot set be denoted as The drone's flight trajectory is then approximated as follows:

[0157]

[0158] in It is the three-dimensional position of the UAV in time slot n. h[n] is the horizontal position of the UAV, and h[n] is the flight altitude of the UAV in time slot n;

[0159] Define the maximum horizontal flight speed of the UAV during the entire mission cycle as The maximum vertical flight speed is And the drone needs to start from a given location The drone departs and returns to the same point at the end of its flight cycle, with its maximum and minimum flight altitudes recorded as h. max and h min The trajectory of the drone is subject to the following constraints:

[0160] q u [1] = q U ;

[0161]

[0162]

[0163]

[0164]

[0165]

[0166] Let p[n] be the transmit power of the UAV in time slot n. Then the transmit power meets the following peak and total value constraints:

[0167]

[0168]

[0169] Among them, P max The instantaneous maximum transmit power, This represents the total power that the drone can launch;

[0170] Model the probability of a direct link between the drone and communication user k:

[0171]

[0172] Where a and b are the parameters of the S-curve. Let $\frac{ ... The distance from the drone to the communication user k in time slot n;

[0173] Define a binary variable Used to represent the scheduling status of UAVs to communication users in time slot n, when When the drone communicates with user k, otherwise...

[0174] The following communication flight elevation angle constraints also exist for drones:

[0175]

[0176] in, It is the probability of acceptable line-of-sight. The corresponding angle of elevation;

[0177] If a drone is limited to communicating with a maximum of one user at a time, then:

[0178]

[0179]

[0180] According to the free-space loss model, the channel gain from the UAV to the communication user k is:

[0181]

[0182] Where β0 is the channel power gain when the reference distance is 1m;

[0183] The baseband equivalent channel from the drone to communication user k is represented as:

[0184]

[0185] Where, λ c The wavelength of the transmitted signal, Let be the guidance vector from the UAV to the communication user k. It is the elevation angle of the drone to the communication user k in time slot n;

[0186] Let the information sent by the drone to user k be s. k [n], s k It follows a standard normal distribution, and the information sent to each user is independent of each other, that is... The signal transmitted by the drone is Where w[n] is the precoding vector in time slot n, which meets the following constraints:

[0187]

[0188] Therefore, the throughput of user k in time slot n is:

[0189]

[0190] Where B is the signal bandwidth, σ 2 The power of augmented Gaussian white noise;

[0191] Let R be the throughput threshold for communication users. th Then the throughput constraint for communication users is:

[0192]

[0193] Define another binary variable This is used to represent the perception scheduling of a UAV to target point v in time slot n. The UAV only... When sensing the target point v, the sensing elevation angle constraint is expressed as:

[0194]

[0195] in, It is the distance from the drone to the target point v. It is the acceptable flight elevation angle threshold during perception.

[0196] If a drone is required to hover while performing sensing operations, then there are hovering constraints:

[0197]

[0198] The beam intensity reaching the sensing point is denoted as... in It is the guidance vector of the UAV to the target point v. If the elevation angle of the UAV to the target point v is the beam angle, then the beam strength must be greater than the given beam strength threshold Γ. th ,Right now:

[0199]

[0200] If the drone is limited to sensing only one target at a time, and each target must be sensed once during the entire flight cycle, then:

[0201]

[0202]

[0203]

[0204] Let R be the amount of information generated in a single perception. s The time when the drone completes transmitting the perceived information back to the sensing user is recorded as The throughput constraint for perceived users is then expressed as:

[0205]

[0206]

[0207] The scheduling variable meets the following causality constraints:

[0208]

[0209]

[0210]

[0211] In a preferred embodiment, refer to Figure 2 This is a schematic diagram of the structure of an ISAC system assisted by an unmanned aerial vehicle (UAV) according to an embodiment of the present invention, including an ISAC UAV, a communication user, a sensing user, and a sensing target.

[0212] N-type carrier with communication and sensing functions t An ISAC (Independent Sensing and Detection) drone with one antenna flies over a city. Within a given flight period T, it completes communication metrics for U communication users and assists V sensing users in their sensing activities. Assume that due to obstacles, there is no direct link between the sensing users and their corresponding targets. The ISAC drone needs to assist the sensing users in their sensing activities, including sensing targets and transmitting the collected sensing data back to the sensing users. Simultaneously, the drone also needs to handle communication with the communication users.

[0213] Define user set Where K = V + U, and the first V users are sensing users, and the remaining users are communication users. The set of target points is defined as follows: Each of these elements is related to The first V users in the list correspond one-to-one.

[0214] Considering a three-dimensional coordinate system, let the initial coordinates of user k and the coordinates of the target point v be denoted as follows: and Let the flight period of the UAV be T, and divide it evenly into N time slots, where the length of each time slot is τ = T / N. Let the set of time slots be denoted as... The flight trajectory of the drone can then be approximated as follows: in It is the three-dimensional position of the UAV in time slot n. h[n] is the horizontal position of the UAV in time slot n, and h[n] is the flight altitude of the UAV in time slot n.

[0215] The maximum horizontal flight speed of the UAV during the entire mission cycle is denoted as: The maximum vertical flight speed is The drone needs to start from a given location Set off, It is a given horizontal position. The maximum and minimum flight altitudes of the UAV are given, and the UAV returns to that point at the end of the flight cycle. Let h be the maximum and minimum flight altitudes of the UAV. max and h min The trajectory of the drone is subject to the following constraints:

[0216] q u [1] = q I ;

[0217]

[0218]

[0219]

[0220]

[0221]

[0222] Let p[n] be the transmit power of the UAV in time slot n. Then the transmit power should meet the following peak and total value constraints:

[0223]

[0224]

[0225] Where P max The instantaneous maximum transmit power, This represents the total power that the drone can launch.

[0226] The ISAC UAV in this embodiment operates in the millimeter-wave band. Since millimeter-wave signals suffer severe signal loss at non-line-of-sight (Line-of-Sight) distances, this preferred embodiment considers that the UAV cannot communicate or sense signals at non-Line-of-Sight distances.

[0227] Therefore, consider the following line-of-sight link probability model to model the line-of-sight link probability between the UAV and user k:

[0228]

[0229] Where a and b are S-curve parameters, given by the environment. Let n be the flight elevation angle of the UAV to user k in time slot n. Let be the distance from the drone to user k in time slot n.

[0230] Define a binary variable Used to represent the scheduling status of UAVs to users in time slot n, when When the drone communicates with user k, otherwise... To ensure that the UAV can communicate in a line-of-sight channel, the following communication flight elevation angle constraints apply:

[0231]

[0232] in It is the probability of acceptable line-of-sight. The corresponding angle of elevation.

[0233] Assuming the drone can communicate with at most one user at a time, then:

[0234]

[0235]

[0236] Based on the assumptions about line-of-sight links above, and using the free-space loss model, the channel gain from the UAV to user k in time slot n is: Where β0 is the channel power gain when the reference distance is 1m.

[0237] Assuming the antenna array carried by our drone is a linear array, the baseband equivalent channel from the drone to user k can be expressed as:

[0238]

[0239] Where, λ c The wavelength of the transmitted signal, Let [ ] be the guidance vector from the drone to user k. H This represents the conjugate transpose of a vector. This is the elevation angle of the drone to user k in time slot n. Assume the information sent by the drone to user k in time slot n is s. k [n], s k [n] follows a standard normal distribution, and the information sent to each user is independent. Then the signal transmitted by the drone is... Where w[n] is the precoding vector in time slot n, which meets the following constraints:

[0240]

[0241] Therefore, the throughput of user k in time slot n is:

[0242]

[0243] Where B is the signal bandwidth, σ 2 R is the power of the additive white Gaussian noise. The throughput threshold for communication users is denoted as R. th Then the throughput constraint for communication users is:

[0244]

[0245] While communicating with communication users, ISAC UAVs also need to provide services to sensing users, including: detecting target points and transmitting the collected sensing information back to the corresponding sensing users. Thanks to the ISAC structure, the UAV can sense target points by receiving the echo signal of the communication signal x[n] at the target point. Sensing activities are highly dependent on the line-of-sight link; therefore, target points can only be sensed when the flight elevation angle is greater than a given threshold.

[0246] variables Let this be denoted as the perception and scheduling of the UAV to target point v in time slot n. The UAV only... The target point v is sensed at that time. Therefore, the sensing elevation angle constraint can be expressed as:

[0247]

[0248] in It is the distance from the drone to the target point v. It is the acceptable flight elevation angle threshold during perception.

[0249] To avoid the drone failing to accurately receive echo signals, we assume that the drone needs to hover during perception operations, thus incurring hovering constraints:

[0250]

[0251] Let the beam intensity in the direction of the sensing point v be denoted as in It is the guidance vector of the UAV to the target point v. This is the elevation angle of the UAV to the target point v. To ensure the UAV receives a valid echo signal, we require the beam strength to be greater than a given beam strength threshold Γ. th ,Right now:

[0252]

[0253] Assuming the drone can only sense one target at a time, and each target must be sensed once during a complete flight cycle, then:

[0254]

[0255]

[0256]

[0257] Let R be the amount of information generated in a single perception. s The time point at which the drone completes transmitting the perceived information back to the sensing user v is denoted as v. The throughput constraint for perceived users can then be expressed as:

[0258]

[0259]

[0260] Since drones can only transmit perception information after sensing the target point, the scheduling variables should meet the following causal constraints:

[0261]

[0262]

[0263]

[0264] For step S2, specifically, the non-convex optimization problem of constructing the minimum average information age of the perceived information specifically includes:

[0265] By unifying constraints (12) and (19), we obtain the following constraints:

[0266]

[0267]

[0268] in, It is the throughput threshold for user k;

[0269] Let the time from the generation of perceived information to the completion of the feedback of perceived information be defined as minimizing the average information age of the perceived information. Then, the average information age of minimizing the perceived information is:

[0270]

[0271] in, The moment when the drone senses the target;

[0272] If minimizing the average information age of perceived information is taken as the optimization objective, then the following non-convex optimization problem exists:

[0273]

[0274] st(1)-(10),(13)-(18),(20)-(25).

[0275] In fact, problem (P1) is a mixed integer nonconvex optimization problem. Meanwhile, due to constraints (7), (13)-(15), schedule A... p A c The deep coupling with trajectory Q, power P, and beamforming W makes it impossible to solve the problem directly (P1). Meanwhile, in constraint (25), the sensing task completion time N... F Affecting communication scheduling c The range of values ​​makes it difficult to obtain a value for N. F The closed-form solution.

[0276] Therefore, for step S3, specifically, solving the non-convex optimization problem includes:

[0277] The non-convex optimization problem is simplified by applying constraints to the closed-form solution of the beamforming to obtain a simplified problem.

[0278] The simplified problem is broken down into two sub-problems and solved separately.

[0279] Furthermore, the simplified problem obtained by simplifying the non-convex optimization problem based on the closed-form solution of the beamforming specifically includes:

[0280] Given a communication schedule A c Perception Scheduling A p Flight trajectory Q, transmission power P, mission completion time N F In this case, optimize the beamforming W to obtain all closed-form solutions of the beamforming and the lower bound R of the throughput of communication user k in time slot n. k [n];

[0281] Based on all the closed-form solutions and the lower bound R of the throughput k [n], constraints (15) and (24) can be expressed as:

[0282]

[0283]

[0284] The simplified problem (P1) after simplifying the constraints is:

[0285]

[0286] st(1)-(9),(13),(14),(16)-(18)

[0287] (20)-(23),(25),(36),(37).

[0288] Preferably, the optimized beamforming W yields all closed-form solutions to the beamforming and their corresponding lower bounds for throughput R. k [n] specifically includes:

[0289] When the drone communicates with communication user k only in time slot n To maximize the signal strength received by the user To optimize the objective, the first closed-form solution is obtained as follows:

[0290]

[0291] Substituting equation (28) into equation (11), we can obtain the throughput of the UAV when it is only communicating:

[0292]

[0293] in, d ck The distance from the drone to the communication user;

[0294] When the drone senses target point v in time slot n To maximize beam strength To optimize the objective, the second closed-form solution is obtained as follows:

[0295]

[0296] And the throughput of the drone when it only senses the target point is:

[0297]

[0298] When the UAV senses target point v while communicating with communication user k in time slot n, Under constraints (10) and (15), the goal is to maximize the signal strength received by the user. To optimize the objective, consider the following questions:

[0299]

[0300] st(10),(15)

[0301] Solution (P2) yields the third closed-form solution w3 for the beamforming. * [n] and its lower bound on throughput for:

[0302]

[0303]

[0304] in,

[0305] After rearranging equations (29), (31), and (34), the lower bound of the throughput of the communication user k in time slot n can be obtained. R k [n] is:

[0306]

[0307] In a preferred embodiment, when given a communication schedule A c Perception Scheduling A p Flight trajectory Q, transmission power P, mission completion time N F When, the objective function It remains unchanged. Therefore, when optimizing beamforming W, our aim is to provide a larger solution space for optimizing other variables. Let the closed-form solution of beamforming be... The beamforming for each time slot will be discussed below based on different scheduling scenarios:

[0308] 1. The UAV communicates with user k only once in time slot n, that is... Based on constraint (24) and expression (11), we can maximize the signal strength received by user k. To obtain a larger (P1) solution space. In order to maximize... We can easily obtain:

[0309]

[0310] remember Substituting (28) into (11) yields the throughput for communication only, denoted as

[0311]

[0312] 2. The UAV only senses target point v in time slot n, that is... According to constraint (15), we can find that when the beam intensity The larger the value, the larger the (P1) solution space can be obtained. To maximize... We can easily obtain:

[0313]

[0314] 3. The UAV senses target point v while communicating with user k in time slot n, i.e. Based on the above discussion of the case of communication only, we can maximize under constraints (10) and (15). Therefore, we consider the following questions:

[0315]

[0316] st(10),(15)

[0317] The problem (P2) is studied in "Throughput Maximization for UAV-enabled Integrated Periodic Sensing and Communication", and its closed-form solution w is obtained. * Lower bound of throughput for user k R k

[0318]

[0319]

[0320] in,

[0321] By rearranging (29) and (34), we can obtain a unified expression for the lower bound of throughput, denoted as R k [n]:

[0322]

[0323] in, R k [n] is the lower bound of the throughput of user k in time slot n.

[0324] Based on the lower bounds of throughput corresponding to the three closed-form solutions of beamforming, constraints (15) and (24) can be expressed as:

[0325]

[0326]

[0327] At this point, the original problem (P1) can be equivalent to:

[0328]

[0329] st(1)-(9),(13),(14),(16)-(18)

[0330] (20)-(23),(25),(36),(37).

[0331] In fact, the variables in problem (P3) are coupled together, and (P3) contains many non-convex constraints, making it difficult to solve.

[0332] Therefore, further, the simplified problem is broken down into two sub-problems, specifically including:

[0333] Based on the objective function of (P3), the simplified problem is decomposed into two sub-problems;

[0334] In subproblem one, by optimizing communication scheduling A c Flight trajectory Q, transmission power P, and sensing mission completion time N F To maximize total perception time

[0335] In subproblem two, by optimizing communication scheduling A c Flight trajectory Q and transmission power P are used to minimize the total sensing mission completion time.

[0336] Furthermore, refer to Figure 3 The flowchart illustrating the process of maximizing total sensing time according to an embodiment of the present invention includes:

[0337] S4: Initialize variables; wherein, the variables include communication scheduling, perception scheduling, flight trajectory, transmission power, and perception mission completion time;

[0338] S5: Start iteration, update the initial solution of the CB algorithm, and run the CB algorithm once; wherein, updating the initial solution of the CB algorithm includes updating the communication schedule through a greedy algorithm and updating the flight trajectory and the transmission power through continuous convex optimization techniques;

[0339] S6: Perform the next iteration, update the initial solution of the CB algorithm, and run the CB algorithm again;

[0340] S7: When the solutions for the total sensing time obtained in two consecutive iterations are equal, the iteration ends and the solution for the total sensing time obtained in the last iteration is output as the maximum total sensing time.

[0341] Preferably, the CB algorithm is a linked list-based iterative algorithm, specifically comprising:

[0342] The iteration begins by selecting any one of the sensing targets as the optimization object, optimizing the sensing time with the goal of maximizing the sensing time of the optimization object, and then updating the sensing schedule, communication schedule, flight trajectory, transmission power, and sensing task completion time based on the obtained optimized sensing time using a linked list structure.

[0343] After one iteration is completed, any unoptimized perception target is selected as the optimization object for the next iteration;

[0344] Once all the perception targets have been optimized, the optimized perception times for all the perception targets are summed to obtain the total perception time.

[0345] In a preferred embodiment, according to (P3), the problem of maximizing the total sensing time can be formulated as follows:

[0346]

[0347] st(1)-(9),(13),(14),(16)-(18)

[0348] (20)-(23),(25),(36),(37)

[0349] To solve problem (P4), a chain-based algorithm, referred to as the CB algorithm, is proposed. Executing the CB algorithm once yields a solution for the total sensing time. During the solution process, the CB algorithm needs to be executed repeatedly, with the initial solution reset after each execution and before the next execution, until the total sensing time solutions obtained after two consecutive executions of the CB algorithm are equal.

[0350] The CB algorithm is actually an iterative algorithm. After each iteration of the CB algorithm, the perception time corresponding to a perception target can be obtained. After the iteration ends, the perception time corresponding to all perception targets can be obtained. The total perception time can be obtained by summing these perception times.

[0351] The following section will explain how the CB algorithm constructs the solution for the current iteration based on the solution obtained in the previous iteration. For ease of explanation, the r-th iteration will be referred to as... c The j-th iteration of the algorithm during its execution is denoted as the (r)-th iteration. c (j) iterations.

[0352] In the (r) c In iteration j), the goal of the CB algorithm is to maximize This can be divided into two steps: ① Select the current optimal... Recorded as ②According to For the (r)thc The variables obtained from the (j-1)th iteration are adjusted to obtain the (r)th iteration. c The variables in the j)th iteration.

[0353] 1. Given In the case of obtaining the (r)th c ,j) iterations of variables

[0354]

[0355] and For the (r)th c The communication scheduling, sensing scheduling, flight trajectory, transmission power, and sensing task completion time in the (r)th iteration are also considered. c In the (j)th iteration, the initial and new perception times for target j are denoted as ,j) and ,j) respectively. and

[0356] First, for the transition from time slot 1 to time slot 2... From time slot For variables in time slot N, we maintain the relationship with the (r)th time slot. c The results of the (j-1) iterations are consistent.

[0357]

[0358]

[0359]

[0360]

[0361] For those located and The variables between (r) c The variable from the j-1)th iteration is shifted forward by one time slot to obtain

[0362]

[0363]

[0364]

[0365]

[0366] Finally, due to hovering constraints, the perceived position is always equal to the position of the next time slot, therefore...

[0367]

[0368] (41) This will cause a change in the drone's position when it senses the environment, thus affecting whether the drone can communicate with the user. Communication scheduling should integrate the (r) c Communication scheduling in the (j-1)th iteration and the (r)th iteration c The position in the j)th iteration is obtained

[0369]

[0370] in This indicates whether the UAV meets the elevation angle requirement for communication in time slot n. and They represent the (r)th c The drone's flight altitude and distance to user k are determined in the (j)th iteration. Power and perception scheduling can be maintained in relation to the (r)th iteration. c The results of the (j-1)th iteration are consistent.

[0371]

[0372]

[0373] Finally, we can obtain the task completion time based on the communication scheduling:

[0374]

[0375] 2. According to the (r)th c The optimal result is selected from the results of (j-1) iterations.

[0376] The variable shifts in (42)-(45) may violate constraints (13), (23), (36), and (37). We denote the communication object when the UAV perceives target v as... In order to avoid violating the constraints, The following conditions must be met:

[0377]

[0378]

[0379]

[0380]

[0381] in The time when the drone begins to transmit perceived information back to user j. For the (r)th c The throughput of user k in the nth time slot during the j)th iteration. The left side of (50) represents the throughput of user k in the (r)th time slot.c In the j-1)th iteration, the user The surplus throughput, (50) on the right represents the throughput from the (r)th... c The iterations from (j-1) to the (r)th iteration c The throughput reduction caused by the (j)th iteration, (50) means that after completing the (r)th iteration, c After j) iterations, the user The throughput is still greater than the given threshold. Thanks to condition (47), we can search for the optimal value within a small range based on conditions (48)-(50). That is, in and Search between them.

[0382] Since the CB algorithm runs iteratively, the given initial solution has a significant impact on its performance. Therefore, the updating of the initial solution for the CB algorithm will be introduced below.

[0383] At the rth c Before executing the CB algorithm for the rth time, we based on the rth... c The result obtained from the -1th execution of the CB algorithm is adjusted to obtain the rth... c The initial solution of the CB algorithm is executed once. We note that the CB algorithm selects... The conditions are (48), (49), and (50), and these three conditions are related to {A}. c It relates to P and Q. Therefore, let And for {A c Optimize P,Q to obtain

[0384] First, we introduce slack variables. Represents respectively in The left-hand side of (48), (49), and (50) is...

[0385]

[0386]

[0387]

[0388] It is not difficult to see that it can be maximized To maximize the perception time of each target point by 1, consider the following problem to optimize {A}. c ,P,Q}

[0389]

[0390]

[0391] (1)-(9),(13),(14),(22),(23),(25),

[0392] (36),(37),(51),(52),(53)

[0393] in,

[0394] when as well as When large enough, It can be set to Therefore, in order to avoid wasting resources, (54a) is introduced to constrain the upper bound of the slack variables.

[0395] However, (P5) is a non-convex problem, and the variables are coupled with each other. To overcome this difficulty, it is necessary to apply an alternating optimization method to divide the variables into {A}. c Given {P,Q}, solve the following two problems: 1. Optimize {A} with {P,Q} fixed. c};2. In the case of fixed {A c Optimize {P,Q} in the case of}.

[0396] The following section will introduce how to optimize {A} given a fixed {P,Q}. c}

[0397] With {P,Q} fixed, problem (P5) can be simplified to:

[0398]

[0399]

[0400]

[0401] in, It is the set of users that can communicate in time slot n, i.e. (55c) is a unified expression of constraints (7), (8), (22), (23) and (25).

[0402] Since the communication schedule is repeatedly optimized, the communication schedule obtained in the previous step can be adjusted to quickly obtain the solution to problem (P6). That is, check each time slot; if in time slot n, Based on (37), determine whether the current communication user can be changed to Finally, S is calculated based on (53) and (55b). R .

[0403] The following describes how to fix {A} c Optimize {P,Q} in the case of}.

[0404] When {A} is fixed c When}, problem (P5) can be simplified to:

[0405]

[0406] st(1)-(7),(13),(14),(36),

[0407] (37),(51),(52),(53),(54b)

[0408] Among them, constraints (37), (53) and (54b) are non-convex with respect to Q.

[0409] This problem can be solved using continuous convex optimization methods:

[0410] First, introduce slack variables.

[0411]

[0412] How to R k The lower bound of [n] is denoted as

[0413]

[0414] in, and They are respectively and The lower bound of q is determined by q. u(r) The first-order Taylor expansion of [n].

[0415]

[0416]

[0417] in,

[0418] Finally, with Alternative R k [n], then problem (P7) can be transformed into:

[0419]

[0420]

[0421]

[0422]

[0423] (1)-(7),(13),(14),(36),(51),(52),(57)

[0424] Problem (P8) is a convex problem, which can be solved using a convex optimization solver such as CVX.

[0425] Furthermore, refer to Figure 4 The above is a flowchart illustrating a method for minimizing the total completion time of a sensing task according to an embodiment of the present invention, comprising:

[0426] S8: Initialize the loop variable and input the loop variable into the outer loop; wherein, the loop variable includes the first communication schedule, the first flight trajectory, the first transmission power, and the first sensing task completion time;

[0427] S9: Start the outer loop, update the first flight trajectory according to the first communication schedule and the first sensing task completion time to obtain the second flight trajectory, and then input the first communication schedule, the first transmission power, the first sensing task completion time and the second flight trajectory into the inner loop;

[0428] S10: Start the inner loop, optimize the first communication scheduling and the first transmission power through the Lagrange duality method, and optimize the second flight trajectory through the continuous convex optimization technique to obtain the second communication scheduling, the second transmission power and the third flight trajectory;

[0429] S11: Update the completion time of the first sensing task according to the second communication schedule, the second transmission power and the third flight trajectory to obtain the completion time of the second sensing task, and output the second communication schedule, the second transmission power, the third flight trajectory and the completion time of the second sensing task from the inner loop;

[0430] S12: Compare the completion time of the first sensing task with the completion time of the second sensing task. If they are inconsistent, re-input the second communication scheduling, the second transmission power, the third flight trajectory, and the completion time of the second sensing task into the inner loop until the completion time of the sensing task input into the inner loop is equal to the completion time of the sensing task output into the inner loop. Then stop the inner loop and output the inner loop communication scheduling, inner loop transmission power, inner loop flight trajectory, and inner loop sensing task time obtained in the last inner loop.

[0431] S13: Compare the inner loop sensing task time with the sensing task completion time input to the outer loop. If the two are inconsistent, re-input the inner loop communication scheduling, inner loop transmission power, inner loop flight trajectory, and inner loop sensing task time into the outer loop until the sensing task completion time input into the outer loop is equal to the sensing task completion time output from the outer loop, and then stop the outer loop.

[0432] In a preferred embodiment, the fixed sensing time slot A p Task completion time slot N F Let the throughput surplus of the UAV to the sensing user k during the entire flight cycle be -η. k bit, i.e. To ensure that the drone can meet the communication threshold with the perceived user, the following applies:

[0433] η k ≤0,V <kK (59)

[0434] The problem of maximizing the total throughput of perceived users can then be expressed as:

[0435]

[0436]

[0437] When (P9) obtains the optimal solution value hour, Problem (P9) is given N F The solution is obtained under the condition that N is the optimal solution value. F The function. When At that time, the drone completed the transmission of the perception information from user k. Therefore, when fixed A... p Then, problem (P3) can be transformed into the following problem:

[0438]

[0439]

[0440] Given and Let's assume According to constraint (25), The corresponding feasible region ratio The corresponding feasible region is larger, therefore the solutions to problem (P9) are respectively in and The optimal solution value obtained below and The following relationship must be satisfied: That is, the optimal solution to problem (P9) yes It is a monotonically non-increasing function.

[0441] N can be fixed F Solve the problem (P9) and then perform a one-dimensional search for the optimal N. F ,Right now:

[0442]

[0443] for The current optimal solution. Therefore, the difficulty lies in designing a suitable algorithm to solve problem (P9). Problem (P9) is a mixed integer nonconvex optimization problem with variables that are coupled together. Therefore, based on the block coordinate descent method, (P9) will be divided into two problems: 1. Optimizing {A} with {Q} fixed. c ,P};2. In the case of a fixed {A} c In the case of P, we optimize {Q}. We solve these two problems alternately and update the perception time N according to (62). F Until the perception time N F No further updates possible.

[0444] The initial trajectory design process for problem (P9) is described below.

[0445] It can be observed that due to constraint (7), communication scheduling and UAV flight trajectory are deeply coupled. Therefore, based on the previously obtained trajectory, an initial solution that can provide a larger feasible solution space can be designed for problem (P9).

[0446] First, we introduce the slack variable C. k Representing drones The difference between the flight pitch angle and the pitch angle threshold for user k, i.e.

[0447]

[0448] Consider optimizing the trajectory and transmit power to maximize To enable drones to operate in time slots as much as possible The flight elevation angle requirement for user k is satisfied. Therefore, given A... c In this case, consider the following questions:

[0449]

[0450] stC k ≤0,k≤V (64b)

[0451]

[0452] (1)-(7),(13),(14),(36),(51)(63)

[0453] Where, C = {C k ,k≤V}. When C k When the time is 0, the drone can be in the time slot. To communicate with user k, constraint (64b) is introduced to avoid wasting drone trajectory resources.

[0454] The optimization process of communication scheduling and transmission power in problem (P9) is described below.

[0455] Given Q, optimize {A} c When P}, problem (P9) can be rewritten as:

[0456]

[0457]

[0458]

[0459]

[0460]

[0461] (6),(55c),(59),(60b)

[0462] Among them, constraints (65b) and (65c) are transformations of constraints (15) and (36), and constraint (65d) is equivalent to (9). Then, the partial Lagrangian function of (P12) is:

[0463]

[0464] Among them, ε, Let be the Lagrange multipliers corresponding to constraints (6), (60b), and (65b), respectively. Then the Lagrange dual function of (P12) can be expressed as:

[0465]

[0466] For g(ε,λ,μ) to be bounded, we should have: λ k ≥1, k≤V. Then the dual problem of (P12) is:

[0467]

[0468] stλ k ≥1, k≤V(68b)

[0469]

[0470]

[0471] ε≥0(68e)

[0472] For a given {ε,λ,μ}, η can be easily obtained. k optimal solution However, problem (67) is nonconvex. To solve (67), we first consider the given A. c Optimize P under the given conditions.

[0473] The power of each time slot will be discussed next:

[0474] 1. If the drone only communicates with user i in the nth time slot, we can obtain the optimal power at this time using the water-filling method.

[0475]

[0476] in, And [x] + =max(x,0).

[0477] 2. If the UAV communicates with user i and senses target j in the nth time slot, we can similarly obtain:

[0478]

[0479] in,

[0480] 3. If the UAV only senses target j in the nth time slot, we can obtain:

[0481]

[0482] By rearranging (69), (70), and (71), we obtain the power expression for any time slot n:

[0483]

[0484] in, Substituting the optimal power (72) into (35) yields the optimal throughput of user k in time slot n, denoted as...

[0485]

[0486] Then (67) can be restated as:

[0487]

[0488] in, It is related to The linear function, for a given time slot n, obviously has in,

[0489]

[0490] Therefore, for any given {ε,λ,μ}, the corresponding P can be obtained according to (72) and (75). * and A c* Therefore, the next step is to obtain the optimal {ε,λ,μ}. Since (D12) is a convex non-differentiable problem, the quasi-optimal Lagrange multipliers can be obtained using the subgradient descent method.

[0491] Let ε be the Lagrange multiplier generated in the (r)th iteration. (r) , Then in the (r)th iteration, {ε (r) ,λ (r) ,μ (r) The subgradient corresponding to} is:

[0492]

[0493]

[0494]

[0495] The optimization process of the trajectory in problem (P9) is described below.

[0496] When given {A c When optimizing Q, problem (P9) can be rewritten as:

[0497]

[0498] st(1)-(7),(13),(14),(36),(60b),(59)

[0499] Here, (60b) is a non-convex constraint with respect to Q, which can be approximated using the continuous convex optimization method mentioned above.

[0500] At this point (P13) transforms into:

[0501]

[0502]

[0503] Problem (P14) is a convex problem, which can be solved using CVX.

[0504] Finally, through Figure 5 , Figure 6 , Figure 7 , Figure 8 as well as Figure 9 This invention aims to demonstrate the superiority of an optimization method for an integrated communication and sensing system for unmanned aerial vehicles (UAVs) provided by an embodiment of the present invention.

[0505] Figure 5 R was shown th The 3D flight trajectory of the drone at 5000 Mbit / s shows that, in order to minimize information age within a certain information rate threshold, the drone's flight altitude is not always maintained at high / low altitude, but is adjusted according to the sensing and communication tasks. When the drone is sensing, it tends to fly higher, increasing the flight angle to the sensing point to gain sensing qualification. When the drone is transmitting information back to the sensing user, it tends to fly lower, increasing the instantaneous rate to complete the information transmission quickly. When the drone is performing a communication task, it increases its flight altitude to maintain communication when it is far away from the communication user, and decreases its flight altitude to increase the instantaneous rate when it is close to the communication user.

[0506] Figure 6 , Figure 7 , Figure 8 R was shown respectively th Two-dimensional flight trajectory of a UAV at speeds of 0, 3000, and 5000 Mbit. Where R... th =0Mbit corresponds to the extreme case where communication is completely disregarded, and its solution corresponds to the lower bound of the system's performance. Without considering the communication threshold, it can be observed that after sensing, the drone flies at maximum speed along a straight line towards the sensing user to transmit information back. This characteristic is particularly significant in the 16s→27s and 50s→60s phases. During the sensing-transmission process for user 2, because the drone needs to transmit information back to user 1, the drone's trajectory has a turning point during the 25s→36s phase. However, after completing the information transmission to user 1, the drone flies at maximum speed along a straight line towards user 2.

[0507] Comparing the drone trajectories under different signal rates, it can be observed that as the signal rate requirement increases, the drone tends to approach the communication user more closely to complete the communication task, and its sensing location and task completion location gradually move away from R. th The optimal position at 0 Mbit.

[0508] Figure 9 The average perceived age obtained from different schemes varies with R. thThe change curve shows that as the communication threshold increases, the information age of each scheme shows a monotonically increasing trend. As analyzed earlier, the increase in the communication threshold will cause the UAV to deviate more and more from the straight flight scheme, resulting in an increase in information age. Furthermore, the algorithm proposed in this embodiment is always superior to the Fix-Tra. and Uni.-Pow. schemes, which illustrates the necessity of jointly optimizing the trajectory and power.

[0509] Among them, the information age of Fix-Tra changes very gradually with the communication threshold. This is because there is a deep coupling between trajectory and perception / communication scheduling. When the trajectory remains unchanged, the UAV cannot seek later perception scheduling and earlier communication scheduling, thus leading to the solidification of the information age. When the communication threshold decreases, simple power distribution can also meet the communication threshold requirements without limiting the optimization of communication scheduling and flight trajectory. At the same time, the UAV can allocate more power for perception. Therefore, as the communication threshold decreases, the Uni.-Pow. scheme becomes increasingly closer to the scheme proposed in the embodiments of this invention.

[0510] Meanwhile, it can be observed that the perceived information age obtained by the algorithm proposed in the embodiments of the present invention is consistent with that obtained by the Non-Lagr. scheme, indicating that the application of the Lagrange multiplier method to solve the problem (P12) can not only obtain a solution quickly, but also has little impact on the overall information age.

[0511] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the implementation of the present invention. Those skilled in the art can make other variations or modifications based on the above description. It is neither necessary nor possible to exhaustively describe all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. An optimization method for an integrated communication and sensing system for unmanned aerial vehicles (UAVs), characterized in that, include: Constructing an integrated communication and sensing system model assisted by unmanned aerial vehicles (UAVs); wherein the system model includes several constraints, and the construction of the integrated communication and sensing system model assisted by UAVs specifically includes: Define user set Among them, the user set before One user is the perceived user, the rest... Each user is a communication user, and the set of perceived targets is defined as follows: Each of its elements is related to The front of the middle Each user corresponds to one other user; Define the number of antennas carried by the drone as The drone's flight cycle is And divide it evenly into There are 10 time slots, each with a length of 10 ... ; Let the time slot set be denoted as The flight trajectory of the drone is then approximated as: in It is a drone in a time slot The three-dimensional position, It is the horizontal position of the drone. It is a drone in a time slot Flight altitude; Define the maximum horizontal flight speed of the UAV during the entire mission cycle as The maximum vertical flight speed is And the drone needs to be from a given location The drone departs and returns to that point at the end of its flight cycle, while recording its maximum and minimum flight altitudes as follows: and The trajectory of the drone is subject to the following constraints: drones in time slots The transmit power is denoted as Then the transmit power meets the following peak and total value constraints: in, The instantaneous maximum transmit power, This represents the total power that the drone can launch; From drones to communication users Model the probability of direct links between them: in, and For S-curve parameters, From drones to communication users In the time slot At the flight angle, From drones to communication users In the time slot Distance at time; Define a binary variable , used to represent time slots The scheduling status of drones for communication users, when At that time, drones and communication users To communicate, otherwise ; The following communication flight elevation angle constraints also exist for drones: in, It is the probability of acceptable line-of-sight. The corresponding angle of elevation; If a drone is limited to communicating with a maximum of one user at a time, then: According to the free space loss model, the communication from the drone to the user... The channel gain is: in, This represents the channel power gain at a reference distance of 1m. From drones to communication users The baseband equivalent channel is represented as: in, The wavelength of the transmitted signal, From drones to communication users The guiding vector, From drones to communication users In the time slot The angle of elevation at that time; Set the drone to send to the user The information is , It follows a standard normal distribution, and the information sent to each user is independent of each other, that is... The signal transmitted by the drone is ,in In time slot The time-precoded vectors must meet the following constraints: Therefore, users In the time slot The throughput is: in For signal bandwidth, The power of augmented Gaussian white noise; The throughput threshold for communication users is denoted as... Then the throughput constraint for communication users is: Define another binary variable Used to indicate the drone in a time slot For target point Perception and scheduling; drones can only be used for... Time to target point If sensing is performed, the sensing elevation angle constraint is expressed as: in, It is the drone to the target point distance, It is the acceptable flight elevation angle threshold during perception; If a drone is required to hover while performing sensing operations, then there are hovering constraints: The beam intensity reaching the sensing point is denoted as... ,in It is the drone to the target point The guiding vector, It is the drone to the target point At the elevation angle, the beam strength must be greater than the given beam strength threshold. ,Right now: If the drone is limited to sensing only one target at a time, and each target must be sensed once during the entire flight cycle, then: The amount of information generated in a single perception is denoted as . The time when the drone completes transmitting the perceived information back to the sensing user is recorded as The throughput constraint for perceived users is then expressed as: The scheduling variable meets the following causality constraints: ； Under the aforementioned constraints, a non-convex optimization problem is constructed that minimizes the average information age of the sensed information by jointly optimizing the sensing scheduling, communication scheduling, UAV flight trajectory, UAV transmission power, and beamforming. The non-convex optimization problem is solved repeatedly to keep the average information age updated until it is determined that the average information age has no change, at which point the optimization ends.

2. The optimization method for an integrated communication and sensing system for unmanned aerial vehicles (UAVs) according to claim 1, characterized in that, The non-convex optimization problem of constructing the minimum average information age of perceived information specifically includes: By unifying constraints (12) and (19), we obtain the following constraints: in, , is the user The throughput threshold; Let the time from the generation of perceived information to the completion of the feedback of perceived information be defined as minimizing the average information age of the perceived information. Then, the average information age of minimizing the perceived information is: in, The moment when the drone senses the target; If minimizing the average information age of perceived information is taken as the optimization objective, then the following non-convex optimization problem exists: 。 3. The optimization method for an integrated communication and sensing system for unmanned aerial vehicles (UAVs) according to claim 2, characterized in that, Solving the nonconvex optimization problem specifically includes: The non-convex optimization problem is simplified by applying constraints to the closed-form solution of the beamforming to obtain a simplified problem. The simplified problem is broken down into two sub-problems and solved separately.

4. The optimization method for an integrated communication and sensing system for unmanned aerial vehicles (UAVs) according to claim 3, characterized in that, The simplified problem obtained by simplifying the non-convex optimization problem based on the closed-form solution of the beamforming includes: In a given communication schedule Perception and scheduling Flight trajectory Transmission power Task completion time In the case of optimizing beamforming To obtain all closed-form solutions of the beamforming and communication users In the time slot lower bound of throughput ; Based on all the closed-ended solutions and the lower bound of the throughput Constraints (15) and (24) can be expressed as: The simplified problem (P1) after simplifying the constraints is: 。 5. The optimization method for an integrated communication and sensing system for unmanned aerial vehicles (UAVs) according to claim 4, characterized in that, The optimized beamforming This yields all closed-form solutions for the beamforming and their corresponding lower bounds for throughput. Specifically, it includes: When the drone is in the time slot n Only with communication users During communication, To maximize the signal strength received by the user To optimize the objective, the first closed-form solution is obtained as follows: Substituting equation (28) into equation (11), we obtain the throughput of the UAV when it is only communicating: in, , The distance from the drone to the communication user; When the drone is in the time slot n Only perceive the target point hour To maximize beam strength To optimize the objective, the second closed-form solution is obtained as follows: And the throughput of the drone when it only senses the target point is: When the drone is in the drone time slot With communication users Simultaneous communication and target point sensing hour, Under constraints (10) and (15), the goal is to maximize the signal strength received by the user. To optimize the objective, consider the following questions: untie The third closed-form solution for the beamforming is obtained. and its throughput lower bound for: in, , , , , , , ; After rearranging equations (29), (31), and (34), the communication user can be obtained. In the time slot lower bound of throughput for: 。 6. The optimization method for an integrated communication and sensing system for unmanned aerial vehicles (UAVs) according to claim 4, characterized in that, The process of breaking down the simplified problem into two sub-problems specifically includes: Based on the objective function in (P3), the simplified problem is decomposed into two sub-problems; In subproblem one, communication scheduling is optimized. Flight trajectory Transmission power and the time to perceive task completion To maximize total perception time ; In subproblem two, communication scheduling is optimized. Flight trajectory Transmission power To minimize the total perception task completion time .

7. The optimization method for an integrated communication and sensing system for unmanned aerial vehicles (UAVs) according to claim 6, characterized in that, The solution process for sub-problem one is as follows: Initialize variables; wherein, the variables include communication scheduling, sensing scheduling, flight trajectory, transmission power, and sensing mission completion time; The iteration begins, the initial solution of the CB algorithm is updated, and the CB algorithm is run once; wherein, updating the initial solution of the CB algorithm includes updating the communication schedule using a greedy algorithm and updating the flight trajectory and the transmission power using a continuous convex optimization technique; In the next iteration, update the initial solution of the CB algorithm and run the CB algorithm again; When the solutions for the total sensing time obtained in two consecutive iterations are equal, the iteration ends, and the solution for the total sensing time obtained in the last iteration is output as the maximum total sensing time.

8. The optimization method for an integrated communication and sensing system for unmanned aerial vehicles according to claim 7, characterized in that, The CB algorithm is an iterative algorithm based on linked lists, specifically including: The iteration begins by selecting any one of the sensing targets as the optimization object, optimizing the sensing time with the goal of maximizing the sensing time of the optimization object, and then updating the sensing schedule, communication schedule, flight trajectory, transmission power, and sensing task completion time based on the obtained optimized sensing time using a linked list structure. After one iteration is completed, any unoptimized perception target is selected as the optimization object for the next iteration; Once all the perception targets have been optimized, the optimized perception times for all the perception targets are summed to obtain the total perception time.

9. The optimization method for an integrated communication and sensing system for unmanned aerial vehicles (UAVs) according to claim 6, characterized in that, The solution process for sub-problem two is as follows: Initialize the loop variable and input the loop variable into the outer loop; wherein, the loop variable includes the first communication schedule, the first flight trajectory, the first transmission power, and the first sensing task completion time; The outer loop begins, updating the first flight trajectory based on the first communication schedule and the first sensing task completion time to obtain the second flight trajectory. Subsequently, the first communication schedule, the first transmission power, the first sensing task completion time, and the second flight trajectory are input into the inner loop. The inner loop begins, and the first communication scheduling and the first transmission power are optimized by the Lagrange duality method. At the same time, the second flight trajectory is optimized by the continuous convex optimization technique to obtain the second communication scheduling, the second transmission power and the third flight trajectory. The completion time of the first sensing task is updated according to the second communication schedule, the second transmission power, and the third flight trajectory to obtain the completion time of the second sensing task. The second communication schedule, the second transmission power, the third flight trajectory, and the completion time of the second sensing task are then output from the inner loop. If the completion times of the first and second sensing tasks are not consistent, the second communication scheduling, the second transmission power, the third flight trajectory, and the completion time of the second sensing task are re-inputted into the inner loop until the completion time of the sensing task input into the inner loop is equal to the completion time of the sensing task output into the inner loop. Then the inner loop is stopped, and the inner loop communication scheduling, inner loop transmission power, inner loop flight trajectory, and inner loop sensing task time obtained from the last inner loop are output. If the inner loop sensing task time is compared with the sensing task completion time input to the outer loop, and the two are inconsistent, the inner loop communication scheduling, inner loop transmission power, inner loop flight trajectory, and inner loop sensing task time are re-inputted into the outer loop until the sensing task completion time input to the outer loop is equal to the sensing task completion time output to the outer loop, at which point the outer loop stops.

Images