A method for optimizing complex state updates in UAV-assisted IoT under time-varying channels

Abstract

This invention discloses an optimization method for complex state updates in a UAV-assisted Internet of Things (IoT) under time-varying channels. The method involves establishing a state space for the UAV-assisted complex state information update process in a block fading channel, deriving a threshold-based state probability equation, constructing a linear programming problem, obtaining the optimal threshold by solving the linear programming problem under different thresholds, and finally instructing the IoT device and the UAV to choose the behavior determined by the threshold. The proposed optimization method transforms the complex dynamic policy optimization problem into an easily solvable linear programming problem, thereby significantly reducing the computational complexity of obtaining the optimal policy. Simulation results show that the policy under the proposed optimization algorithm can effectively reduce the average information age with low complexity, thus achieving efficient complex state information updates.

Description

Technical Field

[0001] This invention relates to the field of wireless communication state information updating, and in particular to an optimization method for complex state updates of UAV-assisted Internet of Things under time-varying channels. Background Technology

[0002] In recent years, real-time applications in the Internet of Things (IoT) have developed rapidly. In these real-time applications, fresh state information is used to ensure their normal operation. To measure the freshness of state information, SK Kaul et al. proposed the Age of Information (AoI) (SK Kaul, R.D. Yates, and M. Gruteser, “Real-time status: How often should one update?” in Proc. IEEE INFOCOM, 2012, pp. 2731–2735). Existing literature mainly studies state updates involving only communication processes. However, in some emergency applications (such as fire gas monitoring), state information needs to be obtained through complex calculations of raw data, and the state information also needs to be transmitted to the control terminal for further processing through communication. Therefore, this invention defines the state update process, in which state information needs to be obtained through calculations of raw data, as a complex state information update. However, energy-limited IoT devices cannot tolerate the energy consumption pressure caused by performing computational tasks during complex state information updates. In this case, drones equipped with microservers can fly to the vicinity of IoT devices in emergency situations and use edge computing technology to alleviate the energy consumption pressure of IoT devices. Meanwhile, the link between the UAV and ground equipment also has good line-of-sight capability, facilitating the transmission of raw data and status information. Therefore, this invention aims to study UAV-assisted updates of complex status information.

[0003] Three performance metrics need to be considered in the process of updating complex state information with UAV assistance. First, the average information age reflects the average freshness of information over a period of time. Second, IoT devices are energy-constrained devices, and their average power consumption during complex state information updates should be considered. Third, since UAVs are also energy-constrained devices, their average power consumption must also be considered. Furthermore, due to the small amount of raw data and state information, short packet communication is adopted. In short packet communication, the probability of packet errors cannot be ignored.

[0004] In this system, drones fly to the area where IoT devices are located to provide services. To provide high-quality services to a small number of IoT devices, the drones hover throughout the service process, resulting in slow channel state changes. Therefore, the channel between the drone and the IoT devices or control center is block fading, meaning the channel state is fixed within a coherent time (i.e., one time slot) but changes in different time slots. Therefore, the actions of the IoT devices and drones should be adjusted according to the real-time channel state. For example, when the channel quality is poor, the IoT devices or drones can choose to remain silent to avoid a higher probability of packet errors. Thus, dynamic policy optimization for IoT devices and drones is key to improving the performance of complex state information update processes. However, the aforementioned complex state information update process belongs to a Markov Decision Process (MDP) and has a large state and action space, greatly increasing the difficulty and complexity of the solution process. Therefore, obtaining a high-performance dynamic policy with lower complexity is a challenging problem. Summary of the Invention

[0005] The purpose of this invention is to provide a method for optimizing complex state updates of UAV-assisted Internet of Things (IoT) under time-varying channels in short packet communication scenarios, so as to obtain dynamic strategies of IoT devices and UAVs with better performance with lower complexity.

[0006] The technical solution to achieve the purpose of this invention is as follows: A method for optimizing complex state updates in a UAV-assisted Internet of Things under time-varying channels is provided, comprising the following steps:

[0007] Step 1: Establish the state space for the complex state information update process in UAV-assisted Internet of Things under time-varying channels;

[0008] Step 2: Establish a linear programming problem based on an information age threshold;

[0009] Step 3: Set the threshold Δ to 1, the optimal threshold Δ opt The objective function is 1, and the optimal objective function is obj. opt The value is infinity, and the counting variable c is 1;

[0010] Step 4: Obtain the objective function value obj by solving a linear programming problem with a threshold of Δ;

[0011] Step 5: When obj < obj opt , make obj opt =obj,Δ opt =Δ;

[0012] Step 6: Increment the value of the count variable c by 1, and then set the threshold Δ = c;

[0013] Step 7: Return to step 3 until the counter variable c > X. max , where Xmax To control the total number of states in the age state space of the information on the control center side;

[0014] Step 8: Output the optimal threshold Δ opt This enables IoT devices and drones to select behaviors based on threshold-based policies.

[0015] The update and optimization method of the present invention is applicable to the UAV-assisted Internet of Things (IoT) scenario. In this scenario, the IoT device obtains raw data by sensing the environment and transmits the data to the UAV; the UAV obtains state information by processing the raw data and then transmits it to the control center. Compared with the prior art, its significant advantages are: (1) It considers designing a dynamic strategy under time-varying channels, so that the complex state information update process has adaptive behavior characteristics, thereby improving performance; (2) Unlike deep reinforcement learning algorithms that require a long time to train deep neural networks, it proposes a low-complexity threshold-based strategy, thereby transforming the complex dynamic strategy optimization problem into a threshold optimization problem; (3) In order to avoid the long traversal process under a fixed threshold, a state probability equation and a linear programming problem are established, and the optimal threshold is obtained by solving the linear programming problem. Attached Figure Description

[0016] Figure 1 This is a diagram of the UAV-assisted complex state information update model of the present invention.

[0017] Figure 2 This is a schematic diagram illustrating how the average information age on the control center side of the present invention changes with the transmission power of IoT devices.

[0018] Figure 3 This is a schematic diagram illustrating how the average power consumption of the IoT device changes with the transmission power of the IoT device according to the present invention.

[0019] Figure 4 This is a schematic diagram illustrating how the average power consumption of the drone changes with the transmission power of the Internet of Things (IoT) device, according to the present invention.

[0020] Figure 5 This is a schematic diagram showing the weighted sum of the three performance indicators of the present invention as the transmission power of the Internet of Things device changes. Detailed Implementation

[0021] Figure 1This is a model diagram of a drone-assisted complex state information update according to the present invention. During the process of a drone assisting an IoT device in completing complex state information updates, the channel is time-varying and subject to block fading. To efficiently complete complex state information updates under time-varying channel conditions, the IoT device can choose to remain silent or sense the environment and send raw data to the drone based on the real-time system state. Simultaneously, the drone can also choose to remain silent, calculate, or send state information to the control center based on the real-time system state. When the drone successfully receives the raw data, it must perform calculations to convert the raw data into state information. This process belongs to edge computing, which can reduce the energy consumption pressure on IoT devices. When the drone's state information transmission is successful, the drone has no data. At this time, the drone can only choose to remain silent. Furthermore, the amount of raw data obtained by the IoT device through sensing and the state information obtained by the drone through calculation are relatively small; their communication processes both use short packet communication.

[0022] This invention provides a method for optimizing complex state updates in a UAV-assisted Internet of Things (IoT) under time-varying channels, comprising the following steps:

[0023] Step 1: Establish the state space for the UAV-assisted complex state information update process under time-varying channels. The establishment process is as follows:

[0024] Step 1.1: Establish the age state space for the control center side. The states in this state space are integers, and the total number of states is X. max And there are no repeating states, and the minimum information age is N. c +2, the maximum information age is N c +1+X max Among them, N c It is the number of time slots used by the drone to complete the computing tasks of the Internet of Things devices;

[0025] Step 1.2: Establish the UAV data state space. This state space has a total of 3 states: 0, 1, and 2. Specifically, when the UAV data state is 0, the UAV has no data; when the UAV data state is 1, the UAV has raw data; and when the UAV data state is 2, the UAV has state information.

[0026] Step 1.3: Establish the age state space for IoT device information. The states in this state space are integers, and the total number of states is N. c +X max +1, and there are no duplicate states. The minimum information age is 0, and the maximum information age is N. c +X max ;

[0027] Step 1.4: Establish the UAV mission state space. The states in this state space are integers, and the total number of states is N. cAnd there are no repeating states, the minimum state is 0, and the maximum state is N. c -1. When the drone's mission status is 0, the drone is not performing calculations; when the drone's mission status is x and x is not 0, the drone is performing calculations and the calculation time has already occupied x time slots.

[0028] Step 1.5: Establish the channel quality state space between the IoT device and the drone. The states in this state space represent the squared magnitude of the small-scale fading coefficient, and the state space is {0.25, 0.5, 1, 1.5, 2}. Since the small-scale fading coefficient follows a Ricean distribution, the probability of each state in this state space can be calculated.

[0029] Step 1.6: Establish the channel quality state space between the control center and the UAV. The states in this state space represent the squared magnitude of the small-scale fading coefficient, and the state space is {0.25, 0.5, 1, 1.5, 2}. Since the small-scale fading coefficient follows a Ricean distribution, the probability of each state in this state space can be calculated.

[0030] Step 2: Establish a linear programming problem based on the information age threshold. The specific process is as follows:

[0031] Step 2.1: Establish the following state probability equation:

[0032]

[0033] in, The information age of the control center is the first state in its state space. The drone has no data. The information age of the IoT device is 0. The drone has not performed any calculations. The channel state between the IoT device and the drone is the k1th state in the state space. The probability that the channel state between the control center and the drone is the k2th state in the state space. The information age on the control center side is represented by its x-th state in its state space. The drone has state information, and the information age on the IoT device side is N. c +1, the probability that the channel state between the IoT device and the drone is the k1'th state in the state space, and the channel state between the control center and the drone is the k'2th state in the state space, is not calculated. This represents the probability of packet errors when the channel state between the IoT device and the drone is the k1'th state in the state space, where... The signal-to-noise ratio between the IoT device and the drone is given by N0, where N0 is the noise power spectral density and p is the signal-to-noise ratio between the IoT device and the drone. i For the transmit power of IoT devices, β0 is the large-scale fading coefficient between IoT devices and drones, β0 is the average large-scale fading coefficient at a reference distance of 1 meter, and diu It is the Euclidean distance between IoT devices and drones. K represents the small-scale fading coefficient between IoT devices and drones. iu =C1 exp(C2θ) iu θ represents the Rice factor between IoT devices and drones, C1 and C2 are environment-related constants, and θ is the Rice factor. iu It is the elevation angle of the IoT device to the drone. It is the defined line-of-sight channel component between IoT devices and drones. It is a complex Gaussian random variable with a mean of 0 and a variance of 1; D o W represents the amount of raw data; W represents the channel bandwidth. This represents the probability of packet errors when the channel state between the control center and the UAV is the k'2nd state in the state space, where... For the signal-to-noise ratio between the control center and the drone, p u For the drone's transmission power, d represents the large-scale fading coefficient between the control center and the UAV. uc It is the Euclidean distance between the control center and the drone. K represents the small-scale fading coefficient between the control center and the UAV. uc =C1 exp(C2θ) uc θ is the Rice factor between the control center and the drone. uc It is the angle of depression of the drone from the control center. It is the defined line-of-sight channel component between the control center and the UAV. It is a complex Gaussian random variable with a mean of 0 and a variance of 1; This represents the probability that the channel state between the IoT device and the drone is the k1'th state in the state space; This represents the probability that the channel state between the control center and the drone is the k'2nd state in the state space;

[0034] in, The information age of the control center is the first state in its state space. The drone has raw data. The information age of the IoT device is 1. The drone has not performed any calculations. The channel state between the IoT device and the drone is the k1th state in the state space. The probability that the channel state between the control center and the drone is the k2th state in the state space.

[0035] in, Let $\mathbf$ represent the information age of the control center, which is at state $x$ in its state space. The drone has no data. The information age of the IoT device is 0, and the drone has not performed any calculations. The probability that the channel state between the IoT device and the drone is at state $k1$ in its state space, and the channel state between the control center and the drone is at state $k2$ in its state space, is also present. The specific expression is:

[0036] in, The information age of the control center is the (x-1)th state in its state space. The drone has no data. The information age of the IoT device is 0. The drone has not performed any calculations. The channel state between the IoT device and the drone is the k1'th state in the state space. The channel state between the control center and the drone is the k'2th state in the state space. The information age of the control center is the xth state in its state space. The drone has no data. The information age of the IoT device is 0. The drone has not performed any calculations. The channel state between the IoT device and the drone is the k1'th state in the state space. The channel state between the control center and the drone is the k'2th state in the state space. The specific expression is:

[0037]

[0038] in, The information age on the control center side is represented by its y-th state in its state space. The drone has state information, and the information age on the IoT device side is N. c +x, the probability that the channel state between the IoT device and the drone is the k1'th state in the state space, and the channel state between the control center and the drone is the k'2th state in the state space, is not calculated. The age of the information on the control center side represents the Xth state space. max The drone has a state information, and the IoT device's information age is N. c +x, the probability that the channel state between the IoT device and the drone is the k1'th state in the state space, and the channel state between the control center and the drone is the k'2th state in the state space, is not calculated.

[0039]

[0040] in, The information age of the control center is the xth state in its state space. The drone has raw data. The information age of the IoT device is 1. The drone has not performed any calculations. The probability that the channel state between the IoT device and the drone is the k1th state in the state space and the channel state between the control center and the drone is the k2th state in the state space. and The specific expression is as follows:

[0041]

[0042] in, The information age of the control center is the xth state in its state space. The drone has raw data. The information age of the IoT device is z+1. The drone is performing calculations and the calculation time has occupied z time slots. The channel state between the IoT device and the drone is the k1th state in the state space. The probability that the channel state between the control center and the drone is the k2th state in the state space. The information age of the control center is the (x-1)th state in its state space. The drone has raw data. The information age of the IoT device is z. The drone is performing calculations and the calculation time has occupied z-1 time slots. The channel state between the IoT device and the drone is the (k1')th state in the state space. The probability that the channel state between the control center and the drone is the (k'2)th state in the state space. Let the age of the information on the control center side be the x-th state in its state space, the drone have raw data, the age of the information on the IoT device side be z, the drone is currently performing calculations and the calculation time has occupied z-1 time slots, the probability that the channel state between the IoT device and the drone is the k1'-th state in the state space, and the probability that the channel state between the control center and the drone is the k'2-th state in the state space.

[0043]

[0044] in, The information age of the control center is the xth state in its state space. The drone has state information. The information age of the IoT device is y. The drone has not been calculated. The probability that the channel state between the IoT device and the drone is the k1th state in the state space and the channel state between the control center and the drone is the k2th state in the state space. The information age on the control center side is the (x-1)th state in its state space. The drone has raw data. The information age on the IoT device side is y-1. The drone is currently performing calculations, and the calculation time has already occupied N. c -1 time slot, the probability that the channel state between the IoT device and the drone is the k1'th state in the state space, and the channel state between the control center and the drone is the k'2th state in the state space; The information age on the control center side is the xth state in its state space. The drone has raw data. The information age on the IoT device side is y-1. The drone is currently performing calculations, and the calculation time has already occupied N. c-1 time slot, the probability that the channel state between the IoT device and the drone is the k1'th state in the state space, and the channel state between the control center and the drone is the k'2th state in the state space; The information age of the control center is the (x-1)th state in its state space. The drone has state information. The information age of the IoT device is y-1. The drone has not been calculated. The channel state between the IoT device and the drone is the (k1')th state in the state space. The channel state between the control center and the drone is the (k'2)th state in the state space. The information age of the control center is the xth state in its state space. The drone has state information. The information age of the IoT device is y-1. The drone has not been calculated. The probability that the channel state between the IoT device and the drone is the k1'th state in the state space and the channel state between the control center and the drone is the k'2th state in the state space. The information age of the control center is the xth state in its state space. The drone has state information. The information age of the IoT device is y. The drone has not been calculated. The channel state between the IoT device and the drone is the k1'th state in the state space. The channel state between the control center and the drone is the k'2th state in the state space.

[0045] Step 2.2: Establish the following expressions for the average information age of state probabilities, the average power consumption of IoT devices, and the average power consumption of drones:

[0046] in, and These are the average information age, the average power consumption of IoT devices, and the average power consumption of drones, respectively; S represents the overall state, which includes the information age on the control center side, the drone data state, the information age on the IoT device side, the drone computing state, the channel state between IoT devices and drones, and the channel state between the control center and drones. S represents the set of all feasible states as shown in equations (1)-(10); S(1) represents the age of the control center side information in S; μ S p represents the probability that the overall state is S; s For the sensing power of IoT devices; A Δ The Δth information age represents the information age in the state space of the control center; κ is a constant related to the UAV's computing architecture; f UAV S represents the computing power of the UAV; S(2) represents the UAV data state in S;

[0047] Step 2.3: Establish the following threshold-based linear programming problem:

[0048] st (1) - (10), (15)

[0049] Where, ω IoT and ω UAV These are the average power consumption weighting factors for IoT devices and the average power consumption weighting factor for drones, respectively; μ is the set of probabilities for all feasible states;

[0050] Step 3: Set the threshold Δ to 1, the optimal threshold Δ opt The objective function is 1, and the optimal objective function is obj. opt The value is infinity, and the counting variable c is 1;

[0051] Step 4: Obtain the objective function value obj by solving a linear programming problem with a threshold of Δ;

[0052] Step 5: When obj < obj opt , make obj opt =obj,Δ opt =Δ;

[0053] Step 6: Increment the value of the count variable c by 1, and then set the threshold Δ = c;

[0054] Step 7: Return to step 3 until the counter variable c > X. max , where X max To control the total number of states in the age state space of the information on the control center side;

[0055] Step 8: Output the optimal threshold Δ opt This enables IoT devices and drones to select behaviors based on threshold-based policies.

[0056] The strategy is implemented as follows: when the information age of the control center is less than the Δth level of its state space... opt If the information age is less than the Δth information age in its state space, neither the IoT device nor the drone will send information; if the information age on the control center side is not less than the Δth information age in its state space... opt When the information age is reached, the IoT device sends raw data to the drone, and the drone sends state information to the control center when it has state information; specifically, when the drone has raw data, even if the information age on the control center side is not less than the Δth information age in its state space... opt The drone also calculates the age of the individual.

[0057] Example:

[0058] Simulation parameter settings: The coordinates of the IoT device are [0,0], the coordinates of the control center are [3000,30], and the coordinates of the UAV are [1500,100]. The coordinate unit is meters. The channel bandwidth is W = 1MHz, the average large-scale fading coefficient β0 = -50dB, and the original data size is D.o =500 bits, state information data volume D s =300 bits, UAV computing speed f UAV =2.4GHz, raw data and status information packet length n = 1000, number of channels used, calculation architecture correlation coefficient κ = 10 -28 IoT device transmit power p i =10dBm, IoT device sensing power p s =10dBm, UAV transmit power p u =20dBm, environmentally relevant parameters C1=0dB, C2=30dB, IoT device average power consumption weighting factor ω IoT =10, the average power consumption weighting factor ω of the drone UAV =10, noise power spectral density N0 = -174dBm / Hz. Furthermore, to balance simulation performance and simulation time, this invention sets the total number of state-space states for the information age on the control center side to be X. max =15.

[0059] Figure 2 This diagram illustrates the variation of the average information age at the control center with the transmission power of IoT devices, as described in this invention. The random strategy represents the IoT devices and drones randomly selecting feasible behaviors in each system state. The dynamic optimization strategy represents the behavior of the IoT devices and drones being dynamically optimized, no longer limited by thresholds, and performing strategy optimization in each system state. Of course, the computational complexity of obtaining the dynamic optimization strategy is far higher than the complexity of the optimization method proposed in this invention. As can be seen from the figure, the average information age at the control center decreases as the transmission power of the IoT devices increases. This is because increasing the transmission power of the IoT devices reduces the probability of packet errors and improves the success rate of original data transmission, thereby reducing the average information age. Furthermore, compared to the other two strategies, the strategy proposed in this invention achieves the optimal average information age in most cases, demonstrating that the strategy proposed in this invention has better performance in reducing the average information age.

[0060] Figure 3This diagram illustrates the variation of the average power consumption of IoT devices as a function of their transmission power. As can be seen from the diagram, the average power consumption of IoT devices under all three strategies generally increases with increasing transmission power. This is because the power consumption of IoT devices is positively correlated with their transmission power. However, when the transmission power is low, the average power consumption of IoT devices under the dynamic optimization strategy decreases. This is because while the transmission power and transmission success rate increase under the dynamic optimization strategy, the number of transmissions over the long term decreases, leading to a decrease in average power consumption. However, when the transmission power continues to increase, the transmission frequency does not change significantly in order to maintain a low average information age, which increases the average power consumption of IoT devices. Furthermore, the higher average power consumption of IoT devices under the optimization method proposed in this invention indicates a higher transmission frequency.

[0061] Figure 4 This diagram illustrates the variation of the average power consumption of the drone as a function of the transmission power of the IoT device. As can be seen from the diagram, the average power consumption of the drone initially increases with the transmission power of the IoT device, then plateaus. This is because as the transmission power of the IoT device increases, the transmission success rate of the IoT device also increases, thereby increasing the frequency of the drone's computation and transmission. When the number of IoT devices increases to a certain extent, the IoT transmission frequency tends to stabilize, and the drone's power consumption also tends to stabilize. Meanwhile, in the strategy proposed in this invention, the average power consumption of the drone is relatively high, which is related to... Figure 3 The high transmission frequency of the aforementioned IoT devices is highly relevant.

[0062] Figure 5 This is a schematic diagram illustrating the weighted sum of the three performance indicators of this invention as the transmission power of an IoT device changes. (Summary) Figure 2 , Figure 3 , Figure 4 It can be seen that there is a trade-off between the average information age at the control center, the average power consumption of IoT devices, and the average power consumption of drones, which leads to a turning point in the weighted sum curves under all three strategies. This indicates that an appropriate IoT device transmit power needs to be selected to obtain optimal performance. Furthermore, in the strategies under the optimization method proposed in this invention, the weighted sum at the optimal value is superior to the weighted synthesis under the random strategy, demonstrating the performance advantage of the proposed method. In addition, although the performance of the strategies under the optimization method proposed in this invention still lags behind that of the dynamic optimization strategy, the computational complexity of obtaining the dynamic optimization strategy is far greater than that of obtaining the strategies under the optimization method proposed in this invention, thus also demonstrating the effectiveness of the optimization method proposed in this invention.

[0063] The above embodiments are described in detail, but only illustrate one feasible implementation of the present invention, and are not intended to limit the scope of the present invention. It should be noted that researchers and engineers in the art can add several modifications or improvements to these embodiments within the framework of the present invention, but these are all within the protection scope of the present invention, which is determined by the appended claims.

Claims

1. A method for optimizing complex state updates in UAV-assisted Internet of Things (IoT) under time-varying channels, characterized in that, Includes the following steps: Step 1: Establish the state space for the complex state information update process in UAV-assisted Internet of Things under time-varying channels; This includes establishing the information age state space on the control center side, the UAV data state space, the information age state space on the IoT device side, the UAV mission state space, the channel quality state space between the IoT device and the UAV, and the channel quality state space between the control center and the UAV. Step 2: Establish a linear programming problem based on an information age threshold; Step 2 includes: constructing a state probability equation that satisfies the probability of being in each feasible state based on the established state space; establishing expressions for the average information age, average power consumption of IoT devices, and average power consumption of drones based on the probabilities of the states; and constructing a linear programming problem based on an information age threshold with the goal of minimizing the weighted sum of the average information age, the average power consumption of IoT devices, and the average power consumption of drones. Step 3: Set the threshold The optimal threshold is 1. The objective function is 1. Infinity, counting variable =1; Step 4: By solving the threshold... The linear programming problem obtains the objective function value of the problem. ; Step 5: When ,make , ; Step 6: Let the counter variable... The value is increased by 1, and then the threshold is set. ; Step 7: Return to step 4 until the counter variable is reached. ,in, To control the total number of states in the age state space of the information on the control center side; Step 8: Output the optimal threshold This enables IoT devices and drones to select behaviors based on threshold-based policies; The threshold-based strategy is as follows: when the information age of the control center is less than the first value in its state space... If the information age is less than the specified age in its state space, neither IoT devices nor drones will send information; if the information age on the control center side is not less than the specified age in its state space... When the information age is reached, the IoT device sends raw data to the drone, and the drone sends state information to the control center when it has state information; when the drone has raw data, even if the information age on the control center side is not less than the [number]th ... The drone also calculates the age of the individual.

2. The method for optimizing complex states of an unmanned aerial vehicle (UAV)-assisted Internet of Things (IoT) under time-varying channels according to claim 1, characterized in that, Step 1 describes the process of establishing the state space for the complex state information update process in UAV-assisted IoT under time-varying channels. The specific process is as follows: Step 1.1: Establish the age state space for the control center side. The states in this state space are integers and the total number is 1. And there are no duplicate states, the minimum information age is The maximum information age is ;in, It is the number of time slots used by the drone to complete the computing tasks of the Internet of Things devices; Step 1.2: Establish the UAV data state space. The total number of states in this state space is 3, including 0, 1, and 2. When the UAV data state is 0, the UAV has no data. When the UAV data state is 1, the UAV has raw data. When the UAV data state is 2, the UAV has state information. Step 1.3: Establish the age state space of IoT device-side information. The states in this state space are integers and the total number is 1. And there are no duplicate states, the minimum information age is 0, and the maximum information age is ; Step 1.4: Establish the UAV mission state space, where the states are integers and the total number is 1. And there are no repeating states, the minimum state is 0, and the maximum state is When the drone's mission status is 0, the drone does not perform calculations; when the drone's mission status is... and If the value is not 0, then the drone is performing calculations and the calculation time has already been used. One time slot; Step 1.5: Establish the channel quality state space between the IoT device and the drone. The states in this state space represent the squared magnitude of the small-scale fading coefficient, and this state space is... Since the small-scale fading coefficients follow a Rice distribution, the probability of each state in this state space can be calculated. Step 1.6: Establish the channel quality state space between the control center and the UAV. The states in this state space represent the squared magnitude of the small-scale fading coefficient, and this state space is... Since the small-scale fading coefficients follow a Rice distribution, the probability of each state in this state space can be calculated.

3. The method for optimizing complex states of an unmanned aerial vehicle (UAV)-assisted Internet of Things (IoT) under time-varying channels according to claim 2, characterized in that, The process of establishing the linear programming problem based on the information age threshold described in step 2 is as follows: Step 2.1: Establish the following state probability equation: （1） in, The information age on the control center side is the first state in its state space. The drone has no data. The information age on the IoT device side is 0. The drone has not performed any calculations. The channel state between the IoT device and the drone is the [missing information] state space. The state is the channel state between the control center and the UAV, which is the state space of the [number]th state. The probability of each state; The age of the information on the control center side represents its state space. The drone has status information, and the age information on the IoT device side is [age information]. The drone did not perform any calculations; the channel state between the IoT device and the drone was the state space of the [missing information]. The state is the channel state between the control center and the UAV, which is the state space of the [number]th state. The probability of each state; This represents the state space when the channel state between the IoT device and the drone is the first state. The probability of packet errors when an IoT device sends raw data packets to a drone in each state, where... The signal-to-noise ratio between IoT devices and drones. For noise power spectral density, For the transmit power of IoT devices, This represents the large-scale fading coefficient between IoT devices and drones. The average large-scale fading coefficient is given at a reference distance of 1 meter. It is the Euclidean distance between IoT devices and drones. This refers to the small-scale fading coefficient between IoT devices and drones. For the Rice factor between IoT devices and drones, and It is a constant related to the environment. It is the elevation angle of the IoT device to the drone. It is the defined line-of-sight channel component between IoT devices and drones. It is a complex Gaussian random variable with a mean of 0 and a variance of 1; , ; The amount of data in the original dataset; Channel bandwidth; This represents the state space when the channel state between the control center and the drone is the first state. The probability of packet errors when the UAV sends status information to the control center in each state is given by: For the signal-to-noise ratio between the control center and the drone, For the drone's transmission power, This represents the large-scale fading coefficient between the control center and the drone. It is the Euclidean distance between the control center and the drone. The small-scale fading coefficient between the control center and the UAV. For the Rice factor between the control center and the drone, It is the angle of depression of the drone from the control center. It is the defined line-of-sight channel component between the control center and the UAV. It is a complex Gaussian random variable with a mean of 0 and a variance of 1; This represents the state space when the channel state between the IoT device and the drone is the first state. The probability of each state; This represents the state space when the channel state between the control center and the drone is the first state. The probability of each state; （2） in, The information age on the control center side is the first state in its state space. The drone has raw data, the information age on the IoT device side is 1, the drone has not performed any calculations, and the channel state between the IoT device and the drone is the [missing information] state space. The state is the channel state between the control center and the UAV, which is the state space of the [number]th state. The probability of each state; （3） in, The age of the information on the control center side represents its state space. In the current state, the drone has no data, the information age on the IoT device side is 0, the drone is not performing any calculations, and the channel state between the IoT device and the drone is the state space of the [number missing]. The state is the channel state between the control center and the UAV, which is the state space of the [number]th state. The probability of each state. The specific expression is: （4） in, The age of the information on the control center side represents its state space. In the current state, the drone has no data, the information age on the IoT device side is 0, the drone is not performing any calculations, and the channel state between the IoT device and the drone is the state space of the [number missing]. The state is the channel state between the control center and the UAV, which is the state space of the [number]th state. The probability of each state; The age of the information on the control center side represents its state space. In the current state, the drone has no data, the information age on the IoT device side is 0, the drone is not performing any calculations, and the channel state between the IoT device and the drone is the state space of the [number missing]. The state is the channel state between the control center and the UAV, which is the state space of the [number]th state. The probability of each state; The specific expression is: （5） in, The age of the information on the control center side represents its state space. The drone has status information, and the age information on the IoT device side is [age information]. The drone did not perform any calculations; the channel state between the IoT device and the drone was the state space of the [missing information]. The state is the channel state between the control center and the UAV, which is the state space of the [number]th state. The probability of each state; The age of the information on the control center side represents its state space. The drone has status information, and the age information on the IoT device side is [age information]. The drone did not perform any calculations; the channel state between the IoT device and the drone was the state space of the [missing information]. The state is the channel state between the control center and the UAV, which is the state space of the [number]th state. The probability of each state; （6） in, The age of the information on the control center side represents its state space. In the current state, the drone has raw data, the IoT device's information age is 1, the drone has not performed any calculations, and the channel state between the IoT device and the drone is the state space of the [number missing]. The state is the channel state between the control center and the UAV, which is the state space of the [number]th state. The probability of each state; and The specific expression is as follows: （7） （8） （9） in, The age of the information on the control center side represents its state space. In this state, the drone has raw data, and the age of the information on the IoT device side is... The drone is currently performing calculations, and the calculation time has already been used up. In the time slot, the channel state between the IoT device and the drone is the state space _ _. The state is the channel state between the control center and the UAV, which is the state space of the [number]th state. The probability of each state; The age of the information on the control center side represents its state space. In this state, the drone has raw data, and the age of the information on the IoT device side is... The drone is currently performing calculations, and the calculation time has already been used up. In the time slot, the channel state between the IoT device and the drone is the state space _ _. The state is the channel state between the control center and the UAV, which is the state space of the [number]th state. The probability of each state; The age of the information on the control center side represents its state space. In this state, the drone has raw data, and the age of the information on the IoT device side is... The drone is currently performing calculations, and the calculation time has already been used up. In the time slot, the channel state between the IoT device and the drone is the state space _ _. The state is the channel state between the control center and the UAV, which is the state space of the [number]th state. The probability of each state （10） in, The age of the information on the control center side represents its state space. The drone has status information, and the age information on the IoT device side is [age information]. The drone did not perform any calculations; the channel state between the IoT device and the drone was the state space of the [missing information]. The state is the channel state between the control center and the UAV, which is the state space of the [number]th state. The probability of each state; The age of the information on the control center side represents its state space. In this state, the drone has raw data, and the age of the information on the IoT device side is... The drone is currently performing calculations, and the calculation time has already been used up. In the time slot, the channel state between the IoT device and the drone is the state space _ _. The state is the channel state between the control center and the UAV, which is the state space of the [number]th state. The probability of each state; The age of the information on the control center side represents its state space. In this state, the drone has raw data, and the age of the information on the IoT device side is... The drone is currently performing calculations, and the calculation time has already been used up. In the time slot, the channel state between the IoT device and the drone is the state space _ _. The state is the channel state between the control center and the UAV, which is the state space of the [number]th state. The probability of each state; The age of the information on the control center side represents its state space. The drone has status information, and the age information on the IoT device side is [age information]. The drone did not perform any calculations; the channel state between the IoT device and the drone was the state space of the [missing information]. The state is the channel state between the control center and the UAV, which is the state space of the [number]th state. The probability of each state; The age of the information on the control center side represents its state space. The drone has status information, and the age information on the IoT device side is [age information]. The drone did not perform any calculations; the channel state between the IoT device and the drone was the state space of the [missing information]. The state is the channel state between the control center and the UAV, which is the state space of the [number]th state. The probability of each state; The age of the information on the control center side represents its state space. The drone has status information, and the age information on the IoT device side is [age information]. The drone did not perform any calculations; the channel state between the IoT device and the drone was the state space of the [missing information]. The state is the channel state between the control center and the UAV, which is the state space of the [number]th state. The probability of each state; Step 2.2: Establish the following expressions for the average information age of state probabilities, the average power consumption of IoT devices, and the average power consumption of drones: （11） （12） （13） in, , and These are average information age, average power consumption of IoT devices, and average power consumption of drones; It represents the overall status, which includes the information age on the control center side, the drone data status, the information age on the IoT device side, the drone computing status, the channel status between the IoT device and the drone, and the channel status between the control center and the drone. Represents the set of all feasible states as shown in equations (1) to (10); represent Age information in the control center; The overall state is The probability of; The sensing power of IoT devices; The first in the age state space representing the information of the control center side Age information; These are constants related to the drone's computing architecture; The computing power of the drone; represent The status of drone data in the system; Step 2.3: Establish the following threshold-based linear programming problem: （14） st Equations (1)-(10), (15) （16） in, and These are the average power consumption weighting factors for IoT devices and the average power consumption weighting factors for drones, respectively. It is the set of probabilities for all feasible states.

Images