Unmanned aerial vehicle ground-air cooperative networking data communication method

Abstract

The invention relates to the technical field of data communication, and particularly discloses an unmanned aerial vehicle ground-air cooperative networking data communication method, which comprises the following steps: marking an unmanned aerial vehicle used for executing a preset task in an unmanned aerial vehicle cluster as a first unmanned aerial vehicle, and recording a signal obtained by converting data acquired by the first unmanned aerial vehicle as a target signal; a connecting line between the first unmanned aerial vehicle and the center of the ground is obtained and recorded as a target line, a communication link of the first unmanned aerial vehicle is determined based on the target line, and the communication link is formed based on the unmanned aerial vehicle; and keeping to ensure that the unmanned aerial vehicle on the communication link and the first unmanned aerial vehicle are relatively static, transmitting the target signal to a ground center through the communication link, and performing frequency compensation on the received target signal by the ground center. According to the invention, the influence of Doppler effect on data transmission quality can be reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of data communication, and particularly to a data communication method for unmanned aerial vehicle (UAV) ground-air collaborative networking. Background Art

[0002] UAV ground-air collaborative networking is a network architecture that enables the collaboration and complementary advantages between ground networks and aerial UAV systems. UAVs not only exist as individual information collection and transmission platforms but also form a collaborative overall network with ground stations and other UAVs to achieve fast, efficient, and stable information transmission.

[0003] The Doppler effect refers to the phenomenon that when there is relative motion between a wave source and an observer, the frequency of the wave (such as sound waves, light waves, etc.) received by the observer will change. In UAV ground-air collaborative networking, due to the flight of UAVs in the air, there is relative motion between them and ground stations or other platforms, which will cause frequency offset of wireless communication signals and affect data communication quality. Summary of the Invention

[0004] The purpose of the present invention is to provide a data communication method for UAV ground-air collaborative networking to solve the following technical problems: In UAV ground-air collaborative networking, due to the flight of UAVs in the air, there is relative motion between them and ground stations or other platforms, which will cause frequency offset of wireless communication signals and affect data communication quality.

[0005] The purpose of the present invention can be achieved through the following technical solutions: A data communication method for UAV ground-air collaborative networking includes the following steps: Mark the UAVs in the UAV cluster that are used to execute a preset task as the first UAVs, and the signals obtained by converting the data collected by the first UAVs are denoted as target signals; Obtain the connection line between the first UAV and the ground center, denoted as the target line, and determine the communication link of the first UAV based on the target line, where the communication link is composed of the UAVs; Keep the UAVs on the communication link relatively stationary with respect to the first UAV, and transmit the target signals to the ground center through the communication link, and the ground center performs frequency compensation on the received target signals.

[0006] As a further solution of the present invention: determining the communication link of the first UAV based on the target line includes: Step 1: Take the first UAV as the component aircraft, draw a circle with a preset optimal communication distance D as the radius centered on the component aircraft, and mark the UAVs within the circle as candidate aircraft; Obtain the distance between the projection point of the candidate aircraft on the target line and the first unmanned aircraft, denoted as the selection distance, and use the candidate aircraft corresponding to the maximum selection distance as the new component aircraft A; Step 2: Taking the component aircraft A as the center, repeat Step 1 to obtain new component aircraft, and repeat the above steps until the distance between a new component aircraft and the ground center is less than the optimal communication distance, and all the candidate aircraft form the communication link.

[0007] As a further solution of the present invention: The process of keeping the unmanned aircraft on the communication link relatively stationary with the first unmanned aircraft includes: Mark the factors affecting the speed of the unmanned aircraft as target factors, collect the target factors of the first unmanned aircraft, input them into a pre-trained speed decay prediction model, output the speed decay value, set the speed of the first unmanned aircraft as the preset value V, and subtract the speed decay value from the preset value V to obtain the predicted speed V0 of the first unmanned aircraft; Mark the unmanned aircraft on the communication link as the target unmanned aircraft, obtain the speed decay value V1 of the target unmanned aircraft, and set the speed of the target unmanned aircraft as V1 + V0; Ensure that the target unmanned aircraft and the first unmanned aircraft move in the same direction, so as to achieve relative rest between the unmanned aircraft on the communication link and the first unmanned aircraft.

[0008] As a further solution of the present invention: The process of pre-training the speed decay prediction model includes: Obtain a data set, which stores target factors with marked speed decay values, and divide the data set into a training set and a validation set according to a preset ratio; Based on deep learning, establish a speed decay prediction model, and train and validate the speed decay prediction model based on the training set and the validation set to obtain a pre-trained speed decay prediction model.

[0009] As a further solution of the present invention: In the process of transmitting the target signal to the ground center through the communication link, the following steps are further included: Take the last unmanned aircraft on the communication link as the second unmanned aircraft. After the second unmanned aircraft receives the target signal, preprocess the target signal, and the preprocessing includes adaptive filtering and frequency pre-compensation. The second unmanned aircraft transmits the preprocessed target signal to the ground center.

[0010] As a further solution of the present invention: In the process of transmitting the target signal to the ground center through the communication link, the following steps are further included: When the distance between the second drone and the ground center is equal to 0.95D, if the first drone has not completed the task yet, a new drone is activated to join the drone cluster, and the communication link is updated when the distance between the second drone and the ground center is equal to D.

[0011] As a further solution of the present invention: in the process of determining the communication link of the first drone based on the target line, it further includes; When drone i is a part of m communication links and m > n, where n is a preset quantity, the following steps are executed: Take the m communication links corresponding to drone i as pending links, and obtain the distance between drone i and the third drone, denoted as the sorting distance. The third drone is the previous drone adjacent to drone i in the pending links; Sort the pending links in ascending order according to the size of the sorting distance, and drone i is a part of the first n pending links in the sorting.

[0012] As a further solution of the present invention: after determining the communication links of all the first drones, the drones that do not belong to any communication link return to the ground center and enter the sleep state.

[0013] Advantages of the present invention: In this solution, by selecting the first drone in the drone swarm to specifically perform the preset task without participating in the optimization of the communication link, the energy consumption of the first drone is minimized to the greatest extent, ensuring that it can concentrate resources to successfully complete the core task. At the same time, other drones build a communication link around the target line between the first drone and the ground center, and conduct step-by-step selection based on the optimal communication distance. In this way, it is ensured that the data will not experience excessive jumps and unstable links during transmission, and communication delays caused by frequent switching are also avoided. On this basis, the speeds of the drones on the communication link and the first drone remain relatively stationary. Thanks to the flight speed decay compensation setting achieved by the pre-trained speed decay prediction model, this relative stillness can not only make the data transmission process more stable and continuous, but also significantly reduce the Doppler effect caused by the relative speed difference and reduce the influence of frequency offset during communication. In addition, during the process when the transmission is about to end, if the first drone has not completed the task, a new drone will be automatically activated to join the swarm to take over the end of the link, thereby maintaining the ideal communication distance and further reducing the interference caused by sudden Doppler frequency shift. The last drone on the link will perform adaptive filtering and frequency pre-compensation before transmitting the target signal to the ground center. The combination of this pre-compensation measure and the final frequency compensation at the ground center can effectively cancel most of the frequency offsets caused by the flight environment, airflow disturbance or a small amount of remaining relative speed, further reducing the negative impact of the Doppler effect on the communication quality. And when updating the communication link, the second drone changes, that is, the drone that preprocesses the target signal changes, so as to prevent excessive energy consumption of a certain drone. To prevent excessive resource consumption caused by the repeated use of a single drone in multiple communication links, a priority selection mechanism is adopted when the same drone belongs to multiple communication links at the same time, so as to reasonably allocate communication tasks and improve the reliability and coordination of the overall network. All drones that do not belong to any communication link will return to the ground center and enter the sleep state to ensure the effective utilization and expansion of system resources. In summary, through the above method, not only efficient and robust link communication between the drone and the ground center is achieved, but also the negative impact of the Doppler effect is significantly weakened on the premise of reducing the energy consumption of the first drone and ensuring the smooth completion of the task, thereby ensuring the data transmission quality and flight operation efficiency. Brief Description of the Drawings

[0014] The present invention will be further described below with reference to the accompanying drawings.

[0015] Figure 1 It is a schematic flowchart of a method for drone-ground collaborative networking data communication according to the present invention. Detailed Embodiments

[0016] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0017] Please refer to Figure 1 As shown, the present invention is a method for drone ground-air collaborative networking data communication, including the following steps: Mark the drones in the drone cluster used to execute the preset tasks as the first drones, and the signals obtained by converting the data collected by the first drones are denoted as target signals; Obtain the connection line between the first drone and the ground center, denoted as the target line, and determine the communication link of the first drone based on the target line. The communication link is composed of the drones; In a preferred embodiment of the present invention, determining the communication link of the first drone based on the target line includes: Step 1: Take the first drone as the component drone, draw a circle with a radius of the preset optimal communication distance D centered on the component drone, and mark the drones within the circle as candidate drones; Obtain the distance between the projection point of the candidate drone on the target line and the first drone, denoted as the selection distance, and take the candidate drone corresponding to the maximum selection distance as the new component drone A; Step 2: Take the component drone A as the center of the circle, repeat Step 1 to obtain a new component drone, and repeat the above steps until the distance between a certain new component drone and the ground center is less than the optimal communication distance. All the candidate drones form the communication link; It should be noted that first, the three-dimensional space coordinates of all drones and the position information between the first drone and the ground center are obtained in advance, and the line connecting the first drone and the ground center is defined as the target line; then, with the first drone as the initial component drone, based on its three-dimensional coordinates at its location, a spherical area (corresponding to a circular area in a plane scenario) is constructed with a preset optimal communication distance D as the radius, and all drones with coordinates within this area are marked as candidate drones; for each candidate drone, its projection point relative to the target line is calculated, and the distance between this projection point and the position of the first drone is measured, and this distance is the selection distance; among all candidate drones, the drone with the largest selection distance is regarded as the new component drone A, which means it can extend further on the target line than other candidate drones, thus helping to reduce the number of hops of subsequent communication nodes and improve link stability; then, taking the coordinates of this new component drone A as the new center of the sphere, repeat the above steps: construct a spherical area again with the same optimal communication distance D, find the next batch of candidate drones, calculate the projection points, and select the largest selection distance to obtain a new component drone; repeat this process step by step approaching the ground center until in a certain iteration, the distance between the new component drone and the ground center is less than the optimal communication distance D, indicating that the communication link can directly reach the ground center from the first drone without further expansion; in this process, each selected component drone together with the intermediate generated candidate drones forms a continuous drone relay link in the selected order, achieving a segmented and gradually extended network coverage; In a preferred case of this embodiment, during the process of determining the communication link of the first drone based on the target line, it further includes; When drone i is a component of m of the communication links and m > n, where n is a preset quantity, the following steps are executed: Regarding the m communication links corresponding to drone i as pending links, obtain the distance between drone i and the third drone, denoted as the sorting distance, where the third drone is the previous drone adjacent to drone i in the pending links; Sort the pending links in ascending order of the sorting distance, and drone i is a component of the first n pending links in the sorting; It should be noted that, in order to prevent uneven resource consumption or excessive communication burden caused by over-utilization of the same drone, when it is detected that a certain drone i belongs to multiple communication links simultaneously and the number exceeds a pre-set threshold n, the drone will be preferentially retained. Specifically, first, all communication links associated with drone i are identified, denoted as m and regarded as pending links; then, for each pending link, the system calculates a sorting distance based on the distance between drone i and the previous adjacent drone (which can be called the third drone) in this link. Here, the previous adjacent drone can be understood as the node that is before i along the target line between the first drone and the ground center when establishing the communication link. On a certain link, if the third drone is j, the sorting distance is the three-dimensional Euclidean distance or other appropriate measure between i and j. Then, the system arranges these m pending links in ascending order according to the corresponding sorting distances. Generally speaking, the smaller the distance, the more compact and stable the communication configuration of this link, and it also means that the geographical location and topological role of i in this link are more critical. Based on this idea, the system will retain i as a component of the top n ranked links, and exclude i from the links ranked behind, so as to ensure that in the overall communication design of the drone swarm, i only appears in the n most prioritized links.

[0018] Exemplarily, assume that drone i currently belongs to 5 pending links (m = 5), and the preset threshold n = 3; after calculation, the sorting distances of the 5 links are 2, 7, 8, 10, and 15 in sequence; the system will retain i in the three links with the smallest sorting distances of 2, 7, and 8, and exclude i from the two links with sorting distances of 10 and 15. In actual operation, the links after exclusion will try to find alternative drones or re-plan the nodes to ensure communication continuity, while drone i can continue to play its communication role on the most favorable and stable several links; In another preferred embodiment of the present invention, after determining the communication links of all the first drones, the drones that do not belong to any of the communication links return to the ground center and enter the sleep state; It should be noted that this is to avoid wasting energy and spectrum resources due to ineffective waiting in the drone cluster, and at the same time reduce airspace occupation and collision risks, leaving sufficient resources and space for subsequent tasks or new communication requirements of the first drones. If there are indeed new communication links or tasks in the future, the ground center only needs to wake up the standby drones in the sleep state and repeat the previous communication link construction and screening process; Keep the drones on the communication link relatively stationary with the first drone, and transmit the target signal to the ground center through the communication link, and the ground center performs frequency compensation on the received target signal.

[0019] It should be noted that in the digital baseband signal received and down-converted by the ground center, a pre-set pilot sequence or known training symbol is usually combined to estimate frequency deviations such as Doppler frequency shift. First, the deviation is roughly detected through correlation operations or matched filtering, and the detection result is sent to a numerically controlled oscillator or a digital mixing module to perform coarse compensation, so that the carrier center frequency basically returns to the ideal position. Subsequently, fine tracking and phase synchronization are performed through closed-loop methods such as decision feedback or phase-locked loop to further suppress the residual frequency deviation to ensure the accuracy of subsequent demodulation or decoding. If the environment changes violently during the communication process or the Doppler shift fluctuates significantly over time, an adaptive filtering or iterative update algorithm is used to continuously correct the compensation parameters, so that a low frequency deviation and high demodulation reliability can be maintained in a dynamic flight environment. Another preferred embodiment of the present invention, the process of keeping the UAV on the communication link relatively stationary with the first UAV includes: Mark the factors affecting the speed of the UAV as target factors, collect the target factors of the first UAV, input them into a pre-trained speed decay prediction model, output the speed decay value, set the speed of the first UAV as the preset value V, and subtract the speed decay value from the preset value V to obtain the predicted speed V0 of the first UAV; Mark the UAV on the communication link as the target UAV, obtain the speed decay value V1 of the target UAV, and set the speed of the target UAV as V1 + V0; Ensure that the target UAV and the first UAV move in the same direction, and achieve relative static of the UAV on the communication link and the first UAV.

[0020] In a preferred case of this embodiment, the process of pre-training the speed decay prediction model includes: Obtain a data set, which stores the target factors with the marked speed decay values, and divide the data set into a training set and a validation set according to a preset ratio; Based on deep learning, establish a speed decay prediction model, and train and validate the speed decay prediction model based on the training set and the validation set to obtain a pre-trained speed decay prediction model; It should be noted that during the model pre-training stage, historical or experimental data of various factors (target factors) affecting the speed of the drone need to be prepared. For example, it can include various environmental and mission-related elements such as wind speed, wind direction, altitude, temperature and humidity, aircraft load, engine or motor power loss, etc., and each data sample is labeled with its corresponding speed attenuation value. After cleaning and standardizing these raw data, they are divided into a training set and a validation set. Subsequently, a suitable deep learning or machine learning model (such as based on a multi-layer perceptron, LSTM network, or other neural network structures) is selected to iteratively train the training set, and the model accuracy and generalization ability are evaluated on the validation set. If the accuracy meets the standard, a pre-trained speed attenuation prediction model is obtained; In the actual deployment stage, assume there is such a scenario: The first drone F (performing the core task) and several target drones B, C, etc. that form a communication link need to fly together to achieve data relay and transmission. To keep B and C relatively stationary with respect to F during flight, thereby avoiding excessive relative displacement and Doppler effect caused by the speed difference, the following steps can be taken: During flight, the ground command center or the on-board system continuously obtains the current environmental parameters and internal states of each drone, such as the altitude of A, wind speed and direction, external temperature, load weight, engine performance attenuation, and battery power, etc. These key values related to speed are used as target factors and input into the pre-trained speed attenuation prediction model to obtain the speed attenuation value of drone F (denoted as F1); For the first drone F, an ideal operating speed V (such as the cruise speed required for a certain task) is usually pre-given. After obtaining the speed attenuation value F1 of F, it can be directly subtracted from V to obtain the predicted speed V0 of the first drone, that is, the speed that can be maintained under the current target factors (the target factors affect the speed, the set speed is V, but in fact it may only reach V0); For the other drones B, C that form the communication link, their respective target factors also need to be collected and input into the speed attenuation prediction model to obtain their respective attenuation values (which can be denoted as B1 and C1 respectively). Since B and C do not need to bear the load of performing the core task, their attenuation values will often be different from that of F. To achieve their relative static state with respect to F1, theoretically, as long as the actual speeds of B and C are the same as the actual speed of F. Specifically, when implementing, it can be set in the way of "the speed of B = V0 + B1" to make B and F have the same speed (similarly for C); In addition to the matching of speed scalars, it is also necessary to ensure that the heading directions of all target drones are the same as that of A. This requires synchronous adjustment of the attitude and route planning in the on-board controller or autopilot. For example, after receiving the instruction of "follow F" on the flight control system, B and C can automatically correct their azimuth and heading angles to make them on the same route or parallel routes as F, and there is no obvious yaw or side deviation, so as to maintain a relatively static state with respect to F for a long time; In a preferred embodiment of the present invention, during the process of transmitting the target signal to the ground center through the communication link, the following steps are further included: Regarding the last drone on the communication link as the second drone, after the second drone receives the target signal, it preprocesses the target signal. The preprocessing includes adaptive filtering and frequency pre-compensation. The second drone transmits the preprocessed target signal to the ground center; It should be noted that adaptive filtering focuses on real-time suppression of random noise or interference in the air environment, and can dynamically adjust the filtering parameters according to the collected noise characteristics to reduce the clutter in the signal during transmission; frequency pre-compensation preliminarily corrects the carrier frequency according to the detected offset at the air end, so that the signal arriving at the ground center has the characteristics closer to the ideal frequency point. In this way, most of the interference and offset can be eliminated when the ground center performs refined frequency compensation and demodulation, thereby improving the compensation accuracy and reducing the overall bit error rate of the system; In a preferred case of this embodiment, during the process of transmitting the target signal to the ground center through the communication link, the following steps are further included: When the distance between the second drone and the ground center is equal to 0.95D, if the first drone has not completed the task yet, a new drone is activated to join the drone cluster, and when the distance between the second drone and the ground center is equal to D, the communication link is updated; It can be understood that when the target signal is transmitted step by step along the communication link and reaches the last drone (i.e., the second drone), the second drone will first preprocess the received signal and then send the preprocessed signal to the ground center. Specifically, the second drone can be built with an adaptive filtering and frequency pre-compensation function module. By real-time detecting the noise characteristics of the incoming signal or the offset from the carrier reference frequency, it dynamically adjusts the filter coefficients to maximize the signal-to-noise ratio, and performs appropriate pre-compensation according to the monitored frequency offset amount, thus effectively weakening the influence of negative factors such as environmental noise and residual Doppler frequency shift during communication on the signal integrity. After the preprocessing is completed, the second drone transmits the target signal that has been filtered and frequency corrected to the ground center in digital or analog form. The ground center can also perform further frequency compensation or filtering after finally receiving the signal to ensure the demodulation accuracy and reliability of the data. In addition, during this transmission process, if the distance between the second drone and the ground center approaches 0.95D (where D represents the optimal communication distance), the system will detect whether the first drone has completed its task. If the first drone has not ended, a new drone will be launched from the idle or backup drone pool to join the cluster to maintain the integrity of the communication link in the subsequent process. When the second drone continues to approach the ground center and reaches D, the system will trigger an update operation on the existing communication link, that is, without affecting the ongoing data transmission, the newly added drone will replace or be incorporated into the communication link, enabling the entire link to continue to maintain a high-quality signal channel and relatively stable communication coverage without the first drone completing its core task, thus avoiding data interruption or reliability degradation caused by too short distance or excessive attenuation.

[0021] The above has described a detailed embodiment of the present invention, but the content described is only the preferred embodiment of the present invention and cannot be considered as limiting the scope of implementation of the present invention. All equivalent changes and improvements made according to the scope of the present invention application should still fall within the patent coverage scope of the present invention.

Claims

1. A method for data communication in an unmanned aerial vehicle (UAV) ground-air collaborative networking, characterized in that It includes the following steps: Mark the drones in the drone swarm that are used to execute the preset tasks as the first drones, and denote the signals obtained by converting the data collected by the first drones as target signals; Obtain the connection line between the first drone and the ground center, denoted as the target line, and determine the communication link of the first drone based on the target line, where the communication link is composed of the drones; Keep the drones on the communication link relatively stationary with the first drone, and transmit the target signal to the ground center through the communication link, and the ground center performs frequency compensation on the received target signal.

2. The method for data communication of UAV ground-air collaborative networking according to claim 1, wherein, Determining the communication link of the first drone based on the target line includes: Step 1: Take the first drone as the component drone, draw a circle with the component drone as the center and the preset optimal communication distance D as the radius, and mark the drones inside the circle as candidate drones; Obtain the distance between the projection point of the candidate drone on the target line and the first drone, denoted as the selection distance, and take the candidate drone corresponding to the maximum selection distance as the new component drone A; Step 2: Take the component drone A as the center, repeat Step 1 to obtain new component drones, and repeat the above steps until the distance between a certain new component drone and the ground center is less than the optimal communication distance, and all the candidate drones form the communication link.

3. A method for data communication in an unmanned aerial vehicle (UAV) ground-air collaborative networking, according to claim 1, wherein The process of keeping the drones on the communication link relatively stationary with the first drone includes: Mark the factors that affect the speed of the drone as target factors, collect the target factors of the first drone, input them into the pre-trained speed decay prediction model, output the speed decay value, set the speed of the first drone as the preset value V, and subtract the speed decay value from the preset value V to obtain the predicted speed V0 of the first drone; Mark the drones on the communication link as target drones, obtain the speed decay value V1 of the target drones, and set the speed of the target drones as V1 + V0; Ensure that the target drones and the first drone move in the same direction to achieve the relative static state of the drones on the communication link and the first drone.

4. The method for data communication in the ground-air collaborative networking of an unmanned aerial vehicle according to claim 3, wherein The process of pre-training the speed decay prediction model includes: Obtain a data set, where the data set stores the target factors with the marked speed decay values, and divide the data set into a training set and a validation set according to a preset ratio; Establish a speed decay prediction model based on deep learning, and train and validate the speed decay prediction model based on the training set and the validation set to obtain the pre-trained speed decay prediction model.

5. A method for drone-ground collaborative networking data communication according to claim 1, characterized in that, During the process of transmitting the target signal to the ground center through the communication link, the following steps are further included: Take the last drone on the communication link as the second drone. After the second drone receives the target signal, it preprocesses the target signal, where the preprocessing includes adaptive filtering and frequency pre-compensation, and the second drone transmits the preprocessed target signal to the ground center.

6. A method for data communication in an unmanned aerial vehicle (UAV) ground-air collaborative networking, according to claim 5, characterized in that, In the process of transmitting the target signal to the ground center through the communication link, the following steps are further included: When the distance between the second unmanned aerial vehicle (UAV) and the ground center is equal to 0.95D, if the first UAV has not completed the task yet, a new UAV is activated to join the UAV cluster, and the communication link is updated when the distance between the second UAV and the ground center is equal to D.

7. A method for data communication in an unmanned aerial vehicle (UAV) ground-air collaborative networking, according to claim 2, characterized in that, In the process of determining the communication link of the first UAV based on the target line, the following is further included; When the UAV i is a component of m communication links and m > n, where n is a preset quantity, the following steps are executed: Regarding the m communication links corresponding to the UAV i as pending links, obtaining the distance between the UAV i and the third UAV, denoted as the sorting distance, where the third UAV is the previous UAV adjacent to the UAV i in the pending links; Sorting the pending links in ascending order according to the magnitude of the sorting distance, and the UAV i is a component of the first n pending links in the sorting.

8. A method for data communication in an unmanned aerial vehicle (UAV) ground-air collaborative networking, according to claim 1, characterized in that, After determining all the communication links of the first UAVs, the UAVs that do not belong to any of the communication links return to the ground center and enter the sleep state.

Images