Cloud unmanned aerial vehicle control method, device, equipment and medium

Abstract

The application relates to the technical field of cloud unmanned aerial vehicle remote control, and provides a cloud unmanned aerial vehicle control method, device, equipment and medium, which comprises the following steps: maintaining a direction permission field according to the communication channel health conditions of each direction; dividing a motion direction space into multiple sectors, distributing control residuals to corresponding sectors according to instruction directions, respectively calculating differential entropy of residual probability density of each sector, and obtaining a residual entropy time sequence of each direction; calculating a residual entropy correlation mean value between each other at the same time, calculating a communication selectivity index according to the residual entropy correlation mean value, and activating a direction mixed control according to the communication selectivity index; receiving a stick quantity instruction of an operator and extracting a direction angle, querying a direction permission value corresponding to the direction angle from the direction permission field, and generating a mixed speed instruction; and sending the mixed speed instruction to an airborne flight control to perform flight control. The application improves the safety and reliability of cloud unmanned aerial vehicle remote control.

Description

Technical Field

[0001] This application relates to the field of cloud-based drone remote control technology, and in particular to a cloud-based drone control method, device, equipment and medium. Background Technology

[0002] With the rapid development of the low-altitude economy, drones are increasingly used in logistics delivery, agricultural plant protection, power line inspection, emergency rescue, and other fields, and the requirements for the safety and reliability of remote control are also becoming increasingly stringent. In the current mainstream cloud-based drone remote control architecture, the cloud server interacts with the airborne flight control system through multiple communication links, including an uplink for receiving telemetry information such as the drone's position and speed, a downlink for issuing stick control commands to the operator, and an interface link for sending high-level tasks such as fixed-point flight.

[0003] In existing cloud-based drone flight control technologies, ArduPilot PosHold's hybrid flight mode implements independent state machines for the roll and pitch axes. Each axis maintains its own state, including stick control, brake hold, and hovering. The hybrid ratio is driven by the stick amplitude and state transition timing. When the operator only pushes the stick forward, the pitch axis is in stick control mode while the roll axis is in position hold mode, achieving directional differentiation at the Cartesian coordinate level. The technical goal of this solution is to provide a smooth flight experience with centered stick hovering and push-to-fly.

[0004] However, the above-mentioned cloud-based drone flight control solution faces the following problems in actual deployment: during the operation of the drone, the drone may exhibit unpredictable motion responses, which may pose safety hazards. Summary of the Invention

[0005] To address the aforementioned problems in the prior art, this application provides a cloud-based drone control method, apparatus, device, and medium that eliminates unpredictable motion responses and improves the security and reliability of remote control of cloud-based drones.

[0006] To achieve the above objectives, the technical solution adopted in this application is as follows: In a first aspect, this application provides a cloud-based drone control method, including: The directional authority field is maintained based on the health status of the communication channels in each direction. The directional authority field is a continuous function of the horizontal movement direction, and the value of the directional authority field represents the lever control authority in the corresponding direction. The motion direction space is divided into multiple sectors, and the control residuals are allocated to the corresponding sectors according to the command direction. The differential entropy of the residual probability density of each sector is calculated to obtain the time series of residual entropy in each direction. Calculate the mean residual entropy correlation between each pair of residual entropy time series at the same time. Calculate the communication selectivity index based on the mean residual entropy correlation. Determine whether the communication selectivity index exceeds the activation threshold. If so, activate directional mixing control. The communication selectivity index is inversely related to the mean residual entropy correlation. The communication selectivity index is used to distinguish between directional selective communication degradation and directional uniformity environmental interference. When the directional hybrid control is activated, the operator's lever input command is received and its directional angle is extracted. The directional authority value corresponding to the directional angle is queried from the directional authority field. A hybrid speed command is generated based on the directional authority value and the lever input command. The mixed speed command is sent to the airborne flight controller to execute flight control.

[0007] The beneficial effects of this application are as follows: By maintaining the directional authority field through the health status of communication channels in each direction, the directional authority field dynamically changes with the communication status. Full control authority is retained in directions with healthy communication, while control authority is automatically reduced in directions with degraded communication. This eliminates the unpredictable motion response of the UAV when the operator continues to push the stick at full power in a degraded direction, resolving the safety risk of operators being unable to perceive directional communication degradation. Simultaneously, by calculating the average residual entropy correlation between each pair of residual entropy time series in each direction at the same time, and obtaining a communication selectivity index based on the inverse relationship, this index distinguishes between directional selective communication degradation and environmental interference related to directional uniformity. Directional hybrid control is activated only when communication degrades, avoiding false activation caused by environmental interference. When directional hybrid control is activated, the hybrid speed command is calculated by querying the authority value from the directional authority field according to the stick direction and sent to the flight controller. This allows the UAV to maintain flight stability while eliminating some safety hazards caused by directional communication degradation, improving the safety and reliability of cloud-based UAV remote control.

[0008] Optionally, the step of dividing the motion direction space into multiple sectors, allocating the control residuals to the corresponding sectors according to the command direction, and calculating the differential entropy of the residual probability density of each sector to obtain the time series of residual entropy for each direction includes: Divide the space of motion direction into K equal sectors, where K≥2; The control residual is obtained by subtracting the desired velocity from the observed velocity. This control residual is then allocated to the corresponding sector according to the direction angle of the corresponding lever command. The expression for the direction angle is: θ i =atan2(s y,i ,s x,i ); In the formula, s x,i s y,i For the horizontal and vertical components of the i-th lever command; Calculate the differential entropy E of the residual probability density independently for each sector. k : E k =H(R k ) / H max ; In the formula, R k Let H(R) be the set of control residual samples for the k-th sector. k H is the differential entropy that controls the residual sample set. max It is the maximum entropy; The differential entropy calculated for each sector in each control cycle is accumulated over time and smoothed by an exponentially weighted moving average to obtain the time series of residual entropy in each direction.

[0009] Optionally, the calculation of the mean residual entropy correlation between each pair of residual entropy time series in each direction at the same time includes: For each pair of directions formed in the residual entropy time series of each direction, the Pearson correlation coefficient of all directions is calculated within a sliding window and the average value is taken to obtain the mean value of residual entropy correlation. The sliding window is a time window with a fixed period sampling. The residual entropy time series of each sector is updated uniformly on the time grid of the same fixed control period.

[0010] Optionally, the communication selectivity index (CSI) is 1 minus the mean of the residual entropy correlation, and the activation threshold ranges from 0.3 to 0.7.

[0011] Optionally, determining whether the communication selectivity index exceeds the activation threshold, and if so, then activating the direction mixing control, includes: Set activation threshold CSI on and exit threshold CSI off Among them, CSI off <CSI on ; When directional mixing control is not activated, determine whether the communication selectivity index exceeds the activation threshold CSI. on If so, then activate direction blending control; When directional mixing control is activated, it is determined whether the communication selectivity index is lower than the exit threshold CSI. off If yes, then exit directional mixing control; otherwise, maintain the current active state.

[0012] Optionally, generating a hybrid speed command based on the direction permission value and the lever input command includes: Calculate the position-maintaining velocity command v based on the shadow target's position. ph : v ph =K p (p s -p c )+K d (-v c ); In the formula, K p For position gain, p s To track the shadow target position of the drone's current actual position, p c K represents the current actual location of the drone. d For damping gain, v c This is the drone's current speed; A hybrid speed command v is generated based on the direction permission value, the lever input command, and the position hold speed command. cmd : v cmd =w K v s op +(1 w) v ph ; In the formula, w is the direction permission value corresponding to the direction angle retrieved from the direction permission field, and K v The gain relationship between lever quantity and velocity, s op The lever vector for the operator consists of the horizontal and vertical components of the lever command.

[0013] Optionally, it also includes: For the k-th sector, determine whether the time elapsed since the lever command was not updated within the sliding window exceeds the preset timeout threshold T. s If so, then the exponentially weighted moving average residual entropy of that sector is... Decreasing exponentially toward the default entropy value E d convergence: ; In the formula, t represents the current time. last,k E represents the last valid update timestamp of the k-th sector. d The default entropy value is given by λ, where λ is the decay rate constant. When the residual entropy decays to the default entropy value, the permission weight obtained by the sigmoid function tends to 1.

[0014] Secondly, this application provides a cloud-based drone control device, comprising: The permission maintenance module is used to maintain the directional permission field according to the health status of the communication channel in each direction. The directional permission field is a continuous function of the horizontal movement direction, and the value of the directional permission field represents the lever control permission in the corresponding direction. The residual entropy calculation module is used to divide the motion direction space into multiple sectors, allocate the control residuals to the corresponding sectors according to the command direction, calculate the differential entropy of the residual probability density of each sector, and obtain the time series of residual entropy in each direction. The hybrid activation module is used to calculate the mean residual entropy correlation between each pair of residual entropy time series in each direction at the same time. Based on the mean residual entropy correlation, a communication selectivity index is calculated. It is then determined whether the communication selectivity index exceeds the activation threshold. If so, directional hybrid control is activated. The communication selectivity index is inversely related to the mean residual entropy correlation. The communication selectivity index is used to distinguish between directional selective communication degradation and directional uniformity environmental interference. The instruction generation module is used to receive the operator's lever input instruction and extract its direction angle when the direction hybrid control is activated, query the direction authority value corresponding to the direction angle from the direction authority field, and generate a hybrid speed instruction based on the direction authority value and the lever input instruction. The command sending module is used to send the mixed speed command to the airborne flight controller to execute flight control.

[0015] Thirdly, this application provides a computer device, including: one or more processors, and a memory; The memory stores computer-readable instructions that, when executed by the one or more processors, perform the steps of the method as described in the first aspect.

[0016] Fourthly, this application provides a storage medium storing computer-readable instructions that, when executed by one or more processors, cause the one or more processors to perform the steps of the method described in the first aspect.

[0017] The technical effects of the cloud-based drone control device provided in the second aspect, the computer equipment provided in the third aspect, and the storage medium provided in the fourth aspect are described in the relevant description of the cloud-based drone control method provided in the first aspect. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the main process of the cloud-based drone control method according to an embodiment of this application; Figure 2 This is a schematic diagram of the directional authority field involved in an embodiment of this application; Figure 3 This is a schematic diagram of the structure of the cloud-based drone control device according to an embodiment of this application; Figure 4 This is a schematic diagram of the structure of a computer device according to an embodiment of this application. Detailed Implementation

[0019] To better understand the above technical solutions, exemplary embodiments of this application will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of this application are shown in the drawings, it should be understood that this application can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this application can be understood more clearly and thoroughly, and that the scope of this application can be fully conveyed to those skilled in the art.

[0020] The embodiments of this application are applied to cloud-based drone remote control scenarios, and are particularly suitable for cloud-based drone remote control scenarios where communication links degrade and exhibit directional selectivity. In the prior art, during the operation of a drone by an operator, the drone may exhibit unpredictable motion responses, posing safety hazards.

[0021] Therefore, in various embodiments of this application, a directional authority field is maintained based on the health status of the communication channels in each direction. The directional authority field is a continuous function of the horizontal motion direction, and the value of the directional authority field represents the stick control authority in the corresponding direction. The motion direction space is divided into multiple sectors, and the control residuals are allocated to the corresponding sectors according to the command direction. The differential entropy of the residual probability density of each sector is calculated to obtain the time series of residual entropy in each direction. The mean value of the residual entropy correlation between each pair of residual entropy time series at the same time is calculated. The communication selectivity index is calculated based on the mean value of the residual entropy correlation. It is determined whether the communication selectivity index exceeds the activation threshold. If so, the directional hybrid control is activated. The communication selectivity index and the mean value of the residual entropy correlation are inversely related. The communication selectivity index is used to distinguish between communication degradation of directional selectivity and environmental interference of directional uniformity. When the directional hybrid control is activated, the stick command of the operator is received and its direction angle is extracted. The directional authority value corresponding to the direction angle is queried from the directional authority field. A hybrid speed command is generated based on the direction authority value and the stick command. The hybrid speed command is sent to the airborne flight control to execute flight control. Therefore, this application improves the security and reliability of cloud-based drone remote control.

[0022] The present application will now be described in further detail with reference to the accompanying drawings and embodiments.

[0023] This application provides a cloud-based drone control method, such as... Figure 1 As shown, the method includes the following steps: Step S1: Maintain the direction permission field based on the health status of the communication channels in each direction. The direction permission field is a continuous function of the horizontal movement direction, and the value of the direction permission field represents the lever control permission in the corresponding direction.

[0024] like Figure 2As shown, the corresponding values ​​are marked in the angular direction. Meanwhile, the direction permission field is a data structure maintained in the cloud server to store the mapping relationship between horizontal movement directions and lever control permissions. w drc (θ,t)∈[0,1]; In the formula, w drc (θ,t) is the directional control field, whose value varies continuously between 0 and 1. When w drc (θ,t) being 1 indicates that the communication channel is healthy in this direction, and the operator's lever control is the primary control mode for this direction; when w drc (θ,t) being 0 indicates that the communication channel in that direction is degraded, and position holding is the primary control mode for that direction. Θ is the horizontal motion direction angle, ranging from [0, 2π), where 0 represents true north and π / 2 represents true east. T is the current time.

[0025] The value of the direction permission field is determined by the health status of the communication channels in each direction. The cloud server evaluates the health status of the communication channels in each direction by analyzing the change law of the control residual with the direction, and updates the direction permission field accordingly.

[0026] Step S2: Divide the motion direction space into multiple sectors, allocate the control residuals to the corresponding sectors according to the command direction, calculate the differential entropy of the residual probability density of each sector, and obtain the time series of residual entropy for each direction.

[0027] In this embodiment, step S2 specifically includes: Step S21: Divide the motion direction space into K sectors, where K ≥ 2.

[0028] Each sector covers an angle range of 2π / K. The value of the number of sectors K determines the balance between directional resolution and computational overhead. In one example, K is set to 8, which means that the 360° directional space is divided into 8 equal sectors, each with an angle of 45°.

[0029] Step S22: Subtract the desired speed from the observed speed to obtain the control residual. Distribute the control residual to the corresponding sector according to the direction angle of the corresponding lever command. The expression for the direction angle is: θ i =atan2(s y,i ,s x,i ); In the formula, s x,i s y,i Let be the horizontal and vertical components of the i-th lever command.

[0030] Within each control cycle, the cloud server receives lever movement commands from the operator, which include a horizontal component s. xand vertical component s y The cloud server simultaneously acquires the UAV's observation speed v via the OSD (On-Screen Display, airborne telemetry data overlay system) telemetry uplink channel. obs Based on the lever input command and the preset lever input and speed gain K v Calculate the desired velocity v exp =K v ·s op Control residual r=v obs -v exp This represents the deviation between the actual speed and the desired speed of the UAV. This deviation reflects the effectiveness of the control commands and the transmission quality of the communication channel. The control residual for each control cycle is allocated to the corresponding sector according to the direction angle of the corresponding lever command.

[0031] OSD is a general term for airborne telemetry data downlink, and in this embodiment, it refers to the UAV reporting its position, speed, attitude and other status information to the cloud server through the uplink.

[0032] Step S23: Calculate the differential entropy E of the residual probability density independently for each sector. k : E k =H(R k ) / H max ; In the formula, R k Let H(R) be the set of control residual samples for the k-th sector. k H is the differential entropy that controls the residual sample set. max It represents the maximum entropy.

[0033] Each sector independently maintains a set of control residual samples R within a sliding window. k For each sector, calculate the probability density distribution of its residual sample set and obtain the differential entropy H(R). k The larger the value of the differential entropy, the more uncertain and unpredictable the control deviation in that direction, indirectly indicating a worse communication link quality in that direction. To make a fair comparison between different sectors, the differential entropy is divided by the maximum entropy H. max Perform normalization to make the normalized entropy value E k The value range is [0,1]. After smoothing the differential entropy of each sector using an exponentially weighted moving average (EWMA), it also needs to undergo degradation rate regression processing.

[0034] Specifically, degradation rate regression calculates the trend of entropy changes by performing a first-order linear regression on the smoothed entropy time series, capturing the slow degradation or recovery trend of channel quality. The smoothed entropy and its degradation rate are used as input and mapped to sector permission weights between 0 and 1 using the sigmoid function. The sigmoid function is a standard sigmoid function that smoothly maps the input entropy and degradation rate to permission weights. Sectors with low entropy and weak degradation trends receive permission weights close to 1, while sectors with high entropy and strong degradation trends receive permission weights close to 0. In each control cycle, after the permission weights of each sector are determined, linear interpolation is performed on the weights of adjacent sectors, making the directional permission field continuously available at any directional angle, ultimately resulting in a continuous directional permission field that smoothly changes with the directional angle. The interpolation operation is performed in each control cycle as part of the permission field update, ensuring that the directional permission field is continuously maintained and updated over time.

[0035] Step S24: Accumulate the differential entropy calculated for each sector in each control cycle over time, and smooth it by exponential weighted moving average to obtain the time series of residual entropy in each direction.

[0036] The residual entropy time series of each sector is updated on the same fixed control period time grid. At the end of each control period, the system performs an exponentially weighted moving average update on all K sectors, that is, it calculates a new smoothed value using the entropy value calculated in the current period and the smoothed value in the previous period. If no new residual samples are assigned to a sector in the current period, the exponentially weighted moving average of that sector remains unchanged from the previous period.

[0037] Step S3: Calculate the mean residual entropy correlation between each pair of residual entropy time series at the same time. Calculate the communication selectivity index based on the mean residual entropy correlation. Determine whether the communication selectivity index exceeds the activation threshold. If so, activate directional hybrid control. The communication selectivity index and the mean residual entropy correlation are inversely related. The communication selectivity index is used to distinguish between directional selective communication degradation and directional uniformity environmental interference.

[0038] The mean residual entropy correlation is used to measure the temporal synchronicity of residual entropy changes in each direction. When environmental disturbances such as wind and turbulence occur, the residual entropy in all directions rises synchronously at the same time, and the mean residual entropy correlation is high. When directional radio frequency interference only affects a specific direction, only the residual entropy in the affected direction rises, while other directions remain low, and the mean residual entropy correlation is low.

[0039] Among them, the communication selectivity index and the mean residual entropy correlation value are inversely related. The higher the mean residual entropy correlation value, the more synchronous the changes in each direction are, and the more inclined it is to environmental interference; the lower the mean residual entropy correlation value, the more independent the changes in each direction are, and the more inclined it is to communication degradation. When the communication selectivity index exceeds the activation threshold, it is judged as directional selective communication degradation, and directional mixing control is activated; when the communication selectivity index does not exceed the activation threshold, it is judged as directional uniformity environmental interference, and directional mixing control is not activated.

[0040] In this embodiment, step S3 calculates the mean residual entropy correlation between each pair of residual entropy time series in each direction at the same time, including: For each pair of directions formed in the residual entropy time series, the Pearson correlation coefficient of all directions is calculated within a sliding window and the average value is taken to obtain the mean value of residual entropy correlation. The sliding window is a time window with a fixed period of sampling, and the residual entropy time series of each sector is updated uniformly on the time grid of the same fixed control period.

[0041] In this embodiment, the mean residual entropy correlation is calculated based on the Pearson correlation coefficient between all pairs of sectors. Specifically, within a sliding window containing 40 sampling points, the zero-lag Pearson correlation coefficient is calculated for the residual entropy time series of any two different sectors out of K sectors. Then, the correlation coefficients of all directional pairs are averaged. Since the residual entropy time series of all sectors are updated uniformly on the same fixed control period time grid, each time series has the same timestamp, and the correlation coefficient is calculated directly on the aligned sampling points without interpolation. The closer the mean residual entropy correlation is to 1, the more synchronous the changes in residual entropy in each direction; the closer the mean residual entropy correlation is to 0, the more independent the changes in each direction.

[0042] Specifically, the mean residual entropy correlation is calculated using the following formula: ; In the formula, ρ cross E represents the mean of the residual entropy correlation. i (t), E j (t) represents the residual entropy time series of the i-th and j-th sectors, respectively; corr(·,·) is the Pearson correlation coefficient.

[0043] The Pearson correlation coefficient reflects the degree of linear correlation between two variables and is independent of the absolute value of the variables. Therefore, even if the input variables are non-negative, the correlation coefficient may still be negative. In this embodiment, the residual entropy correlation mean ρ crossThe value range is [-1, 1]. The Communication Selectivity Index (CSI) is 1 minus the mean of the residual entropy correlation, so its value range is [0, 2]. The activation threshold ranges from 0.3 to 0.7, for example, 0.5.

[0044] In this embodiment, step S3 determines whether the communication selectivity index exceeds the activation threshold. If so, the activation direction mixing control includes: Step S31: Set the activation threshold CSI on and exit threshold CSI off Among them, CSI off <CSI on .

[0045] Step S32: When directional mixing control is not activated, determine whether the communication selectivity index exceeds the activation threshold CSI. on If so, then activate the direction mixing control.

[0046] Step S33: When directional mixing control is activated, determine whether the communication selectivity index is lower than the exit threshold CSI. off If yes, then exit directional mixing control; otherwise, maintain the current active state.

[0047] To prevent frequent fluctuations in the Communication Selectivity Index (CSI) around a threshold, which could lead to repeated activation and deactivation of directional mixing control, the system employs a dual-threshold hysteresis range. The activation threshold is the CSI value. on Default is 0.6, exit threshold CSI off The default value is 0.45. Activation of directional hybrid control requires the Communication Selectivity Index (CSI) to reach the upper limit of the activation threshold. on Exiting the communication selectivity index (CSI) requires it to reach the lower limit threshold. off When CSI is in CSI off With CSI on During this period, the current active state is maintained, effectively suppressing frequent jitter near the threshold.

[0048] Step S4: When the direction mixing control is activated, receive the operator's lever input command and extract its direction angle, query the direction authority value corresponding to the direction angle from the direction authority field, and generate a mixed speed command based on the direction authority value and the lever input command.

[0049] When directional hybrid control is activated, it indicates the presence of directional selective communication degradation, requiring the system to apply different control modes to different directions. For directions with healthy communication, full lever speed control is maintained; for directions with communication degradation, lever speed control authority is reduced and position hold control weight is increased.

[0050] In this embodiment, step S4, which generates a hybrid speed command based on the direction permission value and the lever input command, includes: Step S41: Calculate the position-maintaining velocity command v based on the shadow target position. ph : v ph =K p (p s -p c )+K d (-v c ); In the formula, K p For position gain, p s To track the shadow target position of the drone's current actual position, p c K represents the current actual location of the drone. d For damping gain, v c This represents the drone's current speed.

[0051] The shadow target position is a virtual target point that is dynamically updated based on the stick position status. It is used to calculate the position-holding velocity command when there is no stick input command or when directional hybrid control is activated. When directional hybrid control is activated, the shadow target position directly tracks the current actual position of the UAV, i.e., p. s =p c The position-holding speed command uses a proportional-derivative control structure: the proportional term K... p (p s -p c This generates a restoring force that pulls the drone toward the shadow target's location; the damping term K... d (-v c This generates a damping force in the opposite direction to the current velocity, and the two work together to make the drone hover smoothly at the position of the shadow target.

[0052] Step S42: Generate a mixed speed command v based on the direction permission value, lever input command, and position holding speed command. cmd : v cmd =w K v s op +(1 w) v ph ; In the formula, w is the direction permission value corresponding to the direction angle retrieved from the direction permission field, and K v The gain relationship between lever quantity and velocity, s opThe lever vector for the operator consists of the horizontal and vertical components of the lever command.

[0053] The hybrid speed command integrates the operator's stick movement rate control commands and position-holding speed commands, weighted by a directional authority value w. When w is close to 1, the hybrid speed command prioritizes stick movement rate control, and the drone actively responds to the operator's intentions in that direction. When w is close to 0, the hybrid speed command prioritizes position holding, and the drone automatically hovers in that direction to avoid drifting due to command loss. v The gain relationship between lever quantity and velocity is used to transform the dimensionless lever quantity vector s. op The expected velocity converted to velocity units.

[0054] Step S5: Send the mixed speed command to the airborne flight controller to execute flight control.

[0055] Among them, the cloud server will calculate the hybrid speed command v cmd The mixed speed command is transmitted to the airborne flight control system via the DRC (Direct Remote Control) downlink channel. After receiving the mixed speed command, the airborne flight control system drives the UAV's actuators to perform the corresponding flight maneuvers according to the command.

[0056] The aforementioned DRC is a downlink control link used by the cloud server to send stick control commands to the airborne flight controller. In this application, it refers to the channel through which the cloud server sends stick speed commands to the UAV via the MQTT (Message Queuing Telemetry Transport) protocol.

[0057] In one embodiment, when a sector receives no new lever command data allocation for an extended period within a sliding window, the exponentially weighted moving average residual entropy of that sector will remain unchanged, potentially leading to a permanent lockout of permissions in that direction. To address this issue, a cloud-based drone control method, based on the above embodiments, further includes: For the k-th sector, determine whether the time elapsed since the lever command was not updated within the sliding window exceeds the preset timeout threshold T. s If so, then the exponentially weighted moving average residual entropy of that sector is... Decreasing exponentially toward the default entropy value E d convergence: ; In the formula, t represents the current time. last,k E represents the last valid update timestamp of the k-th sector. d The default entropy value is given by λ, where λ is the decay rate constant. When the residual entropy decays to the default entropy value, the permission weight obtained by the sigmoid function tends to 1.

[0058] That is, in this embodiment, the last valid update timestamp t is maintained for each sector. last,k When tt last,k Exceeding the silent timeout threshold T s At that time, the exponentially weighted moving average residual entropy of the sector decays exponentially toward the default entropy value E. d Convergence. The default entropy value is E. d For a low entropy value, the silent timeout threshold T s It is equal to twice the length of the sliding window, that is, according to the example above, 80 sampling periods multiplied by 200 milliseconds equals 16 seconds; λ is the decay rate constant, with a default value of 0.1 per second. After the residual entropy decays to the default entropy value, the permission weight obtained by the sigmoid function tends to 1, that is, the full lever permission is restored.

[0059] Therefore, the absence of instruction data means that the operator has no operational need in that direction, making it impossible to determine channel degradation. Full privileges should be conservatively maintained. When the operator pushes the lever in that direction again, the system responds with full privileges, only reducing privileges after collecting sufficient new control residuals. This design avoids the privilege freeze problem that can occur from maintaining historical values; that is, a single, occasional channel degradation event will not permanently lock the sector's privileges at a low value.

[0060] In summary, the various embodiments of this application have the following technical effects: First, by maintaining the directional authority field based on the health status of the communication channels in each direction, the directional authority field dynamically changes with the communication status. In directions with healthy communication, the full lever control authority is retained, while in directions with degraded communication, the lever control authority is automatically reduced. This eliminates the unpredictable motion response of the UAV when the operator continues to push the lever at full power in a degraded direction, and solves the safety risk that the operator cannot perceive directional communication degradation.

[0061] Second, by calculating the mean residual entropy correlation between each pair of residual entropy time series in each direction at the same time, and obtaining the communication selectivity index based on the inverse relationship, this index is used to distinguish between communication degradation of directional selectivity and environmental interference of directional uniformity. Directional hybrid control is activated only when communication is degraded, avoiding false activation caused by environmental interference. This allows the UAV to maintain full lever control in normal environments such as wind disturbances, and prevents unreasonable attenuation of all-directional authority.

[0062] Third, when directional hybrid control is activated, the hybrid speed command is calculated by querying the permission value from the directional permission field according to the stick direction and sent to the flight controller. This enables the UAV to adaptively adjust the response characteristics of each direction according to the directional health status of the communication channel, thereby eliminating the safety hazards caused by the degradation of some directional communication while maintaining flight stability.

[0063] Therefore, this application achieves safe control of cloud-based drones in some directional communication degradation scenarios by mapping the directional health status of the communication channel to directional control permissions and combining directional degradation synchronization criteria to distinguish between communication degradation and environmental interference, thereby improving the security and reliability of remote control of cloud-based drones.

[0064] In one embodiment, such as Figure 3 As shown, this application also provides a cloud-based drone control device, including: The permission maintenance module 301 is used to maintain the directional permission field according to the health status of the communication channel in each direction. The directional permission field is a continuous function of the horizontal movement direction, and the value of the directional permission field represents the lever control permission in the corresponding direction. The residual entropy calculation module 302 is used to divide the motion direction space into multiple sectors, allocate the control residual to the corresponding sector according to the command direction, calculate the differential entropy of the residual probability density of each sector, and obtain the time series of residual entropy in each direction. The hybrid activation module 303 is used to calculate the mean residual entropy correlation between each pair of residual entropy time series at the same time. Based on the mean residual entropy correlation, the communication selectivity index is calculated. It is determined whether the communication selectivity index exceeds the activation threshold. If so, the hybrid control of the direction is activated. The communication selectivity index is inversely related to the mean residual entropy correlation. The communication selectivity index is used to distinguish between communication degradation of direction selectivity and environmental interference of direction uniformity. The instruction generation module 304 is used to receive the operator's lever input instruction and extract its direction angle when the direction hybrid control is activated, query the direction authority value corresponding to the direction angle from the direction authority field, and generate a hybrid speed instruction based on the direction authority value and the lever input instruction. The command sending module 305 is used to send mixed speed commands to the airborne flight controller to execute flight control.

[0065] In one embodiment, such as Figure 4 As shown, this application also provides a computer device 400, comprising: Communication interface 401 allows for information exchange with other devices or network nodes.

[0066] One or more processors 402 are connected to a communication interface 401 to enable information interaction with other devices or network nodes, and to execute the methods provided by one or more technical solutions in the above embodiments when running computer programs.

[0067] Memory 403 is used to store computer-readable instructions that can be executed on processor 402. When the computer-readable instructions are executed by one or more processors 402, the steps of the cloud-based drone control method as described in the above embodiments are performed.

[0068] The computer device 400 in this embodiment only shows a block diagram of a portion of the structure related to the present application solution, and does not constitute a limitation on the computer device to which the present application solution is applied. A specific computer device may include more or fewer components than shown in the figures, or combine certain components, or have different component arrangements, such as power supply, input / output interfaces, etc. Furthermore, the computer device 400 of this embodiment can operate an operating system, such as Windows Server, stored in memory 403. TM Mac OS X TM Unix TM Linux TM FreeBSD TM Or similar.

[0069] In one embodiment, this application also provides a storage medium storing computer-readable instructions, which, when executed by one or more processors, cause the one or more processors to perform the steps of the cloud-based drone control method as described in the above embodiments.

[0070] The computer-readable instructions in the aforementioned computer device 400 and storage medium may be application programs.

[0071] In addition, the specific descriptions of the technical effects and steps corresponding to the cloud-based drone control device, a computer device 400, and a storage medium in the above embodiments are all based on the relevant descriptions in the embodiments of the cloud-based drone control method.

[0072] Since the systems / devices described in the above embodiments of this application are systems / devices used to implement the methods of the above embodiments of this application, those skilled in the art can understand the specific structure and modifications of the system / devices based on the methods described in the above embodiments of this application, and therefore will not be repeated here. All systems / devices used in the methods of the above embodiments of this application fall within the scope of protection of this application.

[0073] Those skilled in the art will understand that embodiments of this application can be provided as methods, apparatus, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0074] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (devices), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions.

[0075] It should be noted that any reference numerals within parentheses in this application should not be construed as limiting the technical solution. The word "comprising" does not exclude the presence of components or steps not listed in the technical solution. The word "a" or "an" preceding a component does not exclude the presence of multiple such components. This application can be implemented by means of hardware comprising several different components and by means of a suitably programmed computer. In technical solutions that list several devices, several of these devices may be embodied by the same hardware. The use of the terms first, second, third, etc., is merely for convenience and does not indicate any order. These terms can be understood as part of the component names.

[0076] Furthermore, it should be noted that in the description of this specification, the terms "one embodiment," "some embodiments," "embodiment," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. Furthermore, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0077] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the scope of protection of this application should be interpreted to include the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0078] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if these modifications and variations fall within the scope of protection of this application and its equivalents, this application should also include these modifications and variations.

Claims

1. A cloud-based unmanned aerial vehicle (UAV) control method, characterized in that, include: The directional authority field is maintained based on the health status of the communication channels in each direction. The directional authority field is a continuous function of the horizontal movement direction, and the value of the directional authority field represents the lever control authority in the corresponding direction. The motion direction space is divided into multiple sectors, and the control residuals are allocated to the corresponding sectors according to the command direction. The differential entropy of the residual probability density of each sector is calculated to obtain the time series of residual entropy in each direction. Calculate the mean residual entropy correlation between each pair of residual entropy time series at the same time. Calculate the communication selectivity index based on the mean residual entropy correlation. Determine whether the communication selectivity index exceeds the activation threshold. If so, activate directional mixing control. The communication selectivity index is inversely related to the mean residual entropy correlation. The communication selectivity index is used to distinguish between directional selective communication degradation and directional uniformity environmental interference. When the directional hybrid control is activated, the operator's lever input command is received and its directional angle is extracted. The directional authority value corresponding to the directional angle is queried from the directional authority field. A hybrid speed command is generated based on the directional authority value and the lever input command. The mixed speed command is sent to the airborne flight controller to execute flight control.

2. The method according to claim 1, characterized in that, The process involves dividing the motion direction space into multiple sectors, allocating the control residuals to the corresponding sectors according to the command direction, calculating the differential entropy of the residual probability density for each sector, and obtaining the time series of residual entropy for each direction, including: Divide the space of motion direction into K equal sectors, where K≥2; The control residual is obtained by subtracting the desired velocity from the observed velocity. This control residual is then allocated to the corresponding sector according to the direction angle of the corresponding lever command. The expression for the direction angle is: i i =atan2(s y,i ,s x,i ); In the formula, s x,i s y,i For the horizontal and vertical components of the i-th lever command; Calculate the differential entropy E of the residual probability density independently for each sector. k : E k =H(R k ) / H max ; In the formula, R k Let H(R) be the set of control residual samples for the k-th sector. k H is the differential entropy that controls the residual sample set. max It is the maximum entropy; The differential entropy calculated for each sector in each control cycle is accumulated over time and smoothed by an exponentially weighted moving average to obtain the time series of residual entropy in each direction.

3. The method according to claim 1, characterized in that, The calculation of the mean residual entropy correlation between each pair of residual entropy time series in each direction at the same time includes: For each pair of directions formed in the residual entropy time series of each direction, the Pearson correlation coefficient of all directions is calculated within a sliding window and the average value is taken to obtain the mean value of residual entropy correlation. The sliding window is a time window with a fixed period sampling. The residual entropy time series of each sector is updated uniformly on the time grid of the same fixed control period.

4. The method according to claim 3, characterized in that, The Communication Selectivity Index (CSI) is 1 minus the mean of the residual entropy correlation, and the activation threshold ranges from 0.3 to 0.

7.

5. The method according to claim 4, characterized in that, The step of determining whether the communication selectivity index exceeds the activation threshold, and if so, the activation direction mixing control, includes: Set activation threshold CSI on and exit threshold CSI off Among them, CSI off <CSI on ; When directional mixing control is not activated, determine whether the communication selectivity index exceeds the activation threshold CSI. on If so, then activate direction blending control; When directional mixing control is activated, it is determined whether the communication selectivity index is lower than the exit threshold CSI. off If yes, then exit directional mixing control; otherwise, maintain the current active state.

6. The method according to any one of claims 1 to 4, characterized in that, The step of generating a hybrid speed command based on the direction permission value and the lever input command includes: Calculate the position-maintaining velocity command v based on the shadow target's position. ph : v ph =K p (p s -p c )+K d (-v c ); In the formula, K p For position gain, p s To track the shadow target position of the drone's current actual position, p c K represents the current actual location of the drone. d For damping gain, v c This is the drone's current speed; A hybrid speed command v is generated based on the direction permission value, the lever input command, and the position hold speed command. cmd : v cmd =w K v s op +(1 w) v ph ; In the formula, w is the direction permission value corresponding to the direction angle retrieved from the direction permission field, and K v The gain relationship between lever quantity and velocity, s op The lever vector for the operator consists of the horizontal and vertical components of the lever command.

7. The method according to any one of claims 1 to 4, characterized in that, Also includes: For the k-th sector, determine whether the time elapsed since the lever command was not updated within the sliding window exceeds the preset timeout threshold T. s If so, then the exponentially weighted moving average residual entropy of that sector is... Decreasing exponentially toward the default entropy value E d convergence: ； In the formula, t represents the current time. last,k E represents the last valid update timestamp of the k-th sector. d The default entropy value is given by λ, where λ is the decay rate constant. When the residual entropy decays to the default entropy value, the permission weight obtained by the sigmoid function tends to 1.

8. A cloud-based unmanned aerial vehicle (UAV) control device, characterized in that, include: The permission maintenance module is used to maintain the directional permission field according to the health status of the communication channel in each direction. The directional permission field is a continuous function of the horizontal movement direction, and the value of the directional permission field represents the lever control permission in the corresponding direction. The residual entropy calculation module is used to divide the motion direction space into multiple sectors, allocate the control residuals to the corresponding sectors according to the command direction, calculate the differential entropy of the residual probability density of each sector, and obtain the time series of residual entropy in each direction. The hybrid activation module is used to calculate the mean residual entropy correlation between each pair of residual entropy time series in each direction at the same time, calculate the communication selectivity index based on the mean residual entropy correlation, and determine whether the communication selectivity index exceeds the activation threshold. If so, the hybrid directional control is activated. The communication selectivity index is inversely related to the mean residual entropy correlation. The communication selectivity index is used to distinguish between directional selectivity communication degradation and directional uniformity environmental interference. The instruction generation module is used to receive the operator's lever input instruction and extract its direction angle when the direction hybrid control is activated, query the direction authority value corresponding to the direction angle from the direction authority field, and generate a hybrid speed instruction based on the direction authority value and the lever input instruction. The command sending module is used to send the mixed speed command to the airborne flight controller to execute flight control.

9. A computer device, characterized in that, include: One or more processors, and memory; The memory stores computer-readable instructions that, when executed by the one or more processors, perform the steps of the method as described in any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium stores computer-readable instructions that, when executed by one or more processors, cause the one or more processors to perform the steps of the method as described in any one of claims 1 to 7.

Images