An unmanned aerial vehicle frequency point determination method and system based on ground cooperative calibration

Abstract

The application discloses a kind of based on ground cooperative calibration's unmanned plane frequency point determination method and system, method includes the channel quality parameter of current communication link of unmanned plane, current frequency band information and the real-time signal strength of each candidate frequency band;According to the model identification of unmanned plane and current frequency band information, determine the mapping relationship of signal power and signal indication intensity, combine the real-time signal strength of each candidate frequency band and interference level estimate value, give the effective signal-to-noise ratio of each candidate frequency band and predict the throughput prediction value of each candidate frequency band;Determine target frequency point and generate frequency point switching information;Through judging the channel quality parameter of current communication link, combine frequency point switching information, complete the switching of unmanned plane frequency point, solve the problem that the signal strength returned by consumer-level unmanned plane SDK is not accurate, spectrum sensing interface is not opened, backhaul bandwidth is limited, airborne computing load is high, realize low-cost, easy deployment accurate frequency point switching on consumer-level unmanned plane platform.

Description

Technical Field

[0001] This invention belongs to the technical field of unmanned aerial vehicle (UAV) communication, specifically relating to a method and system for determining UAV frequency points based on ground-based collaborative calibration. Background Technology

[0002] With the popularization of drone technology, consumer drones (such as DJI Mavic, Mini, and Air series) are increasingly used in fields such as power line inspection, agricultural plant protection, security monitoring, and aerial surveying. In these application scenarios, the stability of the communication link between the drone and the ground station directly affects operational safety and data transmission quality. In particular, when drones are used in complex electromagnetic environments such as power line inspection, the electromagnetic radiation generated by equipment such as transformers, circuit breakers, and reactors in substations can cause serious interference to the communication link, leading to image transmission lag, remote control delays, or even link interruptions.

[0003] Consumer drones typically obtain communication link status information through manufacturer-provided Mobile SDKs (such as DJI's MSDK) and develop various applications based on this information. However, practical engineering applications have revealed some technical limitations in consumer drone SDKs, severely restricting the actual application effectiveness of frequency switching technology. For example, the RSSI value returned by the SDK is a processed value, resulting in low accuracy of the returned signal strength; because consumer drone SDKs do not provide a direct RF front-end sampling interface, they cannot directly measure the interference level of a specified frequency, thus affecting the accuracy of predicting the link quality after switching; and the frequency switching process may cause brief communication interruptions, thereby affecting flight safety.

[0004] Patent application CN116390100A discloses a method and device for interference perception and automatic avoidance of UAV-borne data links. It determines the corresponding interference avoidance method by combining preset multi-dimensional information with the spectrum information of the detection area corresponding to the UAV, and generates corresponding parameter information. The control terminal generates corresponding control commands to enable the UAV to avoid corresponding interference situations. This achieves real-time spectrum perception in complex clutter environments, interference identification technology based on feature parameter extraction, and an intelligent autonomous decision-making engine based on spectrum availability situation maps, enabling the UAV to quickly avoid interference attacks. Patent CN113301654A discloses a UAV communication frequency band allocation system and method based on a semi-supervised learning algorithm. The ground control center performs a spectrum scan on the pre-used frequency band and determines whether interference exists on that band. If interference exists, the system jumps to the next frequency band to continue spectrum scanning. Then, when the drone is working normally, the drone's own onboard control terminal is used to cyclically detect the current drone communication frequency band information and working status, and predict the idle status of other frequency bands. When the current communication frequency band is interfered with, the frequency band can be quickly switched to avoid unnecessary interference to the normal telemetry information and data transmission of the drone, thereby improving the drone's working efficiency and anti-interference ability.

[0005] Including the aforementioned existing technologies, current drone communication technologies do not address issues such as inaccurate signal strength returned by consumer-grade drone SDKs, data acquisition of interference levels, and the impact of interference levels on actual throughput. The challenge lies in improving the accuracy of SDK-returned signal strength, rationally utilizing it during drone frequency switching, considering the impact of interference levels on actual throughput, reducing the computational load on the drone, and achieving high-quality drone frequency switching. Summary of the Invention

[0006] To address the shortcomings of the existing technologies, this invention provides a method and system for determining UAV frequency points based on ground-based collaborative calibration. The method includes: a ground control terminal acquiring channel quality parameters of the current communication link of the UAV, and an airborne terminal acquiring current frequency band information and real-time signal strength of each candidate frequency band; the ground control terminal determining the mapping relationship between signal power and signal indication strength based on the UAV's model identification and current frequency band information, and determining the real-time estimated signal power of each candidate frequency band by combining the real-time signal strength of each candidate frequency band; based on the real-time estimated signal power of each candidate frequency band, providing the effective signal-to-noise ratio of each candidate frequency band, and predicting the throughput of each candidate frequency band; selecting a target frequency point based on the throughput prediction value and generating frequency point switching information; and completing the UAV frequency point switching by judging the channel quality parameters of the current communication link and combining the frequency point switching information.

[0007] By determining the mapping relationship and measuring the interference level estimate, the problems of inaccurate signal strength returned by the consumer-grade drone SDK and the lack of open spectrum sensing interface are solved. The computational load of the drone is reduced in the deployment decision-making process at the ground control end, and low-cost, easy-to-deploy, and accurate frequency switching is achieved on the consumer-grade drone platform.

[0008] In a first aspect, the present invention provides a method for determining the frequency point of a UAV based on ground-based collaborative calibration, specifically including the following steps: The ground control terminal collects the channel quality parameters of the current communication link of the UAV, while the airborne terminal collects the current frequency band information and the real-time signal strength of each candidate frequency band. The ground control terminal determines the mapping relationship between signal power and signal indication strength based on the UAV's model identification and current frequency band information, and determines the real-time estimated value of signal power for each candidate frequency band by combining the real-time signal strength of each candidate frequency band. Based on the real-time estimated signal power of each candidate frequency band, the effective signal-to-noise ratio of each candidate frequency band is given, and the predicted throughput of each candidate frequency band is predicted. Select the target frequency point based on the throughput forecast and generate frequency point switching information; By judging the channel quality parameters of the current communication link and combining them with frequency switching information, the frequency switching of the UAV is completed.

[0009] Furthermore, based on the UAV's model identification and current frequency band information, the mapping relationship between signal power and signal indication strength is determined, specifically including: Based on the drone's model identifier and current frequency band information, the corresponding mapping coefficients and mapping constants are retrieved from a pre-built mapping database; By using the mapping coefficients corresponding to the current frequency band information, the signal indication strength is fused to determine the signal variable terms; By integrating the mapping constants between the signal variable terms and the current frequency band information, the mapping relationship between signal power and signal indication strength is given.

[0010] Furthermore, the mapping relationship between signal power and signal indication strength is specifically expressed as follows:

[0011] Among them, P i,est R is the real-time estimated value of the signal power in frequency band i. i,sdk Let a be the real-time signal strength of frequency band i. i Let b be the mapping coefficient for frequency band i. i Let be the mapping relationship constant for frequency band i.

[0012] Furthermore, the pre-built mapping relationship library is obtained through the following steps: According to the calibration instructions from the ground control terminal, the UAV switches frequency bands and collects the real signal power at different frequency points in each frequency band; Based on the actual signal power at different frequency points in each frequency band, record the signal strength indication value of the UAV at different frequency points; By establishing the correspondence between actual signal power and signal strength indication values, data pairs are formed to construct a set of collected data points; Based on the collected data point set, the actual signal power and signal strength indication values ​​in different frequency bands are fitted to obtain the mapping relationship coefficients and mapping relationship constants for different frequency bands; By combining the drone model identifier, a mapping relationship database is formed by constructing the mapping relationship coefficients and mapping relationship constants between the model identifier and different frequency bands.

[0013] Furthermore, based on the real-time signal power estimates for each candidate frequency band, the effective signal-to-noise ratio for each candidate frequency band is given, and the predicted throughput for each candidate frequency band is determined through the following steps: Calculate the difference between the real-time estimated signal power and the estimated interference level for each candidate frequency band, and give the effective signal power value for each candidate frequency band; The effective signal-to-noise ratio of each candidate frequency band is determined based on the ratio of the effective signal power value to the estimated interference level value of each candidate frequency band. The throughput prediction value of each candidate frequency band is predicted based on the effective signal-to-noise ratio of each candidate frequency band.

[0014] Furthermore, the interference level estimate is determined through the following steps: Based on a preset scanning period, spectrum sensing scans are performed on each candidate frequency band to provide ground-based measurement values ​​of interference levels. The ground measurements of interference levels are correlated and stored with the location of the UAV to construct an area electromagnetic map; The ground-based measurements of interference levels obtained from multiple scans in each candidate frequency band are weighted and filtered to obtain an estimated interference level, which is specifically expressed as follows:

[0015] in, Let be the estimated interference level of candidate frequency point j in candidate frequency band i. The ground measurement value of the current interference level at candidate frequency point j in candidate frequency band i. α is the estimated interference level of candidate frequency point j in candidate frequency band i in the previous scan cycle, and α is the filtering coefficient used to balance real-time performance and stability.

[0016] Furthermore, based on the effective signal-to-noise ratio of each candidate frequency band, the throughput prediction value of each candidate frequency band is calculated, specifically including: Based on the effective signal-to-noise ratio of each candidate frequency band, a corresponding throughput prediction function is matched; Substitute the effective signal-to-noise ratio of each candidate frequency band into the corresponding throughput prediction function to obtain the initial throughput prediction value of each candidate frequency band. Based on the drone's current location and the regional electromagnetic map, the scene correction coefficient for the current area is determined. By using scenario correction coefficients, the initial throughput prediction value is corrected, and the throughput prediction value for each candidate frequency band is given.

[0017] Furthermore, the throughput prediction function is specifically expressed as:

[0018] Among them, T base SNR is the initial throughput prediction value. eff For the effective signal-to-noise ratio, β1 is the first prediction coefficient, T1 is the first prediction constant, γ1 is the first prediction threshold, β2 is the second prediction coefficient, T2 is the second prediction constant, γ2 is the second prediction threshold, β3 is the third prediction coefficient, T3 is the third prediction constant, γ3 is the third prediction threshold, T4 is the fourth prediction constant, γ4 is the fourth prediction threshold, and u(x) is the unit step function.

[0019] Furthermore, by judging the channel quality parameters of the current communication link and combining them with frequency switching information, the UAV frequency switching is completed, specifically including: Based on the preset hysteresis time window and combined with the channel quality parameters of the current communication link, the handover trigger command is determined; Based on the handover trigger command and frequency point handover information, determine whether the subband with the highest predicted throughput in the current frequency band meets the preset service threshold, and give the first judgment result; Based on the first judgment result, the sub-band with the highest predicted throughput value in other frequency bands is determined, and frequency switching is performed.

[0020] Furthermore, each candidate frequency band includes at least one candidate sub-band, which is determined through the following steps: Obtain the real-time signal strength of each undetermined sub-band; Based on the real-time signal strength of each undetermined sub-band, the local communication score of each undetermined sub-band is determined; Based on the preset number of candidates and combined with the local communication score of each candidate sub-band, the candidate sub-bands are screened and the candidate sub-bands are determined. The airborne terminal binds the candidate subbands and their corresponding local communication scores, and generates and sends back data packets in conjunction with the preset data compression format.

[0021] Secondly, the present invention also provides a UAV frequency point determination system based on ground-based collaborative calibration, employing any of the above-mentioned UAV frequency point determination methods based on ground-based collaborative calibration, including a ground control terminal and an airborne terminal: The ground control unit includes a communication quality detection module, a handover decision module, a ground calibration module, a ground spectrum sensing module, a throughput prediction module, and a frequency band handover control module. The communication quality detection module is used to collect the channel quality parameters of the current communication link of the UAV; The ground calibration module is used to determine the mapping relationship between signal power and signal indication strength; The ground-based spectrum sensing module is used to collect interference level estimates for each candidate frequency band. The throughput prediction module is used to combine the real-time signal strength and interference level estimates of each candidate frequency band to predict the throughput of each candidate frequency band. The frequency band switching control module is used to select the target frequency point based on the throughput prediction value and generate frequency point switching information; The handover decision module is used to generate a handover command by judging the channel quality parameters of the current communication link and combining the frequency handover information; The airborne terminal includes a command receiving module, a signal strength reporting module, a compression and transmission module, and a switching execution module; The command receiving module is used to receive switching commands from the ground control terminal; The signal strength reporting module is used to collect the current frequency band information of the UAV and the real-time signal strength of each frequency band; The compression and transmission module is used to perform local scoring on each candidate frequency band based on real-time signal strength and then compress and transmit the data back. The switching execution module is used to perform frequency switching operations according to the switching command.

[0022] The present invention provides a method and system for determining UAV frequency points based on ground-based collaborative calibration, which has at least the following beneficial effects: (1) By determining the mapping relationship between signal power and signal indication strength through the UAV model identification and current frequency band information, as well as measuring the interference level estimate, the problems of inaccurate signal strength returned by the consumer-grade UAV SDK and the closed spectrum sensing interface are solved. The target frequency point is selected by combining the effective signal-to-noise ratio of each candidate frequency band to predict the corresponding throughput and generating frequency switching information to complete the UAV frequency switching. The decision-making process is deployed at the ground control end to reduce the computational load of the UAV and realize low-cost, easy-to-deploy, and accurate frequency switching on the consumer-grade UAV platform.

[0023] (2) The UAV is calibrated by the calibration command of the ground control terminal, and the correspondence between the model identification and the mapping coefficient and mapping constant of different frequency bands is constructed to form a mapping database, so as to improve the accuracy of the real-time signal power estimation and avoid the risk of calibration in flight.

[0024] (3) By binding the candidate subband and the local communication score corresponding to the candidate subband, and generating the return data packet in combination with the preset data compression format, the data volume of the return data packet is compressed, which adapts to the bandwidth limitation of the return channel of consumer drones and does not affect the transmission of instructions. Attached Figure Description

[0025] Figure 1 A flowchart of a UAV frequency point determination method based on ground-based collaborative calibration provided in an embodiment of the present invention; Figure 2 An example diagram illustrating the determination of mapping relationships provided in embodiments of the present invention; Figure 3 This is an example diagram illustrating the structure of the mapping relationship library provided in an embodiment of the present invention; Figure 4 A flowchart for determining the effective signal-to-noise ratio of each candidate frequency band is provided in an embodiment of the present invention; Figure 5 Example diagram of spectrum scanning data provided in an embodiment of the present invention; Figure 6 An example diagram illustrating the relationship between baseline throughput and scene correction coefficients provided in this embodiment of the invention; Figure 7 This is a data example diagram of local communication scoring and data return transmission provided in an embodiment of the present invention; Figure 8 This is an example diagram illustrating the effect of ground testing, calibration, and frequency switching provided in an embodiment of the present invention. Figure 9 This is a structural block diagram of a UAV frequency point determination system based on ground-based collaborative calibration provided in an embodiment of the present invention. Detailed Implementation

[0026] To better understand the above technical solutions, a detailed description of the solutions will be provided below in conjunction with the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0027] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. The singular forms “a,” “the,” and “the” as used in the embodiments of this invention and the appended claims are also intended to include the plural forms, and “multiple” generally includes at least two unless the context clearly indicates otherwise.

[0028] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that an article or device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such an article or device. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the article or device that includes said element.

[0029] The UAV frequency point determination method based on ground-based collaborative calibration provided in this invention is applied to consumer-grade UAVs. Taking the DJI Mavic 3 consumer-grade UAV as an example, it is applied in a substation power inspection scenario. The UAV's onboard unit is a DJI Mavic 3 running MSDK 4.0. The ground control unit is an Android-based remote controller with a screen and a built-in RF power detection chip. The candidate frequency bands are 2.4GHz and 5.8GHz, with a candidate subband bandwidth of 20MHz. The service requirement is high-definition video transmission from the onboard unit, with a stable throughput of ≥10Mbps.

[0030] This invention provides a method and system for determining the frequency point of a UAV based on ground-based collaborative calibration. The method includes: a ground control terminal collecting channel quality parameters of the current communication link of the UAV; and an airborne terminal collecting current frequency band information and real-time signal strength of each candidate frequency band. The ground control terminal determines the mapping relationship between signal power and signal indication strength based on the UAV's model identification and current frequency band information, and determines the real-time estimated signal power of each candidate frequency band by combining the real-time signal strength of each candidate frequency band. Based on the real-time estimated signal power of each candidate frequency band, the effective signal-to-noise ratio of each candidate frequency band is given, and the predicted throughput of each candidate frequency band is predicted. A target frequency point is selected based on the predicted throughput, and frequency point switching information is generated. By judging the channel quality parameters of the current communication link and combining the frequency point switching information, the UAV frequency point is switched. This method solves problems such as inaccurate signal strength returned by consumer-grade UAV SDKs, closed spectrum sensing interfaces, limited backhaul bandwidth, and high onboard computing load, achieving low-cost, easy-to-deploy, and accurate frequency point switching on consumer-grade UAV platforms.

[0031] like Figure 1 As shown in the figure, this embodiment of the invention provides a method for determining the frequency point of a UAV based on ground-based collaborative calibration. The specific steps are as follows: S101: Collects the channel quality parameters of the current communication link of the UAV, the current frequency band information, and the real-time signal strength of each candidate frequency band.

[0032] In one specific implementation, the UAV's channel quality parameters include bit error rate (BER) and signal-to-noise ratio (SNR), while the current frequency band information indicates the frequency band in which the UAV is currently communicating. The ground control unit controls the UAV to sequentially traverse the calibration frequency points of each candidate sub-band within the candidate frequency bands. For example, the calibration frequency points for the 2.4GHz band include 2400MHz, 2420MHz, 2440MHz, 2460MHz, and 2480MHz, and the calibration frequency points for the 5.8GHz band include 5800MHz, 5820MHz, 5840MHz, 5860MHz, and 5880MHz. At each calibration frequency point, the RSSI value (i.e., real-time signal strength) returned by the onboard SDK is recorded. Candidate frequency bands can be preset or filtered based on the real-time signal strength of each band; this is not limited. Due to the limited computing power of the consumer UAV's onboard terminal, the process of acquiring channel quality parameters is integrated into the ground control unit. The onboard terminal acquires the current frequency band information and the real-time signal strength of each candidate frequency band and uploads it to the ground control unit.

[0033] S102: The ground control terminal determines the mapping relationship between signal power and signal indication strength based on the UAV's model identification and current frequency band information, and determines the real-time estimated value of signal power for each candidate frequency band by combining the real-time signal strength of each candidate frequency band.

[0034] Specifically, based on the UAV's model identification and current frequency band information, the mapping relationship between signal power and signal indication strength is determined, including: Based on the drone's model identifier and current frequency band information, the corresponding mapping coefficients and mapping constants are retrieved from a pre-built mapping database; By using the mapping coefficients corresponding to the current frequency band information, the signal indication strength is fused to determine the signal variable terms; By integrating the mapping constants between the signal variable terms and the current frequency band information, the mapping relationship between signal power and signal indication strength is given.

[0035] Furthermore, the mapping relationship between signal power and signal indication strength is specifically expressed as follows:

[0036] Among them, P i,est R is the real-time estimated value of the signal power in frequency band i. i,sdk Let a be the real-time signal strength of frequency band i. i Let b be the mapping coefficient for frequency band i. i Let be the mapping relationship constant for frequency band i. This is a signal variable term.

[0037] In one specific implementation, during the operation of a UAV's flight mission, the UAV reports the signal strength indication value of the current frequency band via the SDK. Based on the UAV's model identifier and the current frequency band information, the corresponding mapping coefficients and mapping constants are invoked. The signal strength indication is fused using the mapping coefficients corresponding to the current frequency band information to determine the signal variable. Furthermore, combined with the mapping constants corresponding to the current frequency band information, a mapping relationship between signal power and signal strength indication is given. The real-time signal strength of each candidate frequency band is substituted into the mapping relationship to estimate the UAV's actual signal power in real time, thus obtaining the actual signal power of each candidate frequency band.

[0038] Furthermore, the pre-built mapping relationship library is obtained through the following steps: According to the calibration instructions from the ground control terminal, the UAV switches frequency bands and collects the real signal power at different frequency points in each frequency band; Based on the actual signal power at different frequency points in each frequency band, record the signal strength indication value of the UAV at different frequency points; By establishing the correspondence between actual signal power and signal strength indication values, data pairs are formed to construct a set of collected data points; Based on the collected data point set, the actual signal power and signal strength indication values ​​in different frequency bands are fitted to obtain the mapping relationship coefficients and mapping relationship constants for different frequency bands; By combining the drone model identifier, a mapping relationship database is formed by constructing the mapping relationship coefficients and mapping relationship constants between the model identifier and different frequency bands.

[0039] Because the RSSI value returned by the consumer-grade drone SDK is not the actual physical layer received power, but rather a normalized score or relative intensity value, the RSSI value returned by the SDK is inaccurate. There is a complex nonlinear relationship between the SDK return value and the actual received power, and the mapping relationship varies for different drone models and frequency bands. Directly using the SDK return value for frequency selection may lead to serious decision-making biases. Therefore, this invention proposes a calibration mechanism based on pre-flight ground testing to improve data accuracy and avoid the risks of in-flight calibration.

[0040] In a specific example, before the drone performs its flight mission, it is placed at a ground testing station, and a communication link is established with the ground control unit to ensure the calibration process is safe and controllable. For example, the drone is 5 meters away from the ground control unit. The ground control unit generates calibration instructions according to a pre-set calibration scheme, instructing the drone to sequentially switch to preset calibration frequencies. For each frequency band, 3-5 representative frequencies are selected (including the band edges and center). The calibration process is completed before takeoff, eliminating the risk of flight loss of control. (Refer to...) Figure 2At each calibration frequency point, the ground control unit measures the actual signal power received from the UAV (i.e., the actual received power) through its built-in radio frequency power detection unit, and simultaneously records the signal strength indication value reported by the UAV via the SDK (i.e., the SDK return value). Each set of actual signal power and signal strength indication values ​​forms a data pair, which ultimately constitutes the collected data point set, specifically represented as follows:

[0041] in, The signal strength indication value at the k-th calibration frequency point and actual signal power The data pairs consist of k, which is the number of the calibration frequency point, and K, which is the total number of calibration frequency points.

[0042] Based on the collected data point set, a mapping relationship between the actual signal power and the signal strength indication value is established. In this example, the least squares method is used for linear fitting to establish the mapping relationship. The nonlinear relationship is implicit in the collected data point set, and the linear fitting can meet the engineering accuracy requirements. Specifically, it is expressed as follows:

[0043] Among them, P i,est R is the real-time estimated value of the signal power in frequency band i. i,sdk Let a be the real-time signal strength of frequency band i. i Let b be the mapping coefficient for frequency band i. i Let be the mapping relationship constant for frequency band i.

[0044] A mapping relationship library is established, linking drone model identifiers with different frequency bands using mapping coefficients and constants. Mapping relationships for the same drone model can be reused across flight missions without requiring recalibration before each flight. The mapping relationship library supports pre-set mapping coefficients and constants for mainstream consumer drone models (such as DJI Mavic, Mini, and Air series). When the same ground control terminal is used with different drone models, the system automatically calls the corresponding function formulas, mapping coefficients, and constants based on the drone model identifier reported by the drone, enabling unified management of multi-model mixed operation scenarios. (Refer to...) Figure 3 The mapping relationship library includes the model identifier, frequency band, mapping relationship coefficients, and mapping relationship constants. In this example, the mapping relationship coefficients and mapping relationship constants include mapping coefficient a and mapping coefficient b. In other examples, they can be adjusted according to the actual situation, and there is no limitation on this.

[0045] In a specific example, when the frequency band is 2.4GHz, the mapping relationship is specifically represented as: P est =0.15·R sdk-86; When the frequency band is 5.8GHz, the mapping relationship is specifically represented as: P est =0.13·R sdk -84.

[0046] By establishing a mapping relationship between signal strength indication values ​​and actual signal power during the ground testing phase before drone takeoff, the risks and efficiency losses of in-flight calibration can be avoided, eliminating the need for complex calculations on the airborne end.

[0047] S103: Based on the real-time estimated signal power of each candidate frequency band, the effective signal-to-noise ratio of each candidate frequency band is given, and the throughput prediction value of each candidate frequency band is predicted.

[0048] Reference Figure 4 Furthermore, based on the real-time signal power estimates for each candidate frequency band, the effective signal-to-noise ratio for each candidate frequency band is given, and the predicted throughput for each candidate frequency band is determined through the following steps: Calculate the difference between the real-time estimated signal power and the estimated interference level for each candidate frequency band, and give the effective signal power value for each candidate frequency band; The effective signal-to-noise ratio of each candidate frequency band is determined based on the ratio of the effective signal power value to the estimated interference level value of each candidate frequency band. The throughput prediction value of each candidate frequency band is predicted based on the effective signal-to-noise ratio of each candidate frequency band.

[0049] Specifically, the effective signal-to-noise ratio of each candidate frequency band is expressed as follows:

[0050] in, Let be the effective signal-to-noise ratio of candidate frequency point j in candidate frequency band i. This is the real-time estimated value of the signal power at candidate frequency point j in candidate frequency band i. The ground control unit measures the estimated interference level of candidate frequency point j in candidate frequency band i.

[0051] It is important to understand that in the embodiments provided by this invention, the units of the real-time signal power estimate and the interference level estimate are dBm. dBm is a logarithmic decibel unit, which is the relative logarithmic dimension of power. When both the real-time signal power estimate and the interference level estimate are in dBm, linear power division is not required; the values ​​are directly subtracted, and the result is the effective signal-to-noise ratio (SNR). In other embodiments, if the units of the real-time signal power estimate and the interference level estimate are not in dBm, then the units need to be standardized before calculating the effective SNR.

[0052] Furthermore, the interference level estimate is determined through the following steps: Based on a preset scanning period, spectrum sensing scans are performed on each candidate frequency band to provide ground-based measurement values ​​of interference levels. The ground measurements of interference levels are correlated and stored with the location of the UAV to construct an area electromagnetic map; The ground-based measurements of interference levels obtained from multiple scans in each candidate frequency band are weighted and filtered to obtain an estimated interference level, which is specifically expressed as follows:

[0053] in, Let be the estimated interference level of candidate frequency point j in candidate frequency band i. The ground measurement of the current interference level at candidate frequency point j in candidate frequency band i. This is the estimated interference level of candidate frequency point j in candidate frequency band i during the previous scan cycle, where α is the filtering coefficient used to balance real-time performance and stability. For example, α∈[0.3,0.7].

[0054] To address the lack of a spectrum sensing interface in consumer-grade drone SDKs, the spectrum sensing function is deployed on the ground control unit. The spectrum analysis module on the ground unit measures the interference level at the drone's current location. It's important to understand that interference sources in power line inspection scenarios are primarily located on ground equipment (transformers, circuit breakers, reactors, etc.). The interference data measured on the ground is highly correlated with the interference received by the airborne unit, thus meeting the engineering accuracy requirements for frequency selection.

[0055] In one specific implementation, refer to Figure 5 At the ground control end, a periodic scanning mechanism is used to scan candidate frequency bands. For example, the period is 10-30 seconds, which can be dynamically adjusted according to the flight speed. The interference levels of each candidate sub-band in the candidate frequency band are collected, and the results of multiple scans are weighted and filtered to eliminate instantaneous interference fluctuations.

[0056] In a specific example, after the drone takes off, the ground control continuously scans the 2.4GHz and 5.8GHz frequency bands at 20-second intervals. When the drone flies into an open area, the ground measurement results show that the interference level in the 2.4GHz band is -75dBm, and the interference level in the 5.8GHz band is -70dBm.

[0057] By deploying spectrum sensing at the ground control terminal, the interference level measured by the ground control terminal can replace airborne measurement, eliminating the need for airborne spectrum sensing hardware. The interference level estimate is used for subsequent throughput prediction, and the measured interference level estimate is stored in association with GPS location, which can be used to construct regional electromagnetic maps.

[0058] Furthermore, based on the effective signal-to-noise ratio of each candidate frequency band, the throughput prediction value of each candidate frequency band is calculated, specifically including: Based on the effective signal-to-noise ratio of each candidate frequency band, a corresponding throughput prediction function is matched; Substitute the effective signal-to-noise ratio of each candidate frequency band into the corresponding throughput prediction function to obtain the initial throughput prediction value of each candidate frequency band. Based on the drone's current location and the regional electromagnetic map, the scene correction coefficient for the current area is determined. By using scenario correction coefficients, the initial throughput prediction value is corrected, and the throughput prediction value for each candidate frequency band is given.

[0059] In one specific implementation, the decision-making algorithm is deployed at the ground control terminal, leveraging its superior computing power for accurate throughput prediction and frequency selection. Based on the calculated effective signal-to-noise ratio (SNR) of each candidate frequency band, a corresponding throughput prediction function is matched, and the effective SNR is substituted to obtain the initial throughput prediction value for each candidate frequency band.

[0060] The throughput prediction function is specifically expressed as:

[0061] Among them, T base SNR is the initial throughput prediction value. eff For the effective signal-to-noise ratio, β1 is the first prediction coefficient, T1 is the first prediction constant, γ1 is the first prediction threshold, β2 is the second prediction coefficient, T2 is the second prediction constant, γ2 is the second prediction threshold, β3 is the third prediction coefficient, T3 is the third prediction constant, γ3 is the third prediction threshold, T4 is the fourth prediction constant, γ4 is the fourth prediction threshold, and u(x) is the unit step function.

[0062] Based on the different effective signal-to-noise ratio ranges, a piecewise linear throughput prediction model is used to predict the baseline throughput of each candidate frequency band. This model includes multiple throughput prediction functions. In a strong electromagnetic interference environment for power equipment, it provides a basis for throughput prediction to select candidate frequency bands / subbands. In a specific example, γ1=5, γ2=15, γ3=25, γ4=35, β1=1.0, β2=1.5, β3=1.0, T1=5, T2=15, T3=30, T4=40, then:

[0063] Among them, T base SNR is the initial throughput prediction value. eff For effective signal-to-noise ratio.

[0064] In other examples, the prediction threshold can be adjusted based on factors such as the communication environment, thereby adaptively adjusting the prediction coefficients and prediction constants, which is not limited here.

[0065] In the context of strong electromagnetic interference during power line inspection, a segmented throughput prediction function is used to adapt to changes in link performance under different interference intensities. The segmented correspondence between throughput and signal-to-noise ratio is consistent with the modulation and coding switching rules of wireless communication links, facilitating engineering verification and adjustment.

[0066] After obtaining the initial throughput prediction value, based on the current position of the UAV, the scene correction coefficient for the corresponding area of ​​the current position is obtained from the preset regional electromagnetic map. The initial throughput prediction value is then adjusted according to the scene correction coefficient to obtain the throughput prediction value for each candidate frequency band, specifically expressed as follows:

[0067] Among them, T pred T is the predicted throughput value. base λ is the initial throughput prediction value. zone This is a scene correction factor.

[0068] In a specific example, the scene correction factor for the main transformer area in the regional electromagnetic map is 0.85, the scene correction factor for the GIS equipment area is 0.9, the scene correction factor for the switchyard area is 0.95, and the scene correction factor for the open area is 1. Figure 6 As shown.

[0069] S104: Select the target frequency point based on the throughput prediction value and generate frequency point switching information.

[0070] In one specific implementation, based on the throughput prediction value, the candidate sub-band with the highest throughput prediction value is selected from each candidate frequency band as the target frequency point, and frequency point switching information is generated. The frequency point switching information includes relevant information of the target frequency point, and the airborne terminal can complete the frequency point switching according to the frequency point switching information.

[0071] In a specific example, after receiving information from each candidate sub-band, the ground control unit makes a precise decision. Taking the first sub-band in the 5.8GHz candidate frequency band as an example: the real-time signal strength is substituted into the mapping relationship to obtain the real-time estimated signal power value P. est =0.42×72-90=-59.76. The estimated interference level in the main transformer area is -82dBm, and the effective signal-to-noise ratio is calculated as SNR. eff=-59.76-(-82)=22.24. Substituting the effective signal-to-noise ratio into the corresponding throughput prediction function, the initial throughput prediction value is 25.86. Combining this with the scenario correction coefficient of 0.85 for the main transformer area, the throughput prediction value is 25.86×0.85=22.0. Similarly, the other candidate subbands are calculated to obtain their corresponding throughput prediction values. If the throughput prediction values ​​of all candidate subbands reach the service threshold, the candidate subband with the highest throughput prediction value is selected as the target frequency.

[0072] S105: By judging the channel quality parameters of the current communication link and combining the frequency switching information, the UAV frequency switching is completed.

[0073] Specifically, by judging the channel quality parameters of the current communication link and combining them with frequency switching information, the UAV frequency switching is completed, including: Based on the preset hysteresis time window and combined with the channel quality parameters of the current communication link, the handover trigger command is determined; Based on the handover trigger command and frequency point handover information, determine whether the subband with the highest predicted throughput in the current frequency band meets the preset service threshold, and give the first judgment result; Based on the first judgment result, the sub-band with the highest predicted throughput value in other frequency bands is determined, and frequency switching is performed.

[0074] In one specific implementation, the hysteresis window is the minimum time interval between two adjacent frequency switchings. Timing begins after a frequency switching is completed. Within the hysteresis window, even if a switching trigger command is received, no frequency switching occurs to avoid communication instability caused by frequent frequency switching. Once the hysteresis window is exceeded, the channel quality parameters of the current communication link are assessed in real time. If the channel quality parameters do not meet the channel quality requirements, a switching trigger command is generated. In a specific example, a switching trigger command is generated if the bit error rate in the channel quality parameters exceeds a preset switching threshold N times consecutively, or the received signal strength indicator value is lower than a preset switching threshold N times consecutively, or the effective signal-to-noise ratio is lower than a preset switching threshold N times consecutively, or if the ground operator manually triggers the switch via a human-machine interface.

[0075] In a specific example, the hysteresis window duration is 5-10 seconds. During this window, no new handover triggers are responded to to avoid the ping-pong effect. If communication is completely interrupted during the window, a handover is forced. For example, after the handover is completed, an 8-second hysteresis window is set. During the hysteresis window, even if communication quality fluctuates slightly, the system does not respond to new handover triggers to avoid frequent handovers. If communication is completely interrupted during the window, a handover is forced.

[0076] After generating the handover trigger command, the subband with the highest predicted throughput value is selected from the current frequency band, and it is determined whether the predicted throughput value meets the preset service threshold. If yes, a frequency switch is performed, switching to the target frequency corresponding to the subband with the highest predicted throughput value in the current frequency band. If not, the subband with the highest predicted throughput value is selected from other frequency bands, and the predicted throughput value is checked. If the preset service threshold is met, the frequency switch operation is performed; if the preset service threshold is not met, a communication warning message is output to remind the operator to pay attention to communication quality.

[0077] In a specific example, when communication quality triggers a handover, the system first assesses whether there are any idle subbands within the current frequency band (2.4GHz) whose predicted throughput meets a preset service threshold (≥10Mbps). The real-time signal strength of each candidate subband in the 2.4GHz band is 58-65, the real-time estimated signal power is approximately -65 to -70 dBm, the effective signal-to-noise ratio is approximately 15-20 dB, and the predicted throughput is approximately 10-15 Mbps. After multiplying by a scenario correction factor of 0.85, this decreases to 8.5-12.8 Mbps. Some candidate subbands may not meet the threshold. If a candidate subband meets the threshold, a first-level intra-band handover is performed (handover time <100ms); otherwise, a second-level cross-band handover is performed.

[0078] Specifically, each candidate frequency band includes at least one candidate sub-band, which is determined through the following steps: Obtain the real-time signal strength of each undetermined sub-band; Based on the real-time signal strength of each undetermined sub-band, the local communication score of each undetermined sub-band is determined; Based on the preset number of candidates and combined with the local communication score of each candidate sub-band, the candidate sub-bands are screened and the candidate sub-bands are determined.

[0079] The candidate subbands and their corresponding local communication scores are bound together, and a return data packet is generated and sent using a preset data compression format.

[0080] To address the issue of limited bandwidth in the backhaul channel of consumer-grade drones, this invention sets up a local pre-scoring and compression backhaul mechanism, compressing the backhaul data volume from hundreds of bytes to less than 50 bytes.

[0081] Reference Figure 7In one specific implementation, the UAV's onboard unit collects the real-time signal strength (i.e., signal strength indication value) of each candidate subband and uses this real-time signal strength as the local communication score (i.e., local score) for each candidate subband. Combined with a preset number of candidates, candidate subbands are selected from the candidate subbands. In a specific example, with 3 candidates, the three candidate subbands with the highest local communication scores are selected as candidate subbands; with 5 candidates, the five candidate subbands with the highest local communication scores are selected as candidate subbands. The returned data packet only includes information about these candidate subbands. With 3 candidates, over 90% of the optimal frequency points are covered; with 5 candidates, over 98% of the optimal frequency points are covered. Increasing the number of candidates linearly increases the data volume of the returned data packet, requiring a balance between accuracy and bandwidth. The data compression format is compact binary format. The transmission of the returned data packet is triggered sequentially: immediately after a frequency switch; periodic transmission to update candidate frequency point information; and transmission when actively requested by the ground control unit.

[0082] After local scoring on the airborne end, the data is compressed and sent back, reducing the amount of data sent back from hundreds of bytes to less than 50 bytes, thus adapting to the bandwidth limitations of the backhaul.

[0083] Reference Figure 8 This invention establishes a mapping relationship between signal power and signal indication strength through pre-flight ground testing and calibration, thereby reducing real-time signal strength errors and fundamentally solving the decision-making bias problem caused by inaccurate signal strength in consumer-grade drone SDKs. Simultaneously, a spectrum sensing unit is deployed in the ground control terminal, using interference levels measured by the ground control terminal to replace airborne measurements. This eliminates the need for the consumer-grade drone to possess spectrum sensing capabilities, overcoming SDK interface limitations. Combined with a local scoring and compressed backhaul mechanism, the data size of the backhaul data packets is compressed from hundreds of bytes to less than 50 bytes, adapting to the bandwidth limitations of the backhaul channel in consumer-grade drones without affecting remote control command transmission.

[0084] By deploying the decision-making process at the ground control station, the airborne terminal only performs lightweight measurement, pre-scoring, and data transmission, eliminating the need to run complex algorithms such as machine learning, thus adapting to the limited computing resources of consumer-grade drones. The mapping relationship library supports pre-configured mapping relationship coefficients and constants for various consumer-grade drone models, allowing the same ground station to adapt to different drone models, making it suitable for large-scale fleet operation scenarios.

[0085] Reference Figure 9 This invention provides a UAV frequency point determination system based on ground-based collaborative calibration, comprising a ground control terminal and an airborne terminal: The ground control unit includes a communication quality detection module, a handover decision module, a ground calibration module, a ground spectrum sensing module, a throughput prediction module, and a frequency band handover control module. The communication quality detection module is used to collect the channel quality parameters of the current communication link of the UAV; The ground calibration module is used to determine the mapping relationship between signal power and signal indication strength; The ground-based spectrum sensing module is used to collect interference level estimates for each candidate frequency band. The throughput prediction module is used to combine the real-time signal strength and interference level estimates of each candidate frequency band to predict the throughput of each candidate frequency band. The frequency band switching control module is used to select the target frequency point based on the throughput prediction value and generate frequency point switching information; The handover decision module is used to generate a handover command by judging the channel quality parameters of the current communication link and combining the frequency handover information; The airborne terminal includes a command receiving module, a signal strength reporting module, a compression and transmission module, and a switching execution module; The command receiving module is used to receive switching commands from the ground control terminal; The signal strength reporting module is used to collect the current frequency band information of the UAV and the real-time signal strength of each frequency band; The compression and transmission module is used to perform local scoring on each candidate frequency band based on real-time signal strength and then compress and transmit the data back. The switching execution module is used to perform frequency switching operations according to the switching command.

[0086] By deploying the decision-making process at the ground control end, the airborne end only performs lightweight measurements and data transmission, reducing the airborne computing load.

[0087] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the described module can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0088] Although preferred embodiments of the invention 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 appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention. Clearly, those skilled in the art can make various alterations and modifications to the invention without departing from its spirit and scope. Thus, if these modifications and modifications of the invention fall within the scope of the claims and their equivalents, the invention is also intended to include these modifications and modifications.

Claims

1. A method for determining the frequency point of a UAV based on ground-based collaborative calibration, characterized in that, include: The ground control terminal collects the channel quality parameters of the current communication link of the UAV, while the airborne terminal collects the current frequency band information and the real-time signal strength of each candidate frequency band. The ground control terminal determines the mapping relationship between signal power and signal indication strength based on the UAV's model identification and current frequency band information, and determines the real-time estimated value of signal power for each candidate frequency band by combining the real-time signal strength of each candidate frequency band. Based on the real-time estimated signal power of each candidate frequency band, the effective signal-to-noise ratio of each candidate frequency band is given, and the predicted throughput of each candidate frequency band is predicted. Select the target frequency point based on the throughput forecast and generate frequency point switching information; By judging the channel quality parameters of the current communication link and combining them with frequency switching information, the frequency switching of the UAV is completed.

2. The UAV frequency point determination method based on ground-based collaborative calibration as described in claim 1, characterized in that, Based on the UAV's model identification and current frequency band information, the mapping relationship between signal power and signal indication strength is determined, specifically including: Based on the drone's model identifier and current frequency band information, the corresponding mapping coefficients and mapping constants are retrieved from a pre-built mapping database; By using the mapping coefficients corresponding to the current frequency band information, the signal indication strength is fused to determine the signal variable terms; By integrating the mapping constants between the signal variable terms and the current frequency band information, the mapping relationship between signal power and signal indication strength is given.

3. The UAV frequency point determination method based on ground-based collaborative calibration as described in claim 2, characterized in that, The mapping relationship between signal power and signal indication strength is specifically expressed as follows: ； Among them, P i,est R is the real-time estimated value of the signal power in frequency band i. i,sdk Let a be the real-time signal strength of frequency band i. i Let b be the mapping coefficient for frequency band i. i Let be the mapping relationship constant for frequency band i.

4. The UAV frequency point determination method based on ground-based collaborative calibration as described in claim 2, characterized in that, The pre-built mapping database is obtained through the following steps: According to the calibration instructions from the ground control terminal, the UAV switches frequency bands and collects the real signal power at different frequency points in each frequency band; Based on the actual signal power at different frequency points in each frequency band, record the signal strength indication value of the UAV at different frequency points; By establishing the correspondence between actual signal power and signal strength indication values, data pairs are formed to construct a set of collected data points; Based on the collected data point set, the actual signal power and signal strength indication values ​​in different frequency bands are fitted to obtain the mapping relationship coefficients and mapping relationship constants for different frequency bands; By combining the drone model identifier, a mapping relationship database is formed by constructing the mapping relationship coefficients and mapping relationship constants between the model identifier and different frequency bands.

5. The UAV frequency point determination method based on ground-based collaborative calibration as described in claim 1, characterized in that, Based on the real-time signal power estimates for each candidate frequency band, the effective signal-to-noise ratio for each candidate frequency band is given, and the predicted throughput for each candidate frequency band is determined through the following steps: Calculate the difference between the real-time estimated signal power and the estimated interference level for each candidate frequency band, and give the effective signal power value for each candidate frequency band; The effective signal-to-noise ratio of each candidate frequency band is determined based on the ratio of the effective signal power value to the estimated interference level value of each candidate frequency band. The throughput prediction value of each candidate frequency band is predicted based on the effective signal-to-noise ratio of each candidate frequency band.

6. The UAV frequency point determination method based on ground-based collaborative calibration as described in claim 5, characterized in that, The interference level estimate is determined through the following steps: Based on a preset scanning period, spectrum sensing scans are performed on each candidate frequency band to provide ground-based measurement values ​​of interference levels. The ground measurements of interference levels are correlated and stored with the location of the UAV to construct an area electromagnetic map; The ground-based measurements of interference levels obtained from multiple scans in each candidate frequency band are weighted and filtered to obtain an estimated interference level, which is specifically expressed as follows: ； in, Let be the estimated interference level of candidate frequency point j in candidate frequency band i. The ground measurement value of the current interference level at candidate frequency point j in candidate frequency band i. α is the estimated interference level of candidate frequency point j in candidate frequency band i in the previous scan cycle, and α is the filtering coefficient used to balance real-time performance and stability.

7. The UAV frequency point determination method based on ground-based collaborative calibration as described in claim 5, characterized in that, Based on the effective signal-to-noise ratio of each candidate frequency band, the throughput prediction value of each candidate frequency band is calculated, specifically including: Based on the effective signal-to-noise ratio of each candidate frequency band, a corresponding throughput prediction function is matched; Substitute the effective signal-to-noise ratio of each candidate frequency band into the corresponding throughput prediction function to obtain the initial throughput prediction value of each candidate frequency band. Based on the drone's current location and the regional electromagnetic map, the scene correction coefficient for the current area is determined. By using scenario correction coefficients, the initial throughput prediction value is corrected, and the throughput prediction value for each candidate frequency band is given.

8. The UAV frequency point determination method based on ground-based collaborative calibration as described in claim 1, characterized in that, By judging the channel quality parameters of the current communication link and combining them with frequency switching information, the frequency switching of the UAV is completed, specifically including: Based on the preset hysteresis time window and combined with the channel quality parameters of the current communication link, the handover trigger command is determined; Based on the handover trigger command and frequency point handover information, determine whether the subband with the highest predicted throughput in the current frequency band meets the preset service threshold, and give the first judgment result; Based on the first judgment result, the sub-band with the highest predicted throughput value in other frequency bands is determined, and frequency switching is performed.

9. The UAV frequency point determination method based on ground-based collaborative calibration as described in claim 1, characterized in that, Each candidate frequency band includes at least one candidate sub-band, which is determined through the following steps: Obtain the real-time signal strength of each undetermined sub-band; Based on the real-time signal strength of each undetermined sub-band, the local communication score of each undetermined sub-band is determined; Based on the preset number of candidates and combined with the local communication score of each candidate sub-band, the candidate sub-bands are screened and the candidate sub-bands are determined. The airborne terminal binds the candidate subbands and their corresponding local communication scores, and generates and sends back data packets in conjunction with the preset data compression format.

10. A UAV frequency point determination system based on ground-based collaborative calibration, characterized in that, The UAV frequency point determination method based on ground-based collaborative calibration as described in any one of claims 1-9 includes a ground control terminal and an airborne terminal: The ground control unit includes a communication quality detection module, a handover decision module, a ground calibration module, a ground spectrum sensing module, a throughput prediction module, and a frequency band handover control module. The communication quality detection module is used to collect the channel quality parameters of the current communication link of the UAV; The ground calibration module is used to determine the mapping relationship between signal power and signal indication strength; The ground-based spectrum sensing module is used to collect interference level estimates for each candidate frequency band. The throughput prediction module is used to combine the real-time signal strength and interference level estimates of each candidate frequency band to predict the throughput of each candidate frequency band. The frequency band switching control module is used to select the target frequency point based on the throughput prediction value and generate frequency point switching information; The handover decision module is used to generate a handover command by judging the channel quality parameters of the current communication link and combining the frequency handover information; The airborne terminal includes a command receiving module, a signal strength reporting module, a compression and transmission module, and a switching execution module; The command receiving module is used to receive switching commands from the ground control terminal; The signal strength reporting module is used to collect the current frequency band information of the UAV and the real-time signal strength of each frequency band; The compression and transmission module is used to perform local scoring on each candidate frequency band based on real-time signal strength and then compress and transmit the data back. The switching execution module is used to perform frequency switching operations according to the switching command.

Images