An unmanned aerial vehicle anti-collision method and system based on inductive processing

Abstract

This invention discloses a drone collision avoidance method and system based on inductive processing, belonging to the field of drone inspection technology. The method includes: converting the magnetic field signal of the line to be inspected into an induced electrical signal via the drone's inductive sensing module and transmitting it to a signal conditioning module; converting the induced electrical signal into a voltage signal via the signal conditioning module and transmitting it to a collision avoidance control module; converting the voltage signal into a digital magnetic field strength signal via the collision avoidance control module; determining the spatial orientation relationship based on the calculated left-right and front-back difference ratios and a preset judgment threshold; matching the corresponding flight control strategy based on the ground altitude data, spatial attitude data, and heading data collected and transmitted by the pose perception module; and calculating and outputting motor control commands based on the pose correction amount mapped by the left-right and front-back difference ratios and the spatial orientation relationship to drive the flight actions corresponding to the flight control strategy. By implementing this invention, the problems of high cost and poor collision avoidance in existing technologies are solved.

Description

Technical Field

[0001] This invention relates to the field of drone inspection technology, and in particular to a drone collision avoidance method and system based on inductive processing. Background Technology

[0002] With the rapid advancement of power grid construction, the coverage of transmission lines has expanded, and a large number of transmission lines traverse complex geographical environments. Traditional manual inspection methods suffer from significant problems such as low operational efficiency and numerous blind spots, making them unable to meet the development needs of intelligent power grid operation and maintenance. In recent years, drones, due to their wide operating range, have been widely used in routine inspections and fault diagnosis of transmission lines. During drone inspections of transmission lines, drones need to fly close to the transmission lines to complete inspection tasks such as image acquisition. Since transmission lines are mostly thin-diameter overhead conductors, collision avoidance and obstacle avoidance capabilities are the core prerequisites for ensuring the safety of drone inspection operations.

[0003] Currently, most power line inspection drones rely on visual sensors, lidar, or ultrasonic sensors to identify power lines and obstacles, thereby achieving collision avoidance and inspection tasks. However, these sensors are limited by weather, lighting conditions, and cost. In visual solutions, cameras are significantly affected by weather and lighting, and their accuracy in detecting small targets is insufficient, leading to false or missed detections of power line targets and a substantial decrease in collision avoidance reliability. LiDAR is expensive, and ultrasonic sensors are easily interfered with by ambient sound waves, making them prone to ranging errors in complex environments and unable to provide collision avoidance warnings. Summary of the Invention

[0004] This invention provides a drone collision avoidance method and system based on inductive processing, which can effectively solve the problems of high inspection costs and poor collision avoidance effect in existing technologies.

[0005] One embodiment of the present invention provides a drone collision avoidance method based on inductive processing, applicable to drones; the drone is equipped with an inductive sensing module, a signal conditioning module, a pose perception module, and a collision avoidance control module; The drone collision avoidance method based on inductance processing includes: The magnetic field signal of the line to be inspected is obtained through the inductive sensing module. After the magnetic field signal is converted into an induced electrical signal, the induced electrical signal is transmitted to the signal conditioning module. The induced electrical signal is converted into a voltage signal by the signal conditioning module and transmitted to the collision avoidance control module. The pose perception module collects and transmits ground altitude data, spatial attitude data, and heading data to the collision avoidance control module. The anti-collision control module performs analog-to-digital conversion on the voltage signal according to the preset sampling frequency to obtain a digital signal of magnetic field strength. Based on the digital signal of magnetic field strength, the left-right difference ratio, which characterizes the lateral deviation of the UAV, and the front-back difference ratio, which characterizes the longitudinal deviation of the UAV, are calculated respectively. Based on the left-right difference ratio, front-back difference ratio, and preset judgment threshold, the spatial orientation relationship relative to the line to be inspected is determined. Based on spatial orientation, corresponding flight control strategies are matched according to ground altitude data, spatial attitude data, and heading data; The left-right difference ratio and the front-back difference ratio are mapped to attitude correction values. After calculation based on the attitude correction values ​​and spatial orientation relationships, motor control commands are output to drive the execution of flight actions corresponding to the flight control strategy.

[0006] Furthermore, the inductive sensing module includes several acquisition channels in various directions, and each acquisition channel includes a magnetic field sensing unit and a frequency selective filtering unit. The magnetic field signal of the line to be inspected is acquired through the inductive sensing module, converted into an induced electrical signal, and then transmitted to the signal conditioning module, including: The magnetic field signal of the line to be inspected is collected by the magnetic field induction unit in the corresponding direction, the magnetic field signal is converted into the induced electrical signal of the corresponding channel, and the induced electrical signal is transmitted to the frequency selective filtering unit. After the electromagnetic interference of non-target frequency in the induced electrical signal is filtered out by the frequency selective filtering unit, the filtered induced electrical signal is transmitted to the signal conditioning module. The target frequency of the frequency selective filter unit is matched with the power frequency AC frequency of the line to be inspected.

[0007] Furthermore, the induced electrical signal is converted into a voltage signal and transmitted to the collision avoidance control module via a signal conditioning module, including: The induced electrical signal is subjected to amplitude limiting protection processing through the signal conditioning module, which limits the voltage amplitude of the induced electrical signal within a preset safety threshold. The induced electrical signal after the amplitude limiting protection is linearly amplified to a preset amplitude range. The linearly amplified induced electrical signal is subjected to level bias processing and low-pass filtering processing. After being superimposed with a preset DC bias voltage, a voltage signal is obtained and transmitted to the anti-collision control module.

[0008] Furthermore, the induced electrical signal is converted into a voltage signal and transmitted to the collision avoidance control module via a signal conditioning module, including: The induced electrical signal is subjected to amplitude limiting protection processing through the signal conditioning module, which limits the voltage amplitude of the induced electrical signal within a preset safety threshold. The induced electrical signal after the amplitude limiting protection is linearly amplified to a preset amplitude range. The linearly amplified induced electrical signal is subjected to level bias processing and low-pass filtering processing. After being superimposed with a preset DC bias voltage, a voltage signal is obtained and transmitted to the anti-collision control module.

[0009] Furthermore, the preset judgment thresholds include: alignment threshold, horizontal offset threshold, and vertical offset threshold; Based on the left-right difference ratio, front-back difference ratio, and preset judgment thresholds, the spatial orientation relative to the line to be inspected is determined, including: The absolute value of the left-right difference ratio is compared with the alignment threshold and the horizontal offset threshold, respectively; the absolute value of the front-back difference ratio is compared with the alignment threshold and the vertical offset threshold, respectively. If the absolute values ​​of both the left-right difference ratio and the front-back difference ratio are less than the alignment threshold, the spatial orientation relationship is determined to be aligned with the center of the line to be inspected. If the absolute value of the left-right difference ratio is not less than the alignment threshold and less than the lateral offset threshold, then the spatial orientation relationship is determined to be a lateral offset to the left or right relative to the line to be inspected, depending on the sign of the left-right difference ratio. If the absolute value of the left-right difference ratio is greater than or equal to the lateral offset threshold, the spatial orientation relationship is determined to be a lateral offset to the left or right relative to the line to be inspected, depending on the sign of the left-right difference ratio. If the absolute value of the front-to-back difference ratio is not less than the alignment threshold and less than the longitudinal offset threshold, then the spatial orientation relationship is determined to be forward longitudinally approaching and backward longitudinally moving away from the line to be inspected, depending on the sign of the front-to-back difference ratio. If the absolute value of the front-to-back difference ratio is greater than or equal to the longitudinal offset threshold, the spatial orientation relationship is determined according to the sign of the front-to-back difference ratio: it is longitudinally approaching the line to be inspected and longitudinally moving away from it.

[0010] Furthermore, based on spatial orientation relationships, and according to ground altitude data, spatial attitude data, and heading data, corresponding flight control strategies are matched, including: Based on spatial orientation, the inspection altitude range is verified using ground altitude data, the flight action boundary is verified using spatial attitude data, and the direction of the inspection route is matched using heading data to determine the current working condition risk level and the current collision risk level. Based on the current operational risk level and the current collision risk level, determine the corresponding flight control strategy.

[0011] Furthermore, after calculating based on the pose correction amount and spatial orientation relationship, motor control commands are output, including: The pose correction amount is matched and verified with the current spatial orientation to determine the pose correction target that is compatible with the current spatial orientation. The attitude correction target is superimposed on the original attitude reference command to obtain the corrected closed-loop control target value; The closed-loop control target value is solved using a proportional-integral-derivative control law to obtain the speed control increment corresponding to each power motor. Motor control commands are generated based on the speed control increment.

[0012] Furthermore, before acquiring the magnetic field signal of the line to be inspected, the process also includes: The anti-collision control module sequentially performs short-circuit detection, open-circuit detection, zero-bias measurement, and temperature drift self-test on each acquisition channel of the inductive sensing module.

[0013] Furthermore, after driving the flight actions corresponding to the flight control strategy, it also includes: The collision avoidance control module collects the digital signal of the magnetic field strength after the flight maneuver is performed, and determines the trend of signal change based on the digital signal of the magnetic field strength after the flight maneuver is performed. The effectiveness of flight maneuvers is assessed based on signal change trends to obtain evaluation results; Adjust the mapping gain of the pose correction amount based on the evaluation results.

[0014] As an improvement to the above solution, another embodiment of the present invention provides a drone collision avoidance system based on inductance processing, comprising: The inductive sensing module is used to acquire the magnetic field signal of the line to be inspected, convert the magnetic field signal into an induced electrical signal, and then transmit the induced electrical signal to the signal conditioning module. The signal conditioning module is used to convert the induced electrical signal into a voltage signal and transmit it to the collision avoidance control module; The pose perception module is used to collect and transmit ground altitude data, spatial attitude data, and heading data to the collision avoidance control module. The collision avoidance control module performs analog-to-digital conversion on the voltage signal according to a preset sampling frequency to obtain a digital signal of magnetic field strength. Based on the digital signal of magnetic field strength, it calculates the left-right difference ratio, which characterizes the lateral deviation of the UAV, and the front-back difference ratio, which characterizes the longitudinal deviation of the UAV. Based on the left-right difference ratio, the front-back difference ratio, and a preset judgment threshold, it determines the spatial orientation relative to the line to be inspected. Based on the spatial orientation, it matches the corresponding flight control strategy according to the ground altitude data, spatial attitude data, and heading data. It maps the left-right difference ratio and the front-back difference ratio into attitude correction values, and outputs motor control commands after solving the attitude correction values ​​and spatial orientation to drive the flight actions corresponding to the flight control strategy.

[0015] By implementing this invention, at least the following beneficial effects are achieved: This invention provides a method and system for drone collision avoidance based on inductive processing. The method collects magnetic field signals from the power line to be inspected using an inductive sensing module, achieving non-contact detection using the inherent power frequency magnetic field of the energized transmission line. Magnetic field propagation is unaffected by weather or lighting conditions. The method replaces lidar with a combination of an inductive sensing module and a signal conditioning module, resulting in a simpler structure and significantly reduced overall cost. Simultaneously, the method quantifies the offset state between the drone and the line by left-right and front-back difference ratios, thereby determining the spatial orientation relative to the line. This eliminates the need to identify the outline of small-diameter lines, avoiding missed or false detections. Based on the spatial orientation, the method uses altitude data, spatial attitude data, and heading data from the pose perception module to match corresponding flight control strategies and output motor control commands, achieving efficient and stable power line inspection collision avoidance operations. Therefore, through the interaction of the inductive sensing module, signal conditioning module, pose perception module, and collision avoidance control module, both operating costs and the reliability of drone collision avoidance can be reduced. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating a drone collision avoidance method based on inductance processing according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the I-shaped inductor coil of an inductor sensing module provided in an embodiment of the present invention in a magnetic field; Figure 3 This is a schematic diagram of an LC parallel resonant circuit provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of an inductor-based anti-collision system for unmanned aerial vehicles (UAVs) according to an embodiment of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] See Figure 1 To address the issues of high inspection costs and poor collision avoidance performance in existing technologies, an embodiment of the present invention provides a flowchart of a drone collision avoidance method based on inductive processing. The method is applicable to drones; the drone is equipped with an inductive sensing module, a signal conditioning module, a pose perception module, and a collision avoidance control module. Specifically, the inductive sensing module acquires magnetic field signals based on the law of electromagnetic induction, capturing the alternating power frequency magnetic field emitted by the energized power line to be inspected and converting the magnetic signal into an induced electrical signal, unlike existing visual, laser, and ultrasonic sensing units. The signal conditioning module processes the weak, non-standardized induced electrical signal output from the inductive sensing module, converting it into a standard voltage signal that the collision avoidance control module can accurately sample. The pose perception module integrates altitude, attitude, and heading detection devices to acquire real-time spatial flight status data of the UAV. The collision avoidance control module, the main control chip mounted on the UAV, undertakes the core functions of signal analog-to-digital conversion, data processing, orientation determination, strategy matching, and control command output. The UAV is a quadcopter or hexacopter power line inspection UAV, powered by a brushless DC motor, and the flight control system supports external command superposition control.

[0019] The drone collision avoidance method based on inductance processing includes: S1. The magnetic field signal of the line to be inspected is obtained through the inductive sensing module, the magnetic field signal is converted into an induced electrical signal and then transmitted to the signal conditioning module. Specifically, the line to be inspected is a high-voltage transmission line carrying power frequency alternating current, which is the source of the magnetic field signal. When alternating current flows through the line conductors, according to the principle of electromagnetic induction, a concentric alternating magnetic field is generated around the conductors, such as... Figure 2 As shown, the strength of the alternating magnetic field decreases as the distance from the conductor increases.

[0020] To illustrate, when the drone flies along the line to be inspected, the onboard inductive sensing module captures the power frequency alternating magnetic field signal emitted by the line to be inspected in real time through the electromagnetic induction effect, and simultaneously converts the magnetic signal into a weak induced electrical signal, and transmits the induced electrical signal to the signal conditioning module in real time.

[0021] Preferably, the inductive sensing module includes several acquisition channels in various directions, and each acquisition channel includes a magnetic field sensing unit and a frequency selective filtering unit. The magnetic field signal of the line to be inspected is acquired through the inductive sensing module, converted into an induced electrical signal, and then transmitted to the signal conditioning module, including: The magnetic field signal of the line to be inspected is collected by the magnetic field induction unit in the corresponding direction, the magnetic field signal is converted into the induced electrical signal of the corresponding channel, and the induced electrical signal is transmitted to the frequency selective filtering unit. After the electromagnetic interference of non-target frequency in the induced electrical signal is filtered out by the frequency selective filtering unit, the filtered induced electrical signal is transmitted to the signal conditioning module. The target frequency of the frequency selective filter unit is matched with the power frequency AC frequency of the line to be inspected.

[0022] Specifically, the acquisition channel is an independent magnetic field acquisition and preprocessing branch within the inductive sensing module. Each channel corresponds to a fixed position of the UAV, and each channel operates independently without interference, achieving omnidirectional magnetic field acquisition. The magnetic field induction unit is the core magnetoelectric conversion device unit of the acquisition channel, a high-sensitivity I-shaped inductor that converts alternating magnetic field signals into induced electrical signals based on the law of electromagnetic induction, serving as the source of magnetic field signal acquisition. The frequency selection filtering unit is the pre-processing circuit unit within the acquisition channel, an LC parallel resonant circuit. Its core function is to allow only the target frequency signal, which matches the power frequency of the line under inspection, to pass through, filtering out electromagnetic clutter interference of other frequencies in the environment and improving signal purity. The target frequency is the resonant frequency of the frequency selection filtering unit, perfectly matching the power frequency AC frequency of the line under inspection (50Hz for domestic power grids and 60Hz for overseas power grids). Non-target frequency electromagnetic interference refers to other electromagnetic signals in the environment besides the power frequency signal of the line under inspection, including stray electromagnetic signals generated by radio stations, base stations, and power equipment, which can interfere with the accuracy of magnetic field detection.

[0023] To illustrate, the line to be inspected will generate an alternating magnetic field. The coil of the I-shaped inductor in the inductive sensing module will receive an induced current in the magnetic field, such as... Figure 2 As shown, and through as Figure 3 The LC parallel resonant circuit shown selects the frequency of the induced electromotive force. The I-shaped inductor acts as a magnetic field sensor; when placed in an alternating magnetic field, magnetic lines of force pass through the coil of the I-shaped inductor, generating an induced electromotive force (induced current) at the ends of the coil, converting the ambient magnetic field signal into a weak electrical signal. Each acquisition channel independently senses the magnetic field strength at the drone's current location. When the drone approaches the guide wire, the magnetic field strength increases, and the induced signal increases; when the drone deviates from the guide wire (e.g., to the left), the signal strength of the left sensor is significantly higher than that of the right.

[0024] Specifically, there are four sets of acquisition channels, corresponding to the front, rear, left, and right positions of the drone's fuselage. These channels correspond one-to-one with the left-right and front-rear differential ratios, enabling full-dimensional magnetic field acquisition in both horizontal and vertical directions. The LC parallel resonant circuit of the frequency selection filter unit uses high-precision inductors and capacitors, with a resonant frequency error not exceeding ±1Hz, ensuring accurate matching of the power frequency. For example, at a 50Hz power frequency, an inductor L=10mH and a capacitor C=10μF can be selected, with a resonant frequency... The I-shaped inductor of the magnetic field sensing unit uses a high-permeability manganese-zinc ferrite core with an inductance of 10mH±5%, ensuring high sensitivity acquisition of power frequency magnetic fields.

[0025] Schematic illustration: Multiple acquisition channels of the inductive sensing module operate synchronously and independently. Each channel's magnetic field sensing unit acquires the power frequency alternating magnetic field signal of the line to be inspected at the corresponding location. Based on the electromagnetic induction effect, the magnetic signal is converted into an independent induced electrical signal for the corresponding channel, and this induced electrical signal is transmitted in real time to the frequency-selective filtering unit of the same channel. The frequency-selective filtering unit receives the induced electrical signal from the corresponding channel and, utilizing the frequency selection characteristics of LC parallel resonance, allows only valid signals consistent with the target frequency to pass through, filtering out all environmental electromagnetic interference of other frequencies in the induced electrical signal, resulting in a high-purity power frequency induced electrical signal. The pure induced electrical signals output by the frequency-selective filtering units of each channel are synchronously transmitted to the back-end signal conditioning module, providing a high signal-to-noise ratio input signal for subsequent signal processing.

[0026] In a preferred embodiment of the present invention, one set of acquisition channels is deployed at each of the four positions (front, rear, left, and right) of the UAV fuselage. Each set of channels includes a magnetic field sensing unit with a 10mH I-shaped inductor and a frequency selective filter unit with LC parallel resonant circuitry. The target frequency matches the 50Hz power frequency of the domestic power grid. During the inspection, the magnetic field sensing unit of the left channel acquires the 50Hz power frequency magnetic field of the line, converts it into a 20mV induced electrical signal, and transmits it to the frequency selective filter unit. After filtering out noise from radio stations and base stations such as 100Hz and 433MHz in the environment, a clean 50Hz power frequency induced electrical signal is output to the signal conditioning module to ensure that the subsequent left-right difference ratio calculation is interference-free and distortion-free.

[0027] This embodiment achieves full-dimensional magnetic field perception of the UAV in the horizontal and vertical directions through independent acquisition of four-directional channels, providing an accurate signal source for calculating the left-right difference ratio and front-back difference ratio, and ensuring the omnidirectionality of orientation determination. Through a dedicated power frequency selective filtering unit, environmental electromagnetic noise interference is filtered out from the source, which greatly improves the signal-to-noise ratio of the front-end signal and avoids distortion of the difference ratio calculation and misjudgment of orientation caused by interference signals. At the same time, the hardware architecture is simple and highly stable, requiring no complex calibration, and is suitable for long-term stable operation in complex outdoor inspection scenarios.

[0028] Preferably, before acquiring the magnetic field signal of the line to be inspected, the method further includes: The anti-collision control module sequentially performs short-circuit detection, open-circuit detection, zero-bias measurement, and temperature drift self-test on each acquisition channel of the inductive sensing module.

[0029] Specifically, short-circuit detection involves checking the continuity of the acquisition channel circuit. By measuring the DC resistance at both ends of the channel, it determines whether a short circuit fault has occurred. If the resistance is less than 10Ω, the channel is considered short-circuited. Open-circuit detection involves checking the continuity of the acquisition channel circuit. By measuring the DC resistance at both ends of the channel, it determines whether a circuit break or component desoldering fault has occurred. If the resistance is greater than 1MΩ, the channel is considered open-circuited. Short-circuit and open-circuit detection ensures the integrity of the hardware link. If the sensor is damaged or the wiring is loose, the drone will immediately report an error and prevent takeoff, preventing the collision avoidance system from failing due to sensor failure.

[0030] Specifically, zero-bias measurement involves continuously acquiring the channel's output signal hundreds of times in an environment free from external magnetic field interference, and taking the average value as the zero-point offset value for that channel. This value is used for zero-bias compensation in subsequent signal processing, eliminating detection errors caused by the sensor's inherent zero-bias. Temperature drift self-test involves acquiring the channel's zero-point offset value changes at different temperatures, establishing a temperature-zero-bias compensation model, and using it for temperature drift compensation in subsequent signal processing. This eliminates sensor zero-point drift errors caused by changes in ambient temperature, ensuring consistent detection accuracy in environments with high temperatures in southern summers or low temperatures in northern winters.

[0031] Before acquiring magnetic field signals, hardware continuity testing is used to troubleshoot channel faults, and self-testing of zero bias and temperature drift eliminates inherent sensor errors. This proactively checks for short circuits and open circuits in the acquisition channel, preventing signal loss and misjudgment due to hardware failure, thus avoiding collision accidents at the source and improving the system's hardware reliability. Furthermore, self-testing of zero bias and temperature drift eliminates detection errors caused by inherent sensor zero bias and temperature drift, significantly improving the accuracy of magnetic field signal acquisition and ensuring the precision of subsequent difference ratio calculations and azimuth determination. This is a crucial guarantee for the long-term stable operation of the UAV.

[0032] S2. The induced electrical signal is converted into a voltage signal by the signal conditioning module and transmitted to the anti-collision control module; Schematic illustration: The UAV's signal conditioning module receives the weak induced electrical signal from the front end, performs protection, amplification, and standardization processing on the signal, converts the non-standardized induced electrical signal into a stable voltage signal that the collision avoidance control module can accurately sample, and transmits it to the collision avoidance control module in real time. An active circuit composed of operational amplifiers (Op-Amp) is used to amplify, bias, and limit the weak power frequency analog signal obtained from the inductive sensing module.

[0033] Preferably, the induced electrical signal is converted into a voltage signal and transmitted to the collision avoidance control module via a signal conditioning module, including: The induced electrical signal is subjected to amplitude limiting protection processing through the signal conditioning module, which limits the voltage amplitude of the induced electrical signal within a preset safety threshold. The induced electrical signal after the amplitude limiting protection is linearly amplified to a preset amplitude range. The linearly amplified induced electrical signal is subjected to level bias processing and low-pass filtering processing. After being superimposed with a preset DC bias voltage, a voltage signal is obtained and transmitted to the anti-collision control module.

[0034] Specifically, the amplitude limiting protection is a pre-processing protection step of the signal conditioning module, implemented using Schottky or Zener diodes. Its core function is to limit the instantaneous voltage amplitude of the induced electrical signal within the safe withstand voltage range of the downstream circuitry, preventing the instantaneous high-voltage signal induced near high-voltage transmission lines from burning out the amplifier chip and the main control ADC interface. Linear amplification is achieved through an instrumentation amplifier, linearly amplifying the weak induced electrical signal by a fixed factor without signal distortion. This raises the millivolt-level weak signal to an amplitude range suitable for the downstream ADC sampling, improving sampling accuracy. The preset safety threshold is the maximum withstand voltage of the main control chip's ADC interface, serving as the upper limit voltage for amplitude limiting protection. The preset amplitude range is the voltage range suitable for the main control ADC sampling range, primarily 0-3.3V, representing the target amplitude range for linear amplification. Level biasing is implemented through an adder circuit built with operational amplifiers, superimposing a fixed DC bias voltage onto the alternating positive and negative induced electrical signal, raising the overall alternating signal to a positive voltage range to meet the sampling requirements of a unipolar ADC. Low-pass filtering is achieved through a second-order Butterworth low-pass filter circuit, which filters out high-frequency noise and residual clutter introduced during signal amplification, retaining only the effective power frequency signal and further improving the signal-to-noise ratio. The preset DC bias voltage is half of the ADC sampling range, such as 1.65V for a 3.3V range, ensuring that the superimposed alternating signals are all in the positive voltage range with no negative voltage signals.

[0035] Specifically, the limiting protection uses an SMBJ3.3CA bidirectional TVS diode with a preset safety threshold of 3.3V, which is compatible with the voltage withstand requirements of the ADC of mainstream STM32 main control; the linear amplification uses an INA128 instrumentation amplifier with a magnification factor of 100, which can amplify a weak 20mV signal to 2V, falling within the preset amplitude range of 0-3.3V; the low-pass filter has a cutoff frequency of 100Hz, retaining only the 50Hz power frequency effective signal and filtering out high-frequency noise introduced during amplification; the level bias uses an OP07 operational amplifier to build an adder circuit with a preset DC bias voltage of 1.65V, ensuring that the amplified alternating signals are all within the 0-3.3V positive voltage range after superposition.

[0036] Schematic illustration: The signal conditioning module receives the induced electrical signal output from the inductor sensing module. First, it uses a bidirectional TVS diode to limit the signal amplitude, strictly restricting the instantaneous voltage amplitude within a preset safety threshold. This provides overvoltage protection for the downstream circuitry, preventing instantaneous high voltage from burning out components. The limited-amplitude induced electrical signal is then linearly amplified by an instrumentation amplifier, expanding the weak millivolt-level alternating induced electrical signal to a preset amplitude range. This ensures the signal amplitude matches the sampling range of the downstream ADC, resolving the issue of inaccurate sampling of weak signals. The linearly amplified induced electrical signal is then superimposed with a preset DC bias voltage through an adder circuit to perform level biasing, converting the alternating positive and negative AC signal into a unipolar positive voltage signal. A low-pass filter circuit then filters out high-frequency noise and residual clutter, ultimately obtaining a high signal-to-noise ratio standard voltage signal suitable for ADC sampling. This processed standard voltage signal is then transmitted in real-time and synchronously to the ADC sampling interface of the collision avoidance control module.

[0037] In a preferred embodiment of the present invention, the initial amplitude of the induced electrical signal output by the inductor sensing module is 20mV, and the instantaneous peak value can reach up to 5V. After the signal enters the signal conditioning module, it is first limited by an SMBJ3.3CA TVS diode to lock the voltage peak value within 3.3V; then it is linearly amplified by an INA128 instrumentation amplifier by 100 times, and the signal amplitude is increased to about 2V; then a 1.65V DC bias is superimposed by an adder circuit to convert the alternating signal into a unipolar positive voltage signal of 0.65V-3.3V; finally, it is filtered by a second-order low-pass filter with a 100Hz cutoff frequency to remove high-frequency noise and obtain a standard voltage signal, which is transmitted in real time to the ADC interface of the collision avoidance control module, and the main controller can complete the signal sampling without distortion and with high precision.

[0038] By implementing this embodiment, the amplitude limiting protection completely avoids the risk of hardware burnout caused by close-range high-voltage induction, improving the hardware safety and stability of the system; linear amplification and level bias processing adapt to the ADC sampling requirements of the main controller, solving the core problem that weak induction signals cannot be accurately sampled, and significantly improving the resolution and accuracy of signal sampling; post-low-pass filtering further eliminates noise introduced during signal processing, improves the signal-to-noise ratio, and avoids subsequent calculation errors caused by sampling distortion.

[0039] S3. Through the pose perception module, collect and transmit ground altitude data, spatial attitude data and heading data to the collision avoidance control module; To illustrate, the UAV's attitude perception module collects the UAV's ground altitude data, pitch, roll, and yaw spatial attitude data, and flight heading data in real time, and transmits all attitude data synchronously to the collision avoidance control module to provide a flight status verification basis for subsequent collision avoidance strategy matching.

[0040] Specifically, the pose perception module uses a barometer and a laser rangefinder to assist in altitude positioning, and an IMU to assist in orientation positioning. The barometer utilizes the physical property that air pressure decreases with increasing altitude (pressure-altitude formula), and incorporates a high-precision piezoelectric or capacitive pressure sensing element to monitor changes in ambient atmospheric pressure in real time. The laser rangefinder employs the time-of-flight (ToF) method, emitting laser pulses downwards or in the direction of a guide, and measuring the time interval between the pulse's emission and return. Using the formula (Where c is the speed of light) Calculate distance d. The IMU (Inertial Measurement Unit) integrates a three-axis accelerometer and a three-axis gyroscope, and is usually paired with a magnetometer (electronic compass). The gyroscope measures the drone's angular velocity, the accelerometer senses the gravity vector, and complementary filtering or Kalman filtering algorithms are used to calculate the drone's pitch, roll, and yaw angles in real time. The magnetometer determines the drone's absolute heading relative to true north by sensing the Earth's magnetic field. In environments where the drone sways due to strong winds or other factors, the IMU can compensate for attitude deviations in real time, ensuring that the collision avoidance algorithm does not misjudge due to drone vibrations.

[0041] S4. Through the anti-collision control module, the voltage signal is converted from analog to digital according to the preset sampling frequency to obtain the magnetic field strength digital signal. Based on the magnetic field strength digital signal, the left-right difference ratio and the front-back difference ratio, which are used to characterize the lateral deviation of the UAV, are calculated respectively. Specifically, the left-right difference ratio is calculated using the magnetic field signals from the left and right sides of the UAV. It uniquely characterizes the lateral offset direction and degree of the UAV relative to the line to be inspected. Positive and negative values ​​correspond to the offset direction, and the absolute value corresponds to the offset degree. The forward-backward difference ratio is calculated using the magnetic field signals from the forward and backward sides of the UAV. It uniquely characterizes the longitudinal offset / approach direction and degree of the UAV relative to the line to be inspected. Positive and negative values ​​correspond to the approach / remote direction, and the absolute value corresponds to the approach / remote degree. The analog-to-digital conversion sampling precision is no less than 12 bits to ensure the sampling resolution of the magnetic field signals.

[0042] Schematic illustration: The drone's collision avoidance control module performs analog-to-digital conversion on the received voltage signal according to a preset sampling frequency, obtaining a digitized magnetic field strength digital signal. Based on the magnetic field strength digital signals from each direction, it calculates two core characteristic quantities: the left-right difference ratio, which characterizes lateral offset, and the front-back difference ratio, which characterizes longitudinal offset. The preset sampling frequency is an integer multiple of the power frequency of the line to be inspected; for example, a 50Hz power frequency corresponds to a 100Hz / 200Hz sampling frequency, which conforms to the Nyquist sampling theorem and ensures distortion-free sampling.

[0043] Preferably, the left-right difference ratio, which characterizes the lateral deviation of the UAV, and the front-back difference ratio, which characterizes the longitudinal deviation of the UAV, are calculated based on the digital signal of magnetic field strength, including: Based on the digital magnetic field strength signals, the corresponding digital magnetic field strength signals of the front, back, left, and right acquisition channels are determined respectively. The digital magnetic field strength signals of each acquisition channel are normalized to obtain the characteristic values ​​of the magnetic field strength of each acquisition channel. Based on the preset adjustment parameters and the magnetic field strength characteristic values ​​of the left and right acquisition channels, the left-right difference ratio, which is used to characterize the lateral offset of the UAV, is calculated. Based on the preset adjustment parameters and the magnetic field strength characteristic values ​​of the front and rear acquisition channels, the front-rear difference ratio, which is used to characterize the longitudinal offset of the UAV, is calculated. Among them, the positive and negative values ​​and magnitudes of the left-right difference ratio correspond to the lateral offset direction and degree of the UAV relative to the line to be inspected, respectively; the positive and negative values ​​and magnitudes of the front-back difference ratio correspond to the longitudinal offset direction and degree of the UAV relative to the line to be inspected, respectively.

[0044] Specifically, the normalization process employs an extremum normalization algorithm to map the digital magnetic field strength signals from different acquisition channels to a unified numerical range of [0,1]. This eliminates the overall differences in magnetic field amplitude caused by factors such as the wire diameter, current intensity, and absolute distance between the drone and the line being inspected, retaining only the relative difference characteristics between each channel. The magnetic field strength characteristic value is the standardized value obtained after normalizing the digital magnetic field strength signals from each acquisition channel. The preset adjustment parameter, denoted by η, is a fixed, extremely small positive number. Its core functions are twofold: to prevent the denominator from being zero during difference ratio calculation, thus avoiding calculation errors; and to fine-tune the sensitivity of the difference ratio calculation to adapt to the detection needs of different inspection scenarios.

[0045] Specifically, the lateral offset direction refers to the horizontal left-right deviation of the drone relative to the line to be inspected, categorized as leftward or rightward deviation, corresponding to a positive or negative left-right difference ratio. The degree of lateral offset is the magnitude of the horizontal deviation of the drone relative to the center of the line to be inspected, corresponding to the absolute value of the left-right difference ratio; the larger the absolute value, the more severe the deviation. The longitudinal offset direction refers to the forward / backward movement of the drone relative to the line to be inspected, categorized as forward approaching the line or backward moving away from the line, corresponding to a positive or negative forward / backward difference ratio. The degree of longitudinal offset is the magnitude of the forward / backward movement of the drone relative to the line to be inspected, corresponding to the absolute value of the forward / backward difference ratio; the larger the absolute value, the more severe the approach / departure. A positive left-right difference ratio indicates the drone is shifting to the right and the line is to the left of the drone; a negative left-right difference ratio indicates the drone is shifting to the left and the line is to the right of the drone; a positive forward / backward difference ratio indicates the drone is approaching the line; a negative forward / backward difference ratio indicates the drone is moving away from the line.

[0046] Specifically, the left-right difference ratio The calculation formula is: Before and after difference ratio The calculation formula is: L, R, F, and B are the normalized magnetic field strength characteristic values ​​of the left, right, front, and rear channels, respectively, and η is a preset adjustment parameter. The preset adjustment parameter η ranges from 0.001 to 0.01. In a normal inspection scenario, the default value is 0.001, which effectively prevents the denominator from being 0 and does not have a substantial impact on the difference ratio calculation result.

[0047] Schematic illustration: The collision avoidance control module splits the digital magnetic field strength signal obtained from analog-to-digital conversion according to the acquisition channel, extracting independent digital magnetic field strength signals corresponding to the four acquisition channels: front, rear, left, and right. The magnetic field strength signals of the four channels are uniformly normalized to eliminate overall amplitude differences caused by line diameter, current magnitude, and the absolute distance between the UAV and the line, obtaining standardized magnetic field strength characteristic values ​​for each channel. The magnetic field strength characteristic values ​​of the left and right acquisition channels are combined with a preset adjustment parameter η and substituted into the difference ratio calculation formula to calculate the left-right difference ratio, which characterizes the lateral offset direction and degree of the UAV. Similarly, the magnetic field strength characteristic values ​​of the front and rear acquisition channels are combined with the preset adjustment parameter η and substituted into the difference ratio calculation formula to calculate the front-rear difference ratio, which characterizes the longitudinal offset direction and degree of the UAV.

[0048] In a preferred embodiment of the present invention, the anti-collision control module samples the magnetic field strength digital signals of the four channels (front, rear, left, and right) as 2.0V, 1.2V, 1.0V, and 2.2V, respectively. After extreme value normalization, the characteristic values ​​of the magnetic field strength of the four channels are obtained as 0.5, 0.1, 0, and 1.0, respectively. The preset adjustment parameter η = 0.001 is substituted into the formula to calculate the left-right difference ratio W. LR =(0-1.0) / (0+1.0+0.001)≈−0.999, the ratio of the differences before and after is W FB =(0.5-0.1) / (0.5+0.1+0.001)≈0.666; According to the calibration rules, it is determined that the UAV deviated significantly to the left of the line while approaching the line forward. The degree of deviation and approach is relatively serious, which provides accurate quantitative basis for subsequent orientation determination and strategy matching.

[0049] By implementing this embodiment, the difference ratio calculation method is adopted, and only the relative offset feature is retained. Regardless of the distance between the UAV and the line, the relative offset can be accurately characterized, and the detection robustness is greatly improved. The preset adjustment parameters avoid the calculation error of the denominator being 0, ensuring the stability of the system operation. At the same time, the detection sensitivity can be flexibly adjusted to adapt to different inspection scenarios.

[0050] Schematic illustration: After obtaining the digital signal of magnetic field strength, it needs to be preprocessed to ensure that the collision avoidance judgment is not triggered by electrode vibration or transient electromagnetic pulses. Preprocessing includes extremum removal, median filtering, and rolling mean filtering. Extremum removal (removing single spikes) uses an amplitude-limiting comparison method, with a preset sliding window (e.g., N=5 sampling points), setting the current sampled value... The average value of this window Compare them. If (threshold) If the background noise level is dynamically set, then the point is determined to be a sudden pulse interference, the point is removed, and replaced with the effective value or window mean of the previous moment.

[0051] Median filtering (eliminating impulse noise) sorts M consecutively acquired sample values ​​(e.g., M=3 or 5) by numerical value and takes the median value as the representative value for the current moment. This operation effectively filters isolated noise points generated by UAV motor commutation or electrostatic discharge without causing phase shift at the waveform edges. Rolling mean filtering (smoothing the envelope) establishes a FIFO (First-In-First-Out) circular queue in memory. By iteratively calculating the arithmetic mean within the window, it filters out high-frequency white noise in the signal, smoothing the final output induction intensity curve and extracting a stable magnetic field strength envelope.

[0052] S5. Determine the spatial orientation relative to the line to be inspected based on the left-right difference ratio, front-back difference ratio, and preset judgment threshold. Specifically, the spatial orientation relationship refers to the relative position and attitude of the UAV with respect to the power transmission line to be inspected, including four core states: alignment with the line center, lateral offset, longitudinal approach, and longitudinal distance. The preset judgment thresholds are pre-calibrated based on the UAV's safe inspection distance, normal flight speed, and line voltage level, and can be adjusted according to the on-site conditions before the inspection.

[0053] To illustrate, the drone's collision avoidance control module compares the calculated left-right difference ratio and front-back difference ratio with preset judgment thresholds. Based on the comparison results, it determines the drone's current spatial orientation relative to the line to be inspected, and clarifies whether the drone is aligned with the line, the direction of deviation, the degree of deviation, and the risk of collision.

[0054] Preferably, the preset judgment thresholds include: an alignment threshold, a horizontal offset threshold, and a vertical offset threshold; Based on the left-right difference ratio, front-back difference ratio, and preset judgment thresholds, the spatial orientation relative to the line to be inspected is determined, including: The absolute value of the left-right difference ratio is compared with the alignment threshold and the horizontal offset threshold, respectively; the absolute value of the front-back difference ratio is compared with the alignment threshold and the vertical offset threshold, respectively. If the absolute values ​​of both the left-right difference ratio and the front-back difference ratio are less than the alignment threshold, the spatial orientation relationship is determined to be aligned with the center of the line to be inspected. If the absolute value of the left-right difference ratio is not less than the alignment threshold and less than the lateral offset threshold, then the spatial orientation relationship is determined to be a lateral offset to the left or right relative to the line to be inspected, depending on the sign of the left-right difference ratio. If the absolute value of the left-right difference ratio is greater than or equal to the lateral offset threshold, the spatial orientation relationship is determined to be a lateral offset to the left or right relative to the line to be inspected, depending on the sign of the left-right difference ratio. If the absolute value of the front-to-back difference ratio is not less than the alignment threshold and less than the longitudinal offset threshold, then the spatial orientation relationship is determined to be forward longitudinally approaching and backward longitudinally moving away from the line to be inspected, depending on the sign of the front-to-back difference ratio. If the absolute value of the front-to-back difference ratio is greater than or equal to the longitudinal offset threshold, the spatial orientation relationship is determined according to the sign of the front-to-back difference ratio: it is longitudinally approaching the line to be inspected and longitudinally moving away from it.

[0055] Specifically, the alignment threshold is the smallest critical value among the preset judgment thresholds, used to determine whether the drone is aligned with the center of the route to be inspected. When the absolute values ​​of both the left-right difference ratio and the front-back difference ratio are less than this threshold, the drone is determined to be in a center-aligned state. The lateral offset threshold is a critical value among the preset judgment thresholds used to determine the degree of lateral offset of the drone. Its value is greater than the alignment threshold, used to distinguish between slight lateral offset and significant lateral offset. The longitudinal offset threshold is a critical value among the preset judgment thresholds used to determine the degree of longitudinal approach / distance of the drone. Its value is greater than the alignment threshold, used to distinguish between slight longitudinal approach / distance and significant longitudinal approach / distance.

[0056] Schematic, the alignment threshold is less than the horizontal offset threshold, and the alignment threshold is less than the vertical offset threshold.

[0057] Schematic: If the absolute values ​​of both the left-right difference ratio and the front-back difference ratio are less than the alignment threshold, the UAV's current spatial orientation is determined to be aligned with the center of the inspection route, in a standard inspection pose. If the absolute value of the left-right difference ratio is not less than the alignment threshold and less than the lateral offset threshold, the spatial orientation is determined to be either slightly offset to the left or slightly offset to the right relative to the inspection route, depending on the sign of the left-right difference ratio. If the absolute value of the left-right difference ratio is greater than or equal to the lateral offset threshold, the spatial orientation is determined to be significantly offset to the left or slightly offset to the right relative to the inspection route, depending on the sign of the left-right difference ratio. If the absolute value of the front-to-back difference ratio is not less than the alignment threshold and less than the longitudinal offset threshold, the spatial orientation relationship is determined as either slightly approaching or slightly moving away from the line to be inspected, depending on the sign of the front-to-back difference ratio. If the absolute value of the front-to-back difference ratio is greater than or equal to the longitudinal offset threshold, the spatial orientation relationship is determined as either significantly approaching or significantly moving away from the line to be inspected, depending on the sign of the front-to-back difference ratio. If multiple conditions for lateral and longitudinal offset are met simultaneously, the composite spatial orientation relationship of the UAV is determined by superimposing these conditions, such as a significant lateral offset to the right and a slight forward longitudinal approach.

[0058] Schematic, the alignment threshold is 0.2, the lateral offset threshold is 0.5, and the longitudinal offset threshold is 0.5, which can be adjusted on-site according to the inspection safety distance and flight speed. The logical relationship of the threshold intervals is as follows: [0, alignment threshold) is the alignment interval, [alignment threshold, offset threshold) is the slight offset interval, and [offset threshold, +∞) is the significant offset interval.

[0059] In a preferred embodiment of the present invention, the preset alignment threshold is 0.2, the lateral offset threshold is 0.5, and the longitudinal offset threshold is 0.5. The anti-collision control module calculates the left-right difference ratio as 0.6 and the front-back difference ratio as 0.1. Comparison shows that the absolute value of the left-right difference ratio (0.6) is greater than or equal to the lateral offset threshold (0.5) and is positive, indicating a significant rightward lateral offset; the absolute value of the front-back difference ratio (0.1) is less than the alignment threshold (0.2), indicating no longitudinal offset. The final determination of the UAV's spatial orientation is: longitudinally aligned with the inspection route and significantly laterally offset to the right, providing a clear decision-making basis for subsequent strategy matching.

[0060] By implementing this embodiment, the degree of offset / approach is divided into levels based on threshold intervals, providing a direct basis for subsequent graded collision avoidance strategy matching and realizing a one-to-one correspondence between risk level and orientation status. The logic of absolute value comparison and positive and negative direction determination simplifies the determination process, improves determination efficiency, and ensures the real-time performance of control.

[0061] S6. Based on spatial orientation, match the corresponding flight control strategy according to ground altitude data, spatial attitude data and heading data; To illustrate, the collision avoidance control module of the drone uses the determined spatial orientation relationship as the core decision-making basis, and combines synchronously collected ground altitude data, spatial attitude data, and heading data to perform multi-dimensional safety verification, and matches a flight control strategy that is fully adapted to the current working conditions and collision risk level.

[0062] Preferably, based on spatial orientation relationships, and according to ground altitude data, spatial attitude data, and heading data, a corresponding flight control strategy is matched, including: Based on spatial orientation, the inspection altitude range is verified using ground altitude data, the flight action boundary is verified using spatial attitude data, and the direction of the inspection route is matched using heading data to determine the current working condition risk level and the current collision risk level. Based on the current operational risk level and the current collision risk level, determine the corresponding flight control strategy.

[0063] Specifically, the inspection altitude range is a preset safe inspection altitude range for drones, divided into a minimum safe altitude and a maximum inspection altitude, to prevent drones from colliding with poles at too low an altitude or deviating from the inspection position at too high an altitude. The flight maneuver boundaries are preset safe attitude ranges for drones, including the maximum safe angles for pitch, roll, and yaw, to prevent drones from crashing while performing obstacle avoidance maneuvers when their attitude is out of control. The inspection route alignment is determined by matching heading data with the route alignment to ensure that the drone's correction maneuvers always follow the route and do not deviate from the inspection route. The operational risk level is a risk level based on the drone's altitude above the ground, spatial attitude, heading, and other flight states, representing the safety of the drone's own flight state. The collision risk level is a collision risk level based on the spatial orientation of the drone relative to the route, representing the probability and severity of a collision between the drone and the route. The flight control strategy is a standardized collision avoidance flight plan developed for different risk levels, clearly defining the flight speed, flight maneuvers, correction amplitude, and safety protection rules under different operating conditions.

[0064] To illustrate, ground altitude data is used to verify whether the UAV is within a preset inspection altitude range to determine its altitude safety status; spatial attitude data is used to verify whether the UAV is within preset flight action boundaries to determine its attitude stability status; and heading data is used to match the direction of the inspection route to determine its heading compliance status. Combining the collision risk based on spatial orientation and the operational risk from the three verifications, a comprehensive determination is made of the current operational risk level and the corresponding collision risk level. These two risk levels are categorized into five levels: no risk, low risk, medium risk, high risk, and critical risk. Based on the determined current operational risk level and collision risk level, a flight control strategy corresponding to each risk level is matched to ensure that the strategy is fully adapted to the current operational condition and collision risk, achieving graded collision avoidance control.

[0065] Schematic illustration: The current operational risk level corresponds one-to-one with the current collision risk level, divided into five levels: no risk, low risk, medium risk, high risk, and critical risk. The triggering conditions and corresponding flight control strategies for each level are as follows: When the spatial orientation is aligned with the center of the line to be inspected, the ground altitude data is within the preset inspection altitude range, and the spatial attitude data is within the stable flight range, it is determined to be at the no-risk level. The tracking alignment mode is matched to maintain the center relative attitude between the UAV and the line to be inspected, and routine inspection flight is performed. When the spatial orientation is laterally offset... When the flight path shifts or approaches longitudinally, and the offset or approach amount is within a slight range between the alignment threshold and the corresponding lateral offset threshold and longitudinal offset threshold, without any attitude or altitude anomalies exceeding the safety boundary, it is judged as a low-risk level. A warning correction mode is then applied. First, the direction of the route to be inspected is locked based on the heading data. The flight speed is reduced to a preset low inspection speed, and simultaneously, a small pose correction command opposite to the offset direction is output to gradually correct the flight trajectory to the alignment state with the route center. When the spatial orientation relationship is a lateral offset or a forward longitudinal approach, and the offset or approach amount reaches or exceeds the lateral offset threshold... When there are no abnormal attitudes or altitudes exceeding the safety boundaries, the risk level is determined to be medium. Active obstacle avoidance mode is then activated, combining ground altitude data and spatial attitude data to lock a safe obstacle avoidance path. A large-stroke attitude correction command is output, driving the UAV to perform lateral translation, vertical climb, or directional backward movement to quickly move away from the inspection route and eliminate collision risk. When the spatial orientation relationship is a lateral offset or forward / vertical approach, and the offset or approach exceeds the lateral offset threshold or longitudinal offset threshold by two times or more, or when the ground altitude data is below the preset safe altitude or the spatial attitude data exceeds the stable flight range, the risk level is determined to be high. Forced avoidance mode is activated, immediately cutting off the inspection mission command, locking the current safe heading, and driving the UAV to perform a vertical climb to a safe altitude and then hover, simultaneously outputting a risk warning signal. When sensor data is continuously abnormal, the spatial orientation relationship determination exceeds the preset safety boundary for three or more consecutive sampling periods, or the collision risk reaches the preset critical value, the risk level is determined to be critical. Emergency safety protection mode is activated, driving the UAV to perform hovering, automatic return, or emergency power-off safety protection actions.

[0066] In a preferred embodiment of the present invention, the risk-free level corresponds to the tracking alignment mode, which is suitable for center alignment and normal flight conditions; the low-risk level corresponds to the warning correction mode, which is suitable for slight deviation / approach and normal flight conditions; the medium-risk level corresponds to the active obstacle avoidance mode, which is suitable for significant deviation / approach and normal flight conditions; the high-risk level corresponds to the forced avoidance mode, which is suitable for severe deviation / approach or abnormal flight conditions; and the critical risk level corresponds to the emergency safety protection mode, which is suitable for sensor malfunction and critical collision risk conditions.

[0067] By implementing this embodiment, corresponding strategies are matched for different risk levels, ensuring the continuity of inspection operations while maximizing collision avoidance safety and avoiding over-correction or under-correction. Multi-dimensional safety verification is performed by combining pose data to ensure that the UAV is in a safe flight state when collision avoidance actions are executed, avoiding crashes caused by obstacle avoidance actions when attitude is out of control or altitude is abnormal, thus improving the overall safety of the system. Using spatial orientation relationship as the core matching basis and pose data as the auxiliary verification boundary, the logic is clear, the decision-making efficiency is high, and the real-time matching of strategies is guaranteed.

[0068] S7. Map the left-right difference ratio and the front-back difference ratio into attitude correction values. After calculating based on the attitude correction values ​​and spatial orientation relationships, output motor control commands to drive the execution of flight actions corresponding to the flight control strategy.

[0069] Specifically, the pose correction is a numerical adjustment of the UAV's position and attitude based on the left-right difference ratio and front-back difference ratio mapping, including lateral translation, longitudinal translation, and attitude adjustment, used to correct the UAV's flight trajectory and avoid collision risks. The motor control command is a standardized PWM control signal that drives the UAV's power motor, used to control the UAV's ascent, descent, translation, turning, speed adjustment, and other flight actions, enabling the implementation of the collision avoidance strategy.

[0070] Specifically, the nonlinear difference ratio signal is converted into a linear pose correction. This is achieved through mapping functions using either linear gain or piecewise functions. For example, the roll angle correction... left-right difference ratio index The mapping formula is: ,in, This is the mapping ratio, preset based on the drone's sensitivity and collision avoidance distance. The saturation limiting function ensures that the correction amount for a single channel does not exceed the maximum safe tilt angle of the drone (e.g., To prevent rollover due to interference, This is the maximum roll angle. A threshold is set to prevent frequent motor vibration caused by minor disturbances during inspections directly above the conductor. Only when Only then does the system generate a non-zero pose correction, ensuring the stability of the inspection flight.

[0071] To illustrate, the collision avoidance control module of the UAV maps the left-right difference ratio and the front-back difference ratio into the pose correction amount in the corresponding direction. It performs closed-loop control calculation on the pose correction amount in combination with the spatial orientation relationship, and finally outputs standardized motor control commands to drive the power actuator of the UAV to execute the flight actions corresponding to the matched flight control strategy, so as to achieve trajectory correction and collision avoidance.

[0072] Preferably, the motor control command is output after calculation based on the pose correction amount and spatial orientation relationship, including: The pose correction amount is matched and verified with the current spatial orientation to determine the pose correction target that is compatible with the current spatial orientation. The attitude correction target is superimposed on the original attitude reference command to obtain the corrected closed-loop control target value; The closed-loop control target value is solved using a proportional-integral-derivative control law to obtain the speed control increment corresponding to each power motor. Motor control commands are generated based on the speed control increment.

[0073] Specifically, the attitude correction target is the UAV target attitude and position determined by combining the attitude correction amount and the spatial orientation relationship, including lateral translation target, longitudinal translation target, and attitude adjustment target, which is the target value for control. The original attitude reference command is the routine inspection flight attitude command output by the UAV's underlying layer, which is the basic control command for normal UAV inspection. The closed-loop control target value is the final control target value obtained by superimposing the attitude correction target onto the original attitude reference command. It retains the flight command for normal inspection and superimposes the correction command for collision avoidance and correction, realizing the synchronous execution of inspection and collision avoidance. The proportional-integral-derivative control law is a PID closed-loop control algorithm commonly used in the field of industrial control. This invention adopts a position-type cascade PID control law to solve the closed-loop control target value into the motor speed control increment, realizing smooth and oscillation-free closed-loop control. The speed control increment is the speed adjustment value of each power motor of the UAV, corresponding to the four control channels of the UAV: ​​roll, pitch, yaw, and throttle, which directly determines the flight behavior of the UAV. The motor control command is a standardized PWM signal with a frequency of 50Hz corresponding to the speed control increment. It is directly output to the UAV's ESC and brushless motor to drive the motor to perform the corresponding speed adjustment.

[0074] Specifically, the UAV receives the raw attitude reference (Target Attitude) from the remote controller or mission payload. The calculated pose correction is then... (Roll angle correction) or The pitch correction is added to the original command in real time to obtain the corrected attitude reference. , The initial attitude is the original attitude reference command. The final attitude reference input is fed into the flight controller's built-in PID (Proportional-Integral-Derivative) controller: P (Proportional) responds in real time to changes in the differential ratio, generating an instantaneous reverse obstacle avoidance torque. I (Integral) eliminates static position deviations caused by continuous wind yaw or guide wire attraction. D (Derivative) senses the rate at which the UAV approaches the guide wire (rate of change of the differential ratio), suppressing overshoot through damping to prevent excessive obstacle avoidance. The controller's final output is a corrected four- or six-channel motor PWM duty cycle command or torque increment. These commands directly change the motor speed, causing the UAV to generate a horizontal acceleration away from the guide wire, thus achieving non-contact automatic obstacle avoidance.

[0075] Indicatively, the mapped pose correction is matched and verified with the currently determined spatial orientation to confirm that the direction and magnitude of the pose correction are fully adapted to the current orientation state. This identifies the attitude correction target that precisely matches the current spatial orientation, avoiding incorrect correction direction or excessive correction magnitude. The determined attitude correction target is superimposed on the original attitude reference command of the UAV flight control system, fusing to obtain the corrected closed-loop control target value. This achieves synchronous fusion of normal inspection commands and collision avoidance correction commands without interrupting normal inspection operations. The fused closed-loop control target value is input into a preset proportional-integral-derivative control law for flight control calculation, obtaining the speed control increments corresponding to the four power motors of the UAV, realizing the conversion from control target to motor action. Based on the calculated speed control increments of each motor, standardized motor PWM control commands are generated and output to the UAV's ESC and power actuators.

[0076] To illustrate, a second-order ramp function or S-curve limiting is used to limit the variation in motor torque within each frame (typically 4ms). Even with extremely large correction commands, the system smoothly distributes them across multiple control cycles, thus avoiding rack structure vibration, high battery current discharge, or IMU sensor data overflow caused by instantaneous acceleration.

[0077] By implementing this embodiment, normal inspection and collision avoidance correction are executed simultaneously through instruction superposition and fusion, without interrupting the inspection operation, which greatly improves the inspection efficiency. The industrial-grade PID closed-loop control law is adopted, and the control action is smooth without shock or oscillation, and the trajectory correction is accurate, which greatly improves the flight stability of the UAV and avoids the risk of crash due to excessive actions.

[0078] Preferably, after driving the flight maneuver corresponding to the flight control strategy, the method further includes: The collision avoidance control module collects the digital signal of the magnetic field strength after the flight maneuver is performed, and determines the trend of signal change based on the digital signal of the magnetic field strength after the flight maneuver is performed. The effectiveness of flight maneuvers is assessed based on signal change trends to obtain evaluation results; Adjust the mapping gain of the pose correction amount based on the evaluation results.

[0079] Specifically, the signal change trend refers to the numerical changes in the digital signals of magnetic field strength, left-right difference ratio, and front-back difference ratio of each channel after the flight maneuver is executed. This trend is used to determine whether the flight maneuver has achieved the expected correction effect. Based on the signal change trend, the correction effect of this flight maneuver is quantitatively evaluated and categorized into four levels: effective, partially effective, ineffective, and overshoot. The evaluation criteria are whether the absolute value of the difference ratio parameter changes in a decreasing direction and whether the magnitude of the change meets expectations. The mapping gain is the conversion ratio coefficient between the left-right difference ratio, front-back difference ratio, and attitude correction amount. It directly determines the magnitude of the attitude correction amount under the same difference ratio parameter, i.e., the control sensitivity. The larger the gain, the higher the control sensitivity and the larger the correction action amplitude. The evaluation results are used to guide the adaptive adjustment of the mapping gain, achieving closed-loop optimization of the control parameters.

[0080] Schematic illustration: After the collision avoidance control module drives the UAV to perform flight maneuvers, it again collects the digital signal of the magnetic field strength after the flight maneuver through the inductive sensing module and signal conditioning module to obtain the latest magnetic field data. By comparing the changes in the digital signal of magnetic field strength, the left-right difference ratio, and the front-back difference ratio before and after the flight maneuver, the signal change trend is determined, and it is judged whether the absolute value of the difference ratio parameter is changing in a decreasing direction and whether the change magnitude meets expectations. Based on the signal change trend, the execution effect of this flight maneuver is quantitatively evaluated to obtain a clear evaluation result, determining whether the correction action is effective, partially effective, ineffective, or overshoot. Based on the evaluation result, the mapping gain of the pose correction amount is adaptively adjusted: if the action is ineffective, the gain is appropriately increased to improve control sensitivity; if the action is overshoot, the gain is appropriately decreased to reduce control sensitivity; if the action is partially effective, the gain is fine-tuned to the optimal value to achieve closed-loop optimization of control parameters.

[0081] Schematic, the correction effect is evaluated in real time by comparing the signal gradient (rate of change) before and after the instruction is executed. After outputting the pose correction amount for a rightward translation, the expected induced voltage in the right inductor channel is... It should be gradually reduced, and the difference between the left and right sides should be... It should revert to zero. If If the difference ratio is decreasing, the current obstacle avoidance strategy is deemed effective, and the current control gain is maintained. If, after executing the obstacle avoidance command, the total sensing intensity... Still rising, or The change is extremely slow, indicating that environmental interference is too strong, such as strong winds offsetting the drone's power, or that the current obstacle avoidance capabilities are insufficient. If the signal rapidly crosses zero and surges in the opposite direction, it is considered an overshoot, and the system must immediately reduce the correction amount.

[0082] Schematic illustration: Based on the above evaluation results, the underlying algorithm parameters are dynamically modified to achieve adaptive control. If the ambient background noise suddenly increases, causing a drastic jump in the difference ratio, the system automatically increases the adjustment parameter in the formula. . Increasing the value of the mapping ratio is equivalent to increasing the weight of the denominator, which can effectively smooth out instantaneous fluctuations caused by electromagnetic noise and improve the system's stability in complex environments. If the assessment determines that the correction is insufficient, the system will automatically lower the safety threshold. This allows the system to trigger obstacle avoidance actions earlier in subsequent detections, sacrificing some inspection distance for higher safety redundancy. If the trend of change lags behind, the system synchronously increases the mapping ratio coefficient. This allows the drone to output a larger tilt angle in the next control cycle, enabling it to perform obstacle avoidance maneuvers with greater power.

[0083] In a preferred embodiment of the present invention, after the UAV performs a leftward lateral correction maneuver, the collision avoidance control module collects the magnetic field signal after the maneuver and calculates that the left-right difference ratio decreases from 0.6 to 0.3, which is evaluated as partially effective. Based on the evaluation result, the mapping gain is finely adjusted from the initial value of 1.0 to 1.1 to improve control sensitivity. In the next control cycle, after performing the correction maneuver again, the left-right difference ratio decreases to 0.15, enters the alignment range, is evaluated as effective, and the closed-loop optimization is completed, with a significant improvement in the correction effect.

[0084] In a preferred embodiment of the present invention, a 10kV power distribution line inspection quadcopter drone is equipped with an inductive sensing module, a signal conditioning module, a posture perception module, and a collision avoidance control module. The drone's normal inspection flight speed is 3m / s, and the safe inspection distance from the line is 1.5m. During the inspection, the inductive sensing module captures the 50Hz power frequency magnetic field of the line to be inspected in real time, converts it into an induced electrical signal, and outputs a 3.3V standard voltage signal through the signal conditioning module. The collision avoidance control module performs ADC sampling at a frequency of 200Hz, calculates the left-right difference ratio as 0.6 and the front-back difference ratio as 0.1. After comparing with a preset judgment threshold, it is determined that the drone has deviated significantly to the right from the line. Combined with the posture data of an altitude of 8m above the ground, stable attitude, and heading consistent with the line direction, a warning correction strategy is matched. The left-right difference ratio and the front-back difference ratio are mapped to a lateral posture correction amount to the left. After calculation, a motor control command is output to drive the drone to translate to the left and return to the center position of the line. The entire process is unaffected by cloudy or backlit environments, with no line misjudgment or missed inspection, and the trajectory correction is completed smoothly.

[0085] By implementing this embodiment, the magnetic field signal of the power line to be inspected is collected by the inductive sensing module. Non-contact detection is achieved using the inherent power frequency magnetic field of the energized transmission line. Magnetic field propagation is unaffected by weather and lighting conditions. The combination of the inductive sensing module and signal conditioning module replaces lidar, resulting in a simpler structure and significantly reduced overall cost. Simultaneously, the offset state between the UAV and the line to be inspected is quantified by left-right and front-back difference ratios, thereby determining the spatial orientation relative to the line. This eliminates the need to identify the outline of small-diameter lines, thus avoiding missed or false detections. Based on the spatial orientation, the altitude, attitude, and heading data from the pose perception module can be used to match corresponding flight control strategies and output motor control commands, achieving efficient and stable power line inspection collision avoidance operations. Therefore, the interaction of the inductive sensing module, signal conditioning module, pose perception module, and collision avoidance control module reduces operating costs and improves the reliability of UAV collision avoidance.

[0086] See Figure 4 This is a schematic diagram of a drone collision avoidance system based on inductance processing according to an embodiment of the present invention, comprising: The inductive sensing module is used to acquire the magnetic field signal of the line to be inspected, convert the magnetic field signal into an induced electrical signal, and then transmit the induced electrical signal to the signal conditioning module. The signal conditioning module is used to convert the induced electrical signal into a voltage signal and transmit it to the collision avoidance control module; The pose perception module is used to collect and transmit ground altitude data, spatial attitude data, and heading data to the collision avoidance control module. The collision avoidance control module performs analog-to-digital conversion on the voltage signal according to a preset sampling frequency to obtain a digital signal of magnetic field strength. Based on the digital signal of magnetic field strength, it calculates the left-right difference ratio, which characterizes the lateral deviation of the UAV, and the front-back difference ratio, which characterizes the longitudinal deviation of the UAV. Based on the left-right difference ratio, the front-back difference ratio, and a preset judgment threshold, it determines the spatial orientation relative to the line to be inspected. Based on the spatial orientation, it matches the corresponding flight control strategy according to the ground altitude data, spatial attitude data, and heading data. It maps the left-right difference ratio and the front-back difference ratio into attitude correction values, and outputs motor control commands after solving the attitude correction values ​​and spatial orientation to drive the flight actions corresponding to the flight control strategy.

[0087] This invention provides a drone collision avoidance system based on inductive processing. The system uses an inductive sensing module to acquire the magnetic field signal of the route to be inspected, converts the magnetic field signal into an induced electrical signal, and transmits the induced electrical signal to a signal conditioning module. The signal conditioning module then converts the induced electrical signal into a voltage signal and transmits it to a collision avoidance control module. A posture perception module collects and transmits ground altitude data, spatial attitude data, and heading data to the collision avoidance control module. In the collision avoidance control module, the voltage signal is converted from analog to digital at a preset sampling frequency to obtain a digital magnetic field strength signal. Based on this digital magnetic field strength signal, the left-right difference ratio (representing the drone's lateral deviation) and the front-back difference ratio (representing the drone's longitudinal deviation) are calculated. Based on the left-right difference ratio, the front-back difference ratio, and a preset judgment threshold, the spatial orientation relative to the route to be inspected is determined. Based on this spatial orientation, a corresponding flight control strategy is matched according to the ground altitude data, spatial attitude data, and heading data. The left-right difference ratio and the front-back difference ratio are mapped to posture correction values. After calculation based on the posture correction values ​​and the spatial orientation, motor control commands are output to drive the flight actions corresponding to the flight control strategy.

[0088] The system collects magnetic field signals from the power transmission line under inspection using an inductive sensing module. This allows for non-contact detection by utilizing the inherent power frequency magnetic field of the energized transmission line. Magnetic field propagation is unaffected by weather or lighting conditions. The combination of the inductive sensing module and signal conditioning module replaces lidar, resulting in a simpler structure and significantly reduced overall cost. Simultaneously, the system quantifies the offset between the UAV and the line under inspection by using left-right and front-back difference ratios, thereby determining the spatial orientation relative to the line. This eliminates the need to identify the outline of small-diameter lines, preventing missed or false detections. Based on this spatial orientation, the system uses altitude, attitude, and heading data from the pose perception module to match corresponding flight control strategies and output motor control commands, achieving efficient and stable power line inspection collision avoidance. Therefore, the interaction of the inductive sensing module, signal conditioning module, pose perception module, and collision avoidance control module reduces operating costs and improves the reliability of UAV collision avoidance.

[0089] It should be noted that the system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the system embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0090] Those skilled in the art will understand that, for convenience and brevity, the specific working process of the system described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0091] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. An unmanned aerial vehicle anti-collision method based on inductive processing, characterized in that, Suitable for drones; The drone is equipped with an inductive sensing module, a signal conditioning module, a pose perception module, and a collision avoidance control module; The drone collision avoidance method based on inductance processing includes: The magnetic field signal of the line to be inspected is obtained through the inductive sensing module. After the magnetic field signal is converted into an induced electrical signal, the induced electrical signal is transmitted to the signal conditioning module. The induced electrical signal is converted into a voltage signal by the signal conditioning module and transmitted to the collision avoidance control module. The pose perception module collects and transmits ground altitude data, spatial attitude data, and heading data to the collision avoidance control module. The anti-collision control module performs analog-to-digital conversion on the voltage signal according to the preset sampling frequency to obtain a digital signal of magnetic field strength. Based on the digital signal of magnetic field strength, the left-right difference ratio, which characterizes the lateral deviation of the UAV, and the front-back difference ratio, which characterizes the longitudinal deviation of the UAV, are calculated respectively. Based on the left-right difference ratio, front-back difference ratio, and preset judgment threshold, the spatial orientation relationship relative to the line to be inspected is determined. Based on spatial orientation, corresponding flight control strategies are matched according to ground altitude data, spatial attitude data, and heading data; The left-right difference ratio and the front-back difference ratio are mapped to attitude correction values. After calculation based on the attitude correction values ​​and spatial orientation relationships, motor control commands are output to drive the execution of flight actions corresponding to the flight control strategy.

2. The unmanned aerial vehicle anti-collision method based on inductive processing of claim 1, wherein, The inductive sensing module includes several acquisition channels in various directions, and each acquisition channel includes a magnetic field sensing unit and a frequency selective filtering unit. The magnetic field signal of the line to be inspected is acquired through the inductive sensing module, converted into an induced electrical signal, and then transmitted to the signal conditioning module, including: The magnetic field signal of the line to be inspected is collected by the magnetic field induction unit in the corresponding direction, the magnetic field signal is converted into the induced electrical signal of the corresponding channel, and the induced electrical signal is transmitted to the frequency selective filtering unit. After the electromagnetic interference of non-target frequency in the induced electrical signal is filtered out by the frequency selective filtering unit, the filtered induced electrical signal is transmitted to the signal conditioning module. The target frequency of the frequency selective filter unit is matched with the power frequency AC frequency of the line to be inspected.

3. The unmanned aerial vehicle anti-collision method based on inductive processing of claim 1, wherein, The signal conditioning module converts the induced electrical signal into a voltage signal and transmits it to the collision avoidance control module, including: The induced electrical signal is subjected to amplitude limiting protection processing through the signal conditioning module, which limits the voltage amplitude of the induced electrical signal within a preset safety threshold. The induced electrical signal after the amplitude limiting protection is linearly amplified to a preset amplitude range. The linearly amplified induced electrical signal is subjected to level bias processing and low-pass filtering processing. After being superimposed with a preset DC bias voltage, a voltage signal is obtained and transmitted to the anti-collision control module.

4. The unmanned aerial vehicle anti-collision method based on inductive processing of claim 2, wherein, Based on the digital magnetic field strength signal, the left-right difference ratio, which characterizes the lateral deviation of the UAV, and the front-back difference ratio, which characterizes the longitudinal deviation of the UAV, are calculated respectively, including: Based on the digital magnetic field strength signals, the corresponding digital magnetic field strength signals of the front, back, left, and right acquisition channels are determined respectively. The digital magnetic field strength signals of each acquisition channel are normalized to obtain the characteristic values ​​of the magnetic field strength of each acquisition channel. Based on the preset adjustment parameters and the magnetic field strength characteristic values ​​of the left and right acquisition channels, the left-right difference ratio, which is used to characterize the lateral offset of the UAV, is calculated. Based on the preset adjustment parameters and the magnetic field strength characteristic values ​​of the front and rear acquisition channels, the front-rear difference ratio, which is used to characterize the longitudinal offset of the UAV, is calculated. Among them, the positive and negative values ​​and magnitudes of the left-right difference ratio correspond to the lateral offset direction and degree of the UAV relative to the line to be inspected, respectively; the positive and negative values ​​and magnitudes of the front-back difference ratio correspond to the longitudinal offset direction and degree of the UAV relative to the line to be inspected, respectively.

5. The unmanned aerial vehicle anti-collision method based on inductive processing of claim 4, wherein, The preset judgment thresholds include: alignment threshold, horizontal offset threshold, and vertical offset threshold; Based on the left-right difference ratio, front-back difference ratio, and preset judgment thresholds, the spatial orientation relative to the line to be inspected is determined, including: The absolute value of the left-right difference ratio is compared with the alignment threshold and the horizontal offset threshold, respectively; the absolute value of the front-back difference ratio is compared with the alignment threshold and the vertical offset threshold, respectively. If the absolute values ​​of both the left-right difference ratio and the front-back difference ratio are less than the alignment threshold, the spatial orientation relationship is determined to be aligned with the center of the line to be inspected. If the absolute value of the left-right difference ratio is not less than the alignment threshold and less than the lateral offset threshold, then the spatial orientation relationship is determined to be a lateral offset to the left or right relative to the line to be inspected, depending on the sign of the left-right difference ratio. If the absolute value of the left-right difference ratio is greater than or equal to the lateral offset threshold, the spatial orientation relationship is determined to be a lateral offset to the left or right relative to the line to be inspected, depending on the sign of the left-right difference ratio. If the absolute value of the front-to-back difference ratio is not less than the alignment threshold and less than the longitudinal offset threshold, then the spatial orientation relationship is determined to be forward longitudinally approaching and backward longitudinally moving away from the line to be inspected, depending on the sign of the front-to-back difference ratio. If the absolute value of the front-to-back difference ratio is greater than or equal to the longitudinal offset threshold, the spatial orientation relationship is determined according to the sign of the front-to-back difference ratio: it is longitudinally approaching the line to be inspected and longitudinally moving away from it.

6. The unmanned aerial vehicle anti-collision method based on inductive processing of claim 5, wherein, Based on spatial orientation relationships, and according to ground altitude data, spatial attitude data, and heading data, corresponding flight control strategies are matched, including: Based on spatial orientation, the inspection altitude range is verified using ground altitude data, the flight action boundary is verified using spatial attitude data, and the route to be inspected is matched using heading data to determine the current working condition risk level and the current collision risk level. Based on the current operational risk level and the current collision risk level, determine the corresponding flight control strategy.

7. The drone collision avoidance method based on inductance processing as described in claim 1, characterized in that, After calculating the pose correction amount and spatial orientation relationship, the motor control commands are output, including: The pose correction amount is matched and verified with the current spatial orientation to determine the pose correction target that is compatible with the current spatial orientation. The attitude correction target is superimposed on the original attitude reference command to obtain the corrected closed-loop control target value; The closed-loop control target value is solved using a proportional-integral-derivative control law to obtain the speed control increment corresponding to each power motor. Motor control commands are generated based on the speed control increment.

8. The drone collision avoidance method based on inductance processing as described in claim 2, characterized in that, Before acquiring the magnetic field signal of the line to be inspected, the following steps are also included: The anti-collision control module sequentially performs short-circuit detection, open-circuit detection, zero-bias measurement, and temperature drift self-test on each acquisition channel of the inductive sensing module.

9. The drone collision avoidance method based on inductance processing as described in claim 1, characterized in that, After driving and executing the flight actions corresponding to the flight control strategy, it also includes: The collision avoidance control module collects the digital signal of the magnetic field strength after the flight maneuver is performed, and determines the trend of signal change based on the digital signal of the magnetic field strength after the flight maneuver is performed. The effectiveness of flight maneuvers is assessed based on signal change trends to obtain evaluation results; Adjust the mapping gain of the pose correction amount based on the evaluation results.

10. A drone collision avoidance system based on inductive processing, characterized in that, include: The inductive sensing module is used to acquire the magnetic field signal of the line to be inspected, convert the magnetic field signal into an induced electrical signal, and then transmit the induced electrical signal to the signal conditioning module. The signal conditioning module is used to convert the induced electrical signal into a voltage signal and transmit it to the collision avoidance control module; The pose perception module is used to collect and transmit ground altitude data, spatial attitude data, and heading data to the collision avoidance control module. The anti-collision control module is used to perform analog-to-digital conversion on the voltage signal according to the preset sampling frequency to obtain a digital signal of magnetic field strength. Based on the digital signal of magnetic field strength, it calculates the left-right difference ratio, which represents the lateral deviation of the UAV, and the front-back difference ratio, which represents the longitudinal deviation of the UAV. Based on the left-right difference ratio, the front-back difference ratio, and the preset judgment threshold, it determines the spatial orientation relative to the line to be inspected. Based on spatial orientation, the system matches the corresponding flight control strategy according to ground altitude data, spatial attitude data, and heading data; it maps the left-right difference ratio and the forward-backward difference ratio into attitude correction values; and outputs motor control commands after calculating the attitude correction values ​​and spatial orientation to drive the flight actions corresponding to the flight control strategy.

Images