Fixed-wing unmanned aerial vehicle landing phase relative height fusion method, system and readable storage medium

Abstract

The application discloses a fixed-wing unmanned aerial vehicle landing stage relative height fusion method and system and a storage medium. Satellite navigation, air pressure and ranging equipment original heights are acquired and preprocessed into relative heights; air pressure and satellite navigation fusion, gradual transition fusion to the ranging equipment are sequentially completed, and then complementary filtering is performed in combination with inertial navigation skyward velocity; and for satellite navigation positioning state jump, softening function smoothing processing is adopted to output stable height. Through multi-source height gradual fusion, complementary filtering and positioning jump softening, the scheme solves the problems of large height fluctuation and insufficient precision under complex terrain, and improves the smoothness, precision and robustness of unmanned aerial vehicle landing height estimation.

Description

Technical Field

[0001] This invention relates to the field of unmanned system navigation technology, and in particular to a method, system, and readable storage medium for relative altitude fusion during the landing phase of a fixed-wing unmanned aerial vehicle. Background Technology

[0002] During the landing and flight of a fixed-wing UAV, altitude estimation accuracy has a significant impact on flight safety. When satellite navigation is in differential positioning mode, it can provide centimeter-level altitude positioning accuracy, meeting landing requirements. However, in actual missions, satellite navigation equipment may enter single-point positioning mode due to signal obstruction or interference, leading to significant deviations in relative altitude estimation, with measured deviations reaching up to ten meters. When the UAV lands in this state, directly using the satellite navigation altitude may result in premature touchdown or overrunning the runway. To avoid this problem, radar ranging is currently used as a ground-based altitude estimation compensation, but its measurement results are greatly affected by terrain undulations, and direct use will cause longitudinal fluctuations in the aircraft, reducing landing stability. In addition to satellite navigation and radar, atmospheric altimeters are also commonly used altitude sensors. However, atmospheric altimeters are based on barometric pressure calculations and are greatly affected by environmental factors such as temperature and humidity, resulting in stable output but insufficient accuracy.

[0003] Existing patents, such as CN108955634A, propose a method for fusing altitude sensors in UAVs. This method fuses barometric altitude and radio altitude data, adjusting the barometric altitude to follow the radio altitude at a certain rate when data inconsistencies arise, and automatically stopping fusion when data becomes invalid. This method improves the continuity and reliability of altitude data. However, in complex terrain conditions, radio altitude data fluctuates significantly with the terrain, leading to unstable UAV altitude control. Therefore, there is an urgent need for a ground altitude estimation scheme that can fuse multi-source sensor information and output a smooth and accurate result that meets landing requirements under satellite-guided single-point positioning, thus ensuring the safety of UAV landings.

[0004] Therefore, it is necessary to study a relatively highly integrated method, system, and readable storage medium for the landing phase of a fixed-wing UAV to address the shortcomings of existing technologies and to solve or mitigate one or more of the aforementioned problems. Summary of the Invention

[0005] In view of this, the present invention provides a method, system and readable storage medium for relative altitude fusion during the landing phase of a fixed-wing UAV. It adopts a progressive transition fusion of altitude source from air pressure to radar, combined with complementary filtering of skyward velocity, and a softening mechanism for positioning state transitions. This solves the operational problems of altitude fluctuation and insufficient accuracy in complex terrain. Such UAV landing altitude fusion schemes for satellite navigation anomalies are relatively scarce and have obscure details.

[0006] On one hand, the present invention provides a method for relative altitude fusion during the landing phase of a fixed-wing unmanned aerial vehicle (UAV), the method comprising the following steps: S1: Acquire raw altitude data and status information collected by satellite navigation equipment, barometric pressure measuring equipment and ranging equipment, preprocess the raw altitude data to obtain relative altitude data relative to the target landing area; S2: Perform height fusion based on the relative height data to obtain first fused height data, and use the first fused height data to characterize the fusion result of the barometric measurement device and the satellite navigation device; S3: The relative height data is fused with the first fused height data to obtain second fused height data, which represents the height result of the transition to the ranging device; S4: Obtain the celestial velocity data output by the inertial navigation device, and perform complementary filtering and fusion with the celestial velocity data and the second fused altitude data to obtain the third fused altitude data; S5: Detect the status information obtained by the satellite navigation device. If the status information changes between the first positioning state and the second positioning state, then soften the third fused altitude data through a softening function to obtain the final output altitude data. The first positioning state is a differential or high-precision positioning state, and the second positioning state is a single-point or low-precision positioning state. The ranging device is a radar altimeter or a laser rangefinder.

[0007] In addition to the aspects and any possible implementations described above, a further implementation is provided in which the height fusion in step S2 based on the relative height data to obtain the first fused height data specifically includes: Acquire the first relative altitude data collected by the barometric pressure measuring device and the second relative altitude data collected by the satellite navigation device; Obtain the absolute value of the positioning altitude error of the satellite navigation device in the second positioning state; The first fused height data is determined based on the first relative height data, the second relative height data, and the absolute value of the positioning height error. The first fused height data is the median value among the first relative height data, the second relative height data, and the absolute value of the positioning height error.

[0008] In addition to the aspects and any possible implementations described above, a further implementation is provided in which the second fused height data is obtained by fusing the relative height data with the first fused height data in step S3 specifically includes: Acquire the third relative height data collected by the ranging device; The fusion coefficient is determined based on the preset upper and lower limits of the height source transition. The second fusion height data is calculated based on the fusion coefficient, the first fusion height data, and the third relative height data.

[0009] In addition to the aspects and any possible implementations described above, a further implementation is provided, wherein determining the fusion coefficient in S3 based on the preset upper limit and lower limit of the height source transition specifically includes: Obtain the current flight altitude data of the aircraft; Determine whether the flight altitude data is greater than the altitude source transition upper limit. If the flight altitude data is greater than the altitude source transition upper limit, then determine that the fusion coefficient is a first preset value. Determine whether the flight altitude data is less than the altitude source transition lower limit. If the flight altitude data is less than the altitude source transition lower limit, then determine that the fusion coefficient is a second preset value. If the flight altitude data is between the upper limit of the altitude source transition and the lower limit of the altitude source transition, then the fusion coefficient is calculated by linear interpolation based on the flight altitude data, the upper limit of the altitude source transition, and the lower limit of the altitude source transition. The fusion coefficients are output to the fusion calculation module for later use.

[0010] In addition to the aspects and any possible implementations described above, a further implementation is provided in which the third fused height data is obtained by complementary filtering and fusion of the celestial velocity data and the second fused height data in step S4 specifically includes: Obtain the preset time constant; Based on the time constant, the celestial velocity data, and the second fused altitude data, a complementary filtering model is constructed; The initial third fusion height data is calculated using the complementary filtering model. Obtain the position information of the aircraft relative to the target landing area, and determine whether the aircraft has passed over the target landing area; If the aircraft passes over the target landing area, the pressure altitude contribution weight or position correction feedback gain in the third fused altitude data is set to zero, resulting in updated third fused altitude data.

[0011] In addition to the aspects and any possible implementations described above, a further implementation is provided in which, if the state information changes between the first positioning state and the second positioning state in step S5, the third fused height data is softened using a softening function to obtain the final output height data. Specifically, this includes: Determine whether the status information changes from the first positioning state to the second positioning state; If the status information changes from the first positioning state to the second positioning state, then the maximum allowable single-step altitude change value of the aircraft is obtained; A first softening function is constructed based on the maximum value of the single-step height change. The third fused height data is then calculated using the first softening function to obtain the final output height data.

[0012] In addition to the aspects described above and any possible implementation, a further implementation is provided, wherein determining whether the state information changes from the first positioning state to the second positioning state in step S5 includes: obtaining the first state identifier of the satellite navigation device at the current moment and the second state identifier at the previous moment; comparing the first state identifier with the second state identifier; if the first state identifier represents the second positioning state and the second state identifier represents the first positioning state, then determining that the state information changes from the first positioning state to the second positioning state.

[0013] In addition to the aspects and any possible implementations described above, a further implementation is provided in which, if the state information changes between the first positioning state and the second positioning state in step S5, the third fused height data is softened using a softening function to obtain the final output height data. Specifically, this includes: Determine whether the status information changes from the second positioning state to the first positioning state; If the status information changes from the second positioning state to the first positioning state, then the maximum allowable single-step altitude change value of the aircraft is obtained; A second softening function is constructed based on the maximum value of the single-step height change. The third fused height data is then calculated using the second softening function to obtain the final output height data.

[0014] As described above and in any possible implementation, a relative altitude fusion system for the landing phase of a fixed-wing unmanned aerial vehicle (UAV) is further provided. Based on the aforementioned relative altitude fusion method for the landing phase of a fixed-wing UAV, the relative altitude fusion system for the landing phase of a fixed-wing UAV includes: The data acquisition and preprocessing module is used to acquire raw altitude data and status information collected by satellite navigation equipment, barometric pressure measurement equipment and ranging equipment, and to preprocess the raw altitude data to obtain relative altitude data relative to the target landing area. The first fusion module is used to perform height fusion based on the relative height data to obtain first fused height data, which represents the fusion result of the barometric measurement device and the satellite navigation device. The second fusion module is used to fuse the relative height data collected by the ranging device with the first fused height data to obtain the second fused height data, which represents the height result of the transition to the ranging device. The complementary filtering and fusion module is used to acquire the celestial velocity data output by the inertial navigation device, and to perform complementary filtering and fusion of the celestial velocity data with the second fused altitude data to obtain the third fused altitude data; The softening processing module is used to detect the status information of the satellite navigation device. When the status information changes between the first positioning state and the second positioning state, the third fused altitude data is softened by the softening function to obtain the final output altitude data.

[0015] In addition to the aspects described above and any possible implementations, a computer-readable storage medium is further provided, the computer-readable storage medium storing a computer program that, when executed by a processor, implements the aforementioned method for relative altitude fusion during the landing phase of a fixed-wing unmanned aerial vehicle.

[0016] Compared with the prior art, the present invention can achieve the following technical effects: 1) Multi-source sensor fusion significantly improves robustness. By integrating data from four types of sensors—satellite navigation, barometric measurement, ranging, and inertial navigation—a multi-level redundant sensing system is constructed. This system can withstand anomalies such as single sensor failure, signal interference, and terrain undulations, and can still stably output altitude even in scenarios with satellite signal blockage, electromagnetic interference, and complex terrain. 2) Gradual integration in stages to fully leverage the advantages of sensors The high-altitude system primarily relies on air pressure and satellite navigation, while the low-altitude system smoothly transitions to ranging altitude. Near the ground, it completely depends on ranging equipment, balancing long-term stability at high altitudes with high-precision ground reference at low altitudes. Altitude source switching is seamless and oscillating. 3) Complementary filtering enhances dynamic performance Introducing inertial navigation axial velocity complementary filtering compensates for the slow dynamic response and filtering delay of position sensors, accurately captures instantaneous disturbances in the vertical direction, and improves the real-time performance and dynamic tracking capability of altitude data; 4) Softened positioning status transitions ensure flight safety; To address the altitude jumps caused by satellite navigation differential / point positioning transitions, a softening function is used to smoothly decompose the changes according to the maximum single-step change, completely eliminating the impact of altitude changes on flight control and avoiding risks such as severe attitude fluctuations and premature grounding. 5) The algorithm is highly modular and has wide adaptability. With a parametric and modular design, it can be adapted to fixed-wing UAVs of different sizes and dynamic characteristics by only adjusting parameters such as transition range, time constant, and softening step size, and is compatible with multiple sensor configurations; 6) Highly efficient in calculation and highly applicable to engineering projects. The algorithm has no complex matrix operations or high-order iterations, has a short computation time, and can run in real time on low-power embedded flight control chips, meeting the high real-time navigation requirements of UAVs. 7) Near-ground altitude accuracy is improved step by step The closer to the ground, the higher the weight of the ranging equipment, and the altitude error during the ground contact phase is reduced to the centimeter level, significantly improving the accuracy of fixed-point landing and the stability of ground contact.

[0017] Of course, any product implementing this invention does not necessarily need to achieve all of the technical effects described above at the same time. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is an algorithm flowchart provided in one embodiment of the present invention; Figure 2 This is a schematic diagram of a phased fusion scheme for highly information sources provided in an embodiment of the present invention; Figure 3 This is the height information with added noise provided in one embodiment of the present invention; Figure 4 This is a schematic diagram of the fusion height effect in a single-point state provided by an embodiment of the present invention; Figure 5 This is a schematic diagram of the fusion height effect when the satellite positioning state changes, provided by an embodiment of the present invention; Figure 6 This is a schematic diagram of the fused altitude effect during actual flight without relying on satellite navigation data, provided by an embodiment of the present invention. Detailed Implementation

[0020] To better understand the technical solution of the present invention, the embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0021] It should be understood that the described embodiments are merely some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0022] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. The singular forms “a,” “the,” and “the” as used in the embodiments of this invention and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0023] This invention provides a relative altitude fusion method for the landing phase of a fixed-wing unmanned aerial vehicle (UAV). During the landing process, the accuracy and continuity of altitude information are crucial for ensuring flight safety and achieving precise landing. Since a single sensor has limitations in different flight environments, altitude ranges, and motion states, this solution constructs a highly reliable, smooth, and automatically weighted altitude calculation system through deep fusion of multi-source sensor data.

[0024] Step S1: Obtain the raw altitude data and status information collected by the satellite navigation equipment, barometric pressure measurement equipment and ranging equipment, and preprocess the raw altitude data to obtain the relative altitude data relative to the target landing area.

[0025] Specifically, before the aircraft enters the landing phase, the system initiates a multi-channel data acquisition task. Satellite navigation equipment typically uses the Global Positioning System (GPS) or the BeiDou Navigation Satellite System, and its raw altitude data is usually based on the geographic altitude of the Earth's ellipsoid model. Barometric pressure measurement equipment calculates altitude by sensing changes in atmospheric pressure, and its raw output is a pressure altitude based on a standard atmospheric pressure model. Distance measuring equipment, such as radar altimeters or laser rangefinders, directly measures the geometric distance between the aircraft and obstacles or the ground below by emitting electromagnetic waves or coherent beams and receiving reflected signals from the ground.

[0026] While acquiring this raw data, the system also simultaneously reads the status information of each device. This status information includes, but is not limited to, the number of visible satellites, horizontal accuracy factor, vertical accuracy factor, and positioning mode identifier for satellite navigation devices; the sensor health status and temperature compensation status for barometric pressure measurement devices; and the signal strength and confidence level for ranging devices. This status information provides a basis for subsequent preprocessing and weight allocation decisions.

[0027] Preprocessing the raw altitude data is a prerequisite for ensuring fusion accuracy. Since aircraft landing focuses on the vertical distance relative to the runway or landing point, all absolute altitudes must be converted to relative altitudes. For satellite navigation data, the system pre-stores the geographic coordinates and altitude of the target landing area. The relative altitude is obtained by subtracting the landing point's altitude from the satellite's output geographic altitude, taking into account Earth's curvature correction. For barometric altitude, since atmospheric pressure is significantly affected by weather conditions, the system is calibrated based on corrected sea-level or surface pressure provided by the landing site. In practice, the system uses a barometric pressure trend analysis algorithm to eliminate interference from localized pressure fluctuations caused by aircraft maneuvers, ensuring the relative stability of barometric altitude over a short period.

[0028] For ranging devices, the preprocessing process is more complex. Because ranging devices may receive echoes from lateral obstacles during high-speed flight or steep turns, the system needs to combine the aircraft's attitude angle information to project the slant range onto the vertical direction. Simultaneously, the system establishes a sliding window midpoint filter to denoise the raw ranging values, identifying and removing jumps caused by ground vegetation, buildings, or complex terrain. Through this preprocessing, the system obtains three sets of relative altitude data in the same reference frame, laying the foundation for subsequent hierarchical fusion.

[0029] Step S2: Perform height fusion based on the relative height data to obtain first fused height data, which represents the fusion result of the barometric pressure measurement device and the satellite navigation device.

[0030] During the initial descent phase at higher altitudes, ranging equipment often exceeds its effective range or has low accuracy. At this time, altitude information mainly relies on barometric altimeters and satellite navigation equipment. However, barometric altimeters exhibit time-varying characteristics, while satellite navigation altitude, although without cumulative error, is susceptible to ionospheric interference or multipath effects, resulting in abrupt changes. To combine the advantages of both and suppress anomalous fluctuations, this step employs an intermediate value selection strategy based on error boundaries.

[0031] Step S21: Obtain the first relative altitude data collected by the barometric pressure measuring device and the second relative altitude data collected by the satellite navigation device.

[0032] In actual operation, the first relative altitude data (barometric altitude) has an extremely high sampling frequency and good short-term smoothness, enabling it to keenly detect even minute ascending and descending trends of the aircraft. The second relative altitude data (satellite altitude) provides a long-term absolute reference. Within each calculation cycle, the system extracts these two sets of data from the data bus after time alignment. Due to the different time delays of different sensors, the system utilizes a caching mechanism to perform interpolation compensation based on the sampling timestamps of each device, ensuring that the data participating in the fusion are synchronized in the time dimension.

[0033] Step S22: Obtain the absolute value of the positioning altitude error of the satellite navigation device in the second positioning state.

[0034] The second positioning state here typically refers to a state where the satellite navigation system is in a non-high-precision positioning mode, or where positioning accuracy has decreased. The system monitors the vertical positioning accuracy index output by the satellite navigation equipment in real time. This index reflects the potential error range of the altitude data under the current satellite geometry and signal quality. The absolute value of the positioning altitude error is a dynamically changing parameter, representing the reliability boundary of the satellite altitude at the current moment. When the satellite signal is blocked or interfered with, this absolute value of error will increase rapidly.

[0035] Step S23: Determine the first fused height data based on the first relative height data, the second relative height data, and the absolute value of the positioning height error. The first fused height data is the median value among the first relative height data, the second relative height data, and the absolute value of the positioning height error.

[0036] This logic is one of the core innovations of this solution. In traditional weighted average algorithms, if the satellite altitude changes drastically, the fusion result will inevitably be affected and become biased. However, using intermediate value logic can effectively suppress such abrupt changes. Specifically, the system sorts the relative atmospheric pressure altitude, the relative satellite altitude, and the reference value determined by the error boundary.

[0037] For example, suppose the barometric altitude is 100 meters, and the satellite altitude suddenly jumps to 120 meters due to multipath interference. The absolute value of the positioning altitude error calculated based on the satellite status (combined with the upper limit of the reasonable range determined by the stable value at the previous moment) is 105 meters. Of these three values, 105 meters is output as the intermediate value. This preserves the correction function of satellite navigation while preventing the fused altitude from being skewed by erroneous satellite altitude jumps. This processing method essentially acts as a non-linear limiting mechanism, allowing the first fused altitude data to gradually approach the high-precision satellite altitude while maintaining the smoothness of the barometric altitude, without producing instantaneous and drastic fluctuations.

[0038] Step S3: The relative height data is fused with the first fused height data to obtain second fused height data, which represents the height result of the transition to the ranging device.

[0039] As the aircraft descends, the ranging equipment enters its optimal operating range. The relative altitude provided by the ranging equipment is a geometric altitude directly relative to the ground, which is crucial for landing control in the final stage. However, the switch from barometric / satellite fusion altitude to ranging altitude cannot be abrupt, otherwise it will cause violent elevator maneuvers in the flight control system. Therefore, this step introduces a smooth transition mechanism based on altitude ranges.

[0040] Step S31: Obtain the third relative height data collected by the ranging device.

[0041] The system reads data from a radar altimeter or laser rangefinder in real time. In one possible implementation, the system first checks the validity of the third relative height data. If the ranging device's feedback signal is lost or it is in search mode, the system will maintain the first fused height data as output. Only when the ranging data is stable and has high confidence will it participate in subsequent fusion.

[0042] Step S32: Determine the fusion coefficient based on the preset upper limit and lower limit of the height source transition.

[0043] To achieve seamless switching between different altitude sources, the system defines a transition zone. The upper limit of the altitude source transition is typically set at the critical altitude at which the ranging device begins to operate stably, such as 60 meters; the lower limit of the altitude source transition is set at a near-ground altitude where the ranging data must be fully relied upon, such as 30 meters. The fusion coefficient determines the weight ratio of the ranging data and the first fused altitude data at the current altitude.

[0044] Step S321: Obtain the current flight altitude data of the aircraft.

[0045] The flight altitude data here can use the fused altitude of the previous cycle or the current first fused altitude as a reference to determine the current spatial position of the aircraft.

[0046] Step S322: Determine whether the flight altitude data is greater than the altitude source transition upper limit. If the flight altitude data is greater than the altitude source transition upper limit, then determine the fusion coefficient as a first preset value.

[0047] For example, when the aircraft is above 60 meters in altitude, the ranging equipment may be affected by complex ground reflections, resulting in lower accuracy compared to the fusion of air pressure and satellite data. In this case, setting the fusion coefficient to a first preset value (such as 1.0) means that the second fusion altitude depends entirely on the first fusion altitude, and the weight of the ranging data is zero.

[0048] Step S323: Determine whether the flight altitude data is less than the altitude source transition lower limit. If the flight altitude data is less than the altitude source transition lower limit, then determine the fusion coefficient as the second preset value.

[0049] When the aircraft descends below 30 meters, the accuracy of the ranging device is far superior to that of other sensors, and extremely high ground relative altitude accuracy is required to support a level landing. At this point, setting the fusion coefficient to a second preset value (such as 0.0) means that the second fusion altitude uses data entirely from the ranging device, thereby eliminating the influence of air pressure drift or satellite errors on near-Earth landing.

[0050] Step S324: If the flight altitude data is between the upper limit of the altitude source transition and the lower limit of the altitude source transition, then the fusion coefficient is calculated by linear interpolation based on the flight altitude data, the upper limit of the altitude source transition, and the lower limit of the altitude source transition.

[0051] Within the transition zone (e.g., between 30 and 60 meters), the system calculates a coefficient that varies linearly with height. Specifically, as the height decreases, the fusion coefficient gradually decreases from 1.0 to 0.0. This linear interpolation algorithm ensures the continuity of weight allocation.

[0052] For example, if the current altitude is 45 meters, exactly at the midpoint of the transition zone, the fusion factor is calculated to be 0.5. This means that the second fused altitude consists of half of the first fused altitude and half of the ranging altitude. In this way, the system smoothly transfers the altitude reference point from the atmospheric / space coordinate system to the ground geometric coordinate system during the aircraft's descent.

[0053] Step S325: Output the fusion coefficients to the fusion calculation module for use.

[0054] The calculated fusion coefficients are transmitted to the execution unit in real time. In one embodiment, the system further smooths the coefficients using a low-pass filter to prevent frequent oscillations near the critical point due to highly minute jitter.

[0055] Step S33: Calculate the second fusion height data based on the fusion coefficient, the first fusion height data, and the third relative height data.

[0056] The fusion calculation module uses a weighted summation method to obtain the result. The second fusion height is equal to the fusion coefficient multiplied by the first fusion height, plus (1 minus the fusion coefficient) multiplied by the third relative height. This calculation process is performed once every navigation calculation cycle (e.g., 20 milliseconds).

[0057] It should be noted that this fusion strategy considers not only the static accuracy of the sensors but also the dynamic needs of the aircraft during descent. At higher altitudes, the system utilizes the long-term stability of satellite navigation and the short-term smoothness of the barometer; at lower altitudes, the system automatically switches to high-precision geometric ranging. Through the weighting of linear interpolation, the system effectively solves the problem of systematic deviations between different altitude sources. For example, if there is a fixed deviation of two meters between the barometric altitude and the ranging altitude, through a linear transition within a thirty-meter interval, this two-meter deviation will be distributed throughout the descent. For every meter the aircraft descends, the altitude deviation is corrected by only about 0.06 meters. This tiny correction can be completely absorbed by the inner loop of the flight control system without causing any abnormal shaking of the aircraft.

[0058] In one possible implementation, the system also continuously monitors the validity of the third relative altitude data. If the ranging data suddenly fails within the transition zone, the system will activate a weight backoff mechanism to quickly restore the fusion coefficient to 1.0 and record the current deviation value. A deviation compensation algorithm will then be used to ensure the continuity of the altitude data. This multi-layered protection logic makes the second fused altitude data a composite altitude indicator that possesses both absolute reference value and extremely high local accuracy.

[0059] Specifically, when processing data from the ranging device, the system also considers the influence of terrain slope. If there is a slope in front of the target landing area, the vertical altitude measured by the ranging device will change. The system performs terrain compensation for the third relative altitude by combining the aircraft's horizontal velocity and pre-stored terrain maps. For example, when the aircraft flies over a deep pit or protrusion, the preprocessing algorithm will identify such terrain features of the non-landing area and shield this transient interference by maintaining the fusion weight from the previous moment or increasing the filtering time constant, ensuring that the second fused altitude data always reflects the trend relative to the predetermined landing plane.

[0060] Furthermore, to improve the system's robustness, a rate-of-change limiter is introduced during the calculation in step S33. If the difference between the calculated second fused altitude and the previous value exceeds the physically possible upper limit of descent speed, the system will limit the output value. This multi-verification mechanism ensures that even in extreme sensor failure scenarios, the output altitude data will not exhibit catastrophic jumps that could cause the aircraft to lose control.

[0061] Through the aforementioned hierarchical processing from S1 to S3, the system has completed the evolution from raw multi-source data to preliminary fused altitude, and then to ground reference altitude. The first fused altitude addresses the complementarity issue between air pressure and satellite data, while the second fused altitude solves the problem of a smooth transition from airspace reference to land reference. This provides a high-quality input source for subsequent dynamic enhancement of inertial navigation data and softening processing for satellite state transitions. In practical applications, this fusion architecture can significantly improve the landing success rate of aircraft in complex weather conditions and terrain environments, reduce reliance on a single high-precision sensor, lower system costs, and enhance safety.

[0062] In one embodiment, the system also establishes a temperature and humidity-based compensation model for the barometric altimeter. Since atmospheric density changes during descent do not perfectly conform to a standard atmospheric model, the system uses onboard temperature sensors to correct the barometric altimeter calculation formula in real time. This refined preprocessing ensures that the first relative altitude data possesses high physical accuracy before entering the fusion stage.

[0063] For satellite navigation equipment, the system not only focuses on the altitude value but also on the positioning status word it outputs. In the first positioning state (such as differential positioning or precise point positioning), the satellite altitude is assigned a higher initial confidence level; while in the second positioning state (such as ordinary single-point positioning), the system relies more on the smoothness of barometric altitude. This state-based dynamic adjustment, combined with the intermediate value logic in step S23, constructs an extremely robust primary fusion layer.

[0064] In calculating the transition zone, the system can also dynamically adjust the upper and lower limits of the transition based on the aircraft's descent rate. For example, during rapid descent, to allow sufficient reaction time for the control system, the upper limit of the altitude source transition can be appropriately increased, allowing ranging data to be integrated into the fusion process earlier. During a gentle descent, a standard transition range can be maintained. This adaptive parameter adjustment further optimizes the flexibility of the fusion process.

[0065] Through this rigorous logical chain, the aircraft can obtain altitude information most suitable for the current environment at every stage of landing. From macroscopic positioning at high altitude to microscopic observation distance near the ground, the flow of data and the evolution of weights maintain physical rationality and mathematical continuity. This is not merely a simple numerical superposition, but a deep coupling based on flight mechanics and sensor characteristics, providing solid data support for achieving fully automatic, high-precision autonomous landing.

[0066] In practical engineering implementation, the aforementioned algorithm is deployed in the navigation task module of the flight control computer. Due to the large amount of floating-point operations and logical judgments involved, the system employs efficient vectorization processing technology. The processing time for each frame of data is strictly controlled to the millisecond level to meet the requirements of real-time navigation. Simultaneously, the system also possesses comprehensive fault diagnosis capabilities, capable of monitoring the data stream of each sensor in real time. Once an anomaly is detected, it is immediately isolated at the fusion layer to ensure that the final output altitude data remains under control.

[0067] This fusion method performs exceptionally well in handling complex interference. For example, when encountering strong winds causing fuselage vibrations, barometers may generate high-frequency noise, while satellite navigation may experience brief signal attenuation due to fuselage obstruction. Through intermediate value processing in step S2, these high-frequency noises and transient jumps are effectively filtered out. When the aircraft approaches the ground and the air pressure field is distorted due to the ground effect, ranging fusion in step S3 promptly takes over altitude control, avoiding the air pressure distortion zone and ensuring altitude accuracy at the moment of ground contact.

[0068] Furthermore, the system design takes into account the different physical characteristics of various ranging devices. For radar altimeters, the system focuses on handling multipath reflections and ground clutter; for laser rangefinders, it focuses on the impact of weather conditions such as smoke, rain, and snow on beam scattering. By incorporating a targeted signal quality assessment algorithm in step S31, the system can identify whether the current environment is suitable for the operation of this type of ranging device, thereby dynamically deciding whether to initiate the transition to the ranging altitude. This intelligent decision-making mechanism enables the fusion method to adapt to various landing scenarios, from land-based airports to offshore platforms.

[0069] In a preferred embodiment, the system also records sensor deviation characteristics during each landing. Through machine learning algorithms, the system can learn online the barometer drift patterns and satellite navigation system errors during flight, and pre-compensate them in subsequent preprocessing steps. This self-learning fusion architecture allows the aircraft's navigation accuracy to continuously optimize with the number of flights.

[0070] In summary, the first three main steps construct a robust framework for static and quasi-dynamic altitude fusion. Through a progressively structured logic, it organically combines altitude data based on different physical principles, coordinate references, and precision characteristics. From raw data cleaning and coordinate unification in S1, to error boundary constraints and jump suppression in S2, and then to interval weight allocation and smooth reference frame switching in S3, each step closely follows the actual physical processes of the aircraft's descent phase. This design not only solves the problem of single-sensor susceptibility to interference but also compensates for the discontinuity during sensor switching through mathematical means, providing an extremely stable benchmark for subsequent inertial data fusion incorporating high-frequency dynamic responses.

[0071] When handling linear interpolation in S324, the system can also use higher-order spline curves instead of simple linear interpolation to obtain higher-order derivative continuity. This means that not only is the altitude continuous, but the rate of change of altitude (i.e., vertical velocity) also transitions smoothly at the transition zone boundaries. This is of great significance for advanced autopilot aircraft that pursue ultimate stability. Through this refined mathematical processing, the aircraft's control system can obtain purer and more realistic motion state feedback.

[0072] In actual deployment, to prevent computational overflow or numerical instability, all fusion calculations employ normalization. The calculated fusion coefficients are strictly limited to a closed interval between zero and one. Simultaneously, the system monitors the difference between the first fusion altitude and the third relative altitude in real time. If this difference is too large, exceeding a preset safety threshold, the system triggers an alarm and suspends automatic transition, switching to an emergency fusion strategy based on slow deviation elimination to ensure flight safety. This meticulous consideration of abnormal situations demonstrates the maturity of this fusion method in engineering applications.

[0073] Through this series of complex yet orderly operations, the system successfully integrated the global satellite data, the smoothness of air pressure, and the local accuracy of ranging into a unified altitude index. This index not only reflects the spacecraft's current position but also provides a scientific prediction of its motion trend. During the high-risk landing phase, this high-quality altitude information is the fundamental guarantee that the spacecraft can accurately execute every control command and ultimately land safely.

[0074] In the subsequent processing, the system will incorporate the azimuth velocity data from the inertial navigation system. Inertial data possesses extremely high dynamic response speed, capable of reflecting the instantaneous acceleration and velocity changes of the aircraft in the vertical direction in real time. By performing complementary filtering on the second fused altitude and the inertial azimuth velocity, sensor noise can be further eliminated, and phase lag caused by filtering can be compensated, resulting in a third fused altitude data that possesses both static accuracy and dynamic real-time performance. This process will incorporate geofencing logic regarding whether the aircraft passes over the target landing area for secondary weight optimization.

[0075] Furthermore, to address the potentially significant altitude jumps that may occur when the satellite navigation system switches between different positioning modes (such as from RTK fixed solution to single-point positioning solution), a specialized softening algorithm has been designed. By constructing a softening function based on a single-step rate of change constraint, the system can smoothly expand this transient step deviation over time, transforming it into a weak and controllable trend term, thereby completely eliminating the impact of satellite state jumps on flight control. These subsequent steps will be closely integrated with the aforementioned S1 to S3 steps to form a complete, rigorous, and efficient aircraft landing altitude fusion system.

[0076] In one embodiment, the system also utilizes redundant sensors for cross-validation during step S1. For example, if the aircraft is equipped with dual-redundant barometers, the system first checks the consistency of the data from the two barometers. If the difference is too large, the system combines the vertical cumulative displacement of the inertial navigation to determine which sensor has malfunctioned and removes it. This fault-tolerance mechanism, introduced at the lowest level, greatly improves the starting quality of the entire fusion chain.

[0077] For satellite navigation data, the preprocessing stage also includes the detection of multipath effects. When an aircraft approaches ground buildings or large metal structures, satellite signals may be reflected back to the antenna, causing an error of several meters in altitude calculation. By analyzing the carrier-to-noise ratio fluctuation characteristics of satellite signals, the system can identify this multipath interference and accordingly increase the absolute value of the positioning altitude error in step S22, thereby automatically reducing the weight of satellite altitude at this point through intermediate value logic.

[0078] During step S3, taking into account potential installation deviations in the ranging equipment, the system performs automatic calibration while the aircraft is stationary on the runway. This calibration value is stored in non-volatile memory and introduced as a compensation term in step S31. This attention to detail ensures that the ranging data accurately reflects the actual distance between the landing gear and the ground during the final touchdown phase.

[0079] Through these meticulous technical means, the embodiments of this invention not only provide a theoretical fusion framework but also a practical engineering solution. It can cope with various complex factors such as sensor noise, environmental interference, state transitions, and terrain changes, providing the most reliable altitude support for the aircraft during the landing phase. This transformation process from coarse raw data to refined fusion results demonstrates the powerful capabilities of multi-source information fusion technology, which is an indispensable core component of modern avionics systems.

[0080] In actual flight tests, aircraft using this method showed a significant reduction in the standard deviation of altitude error during the landing phase, especially near the altitude source switching point, where the altitude curve was extremely smooth, without any visible or sensor-detectable steps. This directly improved the performance of the automatic landing system, resulting in a more concentrated distribution of touchdown points and a more stable vertical velocity upon touchdown, greatly enhancing flight comfort and safety.

[0081] This method is also highly versatile. Whether it's a fixed-wing aircraft, a rotary-wing aircraft, or a vertical takeoff and landing (VTOL) drone, as long as the process involves descent from cruising altitude to the landing area, this logic can be applied. For different types of aircraft, only the upper and lower limits of the transition zone and the error boundary parameters need to be adjusted according to their dynamic characteristics and sensor configuration. This modular and parameterized design approach gives this technology extremely high commercial value and application prospects.

[0082] When processing the intermediate value logic of S23, a confidence-based weighting factor can be introduced. If the barometer exhibits extremely high stability after long-term operation, its virtual error range when participating in sorting can be appropriately reduced. This idea of ​​dynamic optimization runs through the entire algorithm design. By continuously evaluating the sensor performance online, the system can always be in an optimal fusion state.

[0083] Furthermore, when the system outputs the fusion coefficients at S325, it also simultaneously outputs a fusion state vector. This vector contains the currently dominant altitude source information and the health status of the fusion process. The flight control system can adjust its control law based on this state vector. For example, when the system fully enters the ranging and altitude-dominant phase, the control system can switch to a higher bandwidth vertical control mode to achieve more precise ground contact control. This deep coordination between navigation and control is key to achieving advanced automatic landing.

[0084] In one possible implementation, the system also models the ground reflection characteristics of the target landing area. The reflection characteristics of the ranging device differ depending on whether it's grass, a concrete runway, or water. By analyzing the waveform characteristics of the echo signal, the system can identify the ground type and adjust the preprocessing parameters in S31 accordingly. This ability to perceive the physical environment makes the fused altitude data more closely reflect actual physical reality.

[0085] In summary, the first part of this scheme successfully constructs a stable, accurate, and highly environmentally adaptable relative altitude benchmark through precise sensor management and a hierarchical fusion strategy. This provides a solid foundation for subsequent dynamic enhancement and anomaly softening processing, and is the core of the entire altitude fusion algorithm. In this way, the aircraft can perceive its distance from the ground clearly throughout the complex descent process, much like it possesses intelligent eyes and sensitive touch.

[0086] In practical applications, this method also considers the impact of electromagnetic interference on sensors. In certain special environments, such as near high-voltage lines or high-power radar stations, satellite navigation and ranging equipment may be simultaneously interfered with. In this case, the system can quickly identify the interference source through spectrum analysis and signal quality monitoring in step S1, and automatically switch to a backup mode based on barometric altitude and inertial estimation. This comprehensive emergency handling logic ensures that the aircraft can still maintain basic altitude awareness under extremely adverse conditions, buying valuable time for manual takeover or emergency landing.

[0087] This highly integrated approach not only improves technical specifications but also solves two major long-standing problems in aircraft automatic landing: altitude jumps and inconsistencies in reference frames. Through ingenious mathematical design, it transforms complex physical problems into calculable and optimizable engineering models, contributing a highly valuable solution to the field of aviation navigation. In future development, with further advancements in sensor technology, this fusion framework can easily integrate more novel sensors, such as visual altimeters or ultrasonic arrays, demonstrating strong scalability and versatility.

[0088] In one embodiment, the system also estimates the static source error of the barometer in real time. When the aircraft is flying at a specific altitude, the system uses the long-term average of satellite altitudes as a reference to calculate the current drift of the barometer and performs real-time compensation in step S21. This closed-loop calibration mechanism eliminates most of the system error before the barometric altitude enters the fusion layer, further improving the quality of the first fused altitude.

[0089] For the ranging equipment, the system also incorporates an attitude compensation algorithm in step S31. Because the aircraft may pitch or roll during descent, the ranging beam may not be pointing directly downwards. The system uses the precise attitude angles provided by inertial navigation to correct the slant range to vertical altitude through trigonometric calculations. This precise geometric correction is essential to ensuring accurate near-ground altitude.

[0090] Through these multi-dimensional and multi-layered technical means, the highly integrated system constructed in this invention strives for perfection in every detail. It not only focuses on numerical accuracy but also on the source of data, the transparency of the process, and the reliability of the results. This rigorous engineering approach ensures that the method can withstand various extreme conditions in actual flight missions, becoming a guardian of safe aircraft landings.

[0091] When processing the linear interpolation of S324, the system dynamically fine-tunes based on the aircraft's current vertical motion characteristics. Linear interpolation is not only for numerical continuity but also for achieving a smooth transition of control weights at the physical level. Within the transition zone, the system monitors the rate of change of the difference between the first fused altitude and the third relative altitude in real time. If this rate of change is too rapid, it indicates a significant dynamic deviation between the two altitude sources. In this case, the system appropriately slows down the rate of change of the fusion coefficient, extending the transition time to achieve greater stability. This approach effectively avoids oscillations in control commands caused by discontinuities in the second derivative of the input signal during altitude source switching.

[0092] Step S4: Obtain the celestial velocity data output by the inertial navigation device, and perform complementary filtering and fusion with the celestial velocity data and the second fused altitude data to obtain the third fused altitude data.

[0093] During the dynamic process of aircraft descent, position information alone is insufficient. Since barometric pressure, satellite, and ranging devices are essentially position sensors, their data update frequency is relatively low, and filtering delays are unavoidable during processing. To improve the real-time performance and dynamic response capability of altitude information, this solution introduces celestial velocity data output from the inertial navigation system. The inertial navigation system obtains velocity by integrating accelerometer data, characterized by extremely fast dynamic response, capable of capturing instantaneous disturbances in the vertical direction, such as turbulence caused by gusts or instantaneous ascent and descent caused by control surface movements. However, inertial velocity exhibits drift errors that accumulate over time. The logic of complementary filtering utilizes the long-term accuracy of the second fused altitude to correct the short-term drift of inertial velocity, while simultaneously using the short-term high-frequency characteristics of inertial velocity to fill the gaps in the sampling interval of the position sensor.

[0094] Step S41: Obtain the preset time constant.

[0095] The time constant is a core parameter in complementary filters, determining the cutoff frequency or frequency at which the filter is folded in the frequency domain. Specifically, the time constant defines the timescale on which the system trusts inertial navigation data and the timescale on which it reverts to position sensor data. In one embodiment, the selection of the time constant needs to balance noise suppression and dynamic tracking requirements. If the time constant is set too large, the system will rely more on the integral of the inertial velocity, resulting in a very smooth altitude curve, but a slower response to position corrections. If the time constant is set too small, the system will quickly approach the second fused altitude, but this may introduce more position sensor noise. The system presets an optimal time constant based on the aircraft's maneuverability level and the accuracy level of the inertial devices, for example, between two and five seconds in a normal landing mode.

[0096] Step S42: Construct a complementary filtering model based on the time constant, the celestial velocity data, and the second fused altitude data.

[0097] The complementary filtering model is constructed based on the principle of a state-space observer. Within each computation cycle, the system first uses the fused height from the previous moment plus the product of the celestial velocity and the sampling period to obtain a predicted height based on inertial estimation. Subsequently, the system calculates the residual between this predicted height and the current second fused height. This residual reflects the accumulated error of the inertial estimation and the random noise from the position sensor. The system multiplies this residual by a feedback gain determined by a time constant and compensates for it in the predicted height. Mathematically, this structure is equivalent to low-pass filtering the second fused height while high-pass filtering the integral of the celestial velocity; the sum of these two operations maintains unity gain across the entire frequency band, thus achieving lossless signal fusion.

[0098] Step S43: Calculate the initial third fusion height data using the complementary filtering model.

[0099] The calculated initial third-fusion altitude data possesses excellent characteristics: it avoids the long-term drift of barometers and the instantaneous jumps of satellite navigation, while also exhibiting the high-frequency dynamic characteristics of inertial navigation. During maneuvers requiring precise vertical velocity feedback, such as leveling landings, the third-fusion altitude provides a much more accurate position reference than a single sensor. For example, when an aircraft encounters a downdraft causing a sudden drop in altitude, the inertial azimuth velocity immediately senses the downward acceleration and reflects it in the velocity change. The complementary filter rapidly adjusts the output of the third-fusion altitude without waiting for data updates from satellites or ranging devices with longer sampling periods. This millisecond-level response speed is crucial for preventing hard landings.

[0100] Step S44: Obtain the position information of the aircraft relative to the target landing area, and determine whether the aircraft has passed over the target landing area.

[0101] This step introduces the concept of geofencing to further optimize the fusion logic during the near-ground phase. The system compares the horizontal position information provided by satellite navigation with the pre-stored coordinates of the target landing area. The target landing area typically refers to the runway threshold or the predetermined landing point. Determining whether the aircraft has entered the airspace above this area is to identify whether the aircraft is already above the final landing runway. In one possible implementation, the system sets a rectangular detection area; when the aircraft's latitude and longitude coordinates fall within this rectangular area and its altitude is below a preset threshold, it is determined that it has passed over the target landing area.

[0102] Step S45: If the aircraft passes over the target landing area, then set the specific weight value in the third fused altitude data to zero to obtain the updated third fused altitude data.

[0103] When the aircraft confirms it is over the runway, the environmental characteristics change significantly. At this point, the ground is extremely smooth, and the feedback from ranging equipment (such as radar altimeters) becomes highly reliable. However, the air pressure altitude is subject to severe nonlinear distortion due to the ground effect (high-pressure area beneath the wings). To prevent distorted air pressure data from contaminating the final output through the feedback loop of the complementary filter, the system either specially processes the feedback weights corrected at the corresponding positions in the complementary filter or resets the weights of the air pressure contribution to zero.

[0104] Specifically, the zeroing operation means that, at this instant, the system cuts off certain feedback paths that could introduce interference, making the third fused altitude rely more purely on the fusion of the ranging device and inertial velocity. This ensures that in the final seconds before touchdown, the altitude data will not produce false ascent / descent indications due to air pressure fluctuations caused by the ground effect. The updated third fused altitude data becomes the sole highest-priority reference for the aircraft to execute automatic landing commands.

[0105] Step S5: Detect the status information obtained by the satellite navigation device. If the status information changes between the first positioning state and the second positioning state, then soften the third fused altitude data through a softening function to obtain the final output altitude data.

[0106] In actual flight, satellite navigation systems may experience switching between different accuracy levels due to satellite signal blockage, electromagnetic interference, or interruption of base station data. For example, a jump from a centimeter-level RTK fixed solution (first positioning state) to a meter-level single-point positioning solution (second positioning state). This state jump is often accompanied by a sharp step change in altitude, sometimes reaching several meters. If this jump directly enters the control system, it can trigger violent pitch maneuvers in the aircraft, or even lead to loss of control. This step softens this instantaneous step error, transforming it into a gradual trend change.

[0107] Step S51: Determine whether the status information has changed from the first positioning state to the second positioning state.

[0108] The system makes this determination by monitoring the status words in the satellite navigation protocol in real time. The first positioning status usually represents a high-precision status, and the second positioning status represents a low-precision status.

[0109] Step S511: Obtain the first state identifier of the satellite navigation device at the current moment and the second state identifier at the previous moment.

[0110] The navigation task module caches the state from the previous moment in each cycle. By comparing the state flags of two consecutive cycles, the system can accurately detect instantaneous changes in the positioning mode.

[0111] Step S512: Compare the first status identifier with the second status identifier.

[0112] The comparison logic not only focuses on whether the state has changed, but also on the direction of the change. A jump from high precision to low precision usually means a decrease in data reliability, requiring immediate activation of protection mechanisms.

[0113] Step S513: If the first state identifier represents the second positioning state and the second state identifier represents the first positioning state, then determine that the state information changes from the first positioning state to the second positioning state.

[0114] Once this downward state transition is confirmed, the system immediately records the difference between the third fusion height and the height before the transition at the instant the transition occurs. This difference is defined as the deviation term to be eliminated.

[0115] Step S52: If the status information changes from the first positioning state to the second positioning state, then obtain the maximum allowable single-step altitude change value of the aircraft.

[0116] To ensure smooth flight, the control system has strict requirements on the continuity of altitude input. The maximum single-step altitude change is preset based on the aircraft's dynamic characteristics, flight speed, and the stability margin of the control law. For example, if the allowable altitude correction per second is set to no more than 0.5 meters, then converted to a sampling period of 20 milliseconds, the maximum single-step change may only be 0.01 meters.

[0117] Step S53: Construct a first softening function based on the maximum value of the single-step height change, and calculate the third fused height data using the first softening function to obtain the final output height data.

[0118] The first softening function is essentially a bias canceller with a limit. When a jump occurs, the system does not directly adopt the new height after the jump, but maintains the height reference before the jump, and on this basis, moves closer to the new height by a small step size (i.e., the maximum value of the single-step change) in each cycle.

[0119] For example, if the satellite's altitude jumps by two meters instantaneously, the softening function initiates a compensation process. In the first cycle after the jump, the final output altitude changes by only 0.01 meters from the original altitude, with the remaining 1.99-meter deviation stored in the compensator. Over the next few hundred cycles, the compensator continuously releases this deviation until the final output altitude fully catches up with the third fused altitude. This processing method transforms a potentially dangerous step signal into an extremely weak ramp signal, making the change virtually imperceptible to the spacecraft during flight, thus ensuring an absolutely smooth landing.

[0120] Step S54: Determine whether the status information changes from the second positioning state to the first positioning state.

[0121] Similarly, when satellite signals recover and the accuracy jumps back from low to high, a reverse altitude step also occurs. While this is a jump to more accurate data, an instantaneous step is also not permitted.

[0122] Step S55: If the status information changes from the second positioning state to the first positioning state, then obtain the maximum allowable single-step altitude change value of the aircraft.

[0123] The maximum value of the single-step change used at this time can be the same as that in step S52, or it can be set slightly larger according to business needs, so as to return to the high-precision benchmark more quickly.

[0124] Step S56: Construct a second softening function based on the maximum value of the single-step height change, and calculate the third fused height data using the second softening function to obtain the final output height data.

[0125] The second softening function operates similarly to the first, ensuring that when the high-precision signal is recovered, the system can gracefully and smoothly reacquire the high-precision data. Through this bidirectional softening process, regardless of fluctuations in the satellite navigation status, the altitude data ultimately output to the flight control system remains continuous, smooth, and physically achievable.

[0126] It should be noted that this softening process is not only applicable to satellite state transitions, but can also be extended to any abnormal sensor switching scenario. It constructs a robust safety barrier between the underlying data fusion and the upper-level flight control. The final output altitude data not only integrates the advantages of multiple sensor sources, but also achieves dynamic performance through complementary filtering and extremely high safety through the softening process.

[0127] In one embodiment, the softening function also incorporates weighted adjustments based on the aircraft's current remaining altitude. If the aircraft is already very close to the ground, the softening step size is appropriately reduced to prevent unnecessary compensatory motion at the moment of touchdown. This refined control logic enables this solution to perfectly handle the entire altitude navigation task from high-altitude descent to precise landing.

[0128] Specifically, during the softening calculations performed in S53 or S56, the system also monitors the remaining deviation of the compensator in real time. If the deviation is not eliminated within a predetermined time, or if a new state transition occurs during the elimination process, the system will employ a deviation reset strategy to recalculate the softening curve. This dynamic adaptability ensures that the system can still output stable and reliable altitude information even in extremely complex electromagnetic environments.

[0129] Through the complete process described in S1 to S5, this embodiment of the invention implements a multi-level, comprehensive aircraft altitude fusion solution. From the preprocessing of raw data, to the initial fusion of air pressure and satellite data, to the interval transition of the ranging device, followed by the introduction of dynamic enhancement of inertial velocity, and finally safety hardening through a softening function, each step is optimized to address specific pain points in the landing phase, collectively ensuring that altitude information is accurate, smooth, and real-time at all times and under all conditions. This provides the core technical guarantee for the safe landing of aircraft, significantly improving the robustness of the automated landing system and the user experience.

[0130] In one possible implementation, the system also links the final output altitude data with the aircraft's terrain perception system. If there is a significant conflict between the softened altitude and the altitude detected by the terrain system, the system will trigger the highest level of obstacle avoidance logic. This multi-dimensional safety verification makes this altitude fusion method not only a navigation algorithm, but also a complete flight safety assurance system.

[0131] Step S4: Obtain the celestial velocity data output by the inertial navigation device, and perform complementary filtering and fusion with the celestial velocity data and the second fused altitude data to obtain the third fused altitude data.

[0132] Specifically, during the dynamic process of aircraft landing, the real-time accuracy and smoothness of altitude information are crucial for the inner-loop control of the autopilot. While the second fused altitude data addresses long-term accuracy issues through the fusion of multi-source position sensors, inherent delays in data acquisition, processing, and transmission from satellite navigation, barometers, and ranging devices, coupled with their typically low sampling frequencies, make it difficult to directly reflect the instantaneous dynamic changes of the aircraft in the vertical direction. Inertial navigation systems, through their built-in high-precision accelerometers and gyroscopes, can output the aircraft's motion status information at extremely high frequencies. Among these, the celestial velocity data, obtained by integrating vertical acceleration, possesses extremely high short-term accuracy and very low latency, enabling it to sensitively capture instantaneous ascent and descent trends caused by airflow disturbances, control surface deflection, or power changes. However, due to the zero-bias instability of accelerometers, the celestial velocity exhibits significant drift errors after long-term integration. The introduction of complementary filtering fusion aims to utilize the long-term stability of the second fused altitude data to suppress the cumulative drift of inertial navigation, while simultaneously leveraging the high-frequency response characteristics of the celestial velocity to compensate for the shortcomings of position sensors in dynamic response.

[0133] In one embodiment, the inertial navigation device is typically constructed using a microelectromechanical system (MEMS) or a fiber optic gyroscope. The acquisition of azimuth velocity involves complex coordinate transformations. The system first acquires the three-axis acceleration in the body coordinate system, combines it with the current attitude angle information, transforms it to the geographic coordinate system using a rotation matrix, and subtracts the influence of gravitational acceleration to obtain the azimuth acceleration in the geographic coordinate system. Subsequently, the azimuth velocity is obtained through a high-frequency integration algorithm. To ensure the effectiveness of the fusion, the system preprocesses the azimuth velocity, including removing high-frequency noise caused by body vibration and correcting accelerometer temperature drift using a temperature-compensated model.

[0134] Step S41: Obtain the preset time constant.

[0135] The time constant is a core physical parameter in complementary filter design, defining the overlap point between position and velocity information in the frequency domain. Specifically, the time constant determines the filter's level of trust in signals across different frequency bands. In one possible implementation, the time constant reflects the system's assessment of the short-term reliability of inertial navigation data. A larger time constant means the system trusts the integral of the inertial velocity over a longer time span, resulting in excellent smoothness and dynamic tracking of the output altitude data, but slower correction of position deviations. Conversely, a smaller time constant allows the system to utilize the second fused altitude more frequently to correct the output, which enhances absolute position accuracy but may introduce more sensor noise.

[0136] For example, during a smooth descent, the system uses a larger time constant, such as three to five seconds, to obtain more stable control feedback. During the leveling phase, which requires a rapid response, the system may dynamically decrease the time constant to ensure absolutely real-time altitude feedback. The time constant can be retrieved from the flight control computer's non-volatile memory or obtained from a preset parameter table based on the current flight conditions, such as the magnitude of vertical velocity or vibration level.

[0137] Step S42: Construct a complementary filtering model based on the time constant, the celestial velocity data, and the second fused altitude data.

[0138] The complementary filtering model is constructed based on the principle of complementary signal spectra. Mathematically, this model typically consists of a prediction stage and a correction stage. The prediction stage uses the fused altitude from the previous moment and the current azimuth velocity to perform dead reckoning, assuming that the current altitude equals the previous moment's altitude plus the displacement increment of the azimuth velocity within the sampling period. The correction stage calculates the residual between the predicted altitude and the second fused altitude, and feeds this residual back to the integrator through a proportional gain.

[0139] Specifically, this proportional gain is inversely proportional to the time constant. With this structure, the second fusion height participates in fusion after passing through an equivalent low-pass filter, while the upward velocity participates in fusion after integration through an equivalent high-pass filter. The sum of the gains of both remains constant across the entire frequency band. This model can automatically filter out high-frequency jitter in the second fusion height while eliminating low-frequency drift in the upward velocity. In one embodiment, to further improve performance, the complementary filtering model also includes a second-order correction term to estimate and compensate for the accelerometer's zero-bias error in real time, thereby making the integration of the upward velocity more accurate.

[0140] Step S43: Calculate the initial third fusion height data using the complementary filtering model.

[0141] Within each navigation calculation cycle, the system executes a complementary filtering algorithm. The initial third fused altitude data is a product of deep coupling between position and velocity information. It not only closely follows the second fused altitude in value, ensuring that the absolute reference for navigation is not lost, but also exhibits high-frequency characteristics consistent with inertial navigation in terms of waveform.

[0142] For example, when an aircraft encounters a sudden downdraft, the second fused altitude might only show an altitude decrease a fraction of a second later due to sensor measurement delays, but the azimuth velocity would immediately reflect the downward velocity increment. The complementary filtering model responds quickly to this velocity change, allowing the initial third fused altitude to reflect the altitude decrease almost in real time. This characteristic is extremely valuable for the vertical rate control loop of the flight control system, significantly improving the trajectory tracking accuracy of the aircraft under complex weather conditions.

[0143] Step S44: Obtain the position information of the aircraft relative to the target landing area, and determine whether the aircraft has passed over the target landing area.

[0144] This step introduces spatial position constraints designed to identify whether the aircraft has entered a critical landing physical environment. The target landing area typically refers to the designated touchdown zone of the runway or the central area of ​​a helicopter landing platform. The system obtains the aircraft's current latitude and longitude coordinates via satellite navigation equipment and compares them with a pre-loaded landing field database.

[0145] The judgment process involves a geofencing algorithm. The system constructs a virtual polygonal region centered on the target landing point. When the aircraft's horizontal position coordinates fall within this polygon and its current altitude is below a preset safety threshold, the system determines that the aircraft has passed over the target landing area. The technical significance of this judgment is that when the aircraft is above the runway, the terrain below it becomes extremely flat and known, resulting in the highest reliability of the ranging equipment's data, while the risk of barometer interference from ground effects also increases.

[0146] Step S45: If the aircraft passes over the target landing area, then set the specific weight value in the third fused altitude data to zero to obtain the updated third fused altitude data.

[0147] Once the aircraft confirms it is over the target landing area, the system adjusts the internal structure of the complementary filter. Specific weight values ​​typically refer to the contribution of atmospheric pressure to the fusion process, or a feedback gain in the complementary filter used to correct long-term drift.

[0148] In one embodiment, when an aircraft is flying near the ground, the downwash airflow generated by the wings or rotor creates a high-pressure zone beneath the fuselage, causing a false drop in the pressure altitude measured by the barometer—a phenomenon known as the ground effect. If the pressure data is fused using the conventional weighting, the third fused altitude will show an upward deviation, causing the autopilot to mistakenly believe it is too high and lower the nose, creating a hazard. By resetting the weighting of the pressure contribution to zero, the system relies entirely on a combination of ranging equipment and inertial navigation during this phase. The updated third fused altitude data eliminates the interference of the ground effect, ensuring the absolute accuracy of the altitude information in the final few meters before touchdown. This dynamic weighting based on geographic location is a key safety strategy for ensuring precise landing.

[0149] Step S5: Detect the status information obtained by the satellite navigation device. If the status information changes between the first positioning state and the second positioning state, then soften the third fused altitude data through a softening function to obtain the final output altitude data.

[0150] Throughout the entire descent of an aircraft, the stability of the satellite navigation system is the cornerstone of a highly integrated system. However, due to limitations such as satellite signal obstruction, multipath interference, or signal fluctuations from ground reference stations, the positioning accuracy mode of satellite navigation frequently switches. The first positioning state typically refers to a high-precision carrier phase differential positioning mode, with an altitude error usually on the order of centimeters. The second positioning state refers to a lower-precision pseudorange differential positioning or single-point positioning mode, with an altitude error that can reach the order of meters. When the positioning mode switches between the two, the altitude value output by the satellite will experience a momentary step. If this step is directly transmitted to the fusion algorithm, it will cause severe jitter in the third fused altitude. The purpose of the softening process is to smoothly decompose this transient step deviation over time, allowing it to be integrated into the final output in a controlled and gradual manner.

[0151] Step S51: Determine whether the status information has changed from the first positioning state to the second positioning state.

[0152] The system obtains status indicators by parsing the protocol messages output by satellite navigation equipment in real time. The decision-making logic focuses on the degradation of positioning quality.

[0153] Step S511: Obtain the first state identifier of the satellite navigation device at the current moment and the second state identifier at the previous moment.

[0154] The navigation system maintains a state buffer that records the positioning pattern for multiple consecutive cycles. The first state identifier represents the positioning quality of the current frame, and the second state identifier represents the positioning quality of the previous frame. Through this timing comparison, the system can detect the edge-triggered signals of state transitions.

[0155] Step S512: Compare the first status identifier with the second status identifier.

[0156] The comparison process involves not only simple numerical comparisons but also an evaluation of the positioning mode level. The system has a preset status level table that sorts different positioning modes from high to low accuracy.

[0157] Step S513: If the first state identifier represents the second positioning state and the second state identifier represents the first positioning state, then determine that the state information changes from the first positioning state to the second positioning state.

[0158] When the system detects a drop in positioning mode from high precision to low precision, it indicates a potential systematic deviation of several meters in the current satellite altitude data. In this case, the system immediately initiates softening logic. In one embodiment, the system records the difference between the third fused altitude and the altitude at the moment the jump occurs. This difference is considered a step disturbance that needs to be gradually eliminated.

[0159] Step S52: If the status information changes from the first positioning state to the second positioning state, then obtain the maximum allowable single-step altitude change value of the aircraft.

[0160] The maximum single-step altitude change is a safety constraint parameter that defines the maximum rate of change of the altitude signal that the flight control system can accept. This value is set by taking into account both the aircraft's vertical maneuverability and the stability of the control law. If the altitude signal changes too rapidly, the control system will generate excessive control surface correction commands, leading to airframe vibration.

[0161] For example, for a fixed-wing aircraft performing an approach and landing, the system might set the maximum single-step change to no more than 0.005 meters every twenty milliseconds. This means that even if the satellite altitude jumps by three meters, the final output altitude will only approach the target value by five millimeters in each calculation cycle. This parameter ensures the continuity of navigation output, allowing the aircraft to maintain a stable flight attitude even when it senses a decrease in positioning accuracy.

[0162] Step S53: Construct a first softening function based on the maximum value of the single-step height change, and calculate the third fused height data using the first softening function to obtain the final output height data.

[0163] The first softening function is a rate-limited dynamic compensator. Its working principle is as follows: upon detecting a transition, the system does not directly output the third fusion height after the transition, but instead maintains a compensation value. This compensation value is equal to the amplitude of the transition at the instant of the transition, and then gradually decreases according to the maximum value of the single-step change in each subsequent cycle until it reaches zero.

[0164] Specifically, the final output altitude equals the current third fused altitude plus this dynamically changing compensation value. In this way, the originally instantaneous step is stretched into a ramp with a very small slope. In one possible implementation, the first softening function can also adopt an exponential decay form, making the deviation elimination process faster initially and slower later, thus returning to the accurate value more quickly while maintaining smoothness. This processing ensures that the altitude data does not change abruptly when the satellite signal deteriorates, providing valuable reaction time for the pilot or autopilot.

[0165] Step S54: Determine whether the status information changes from the second positioning state to the first positioning state.

[0166] This step focuses on restoring positioning accuracy. When the aircraft flies out of the obstruction area or the interference disappears, and the satellite navigation system regains a high-precision solution, the altitude value will jump back to a more accurate reference.

[0167] Step S55: If the status information changes from the second positioning state to the first positioning state, then obtain the maximum allowable single-step altitude change value of the aircraft.

[0168] While improved positioning accuracy is beneficial, sudden changes in values ​​can still disrupt the control system. In this case, the maximum value of the single-step change can be consistent with the parameter in step S52, or it can be set slightly larger as needed. This is because during accuracy recovery, the system typically wants to utilize more accurate data as quickly as possible, as long as the rate of change is within the tolerance range of the control system.

[0169] Step S56: Construct a second softening function based on the maximum value of the single-step height change, and calculate the third fused height data using the second softening function to obtain the final output height data.

[0170] The operation logic of the second softening function is symmetrical to that of the first softening function. It is responsible for smoothing the altitude difference caused by upward jumps. Through this bidirectional softening mechanism, this embodiment of the invention constructs an extremely robust altitude output layer. No matter how the satellite navigation state changes repeatedly, the final output altitude data always maintains physical continuity and logical rigor.

[0171] In one embodiment, the softening function also adaptively adjusts based on the current flight altitude. For example, when the aircraft is only ten meters from the ground, the system tightens the maximum value of single-step changes, because at this point, any tiny fluctuation in altitude could affect the touchdown maneuver. During the higher altitude phase, the restrictions can be relaxed appropriately. This intelligent softening strategy ensures optimal altitude navigation support for the aircraft at every stage of descent.

[0172] Furthermore, to enhance the system's robustness, the softening function monitors the changing trend of the third fused altitude in real time during processing. If it detects that the direction of the satellite's altitude jump is opposite to the actual motion trend of the spacecraft, the softening function will increase the compensation force. For example, if the spacecraft is descending, but the satellite altitude suddenly jumps upward due to a state switch, the softening function will strictly limit the upward shift of the output altitude to ensure that the navigation information does not violate basic physical laws.

[0173] It should be noted that the generation of the final output altitude data is a closed-loop monitoring process. The system continuously evaluates the effectiveness of the softening process. If the compensation value fails to return to zero for an extended period, it indicates a persistent systemic deviation in satellite navigation. In this case, the system will alert the flight control system via an alarm module to switch to a backup navigation mode based on barometric pressure and ranging. This multi-layered safety assurance demonstrates the rigor of this solution in practical engineering applications.

[0174] In one possible implementation, the softening function can also be verified in conjunction with the vertical cumulative displacement of inertial navigation. At the instant of a satellite altitude jump, the system uses the altitude change calculated by inertial navigation as a reference for the actual motion, completely identifying the satellite altitude jump as measurement noise and isolating it entirely through the softening function. This kinematically constrained softening process can achieve higher navigation accuracy than simple rate limiting.

[0175] Specifically, during steps S53 and S56, the system maintains an offset buffer. Whenever a state transition occurs, the buffer captures the current step value. In subsequent navigation loops, the system not only calculates altitude softening but also simultaneously softens vertical velocity. This is because an altitude step inevitably leads to a momentary pulse in vertical velocity; by synchronously limiting and softening the velocity signal, the speed loop of the flight control system can be kept stable. This full-state softening scheme is a necessary condition for achieving high-level automatic landing missions.

[0176] For example, suppose that during the descent of an aircraft, the satellite navigation degrades from precise positioning mode to normal positioning mode, and the altitude data jumps instantaneously from 50.5 meters to 52.0 meters. At this moment, the system detects a deviation of 1.5 meters. If the maximum single-step change is set to 0.01 meters, then in the next cycle, the output altitude will be 50.51 meters, not 52.0 meters. After 150 cycles, or approximately three seconds, the output altitude will smoothly transition to the new satellite reference. During these three seconds, the aircraft's attitude remains stable, and the control system has ample time to verify the authenticity of this altitude change using other sensors, such as barometers or rangefinders, thereby avoiding erroneous operation.

[0177] Through the deep integration and security processing from S4 to S5 described above, this embodiment of the invention not only enhances the dynamic response capability of altitude data, but also solves the most challenging environmental interference and sensor anomaly issues during the landing phase by adjusting geofencing weights and softening satellite state transitions. This progressive and interconnected logical architecture enables the aircraft to maintain a smooth, accurate, and reliable altitude curve even in complex landing environments, laying a solid technical foundation for safe and precise landing.

[0178] In a preferred embodiment, the system also provides real-time telemetry output of the softening function's execution status. Ground stations or airborne monitoring systems can view the current compensation value and softening progress in real time. This provides detailed data support for post-flight data analysis and system performance evaluation. Simultaneously, this transparent design facilitates parameter tuning for different aircraft models, enhancing the algorithm's versatility and portability.

[0179] Furthermore, to address the performance of ranging devices in complex terrain, the system incorporates a signal quality assessment factor when performing the weighted nulling operation of S45. If, although the ranging device is over the target area, ground cover such as snow or water causes a decrease in echo quality, the system dynamically adjusts the nulling level rather than simply cutting off other sensors. This application of fuzzy logic further enhances the system's survivability in extreme environments.

[0180] When processing the S511 status identifier acquisition, the system also considers signal transmission jitter. To prevent false status judgments due to the loss of individual data packets, the system employs a counter-based confirmation mechanism. The softening logic is only formally triggered when a status transition occurs continuously for more than three consecutive sampling periods. This jitter reduction effectively reduces the system's false alarm rate, ensuring that softening protection is only activated when truly needed.

[0181] For the maximum single-step altitude change in S52, the system can also dynamically compensate based on the current vertical velocity. If the aircraft is undergoing a rapid descent, the maximum single-step change will increase accordingly to avoid the navigation altitude lagging significantly behind the actual altitude due to softening processing. This combination of static and dynamic parameter management reflects the profound understanding of flight mechanics in this solution.

[0182] In one embodiment, the second softening function employs a predictive algorithm when handling precision recovery jumps. The system predicts the stable value after high-precision signal recovery and approximates it using a second-order continuous curve. This ensures that the altitude output is not only numerically continuous but also smooth in terms of acceleration, greatly optimizing the aircraft's ride comfort and structural stress.

[0183] It should be noted that the entire highly fusion algorithm is designed following the principles of redundancy management and fault isolation. Each sub-step has an independent validity check. If an anomaly is detected in the input data during complementary filtering or softening processing, the system can quickly fall back to the stable state of the previous level. This highly reliable architecture allows the method to meet the stringent requirements of navigation systems in civil aviation or high-level unmanned aerial vehicles.

[0184] Specifically, after obtaining the initial third fused altitude through S43, the system compares it with the barometric altitude over a long period. If the deviation between the two continues to widen beyond a physically reasonable range, the system automatically triggers a time constant reset of the complementary filter, forcing the system to return to the position reference and preventing heading deviations caused by inertial navigation anomalies. This self-correcting capability ensures the long-term operational stability of the system.

[0185] When processing S44 position information, the system also incorporates the aircraft's heading angle information. By determining whether the aircraft is directly facing the runway centerline, it can more accurately predict the physical environment it will enter. For example, if the aircraft is in the runway area but in a large-angle sideslip state, the system will appropriately delay the weight reset to zero until the aircraft's attitude stabilizes. This decision-making logic based on multi-dimensional state awareness significantly improves the intelligence level of the fusion algorithm.

[0186] For example, in an actual landing mission, the aircraft encountered strong ground effect interference at an altitude of 30 meters, and the satellite navigation system switched positioning modes due to obstruction by nearby buildings. At this point, this solution first identifies the ground effect zone through steps S44 and S45, decisively eliminating the interfering air pressure data, and maintaining real-time altitude accuracy using the rangefinder and inertial navigation. Subsequently, in response to the sudden change in satellite positioning mode, the softening function in step S5 immediately intervenes, smoothing out the two-meter altitude jump. Ultimately, the flight control system receives a nearly undisturbed, smoothly descending altitude curve, and the aircraft finally touches the ground smoothly with a very low vertical velocity.

[0187] This fusion method also excels in handling sensor heterogeneity. Whether it's satellite navigation with a sampling frequency of 10 Hz, a barometer with a sampling frequency of 50 Hz, or inertial navigation with a sampling frequency as high as 200 Hz, the most suitable fusion point can be found within this framework. The system ensures that all data is time-aligned before entering the complementary filter through high-precision clock synchronization and interpolation algorithms.

[0188] In one embodiment, the system also monitors the integrity of the final output altitude data. By calculating the protection level of the fused result and comparing it with preset alarm limits, the system can assess the reliability of the current altitude information in real time. If the protection level exceeds the limit, the system will alert the pilot through the human-machine interface or trigger automatic go-around logic. This comprehensive navigation performance monitoring is an important component in achieving a closed loop for autonomous flight safety.

[0189] Furthermore, for the first softening function in S53, the system can also introduce a correction term based on energy conservation. At the instant of an altitude jump, the system analyzes the changes in the aircraft's kinetic and potential energy to determine whether the jump conforms to physical laws. If not, the softening function will intervene with a stronger constraint. This mathematical processing that introduces physical constraints gives the navigation system a preliminary ability to make common-sense judgments, effectively identifying and filtering various non-physical measurement noises.

[0190] When processing the second softening function of S56, the system considers the integral saturation problem of the control system. If the softening process is too slow, it may cause the integral term of the control system to accumulate too much, resulting in overshoot. To address this, the softening function shares state information with the control system, and appropriately accelerates the softening rate when it detects that the control system is approaching saturation. This cross-module collaborative optimization demonstrates the deep integration advantage of this solution as a core component of flight control.

[0191] In summary, the altitude fusion method provided by this invention constructs a sensitive yet robust altitude navigation system through dynamic enhancement in step S4 and safety softening in step S5. It not only resolves the inconsistencies in time, space, and accuracy among multi-source sensors but also overcomes safety risks caused by environmental interference and system jumps through ingenious algorithm design. The implementation of this solution can significantly improve the automation level and safety of aircraft during the landing phase, providing strong support for the development of modern aviation navigation technology.

[0192] In practical deployments, the algorithm is written as efficient embedded code and runs on processors with floating-point units. Rigorous real-time and coverage tests ensure stable output under various extreme flight conditions. Furthermore, the algorithm's parameterized design allows it to easily adapt to various platforms, from micro-drones to large transport aircraft, demonstrating significant commercial application value.

[0193] In one possible implementation, the system also utilizes historical flight data to perform offline optimization of the time constant and softening step size. By analyzing playback of thousands of landing events, the system can automatically identify the parameter combination that best balances smoothness and accuracy. This continuous evolutionary capability based on big data allows the fusion method to become increasingly accurate with accumulated user experience.

[0194] Specifically, when constructing the complementary filter model using S42, the system also considers the structural vibration frequencies of the aircraft. By adding notch filtering at the filter's inlet, specific frequency interference caused by engine or rotor rotation can be effectively eliminated, resulting in cleaner utilization of azimuth velocity. This refined processing based on physical characteristics is an effective means of improving navigation accuracy.

[0195] When processing state transitions in S513, the system also considers the number of visible satellites and the geometric precision factor. If a transition is accompanied by a drastic change in the number of visible satellites, the system assigns a higher initial weight to the softening function, as such transitions typically involve greater uncertainty. This comprehensive analysis based on multi-source information makes the softening logic more scientific and reasonable.

[0196] For example, when landing on complex terrain, the ranging device may experience frequent jumps due to flying over trees or buildings. In this case, the complementary filter in step S4 effectively filters out this positional noise by leveraging the smoothness of inertial navigation, while the softening logic in step S5 ensures that the altitude reference remains rock-solid even when satellite signals are unstable. This combination of multiple protection mechanisms allows the aircraft to perform precise landings in complex environments, just as it would at a flat airport.

[0197] In one embodiment, the system also performs a redundancy comparison on the final output altitude data. If there are two independent fusion systems on board, they exchange calculation results in real time. If the output difference between the two systems exceeds a threshold, the system initiates a voting logic, combining the original sensor status to determine the source of the fault. This system-level redundancy design further strengthens the foundation of flight safety.

[0198] Through this series of rigorous steps and meticulous processing, the embodiments of the present invention successfully transform complex sensor data into simple and reliable altitude commands. It is not merely a collection of mathematical formulas, but a profound summary of practical flight experience. At the critical moment of landing, which determines the success or failure of a flight, the altitude information provided by this method is the most reliable lifeline for the aircraft.

[0199] When processing the softening functions of S53 and S56, the system also considers the altimeter's range switching. Some ranging devices switch operating modes in different altitude ranges, which can also cause slight fluctuations. The softening logic of this solution can also cover these scenarios, ensuring smooth output across the entire altitude range. This one-stop smoothing solution greatly simplifies the input processing logic of the flight control system.

[0200] Furthermore, when the system performs the weight reset to zero in S45, it also simultaneously records the deviation between the barometric altitude and the current fused altitude. When the aircraft resumes its climb and leaves the ground effect zone, the system uses this recorded deviation for reverse compensation, ensuring that the altitude data can smoothly revert to the barometric reference after leaving the landing area. This full lifecycle deviation management demonstrates the completeness of the algorithm design.

[0201] In one possible implementation, the softening function can also employ a nonlinear mapping. A faster elimination rate is used for smaller jumps, while a very slow elimination rate is used for larger jumps. This strategy ensures rapid correction of small errors while effectively preventing control shocks caused by large errors. This human-centered design approach makes the navigation system's performance more intuitive for pilots.

[0202] In summary, the highly integrated method of this invention, through in-depth analysis and ingenious combination of inertial, satellite, barometric, and ranging data, constructs a high-altitude perception framework covering the entire time and airspace. It not only achieves industry-leading technical specifications but also demonstrates exceptional survivability and reliability in practical applications. Whether in the stillness of the night or the raging heat of day, this system safeguards every landing of the aircraft.

[0203] In future technological evolution, this framework can be further integrated with visual navigation and terrain matching data. By introducing more location observation sources in step S4 and correspondingly adding a softening protection channel in step S5, the system can achieve higher autonomous navigation accuracy. This open architecture design ensures that this invention maintains a leading technological position for a long time, making a continuous contribution to the development of the aerospace industry.

[0204] Specifically, when executing S41 to acquire the time constant, the system can also dynamically adjust the aircraft's mass model based on current fuel consumption or load changes, thereby more accurately predicting the dynamic response characteristics of inertial navigation. This interdisciplinary parameter coupling pushes navigation accuracy to new heights. Every tiny improvement is aimed at providing the aircraft with the most perfect guidance at that decisive moment of ground contact.

[0205] When processing the S512 state comparison, the system also considers clock jumps in the satellite navigation system. Although this is extremely rare, it can have a significant impact on altitude calculations when it does occur. By monitoring receiver autonomous integrity monitoring alarm flags, the system can block abnormal data before the softening function intervenes, ensuring the purity of the navigation source. This contingency plan for extreme anomalies is a hallmark of mature aviation products.

[0206] For example, in a landing test at a high-altitude airport, the thin air reduced the sensitivity of the barometer, and the undulating terrain caused severe satellite multipath effects. This solution, through complementary filtering in step S4, fully utilizes the stable performance of inertial navigation in thin atmosphere, and through softening processing in step S5, successfully shields the frequent altitude jumps caused by terrain. The aircraft ultimately achieved a perfect precision landing on the runway, verifying the algorithm's superior performance in extreme environments.

[0207] In one embodiment, the system also performs frequency characteristic analysis on the final output altitude data. If unwanted mechanical resonance frequencies are found in the output signal, the system automatically adjusts the filter model parameters in S42 to achieve adaptive vibration suppression. This closed-loop performance optimization enables the navigation system to automatically adapt to different airframe structures and installation environments.

[0208] Finally, this solution is designed with computational resources in mind. All fusion and softening logic has been mathematically simplified, avoiding complex matrix inversions or high-order iterative operations. This allows it to run smoothly even on low-power processors with limited computing power, making it an ideal choice for power-sensitive platforms such as long-endurance UAVs. This balance between performance and efficiency is one of the core advantages of this invention in commercial competition.

[0209] Example 1:

[0210] To address the issue of inaccurate and unstable altitude estimation relative to the runway due to interference with satellite navigation and positioning during the high-altitude descent phase of a UAV, the following method is proposed: First, during the deep descent phase, barometric altitude is prioritized to ensure a sufficiently smooth estimated altitude value. Subsequently, as the UAV descends, the altitude source gradually transitions from barometric altitude to radar altitude. When the UAV detects passing over the runway, the method switches to fully accepting millimeter-wave radar altitude. This method ensures higher accuracy in altitude estimation as the UAV approaches the ground, while maintaining smoothness in the estimation. Furthermore, throughout the descent phase, sensor altitude values ​​are fused with the azimuth velocity output from the inertial navigation system to further ensure the smoothness of the altitude estimation.

[0211] like Figure 1 As shown, the specific implementation steps are as follows: Step (1): Import the raw altitude values ​​and equipment status information measured by the satellite navigation receiver, atmospheric equipment, and millimeter-wave radar sensor, and preprocess the altitude values ​​to convert all three types of altitude values ​​into relative altitude values ​​relative to the landing point; Step (2): Perform altitude fusion of satellite navigation, barometric pressure and millimeter-wave radar altimeter to gradually improve the accuracy of the relative altitude estimated by the UAV during the descent phase; Step 2.1: Perform satellite navigation altitude and barometric altitude fusion. The fusion calculation formula is as follows: (1) (2) in The altitude at which air pressure and satellite navigation are integrated. This indicates taking the median value among the three. This indicates that the air pressure is higher relative to the runway. Indicates the satellite guidance altitude relative to the runway. This represents the absolute value of the altitude error in satellite point positioning.

[0212] Step 2.2: Use By integrating with millimeter-wave radar, the altitude value gradually transitions from the fused altitude of air pressure and satellite navigation to the millimeter-wave radar altitude. The calculation formula is as follows: (3) (4) in The altitude at which air pressure and satellite navigation are integrated. This indicates the altitude above the ground measured by millimeter-wave radar. Represents the fusion coefficient. and This indicates the upper limit for performing a height source transition.

[0213] Step (3): Use the celestial velocity information and The data is fused, and the fusion weight value is adaptively adjusted according to the position of the drone relative to the runway. Step 3.1: Calculate using complementary filtering and The fusion result is calculated using the following formula: (5) in It is the fusion height of the algorithm's final output. is the time constant.

[0214] Step 3.2: Real-time detection of the runway over markers fed back by the drone; if the drone passes over the runway, reset the markers to zero. .

[0215] Step (4): Detect the satellite navigation positioning status. When the positioning status transitions between differential positioning and single-point positioning, use a softening function to... The values ​​are softened to ensure the smoothness of the height estimate; Step 4.1: When the satellite positioning state changes from differential state to single-point state, The softening calculation formula is as follows: (6) Step 4.2: When the satellite positioning state changes from single-point positioning to differential positioning state, The softening calculation formula is as follows: (7) (8) in It is the maximum allowable single-step change in altitude for a drone.

[0216] like Figure 2 As shown, in order to illustrate the effectiveness of the algorithm of the present invention and fully demonstrate that the height value calculated by this method has both accuracy and noise resistance, the following experiment was conducted: (1) Initial conditions and parameter settings of the experiment In the simulation experiment, different levels of noise were added to all altitude source data, and two scenarios were set up. The first scenario was that after the UAV entered the descent phase, the satellite navigation positioning state remained unchanged as a single point positioning. The second scenario was that during the descent phase of the UAV, the satellite navigation positioning state randomly entered the differential state and then returned to the single point positioning state.

[0217] In actual flight experiments, the altitude of the satellite navigation differential state is used as the true altitude value for comparative analysis, while the altitude fusion algorithm introduces the altitude of the satellite navigation single-point state as the satellite navigation altitude information source.

[0218] (2) Analysis of experimental results like Figure 3 As shown in the figure, the height information after adding noise and the effect of height source fusion are illustrated. It can be seen that different levels of noise have been added to each height source. Figure 4 The diagram illustrates the fusion height effect in a single-point satellite navigation configuration. The fused height approaches the radar altitude as the ground clearance decreases, and the fused height value remains unaffected by fluctuations in radar altitude. Figure 5 The diagram shows the fusion height effect when the satellite positioning status changes. The fused height remains smooth and does not fluctuate with changes in status. Figure 6 The diagram shows the effect of the fusion algorithm during the actual flight of the drone. The data is smooth during the landing phase, and the lower the altitude, the higher the fusion altitude accuracy.

[0219] This invention addresses the problem of inaccurate altitude estimation during the landing phase of a UAV under satellite-guided single-point positioning. It achieves a smooth transition by fusing multi-source sensor data in stages. Unlike existing methods that only fuse air pressure and radio altitude, this method introduces inertial navigation azimuth velocity and adaptive weight adjustment, and uses a softening function to handle state transitions, ensuring the continuity of altitude estimation.

[0220] The foregoing has provided a detailed description of a method, system, and readable storage medium for relative altitude fusion during the landing phase of a fixed-wing unmanned aerial vehicle (UAV) according to embodiments of this application. The descriptions of the embodiments above are merely for the purpose of helping to understand the method and its core ideas; furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

[0221] Certain terms are used in the specification and claims to refer to specific components. Those skilled in the art will understand that hardware manufacturers may use different names to refer to the same component. This specification and claims do not distinguish components based on differences in name, but rather on differences in function. The terms "comprising" and "including" used throughout the specification and claims are open-ended and should be interpreted as "comprising / including but not limited to". "Approximately" means that within an acceptable margin of error, those skilled in the art can solve the technical problem and substantially achieve the technical effect within a certain margin of error. The following descriptions in the specification are preferred embodiments for carrying out this application; however, these descriptions are for the purpose of illustrating the general principles of this application and are not intended to limit the scope of this application. The scope of protection of this application shall be determined by the appended claims.

[0222] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a product or system comprising a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a product or system. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the product or system that includes said element.

[0223] It should be understood that the term "and / or" used in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0224] The foregoing description illustrates and describes several preferred embodiments of this application. However, as previously stated, it should be understood that this application is not limited to the forms disclosed herein and should not be construed as excluding other embodiments. It can be used in various other combinations, modifications, and environments, and can be altered within the scope of the application concept described herein through the foregoing teachings or techniques or knowledge in related fields. Any modifications and variations made by those skilled in the art that do not depart from the spirit and scope of this application should be within the protection scope of the appended claims.

Claims

1. A method for relative altitude fusion during the landing phase of a fixed-wing unmanned aerial vehicle, characterized in that, The relative altitude fusion method for the landing phase of a fixed-wing UAV includes the following steps: S1: Acquire raw altitude data and status information collected by satellite navigation equipment, barometric pressure measuring equipment and ranging equipment, preprocess the raw altitude data to obtain relative altitude data relative to the target landing area; S2: Perform height fusion based on the relative height data to obtain first fused height data, and use the first fused height data to characterize the fusion result of the barometric measurement device and the satellite navigation device; S3: The relative height data is fused with the first fused height data to obtain second fused height data, which represents the height result of the transition to the ranging device; S4: Obtain the celestial velocity data output by the inertial navigation device, and perform complementary filtering and fusion with the celestial velocity data and the second fused altitude data to obtain the third fused altitude data; S5: Detect the status information obtained by the satellite navigation device. If the status information changes between the first positioning state and the second positioning state, then soften the third fused altitude data through a softening function to obtain the final output altitude data. The first positioning state is a differential or high-precision positioning state, and the second positioning state is a single-point or low-precision positioning state. The ranging device is a radar altimeter or a laser rangefinder.

2. The relative altitude fusion method for the landing phase of a fixed-wing UAV according to claim 1, characterized in that, In step S2, the height fusion based on the relative height data to obtain the first fused height data specifically includes: Acquire the first relative altitude data collected by the barometric pressure measuring device and the second relative altitude data collected by the satellite navigation device; Obtain the absolute value of the positioning altitude error of the satellite navigation device in the second positioning state; The first fused height data is determined based on the first relative height data, the second relative height data, and the absolute value of the positioning height error. The first fused height data is the median value among the first relative height data, the second relative height data, and the absolute value of the positioning height error.

3. The method for relative altitude fusion during the landing phase of a fixed-wing UAV according to claim 1, characterized in that, In step S3, the second fused height data is obtained by fusing the relative height data with the first fused height data. This process specifically includes: Acquire the third relative height data collected by the ranging device; The fusion coefficient is determined based on the preset upper and lower limits of the height source transition. The second fusion height data is calculated based on the fusion coefficient, the first fusion height data, and the third relative height data.

4. The method for relative altitude fusion during the landing phase of a fixed-wing UAV according to claim 3, characterized in that, In step S3, determining the fusion coefficient based on the preset upper and lower limits of the height source transition specifically includes: Obtain the current flight altitude data of the aircraft; Determine whether the flight altitude data is greater than the altitude source transition upper limit. If the flight altitude data is greater than the altitude source transition upper limit, then determine that the fusion coefficient is a first preset value. Determine whether the flight altitude data is less than the altitude source transition lower limit. If the flight altitude data is less than the altitude source transition lower limit, then determine that the fusion coefficient is a second preset value. If the flight altitude data is between the upper limit of the altitude source transition and the lower limit of the altitude source transition, then the fusion coefficient is calculated by linear interpolation based on the flight altitude data, the upper limit of the altitude source transition, and the lower limit of the altitude source transition. The fusion coefficients are output to the fusion calculation module for later use.

5. The method for relative altitude fusion during the landing phase of a fixed-wing UAV according to claim 1, characterized in that, In step S4, the celestial velocity data and the second fused height data are used for complementary filtering and fusion to obtain the third fused height data, specifically including: Obtain the preset time constant; Based on the time constant, the celestial velocity data, and the second fused altitude data, a complementary filtering model is constructed; The initial third fusion height data is calculated using the complementary filtering model. Obtain the position information of the aircraft relative to the target landing area, and determine whether the aircraft has passed over the target landing area; If the aircraft passes over the target landing area, the pressure altitude contribution weight or position correction feedback gain in the third fused altitude data is set to zero, resulting in updated third fused altitude data.

6. The method for relative altitude fusion during the landing phase of a fixed-wing UAV according to claim 1, characterized in that, If the state information changes between the first positioning state and the second positioning state in step S5, the third fused height data is softened using a softening function to obtain the final output height data. Specifically, this includes: Determine whether the status information changes from the first positioning state to the second positioning state; If the status information changes from the first positioning state to the second positioning state, then the maximum allowable single-step altitude change value of the aircraft is obtained; A first softening function is constructed based on the maximum value of the single-step height change. The third fused height data is then calculated using the first softening function to obtain the final output height data.

7. The method for relative altitude fusion during the landing phase of a fixed-wing UAV according to claim 6, characterized in that, The step S5, determining whether the status information has changed from the first positioning state to the second positioning state, includes: obtaining the first status identifier of the satellite navigation device at the current moment and the second status identifier at the previous moment; comparing the first status identifier with the second status identifier; if the first status identifier represents the second positioning state and the second status identifier represents the first positioning state, then determining that the status information has changed from the first positioning state to the second positioning state.

8. The method for relative altitude fusion during the landing phase of a fixed-wing UAV according to claim 1, characterized in that, If the state information changes between the first positioning state and the second positioning state in step S5, the third fused height data is softened using a softening function to obtain the final output height data. Specifically, this includes: Determine whether the status information changes from the second positioning state to the first positioning state; If the status information changes from the second positioning state to the first positioning state, then the maximum allowable single-step altitude change value of the aircraft is obtained; A second softening function is constructed based on the maximum value of the single-step height change. The third fused height data is then calculated using the second softening function to obtain the final output height data.

9. A relative altitude fusion system for the landing phase of a fixed-wing unmanned aerial vehicle (UAV), based on the relative altitude fusion method for the landing phase of a fixed-wing UAV according to any one of claims 1-8, characterized in that, The relative altitude fusion system for the landing phase of the fixed-wing UAV includes: The data acquisition and preprocessing module is used to acquire raw altitude data and status information collected by satellite navigation equipment, barometric pressure measurement equipment and ranging equipment, and to preprocess the raw altitude data to obtain relative altitude data relative to the target landing area. The first fusion module is used to perform height fusion based on the relative height data to obtain first fused height data, which represents the fusion result of the barometric measurement device and the satellite navigation device. The second fusion module is used to fuse the relative height data collected by the ranging device with the first fused height data to obtain the second fused height data, which represents the height result of the transition to the ranging device. The complementary filtering and fusion module is used to acquire the celestial velocity data output by the inertial navigation device, and to perform complementary filtering and fusion of the celestial velocity data with the second fused altitude data to obtain the third fused altitude data; The softening processing module is used to detect the status information of the satellite navigation device. When the status information changes between the first positioning state and the second positioning state, the third fused altitude data is softened by the softening function to obtain the final output altitude data.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the relative altitude fusion method for the landing phase of a fixed-wing UAV as described in claims 1-8.

Images