Flight control system and method for a large payload manned eVTOL aircraft

Abstract

The application discloses a flight control system and method of a large-load manned eVTOL aircraft, relates to the technical field of intelligent control of electric vertical take-off and landing aircraft, and comprises the following steps: S100, acquiring aircraft-carrying cargo parameter information, constructing a reflection risk distribution map covering the sensing range of the aircraft, and using the reflection risk distribution map to mark a high-reflection interference risk space region; S200, based on the reflection risk distribution map, performing angle-domain shielding processing on the laser radar and millimeter wave radar beams, shielding echo signals in the corresponding direction, and obtaining an effective echo region. Through the construction of the reflection risk distribution map and the multi-modal fusion sensing compensation mechanism, the identification and avoidance ability of the aircraft to high-reflection interference in the manned cargo transportation task is improved, closed-loop control of sensing correction and dynamic adjustment of flight control parameters is realized, and flight safety and stability are ensured.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology for electric vertical takeoff and landing (eVTOL) aircraft, specifically to a flight control system and method for a heavy-load, manned eVTOL aircraft. Background Technology

[0002] Flight control of heavy-lift, manned eVTOL aircraft refers to a comprehensive set of control strategies, including attitude stabilization, power distribution, load adaptation, and flight path management, required for electric aircraft with vertical takeoff and landing capabilities that can carry both passengers and heavy cargo during complex flight missions. These aircraft are typically used in low-altitude, short-range urban airlift, emergency delivery, or high-frequency logistics scenarios. Their flight control systems must consider both passenger safety and the dynamic impact of cargo weight changes on flight performance. Under different load structures, passenger / cargo combinations, and environmental disturbances, the system must adjust propulsion system output, center of gravity balancing mechanisms, and trajectory correction strategies in real time to ensure high stability, controllability, and safety redundancy throughout all phases of vertical takeoff and landing, hovering, ferry operations, and landing, meeting the flight control requirements of integrated passenger and cargo transport scenarios.

[0003] The existing technology has the following shortcomings:

[0004] For heavy-load, manned eVTOL aircraft carrying both personnel and heavy cargo in actual flight missions, existing flight control systems generally rely on sensing devices such as lidar and millimeter-wave radar to obtain obstacle distance and terrain information. However, under complex load conditions, especially when the aircraft carries metal cargo with high reflectivity, mirror-like structures, or special industrial packaging materials, the cargo itself may cause unexpected specular reflection, scattering, or secondary reflection interference to the detection beam emitted by the sensing system, leading to misjudgment of radar echo signals. Because such abnormal echoes are highly similar to actual obstacle or ground reflection signals in terms of intensity, time delay, and angular characteristics, the flight control system has difficulty distinguishing them during the signal processing stage, resulting in local distortion of the environmental perception map. Especially during low-altitude obstacle avoidance, hovering in confined areas, or precise landings involving personnel boarding and disembarking, such misjudgments can easily trigger erroneous responses from the flight control system, such as premature avoidance, incorrect descent judgment, or abnormal attitude adjustments. This can not only lead to accidents such as descent impact or landing deviation, but also pose a direct threat to the safety of the occupants.

[0005] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The purpose of this invention is to provide a flight control system and method for a heavy-load, manned eVTOL aircraft. By constructing a reflection risk distribution map and a multimodal fusion perception compensation mechanism, the system enhances the aircraft's ability to identify and avoid high-reflection interference in manned and cargo transport missions, realizes closed-loop control of perception correction and dynamic adjustment of flight control parameters, and ensures flight safety and stability, thereby solving the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a flight control method for a heavy-load, manned eVTOL aircraft, comprising the following steps:

[0008] S100: Acquire cargo parameter information carried by the aircraft and construct a reflection risk distribution map covering the aircraft's perception range to mark spatial areas with high reflection interference risk;

[0009] S200, based on the reflection risk distribution map, performs corner domain masking on the lidar and millimeter-wave radar beams to shield the echo signals in the corresponding directions and obtain an effective echo area.

[0010] S300 performs an environmental echo consistency comparison process on the echo signals within the effective echo area, identifies abnormal echo segments caused by cargo reflection characteristics, and marks the spatial location of the abnormal segments.

[0011] S400 performs three-dimensional environmental compensation on the corner domain occlusion area for the spatial location of the marked abnormal echo segment, and generates a three-dimensional perception fusion map covering the occlusion area and the effective echo area.

[0012] The S500, based on the generated 3D perception fusion map, reconstructs environmental information by combining the credibility level of images and navigation information in abnormal spatial areas, updates obstacle avoidance judgment results and flight altitude recognition parameters, and completes dynamic parameter adjustment of flight control commands.

[0013] Based on the dynamic adjustment of flight control parameters, the S600 extracts the spatial distribution characteristics of abnormal areas during flight and the adaptation performance parameters of alternative information sources, dynamically updates the generation rules of the reflection risk distribution map, and realizes continuous identification of aircraft-sensed interference and forward avoidance closed-loop control.

[0014] Preferably, step S100 includes:

[0015] Obtain the surface material type, surface reflectivity coefficient, smoothness level, structural geometry, and loading attitude angle parameters of the cargo carried by the aircraft;

[0016] Based on the acquired parameters, a spatial interaction model between the outer surface of the cargo and the radar beam is established. Simulation calculation or rule mapping is used to evaluate the electromagnetic wave reflection intensity and reflection direction characteristics of different materials at different incident angles.

[0017] The reflection behavior modeling results are spatially projected and superimposed onto the aircraft's perception field of view coordinate system to construct a spatial interaction model that includes echo intensity prediction, reflection direction distribution, and reflection interference level.

[0018] A reflection risk distribution map is generated in a three-dimensional perception space based on a spatial interaction model.

[0019] Preferably, step S200 includes:

[0020] Read the reflection risk distribution map generated based on the cargo surface material type, geometry and loading angle parameters, and map it to the perception field of view coordinate system under the current attitude of the aircraft;

[0021] Based on the current emission characteristic parameters of the sensor, a beam emission simulation spectrum is constructed to determine the scanning direction, energy coverage area, and scanning timing of the overlapping area of ​​the beam in each emission cycle.

[0022] Based on the comparison between the beam simulation map and the reflection risk distribution map, beam masking operations are performed on the high reflection risk angular region.

[0023] Echo signal data in the unshielded direction is extracted to obtain the effective echo area, eliminate high-reflection interference signals, and achieve accurate output of sensing data.

[0024] Preferably, step S300 includes:

[0025] Extract all echo signal data within the effective echo area during the current flight cycle and perform normalized mapping in the aircraft body coordinate system;

[0026] The airspace echo reference map constructed based on historical flight mission data is called to build an expected model as a reference standard for echo consistency comparison;

[0027] Perform an echo consistency comparison process to identify abnormal echo segments whose echo intensity or shape deviates significantly from the characteristics of the reference spectrum;

[0028] The spatial location of the identified abnormal echo segments is marked, and the interference area caused by cargo reflection is identified by combining the sensor viewpoint, flight attitude and load orientation.

[0029] Preferably, step S400 includes:

[0030] The image features are acquired from the camera and combined with the aircraft's current position, attitude angle, and viewing angle parameters to project the image information onto a three-dimensional spatial coordinate system.

[0031] Based on the position, velocity, and acceleration data provided by the inertial navigation system, the motion trajectory of the aircraft is constructed and the attitude and transformation matrix of the image frame in space are calculated.

[0032] The spatial information generated by image features and inertial trajectories is fused with the effective echo region, and spatial matching, feature similarity verification and weighted processing are used to construct a three-dimensional perception fusion map;

[0033] The accuracy of the 3D perception fusion image is optimized to improve the spatial continuity and resolvability of the perception data, ensuring the integrity and reliability of the perception information of the flight control system.

[0034] Preferably, step S500 includes:

[0035] Extract spatial regions marked as abnormal from the 3D perception fusion map and analyze the spatial attributes and source information provided by each data source;

[0036] Based on the data stability, resolution, and synchronization accuracy of the image information and navigation trajectory, each is assigned a confidence level weight, and a weighted reconstruction operation is performed on the abnormal area to complete the local environment geometric completion.

[0037] The reconstructed geometric feature information is input into the obstacle avoidance judgment and flight altitude recognition process, and the relevant parameters are updated based on the confidence threshold.

[0038] Based on the updated obstacle avoidance results and altitude parameters, the control parameters in the flight control commands, such as target heading, hovering altitude, descent speed, and attitude angle correction, are dynamically adjusted.

[0039] Preferably, step S600 includes:

[0040] By retrospectively analyzing the spatial regions marked as anomalous during the flight cycle, their positional distribution, frequency of occurrence, and attitude relationship in the flight path, spatial features of the interference regions are extracted.

[0041] Evaluate the compensation performance indicators of image information and navigation trajectory in each abnormal area to determine the adaptation capability level;

[0042] Based on the spatial characteristics of abnormal regions and the adaptability of information sources, the spatial boundaries and weighting rules of the reflection risk distribution map are dynamically updated.

[0043] The updated reflection risk distribution map is loaded into the perception control process as the input basis for subsequent occlusion control and perception compensation.

[0044] A flight control system for a heavy-load, manned eVTOL aircraft includes a reflection risk modeling module, a dynamic masking module for the perceived field of view, an abnormal echo identification and marking module, a three-dimensional environment perception fusion module, a flight control dynamic update module, and a perception closed-loop optimization module.

[0045] The reflection risk modeling module acquires cargo parameter information carried by the aircraft and constructs a reflection risk distribution map covering the aircraft's perception range to mark spatial areas with high reflection interference risk.

[0046] The dynamic masking module for sensing the field of view performs corner-domain masking on the lidar and millimeter-wave radar beams based on the reflection risk distribution map, blocking the echo signals in the corresponding directions and obtaining an effective echo area.

[0047] The abnormal echo identification and marking module performs an environmental echo consistency comparison process on the echo signals within the valid echo area, identifies abnormal echo segments caused by the reflection characteristics of the cargo, and marks the spatial location of the abnormal segments.

[0048] The 3D environment perception fusion module performs 3D environment compensation on the corner domain occlusion area for the spatial location of the marked abnormal echo segments, and generates a 3D perception fusion map covering the occlusion area and the effective echo area.

[0049] The flight control dynamic update module, based on the generated 3D perception fusion map, reconstructs environmental information by combining the credibility level of images and navigation information in abnormal spatial areas, and updates obstacle avoidance judgment results and flight altitude recognition parameters to complete the dynamic parameter adjustment of flight control commands;

[0050] The perception closed-loop optimization module, based on the dynamic adjustment of flight control parameters, extracts the spatial distribution characteristics of abnormal areas during flight and the adaptation performance parameters of alternative information sources, dynamically updates the generation rules of the reflection risk distribution map, and realizes continuous identification and forward avoidance closed-loop control of the aircraft's perceived interference.

[0051] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0052] This invention constructs a reflection risk distribution map to achieve forward perception layout of high-reflection interference areas before flight. Combined with corner-domain masking, abnormal echo identification, and three-dimensional environmental compensation, it builds an environment reconstruction mechanism based on multimodal information fusion, significantly improving the reliability of perception data and the accuracy of flight control decisions. Simultaneously, during flight, the system can dynamically extract the adaptability of abnormal areas and information sources, update the risk map generation rules in real time, and form a mission-environment-driven perception adaptive mechanism, realizing closed-loop control throughout the entire process from "interference prediction and perception correction" to "avoidance update." This solution not only improves the obstacle avoidance accuracy and landing safety of aircraft in manned and cargo transport missions but also enhances the system's adaptability to changes in different cargo structures and reflection characteristics, providing key technical support for ensuring flight stability and crew safety. Attached Figure Description

[0053] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0054] Figure 1 This is a flowchart of a flight control method for a heavy-load, manned eVTOL aircraft according to the present invention.

[0055] Figure 2 This is a schematic diagram of the flight control system of a heavy-load, manned eVTOL aircraft according to the present invention. Detailed Implementation

[0056] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0057] This invention provides, for example Figure 1 The flight control system method for a heavy-load, manned eVTOL aircraft, as shown, includes the following steps:

[0058] S100: Obtain the surface material type, geometric structure, and loading angle parameters of the cargo carried by the aircraft; construct a reflection risk distribution map covering the aircraft's perception range based on the obtained parameters; and mark the spatial areas with high reflection interference risk during flight.

[0059] To address the potential for unintended signal interference to lidar and millimeter-wave radar when aircraft carry cargo with high reflectivity, a space reflection interference early warning system based on cargo reflectivity characteristics is proposed. This system is used to construct a reflection risk distribution map covering the entire sensing range of the aircraft before flight or during the initial stage of takeoff. The steps include:

[0060] The system collects multiple static and dynamic characteristic parameters of the cargo carried by the aircraft in the current mission, including but not limited to the material type of the cargo surface (such as polished metal, coated glass, smooth plastic, etc.), surface reflectivity coefficient, smoothness level, structural geometry (such as regular cube, cylinder, irregular structure, etc.), and loading attitude angle relative to the aircraft's coordinate system. This information can be obtained through the mission loading database, on-site scanning, or the aircraft mounting station data interface, ensuring that the extracted parameters are accurately collected and standardized before the start of flight.

[0061] Based on the collected parameter set, a spatial interaction model between the cargo's outer surface and the radar beam is established. This model, through simulation calculations or rule-based mapping, evaluates the electromagnetic wave reflection intensity and direction characteristics of various materials at different incident angles. Furthermore, by combining the cargo's position and angle in the aircraft coordinate system, it analyzes the directional range and energy intensity of potential reflection, scattering, or retracing paths within the aircraft's standard sensing range. This modeling process enables the prediction and quantification of the interaction between the cargo and the sensor, providing a physical basis for subsequent spatial area calibration.

[0062] In the process of perception interference prediction for heavy-load electric vertical takeoff and landing (EVTOL) aircraft carrying highly reflective cargo, a spatial interaction model between the cargo's outer surface and radar beams is constructed based on a set of collected parameters, including cargo material type, surface geometry, and loading angle. The core objective is to simulate and predict the reflection behavior of the cargo on lidar and millimeter-wave radar signals within the aircraft's perception range. Specifically, the model construction steps include the following:

[0063] Based on the electromagnetic properties of the cargo surface material (such as conductivity, dielectric constant, surface roughness, etc.) and the sensor operating frequency band (millimeter wave or laser wavelength), reflectivity functions of various materials at different wavelengths are established as the basic input for electromagnetic wave interaction modeling.

[0064] By using the cargo's geometric model (such as a CAD mesh model) and its loading angle in the aircraft coordinate system, the normal direction and incident angle of each surface unit (surface element) under different flight attitudes and radar views are calculated through spatial rotation transformation, thereby determining the interaction relationship between the beams in each direction and the cargo surface.

[0065] Two technical approaches are introduced to model reflection behavior: simulation calculation and rule mapping. Simulation calculation refers to the detailed simulation of the propagation, reflection and scattering path of radar beams on the surface of complex cargo geometry based on electromagnetic propagation simulation software (such as CST, HFSS or the self-developed Ray-tracing algorithm), and output the echo intensity distribution in each direction. Rule mapping, on the other hand, is based on the typical material-angle-wavelength three-dimensional mapping table obtained by existing experimental measurements or theoretical calculations. It quickly matches the reflection intensity and direction probability distribution under the corresponding incident angle according to predefined rules, which is suitable for scenarios with high real-time requirements.

[0066] The simulation or mapping results are spatially projected and superimposed onto the aircraft's sensing field of view coordinate system to form a spatial interactive model that includes echo intensity prediction, reflection direction distribution, and reflection interference level. This model serves as a basis for generating subsequent reflection risk distribution maps, enabling the aircraft to identify potential sensing interference in spatial directions before flight. This facilitates the development of masking strategies or compensation mechanisms in advance, thereby improving the accuracy of flight control and the robustness of environmental perception.

[0067] In summary, simulation calculations emphasize accuracy and physical realism, making them suitable for scenarios with high reliability requirements, while rule mapping emphasizes computational efficiency and rapid adaptation, making it suitable for quickly generating tasks. Both are key modeling methods for achieving reflection interference prediction capabilities, playing a central role in transforming raw load parameters into spatially perceptible interference prediction information in this step.

[0068] Based on the above modeling results, a spatial mapping relationship of reflection risk is constructed in the three-dimensional perception space of the aircraft. Taking the aircraft as the origin, a perception coverage volume under its standard operating conditions is constructed (e.g., a forward 120-degree horizontal field of view, a vertical 45-degree pitch angle, etc.), and the directional distribution area of ​​high reflection energy potentially caused by cargo is marked within this coverage volume. Each spatial directional unit identified as a high-risk reflection source is assigned a corresponding reflection interference weight value. This value is determined by factors such as material reflectivity, radar wavelength adaptability, relative angle, and the possibility of field of view obstruction, thereby completing a three-dimensional reflection risk data map with high spatial resolution.

[0069] The reflection risk data map is overlaid onto the real-time sensing coordinate system of the aircraft to form a high reflection risk distribution map that can be used by subsequent flight sensing strategies. This distribution map not only marks the spatial regions where the aircraft may sense high-intensity non-environmental echo signals, but also serves as input parameters for multiple flight mission processing processes such as beam dynamic masking control, environmental compensation fusion, and flight control command adjustment. It is worth emphasizing that this reflection risk distribution map has a real-time update mechanism. When the loading angle of the cargo carried by the aircraft or its flight attitude changes, the map can be quickly regenerated based on the modeled data, ensuring that interference prediction always has timeliness and dynamic response capabilities during flight.

[0070] The core function of this step is to identify and warn of potential sensing interference areas caused by the reflective properties of cargo surfaces before or during flight missions, thus providing a basis for subsequent sensing signal processing and flight control strategy adjustments. In the application scenarios of heavy-duty electric vertical takeoff and landing (EVL) aircraft, cargo types are diverse, with varying surface materials, shapes, structures, and loading angles. Some cargo (such as metal components and mirrored equipment) have high electromagnetic wave or laser reflectivity, which can easily cause unexpected reflections or scattering of the beams from sensing devices such as lidar and millimeter-wave radar during flight, leading to distorted echo signals, abnormal sensing data, or false obstacle identification. This step acquires the physical parameters of the cargo, constructs a spatial reflection interaction model between the cargo and sensors, and generates a reflection risk distribution map within the aircraft's three-dimensional sensing coordinate system based on this model. This map clearly identifies which directions and angles of the sensing field of view may be affected by interference, enabling the subsequent sensing system to avoid high-risk areas or perform compensation and rejection processing in advance. This forward modeling approach breaks through the limitations of traditional perception systems that "passively respond" to interference, giving aircraft the ability to "actively predict and perceive risks," significantly improving the stability of the perception system, the safety of flight control, and the overall reliability of mission execution. It is a fundamental and key step in achieving intelligent management and control.

[0071] S200, based on the reflection risk distribution map, performs corner domain masking on the beams of lidar and millimeter-wave radar, and shields the echo signals in the corresponding directions for spatial areas marked as high reflection risk, thereby obtaining the effective echo area after eliminating interference signals.

[0072] To address the echo signal interference caused by highly reflective cargo to sensing equipment, an angle-domain beam masking approach based on a reflection risk distribution map is proposed. This approach actively shields signal input from high-reflection interference areas, ensuring that the sensor receives only reliable and valid echo data. The steps include:

[0073] The system reads the reflection risk distribution map generated earlier based on the cargo surface material type, geometry, and loading angle parameters, and maps it to the perception field of view coordinate system under the aircraft's current attitude. This distribution map describes the reflection interference risk level in each direction in polar coordinates (azimuth and pitch angles) within a three-dimensional spatial range, assigning a quantitative risk value to each directional unit. This mapping step ensures that the aircraft can maintain accurate identification of high-reflection directions under different flight attitudes, payload loading methods, or airframe rotation conditions, thereby achieving dynamically adaptive beam shielding.

[0074] Based on the sensor's current transmission characteristic parameters (such as radar scanning frequency, beamwidth, and scanning mechanism), a beam transmission simulation map is constructed to accurately determine the scanning direction, energy coverage area, and scanning timing of overlapping areas within each transmission cycle. By overlaying and comparing this simulation map with a reflection risk distribution map, the set of scanning angles that will cover high-risk spatial directions in future transmission cycles is identified, forming a masking command queue. This queue is used to control the sensor to adjust its transmission power or perform active signal suppression processing within these angle ranges.

[0075] The specific steps for constructing a beam emission simulation map aim to comprehensively simulate the beam motion trajectory, spatial coverage, and energy distribution characteristics of a radar or lidar sensor during a complete scan cycle at the current moment, thereby achieving spatiotemporal modeling of the actual beam's effective area. This process includes the following steps:

[0076] Obtain the currently used sensor model and its parameter configuration, including beam type (continuous wave or pulsed wave), scanning method (mechanical rotation, solid-state scanning, electronic beam control), scanning frequency, single transmission duration, beamwidth (coverage angle of azimuth and elevation), and transmission power intensity model. This information is typically provided in real time by the sensor driver interface or control platform.

[0077] Based on the above parameters, a polar coordinate space model for beam scanning is established. A three-dimensional spherical coordinate system is defined with the location of the aircraft sensor as the origin. Under this coordinate system, the spatial angular domain covered by each beam within a specific time period is calculated according to the beam emission direction variation law. If the sensor is a rotating scanning type, the beam continuously translates in the horizontal or vertical direction with time; if it is an electronic scanning type, the beam may periodically jump. The beam center direction and boundary range are calculated in each tiny time step to form a precise three-dimensional matrix of "scanning time-angle-energy".

[0078] During the modeling process, an energy attenuation model and main lobe / side lobe structure information are superimposed to further calibrate the effective detection energy density of the beam at different spatial angles, reflecting the reception probability and intensity of echo signals from different directions. For lidar, a Gaussian beam diffusion model can be considered; for millimeter-wave radar, an antenna beam pattern template can be introduced for mapping and fitting.

[0079] The aforementioned angle, time, and energy data are discretized and encoded to form a beam simulation map dataset. This map, indexed by a time axis, provides the beam center angle, coverage area, predicted energy distribution, and multi-beam overlap region at each transmission moment. This simulation map can be dynamically updated during actual flight and serves as the basis for determining whether a high-reflection interference area has been entered, guiding subsequent angular domain shielding decisions and the determination of the effective echo area. The construction of this map not only improves the modeling accuracy of sensor physical behavior but also significantly enhances the forward control capability of the sensing system against interference risks.

[0080] The shielding command is executed to set the identified high-reflection-risk angular regions to a shielded state. In practice, for lidar, its scanning mechanism can be controlled to pause laser beam emission within a specified angle range, or the emission power can be reduced to weaken the reflection intensity; for millimeter-wave radar, the antenna channels in the high-risk direction can be shut down through antenna array control, or a reception threshold can be set to ignore echo signals in this direction. These operations do not alter the sensor's normal detection capability in other directions, thus not weakening the overall sensing capability, but rather improving the quality and reliability of the sensing data.

[0081] The echo signal data obtained through masking operations in the unshielded direction is extracted to construct a perception dataset known as the "effective echo region." This dataset excludes false echo signals, scattered echoes, or secondary return signals caused by high-reflection interference factors, thereby making the input data used for subsequent environmental mapping, obstacle identification, path determination, and flight control command generation more accurate and stable. Simultaneously, this effective echo region can also serve as a reliable basis for subsequent environmental fusion compensation and anomaly detection, giving the entire flight perception process good anti-interference capability and safety margin.

[0082] The purpose of this step is to actively control the signal transmission or reception behavior of lidar and millimeter-wave radar at specific angles and directions, thereby preventing the aircraft from acquiring unexpected interference signals generated by highly reflective cargo surfaces during mission execution, and effectively improving the accuracy and stability of environmental perception data. In the operating environment of heavy-duty electric vertical takeoff and landing (EVTOL) aircraft, the presence of highly reflective metal components, mirrored equipment, or cargo with highly reflective packaging often causes radar sensors to receive a large number of false echo signals reflected back from the cargo itself. The energy, echo delay, and waveform characteristics of these interference signals are often similar to the actual terrain boundaries, obstacles, or target objects, causing recognition errors in the flight control system during path planning, obstacle avoidance decisions, and altitude determination, resulting in serious flight control anomalies such as accidental obstacle avoidance, premature landing, and flight attitude instability.

[0083] This step, by calling a pre-generated reflection risk distribution map, proactively identifies high-risk interference areas in spatial directions before the sensor is ready to perform a transmission scan. Angular domain masking commands are then applied in these directions, controlling the sensor to pause beam transmission, reduce transmission power, or ignore received echo signals within that angular range. Through masking operations precise to the angular unit, the sensing system receives only real environmental echoes from low-reflection-risk directions, thus forming an "effective echo region"—a high-quality set of sensing data with reliable signal sources and consistent spatial response.

[0084] This step eliminates interference at the data source, without relying on post-processing algorithms to correct perceived anomalies. This avoids the propagation of algorithmic errors such as misidentification and misclassification, and also reduces the computational burden of subsequent data filtering and anomaly identification. Simultaneously, this corner-domain masking processing is dynamic and real-time, flexibly adjusting the masking direction during flight based on factors such as attitude changes, load tilt, and mission phase, enabling the aircraft to have adaptive perception capabilities in complex load scenarios. This step not only improves the accuracy and usability of perceived data but also directly enhances the safety and robustness of flight control, making it a core component of the entire perception interference control technology chain.

[0085] The S300 performs an environmental echo consistency comparison process on the echo signals within the effective echo area. By comparing them in real time with the airspace echo reference map constructed based on historical flight data, it identifies abnormal echo segments caused by cargo reflection characteristics and marks the spatial location of the abnormal segments.

[0086] To address the radar echo interference caused by highly reflective cargo, an anomaly echo identification and spatial labeling method based on historical environmental data comparison is proposed. This method is used to accurately filter and identify anomalies in the remaining valid echo areas after excluding high-reflection risk angle regions. It can identify and label local pseudo-signals that may still be caused by the reflective characteristics of cargo in unshielded areas, ensuring a high degree of consistency between sensing data in spatial mapping and flight control input. The method includes the following steps:

[0087] All echo signal data within the effective echo area during the current flight cycle are extracted, including point cloud data detected by lidar and millimeter-wave radar within a unit time window. Each echo signal contains information such as spatial coordinates, intensity value, timestamp, and echo morphology, which are normalized and mapped in the aircraft's coordinate system. This dataset serves as the comparison object, and subsequent real-time comparisons with reference maps will be performed using spatial location and feature values ​​to identify potential anomalies.

[0088] An airspace echo reference map, generated based on historical flight mission data, is invoked. This map is constructed using a spatial rasterization method, with each grid cell containing statistical characteristics of echoes collected at that location during multiple historical flights, including maximum / minimum / average echo intensity, morphological contour features, and echo stability indices. The reference map can be categorized and selected according to labels such as flight mission type, payload status, and airspace location, ensuring that the invoked reference data has similar environmental backgrounds and flight configuration conditions. This step constructs an "expected model" in the environmental perception process, providing a comparative benchmark for subsequent difference detection.

[0089] The "expected model" refers to an airspace echo reference map formed through statistical analysis and summarization of echo data accumulated from historical flight missions under similar airspace conditions. It represents the typical characteristics of the echo signal that a sensor should receive at a specific spatial location under ideal, minimally interfered, or stable environmental conditions. The purpose of constructing the expected model is to effectively compare and analyze the current real-time sensing data during flight, thereby identifying echo segments that deviate from normal performance and exhibit abnormal characteristics. In this step, even after the aircraft carries highly reflective cargo and eliminates most direct reflection interference signals through corner-domain shielding, there may still be spurious signals caused by cargo surface reflection, secondary scattering, or attitude fluctuations. These interferences have a certain degree of uncertainty and non-persistence, making them difficult to completely eliminate through simple physical shielding. At this point, by comparing the current echo signal with the expected model in real time, it can be identified whether the collected echo signal exceeds the "normal fluctuation range" of the environment, thus allowing for precise spatial marking and dynamic removal. The core role of this model here is to serve as a reference standard for difference detection, providing a data-driven "background expectation" judgment mechanism. This allows the perception system to no longer rely on static thresholds or manual rules, but instead to construct dynamic intelligent judgment standards through historical experience data, thereby improving the accuracy, adaptability, and engineering practicality of anomaly identification.

[0090] An echo consistency comparison process is executed. For each echo point acquired during the current flight, its spatial location is matched against the corresponding grid cell in the reference map, and the echo intensity, morphology, and historical statistical characteristics are evaluated for similarity. If the intensity of the current echo point exceeds the reference range threshold, or its morphological contour differs significantly from the reference map, and there is no stable continuous echo support in the surrounding space, then the echo is determined to be a suspected anomalous segment. To eliminate interference from factors such as normal terrain abrupt changes and flight attitude changes, anomaly detection also incorporates a multi-frame data consistency comparison mechanism within a time window to enhance the reliability and robustness of identification.

[0091] For identified anomalous echo segments, spatial location marking is performed. This marking operation not only records the spatial three-dimensional coordinates, timestamp, and anomaly type of the anomaly point, but also infers whether the anomaly might be related to the cargo's reflective characteristics through cross-analysis with sensor viewpoint, flight attitude, and payload orientation. If multiple anomaly points are concentrated in a specific angular direction, or repeatedly appear in key areas such as directly in front of or below the payload, this spatial range is further defined as a "reflection-induced suspected interference zone" and highlighted in the perception environment map, serving as a key input area for subsequent environmental fusion compensation and flight control parameter correction.

[0092] The purpose of this step is to further identify deep anomalies in the remaining effective echo data after the high-reflectivity angular domain masking is completed. This involves identifying and marking localized, irregular pseudo-signals caused by the reflectivity of the cargo surface, thereby ensuring the high accuracy and environmental consistency of the sensory inputs used by the aircraft in environmental perception and flight control decisions. In complex flight missions, even if the main interference directions are avoided through prior reflection risk prediction and angular domain masking, some secondary pseudo-signals caused by indirect reflection or edge scattering are still unavoidable. These signals are usually hidden within the effective echo area, exhibiting intensity and contour features similar to normal obstacles or terrain features. They are easily misidentified by the flight control system as real-world information, leading to flight path deviations, landing misjudgments, or abnormal attitude adjustments, among other flight risks.

[0093] By executing an environmental echo consistency comparison process, the aircraft can compare the currently acquired echo signals with an airspace echo reference map constructed based on historical flight data. This reference map is equivalent to an "expectation model," recording the echo characteristics of a specific airspace under interference-free conditions, including spatial location, signal strength, morphological profile, and temporal stability. When the characteristics of the current echo signal at a specific location differ significantly from the statistical characteristics in the reference map, and there is no reasonable explanation for changes in the natural environment, it can be identified as an abnormal echo segment caused by cargo reflection characteristics. Further spatial location marking of these segments not only allows them to be eliminated or suppressed in the current perception process, but also feeds them back as abnormal area information to subsequent environmental mapping, data fusion, and flight control command generation processes, enabling the flight control system to dynamically avoid and finely control interference areas.

[0094] The implementation of this step not only enhances the perception system's ability to identify anomalies under complex load environments, but also constructs a dynamic verification mechanism that integrates historical data with real-time perception, improving the robustness and safety of flight missions. By introducing a comparison mechanism based on the "expected model," a shift from "rule filtering" to "data-driven judgment" is achieved, enabling the aircraft to possess stronger environmental adaptability and self-correction capabilities.

[0095] S400, for anomalous echo segments marked in space, fuses image feature information acquired by the camera with trajectory information provided by the inertial navigation system, performs three-dimensional environmental compensation for the corner-domain occluded area, and generates a three-dimensional perception fusion map covering the occluded area and the effective echo area.

[0096] To address the issue of missing environmental information caused by partial obstruction of the perception field of view due to highly reflective cargo, a three-dimensional environmental compensation method based on multi-source perception fusion is proposed. This method integrates feature information from camera images with trajectory information provided by the inertial navigation system to spatially complete the area containing marked anomalous echo segments and the surrounding angular obstructed areas, generating a three-dimensional perception fusion map covering the entire field of view. This restores complete and reliable environmental information, improving the flight control system's understanding and decision-making accuracy in low-altitude flight scenarios. The method includes the following steps:

[0097] The system retrieves image sequences captured by cameras during flight and projects the image information into a three-dimensional coordinate system based on the aircraft's current position, attitude angle, and viewing angle parameters. Key feature points extracted from the images (such as edges, corners, and textures) are aligned with echo data in three-dimensional space using visual SLAM (Simultaneous Localization and Mapping) algorithms or feature point matching algorithms to establish a preliminary spatial correlation between image features and radar data. The key to this step is achieving partial image coverage in obscured areas, using visual image information to compensate for blind spots in the radar's field of view.

[0098] By utilizing the continuous position, velocity, and acceleration information obtained by the inertial navigation system during flight, the motion trajectory curves of the aircraft across consecutive moments are constructed, and the absolute attitude and projection transformation matrix of each image frame in space are calculated accordingly. Combining the inertial trajectory and image information, the feature points of the dispersed and asynchronously acquired images can be uniformly mapped to the same three-dimensional reference coordinate system. Spatial interpolation and density modeling are then performed based on the continuity of inertial motion to fill in areas not covered by the images. Through this step, the aircraft can construct preliminary environmental point cloud compensation data in space using the visual and navigation information accumulated during continuous flight.

[0099] The spatial information generated from the aforementioned image features and inertial trajectories is fused with the effective echo region from the original radar echo data. The fusion process employs spatial matching and weighted averaging: overlapping regions undergo feature similarity verification and weighted averaging; obscured regions retain primarily visual-inertial data, and interpolation algorithms ensure continuity and smoothness in boundary regions. The result is a unified 3D environmental map covering both the obscured and effective radar data regions, possessing spatial integrity, source diversity, and redundancy tolerance.

[0100] This method fuses spatial information generated from image features and inertial trajectories with the effective echo regions in the original radar echo data. The specific fusion process utilizes a multi-source data integration method based on spatial alignment, feature complementarity, and weight allocation, aiming to construct a unified, continuous, and reliable 3D environment model. First, the image feature point cloud data, after attitude calculation by the inertial navigation system, is projected onto the aircraft's inertial coordinate system, ensuring fusion with the radar echo data under the same spatial reference. Then, spatial matching and feature verification are performed on the overlapping areas of the two data sources. Similarity scores are calculated using indicators such as feature point position, density, and intensity, and weighted fusion processing is performed based on matching quality: higher weights are assigned to clear and stable image features, and image data is used to complete sparse or missing radar points; when radar data is stable and continuous, it is the primary source, with image information serving as supplementary information. For areas obscured by radar but covered by the image, spatial interpolation is directly used to complete the image point cloud data. The entire fusion process also includes boundary smoothing, time synchronization compensation, point cloud resampling, and noise removal, ultimately forming a spatially continuous 3D perception fusion map with reliable data sources, providing a complete and reliable environmental perception foundation for the flight control system. This fusion process balances sensor complementarity with the integrity of environmental perception, and has significant advantages such as strong robustness, high accuracy, and adaptability to complex scenarios.

[0101] The 3D perception fusion map undergoes precision optimization processing, including noise filtering, boundary smoothing, and altitude consistency adjustment, further improving the spatial continuity and resolvability of the perception data. The optimized 3D perception fusion map can be directly used to update the obstacle avoidance judgment, altitude assessment, and path adjustment inputs of the flight control system, replacing the information gaps caused by the original radar data in obscured areas. This ensures that the flight control system can still issue accurate control commands based on a complete and reliable environmental map even under limited field of view conditions.

[0102] The purpose of this step is to complete the spatial information of the perception blind spot by fusing image information from other sensors (such as cameras) and motion trajectory data from the inertial navigation system after the radar sensor has implemented corner-domain masking due to high reflection risk. This results in a complete and continuous three-dimensional environmental perception map. Because lidar and millimeter-wave radar actively ignore beam echoes from high-reflection-risk directions during masking operations, while effectively eliminating false signal interference, this also leads to a lack of perception data in some spatial areas, forming "information holes" or "blind spots." Without compensation, this incomplete environmental data can cause control errors in the flight control system during obstacle avoidance judgment, altitude recognition, and path planning, and may even lead to flight accidents.

[0103] To address this issue, this step introduces image visual information and inertial navigation trajectory as supplementary data sources. Cameras capture rich 2D image features such as texture, edges, and shapes, while the inertial navigation system provides precise position and attitude change information of the aircraft in 3D space. By spatially aligning image features with the navigation trajectory, key feature points in the image can be restored to a 3D coordinate system, enabling geometric reconstruction of the obscured area. Furthermore, this data is fused with data from the unobscured radar effective echo area, and weighted interpolation and smoothing are applied to boundary regions to ensure the consistency and continuity of the overall environmental model. The resulting 3D perception fusion map not only fills the spatial information gaps in radar blind spots but also improves the accuracy and robustness of the perception results in structural analysis and semantic recognition.

[0104] This step enables the aircraft to autonomously reconstruct its local environment under limited perception conditions, greatly enhancing the integrity and environmental adaptability of the perception system and providing highly reliable data support for subsequent flight control command output. Especially in complex environments such as low-altitude urban areas, areas surrounding buildings, or dynamic terrain, this method effectively reduces flight control errors caused by insufficient radar data, ensuring the safety and accuracy of the flight path, and is a key component in building a robust intelligent flight control architecture.

[0105] The S500, based on a 3D perception fusion map, reconstructs environmental information by combining the credibility level of images and navigation information for spatial regions marked as abnormal, and updates obstacle avoidance judgment results and flight altitude recognition parameters accordingly, thereby completing the dynamic parameter adjustment of flight control commands to ensure the accuracy and stability of the flight control system output commands.

[0106] To address the issue of misinterpretation of flight control commands caused by localized perception interference resulting from highly reflective cargo, a method based on 3D perception fusion mapping for environmental reconstruction and dynamic updating of flight control parameters is proposed. By filtering reliable information, completing the environment, and adjusting control parameters in anomalous spatial regions, the adaptive control capability and safe operation level of the aircraft in complex environments are effectively improved. The method includes the following steps:

[0107] Anomaly-marked spatial regions are extracted from the constructed 3D perception fusion map, and the data types and features provided by different information sources within these regions are analyzed. This 3D perception fusion map contains multi-source spatial information from radar effective echo areas, image feature projection data, and inertial navigation trajectory interpolation. For anomaly regions, each spatial unit possesses basic attributes such as data source labels, spatial coordinates, and timestamps. The goal of this step is to identify reliable data sources that can be used to complete the missing areas of the original perception, providing a basis for subsequent reconstruction operations.

[0108] Based on the data stability, acquisition resolution, and time synchronization accuracy of different information sources, a confidence level weight is assigned to image information and navigation trajectory in each anomalous region. For example, for regions with clear structural textures and high image matching, higher confidence weights can be assigned to image data; while in regions with blurred images or severe occlusion, inertial navigation trajectory interpolation information is used preferentially. Combining these weights, a weighted reconstruction operation is performed on each anomalous region. Through methods such as spatial interpolation, surface fitting, or point cloud expansion, local environmental geometry is completed in a three-dimensional coordinate system, while retaining confidence indicators for subsequent judgment. This step ensures that a relatively reliable environmental model can still be obtained even under conditions of missing perception data, building a stable spatial foundation for flight control input.

[0109] After reconstructing the environmental information, the geometric feature information of the area is input into the obstacle avoidance judgment process and the flight altitude recognition process. By updating parameters such as obstacle boundary contours, terrain surface height, and spatial continuity, the original obstacle avoidance path planning and descent trigger altitude are adjusted to avoid misjudgments caused by false obstacles or void data. For example, in an autonomous landing scenario, if an area is identified as ground in the original echo but is subsequently reconstructed as a void, premature landing can be avoided; during low-altitude hovering, if lateral obstacles are re-identified, the flight control attitude adjustment strategy can be corrected. Before execution, the data confidence level is checked. Parameter updates are only triggered when the confidence level of the environmental reconstruction result is higher than a preset threshold, ensuring that the stability of flight control commands is not disturbed by low-quality data.

[0110] Based on the reconstructed obstacle avoidance results and altitude parameters, key control parameters in the flight control command output are dynamically adjusted, including target heading, hovering altitude, descent speed, and attitude angle correction. This adjustment process is differentiated based on flight mission requirements and current flight status. For example, in precision landing missions, altitude recognition parameters are adjusted first, while obstacle avoidance path judgment is optimized first in complex path flights. The entire process achieves a closed-loop operation from perception data reconstruction to flight control command adjustment, ensuring that the aircraft maintains stable, safe, and intelligent flight control capabilities even in complex reflective environments.

[0111] The purpose of this step is to reconstruct environmental information in flight environments where highly reflective cargo causes partial perception anomalies, based on the spatial information provided by the 3D perception fusion map. This involves reconstructing environmental information in spatial areas marked as having perception uncertainties or false signals, and dynamically correcting the environmental input parameters of the flight control system by analyzing the reliability levels of image and navigation information. This enables accurate adjustment and stable output of flight control commands. Specifically, when the aircraft partially obscures its radar field of view due to high reflective loads, and abnormal segments still exist within the effective echo area, relying solely on radar information is insufficient to fully acquire environmental geometry and terrain structure. In this case, by fusing image features captured by cameras (such as edge contours and texture changes) with spatial trajectories recorded by the inertial navigation system (such as position changes and attitude evolution), a reconstruction model of the missing areas is constructed in the 3D perception fusion map, effectively compensating for spatial perception gaps caused by blind spots.

[0112] After reconstruction, the aircraft can input this updated environmental information into the obstacle avoidance judgment and flight altitude recognition process, enabling dynamic correction of parameters such as obstacle boundaries, ground altitude, and hovering tolerance range. For example, when the reconstruction results show that an area originally judged to be passable actually has obstacle edges, the obstacle avoidance path can be adjusted in real time; when the descent area is re-identified as uneven terrain or an unexpected landing surface, landing commands can be postponed or optimized. Ultimately, these updated environmental parameters will directly affect the calculation process of flight control commands, enabling the replanning of key flight behaviors such as aircraft attitude control, flight path adjustment, and target altitude lock, making the output control commands more closely resemble real environmental conditions.

[0113] This step not only solves the perception blind spot problem caused by sensor masking, but also, by establishing an environment reconstruction and control adjustment mechanism driven by the information source's credibility level, endows the aircraft with intelligent characteristics of "self-perception correction" and "autonomous parameter adjustment," significantly improving the flexibility, stability, and anti-interference capability of flight control decisions. Especially in scenarios with extremely high safety requirements, such as urban low-altitude complex terrain, emergency rescue flights, or large payload delivery, this capability is crucial for ensuring the smooth execution of flight missions and the stable operation of the aircraft platform.

[0114] After the environmental information is reconstructed and the flight control parameters are updated, the S600 extracts the spatial distribution characteristics of the current flight anomaly and the adaptation performance parameters of the alternative information source, updates the generation rules of the reflection risk distribution map, and realizes continuous identification of the aircraft's perceived interference and forward avoidance closed-loop control.

[0115] To further enhance the aircraft's adaptive capability and forward avoidance level against high-reflection interference, a reflection risk map optimization and closed-loop control based on flight experience feedback is proposed. By extracting the spatial distribution characteristics of abnormal echo regions and the adaptation performance of alternative information sources during the compensation process, the generation rules of the reflection risk distribution map are dynamically updated, thereby achieving continuous identification and closed-loop optimization of perceived interference. The process includes the following steps:

[0116] After completing the environmental information reconstruction and flight control parameter update, all spatial regions marked as anomalous during the current flight cycle are retrospectively analyzed, and their spatial distribution characteristics during flight are examined. Specifically, this includes the relative position of the anomalous regions within the flight path (e.g., in front, to the side, or below), their frequency of occurrence, duration, and their correspondence with aircraft attitude and load angles. This information is categorized and statistically analyzed in a three-dimensional coordinate grid format and correlated with the reflection physics model to construct a set of spatial characteristics of the actual interference areas under the current flight mission.

[0117] This study analyzes the adaptation performance metrics of alternative information sources (such as image information or navigation trajectories) during environmental reconstruction in each anomalous region. These performance metrics include, but are not limited to, reconstruction accuracy, boundary integrity, time synchronization deviation, and data confidence. By comparing these metrics with ideal perception accuracy, the actual compensation capability and response stability of different information sources in different types of interference regions are evaluated, and the adaptation quality level is determined accordingly. This evaluation result can serve as a basis for subsequent strategy selection and weight adjustment, providing a reference for optimizing occlusion control strategies.

[0118] By combining the spatial distribution characteristics of the aforementioned abnormal areas with the information source adaptation performance parameters, the generation rules for the original reflection risk distribution map are dynamically adjusted. Specifically, based on the angle, direction, and spatial extent of the actual abnormal areas, the threshold and spatial boundaries of high-reflection-risk areas in the original map are corrected. Simultaneously, based on differences in information source adaptation performance, the forward occlusion weight for areas with poor adaptation capabilities is enhanced, or these areas are automatically marked as high-priority monitoring directions in future missions. This update to the generation rules not only reflects the aircraft's ability to learn from perceived interference during actual flight but also continuously improves the spatial accuracy and coverage of the predicted map through a data feedback mechanism.

[0119] The updated reflection risk distribution map is reloaded into the aircraft's perception and control process, serving as the input for beam blocking, perception compensation, and obstacle avoidance decisions in the next stage, forming a closed-loop control chain that evolves based on actual flight experience. This closed loop not only enhances the aircraft's adaptability to long-term operating environments but also continuously optimizes the perception safety boundary through multiple flight iterations, ultimately achieving a fully self-closed-loop operation of "perception interference—environmental reconstruction—control adjustment—rule update," giving the aircraft high versatility and anti-interference capabilities across mission scenarios and load states.

[0120] The purpose of this step is to establish a self-iterative capability for the perception interference identification and response mechanism based on the actual operational results of flight missions. By performing attribution analysis on abnormal spatial regions appearing during flight and combining the adaptation performance of various alternative information sources in these regions, the rules for generating reflection risk distribution maps for future missions are dynamically optimized. This enables continuous identification and forward avoidance closed-loop control of perceived interference for the aircraft. In practical applications, heavy-load electric vertical takeoff and landing (EVL) aircraft carry diverse cargo types with complex structures and varying surface materials. Some cargo, at specific angles or attitudes, can generate unpredictable high-reflection interference signals to lidar or millimeter-wave radar. Although modeling, corner-domain masking, and perception fusion can address these interferences to some extent in the early stages, the dynamic nature of the aircraft's environment means that interference identification and avoidance strategies for a single flight mission cannot be fully applied to all scenarios.

[0121] Therefore, this step involves retrospectively analyzing the perceived anomaly areas that actually occurred during the current flight, extracting their distribution characteristics in three-dimensional space, such as position density, directional distribution, and relative relationship with the flight path. Simultaneously, it evaluates the performance of alternative information sources used in these areas, such as image data and inertial navigation data, during the environmental reconstruction process, assessing metrics such as reconstruction accuracy, data coverage, and boundary continuity. This forms a knowledge base regarding "which directions are prone to interference" and "which information sources are more adaptable to this interference." Based on this foundation, the aircraft can adjust its original reflection risk map generation rules, optimize the occlusion threshold, expand the prediction range, and improve the response level in specific directions, enabling it to have stronger forward avoidance capabilities before entering similar scenarios in future flight missions.

[0122] Ultimately, this mechanism achieves closed-loop linkage between perception and flight control, enabling the aircraft to move beyond unidirectional responses to interference and instead possess an intelligent cyclical capability of "identification-feedback-optimization-re-avoidance," significantly improving the flight system's continuous adaptability and autonomous flight safety in complex environments. This is not only significant for enhancing the level of intelligent control of individual aircraft but also lays a key technological foundation for future advanced flight modes such as multi-aircraft collaboration, self-learning flight, and long-term operational missions.

[0123] This invention constructs a reflection risk distribution map to achieve forward perception layout of high-reflection interference areas before flight. Combined with corner-domain masking, abnormal echo identification, and three-dimensional environmental compensation, it builds an environment reconstruction mechanism based on multimodal information fusion, significantly improving the reliability of perception data and the accuracy of flight control decisions. Simultaneously, during flight, the system can dynamically extract the adaptability of abnormal areas and information sources, update the risk map generation rules in real time, and form a mission-environment-driven perception adaptive mechanism, realizing closed-loop control throughout the entire process from "interference prediction and perception correction" to "avoidance update." This solution not only improves the obstacle avoidance accuracy and landing safety of aircraft in manned and cargo transport missions but also enhances the system's adaptability to changes in different cargo structures and reflection characteristics, providing key technical support for ensuring flight stability and crew safety.

[0124] This invention provides, for example Figure 2 The flight control system of a heavy-load manned eVTOL aircraft shown includes a reflection risk modeling module, a dynamic masking module for the perceived field of view, an abnormal echo identification and marking module, a three-dimensional environment perception fusion module, a flight control dynamic update module, and a perception closed-loop optimization module.

[0125] The reflection risk modeling module acquires cargo parameter information carried by the aircraft and constructs a reflection risk distribution map covering the aircraft's perception range to mark spatial areas with high reflection interference risk.

[0126] The dynamic masking module for sensing the field of view performs corner-domain masking on the lidar and millimeter-wave radar beams based on the reflection risk distribution map, blocking the echo signals in the corresponding directions and obtaining an effective echo area.

[0127] The abnormal echo identification and marking module performs an environmental echo consistency comparison process on the echo signals within the valid echo area, identifies abnormal echo segments caused by the reflection characteristics of the cargo, and marks the spatial location of the abnormal segments.

[0128] The 3D environment perception fusion module performs 3D environment compensation on the corner domain occlusion area for the spatial location of the marked abnormal echo segments, and generates a 3D perception fusion map covering the occlusion area and the effective echo area.

[0129] The flight control dynamic update module, based on the generated 3D perception fusion map, reconstructs environmental information by combining the credibility level of images and navigation information in abnormal spatial areas, and updates obstacle avoidance judgment results and flight altitude recognition parameters to complete the dynamic parameter adjustment of flight control commands;

[0130] The perception closed-loop optimization module, based on the dynamic adjustment of flight control parameters, extracts the spatial distribution characteristics of abnormal areas during flight and the adaptation performance parameters of alternative information sources, dynamically updates the generation rules of the reflection risk distribution map, and realizes continuous identification and forward avoidance closed-loop control of the aircraft's perceived interference.

[0131] The present invention provides a flight control method for a large-payload, manned eVTOL aircraft, which is implemented through the flight control system of the aforementioned large-payload, manned eVTOL aircraft. For details of the specific method and process of the flight control system of the large-payload, manned eVTOL aircraft, please refer to the embodiment of the flight control method for the aforementioned large-payload, manned eVTOL aircraft, which will not be repeated here.

[0132] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0133] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

[0134] It should be noted that, in this document, the use of relational terms such as "first" and "second" is merely for distinguishing one entity or operation from another, and does not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0135] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0136] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0137] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0138] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0139] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0140] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0141] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A flight control method for a heavy-load, manned eVTOL aircraft, characterized in that, Includes the following steps: S100: Acquire cargo parameter information carried by the aircraft and construct a reflection risk distribution map covering the aircraft's perception range to mark spatial areas with high reflection interference risk; Step S100 includes: Obtain the surface material type, surface reflectivity coefficient, smoothness level, structural geometry, and loading attitude angle parameters of the cargo carried by the aircraft; Based on the acquired parameters, a spatial interaction model between the outer surface of the cargo and the radar beam is established. Simulation calculation or rule mapping is used to evaluate the electromagnetic wave reflection intensity and reflection direction characteristics of different materials at different incident angles. The reflection behavior modeling results are spatially projected and superimposed onto the aircraft's perception field of view coordinate system to construct a spatial interaction model that includes echo intensity prediction, reflection direction distribution, and reflection interference level. Generate a reflection risk distribution map in a three-dimensional perception space based on a spatial interaction model; S200, based on the reflection risk distribution map, performs corner domain masking on the lidar and millimeter-wave radar beams to shield the echo signals in the corresponding directions and obtain an effective echo area. Specifically, this includes: constructing a beam emission simulation map based on the current sensor emission characteristic parameters; comparing the simulation map with the reflection risk distribution map to identify the set of scanning angles that will cover high-risk spatial directions in future emission cycles, forming a masking command queue; extracting echo signal data in the unshielded direction obtained through masking operations; and constructing a perception dataset called the effective echo region. S300 performs an environmental echo consistency comparison process on the echo signals within the effective echo area, identifies abnormal echo segments caused by cargo reflection characteristics, and marks the spatial location of the abnormal segments. Specifically, this includes: identifying abnormal echo segments caused by cargo reflection characteristics by comparing them in real time with an airspace echo reference map constructed based on historical flight data; S400 performs three-dimensional environmental compensation on the corner domain occlusion area for the spatial location of the marked abnormal echo segment, and generates a three-dimensional perception fusion map covering the occlusion area and the effective echo area. Specifically, this includes: for anomalous echo segments marked in space, fusing image feature information acquired by the camera with trajectory information provided by the inertial navigation system, performing three-dimensional environmental compensation on the angular occlusion area, and generating a three-dimensional perception fusion map covering the occlusion area and the effective echo area; The S500, based on the generated 3D perception fusion map, reconstructs environmental information by combining the credibility level of images and navigation information in abnormal spatial areas, updates obstacle avoidance judgment results and flight altitude recognition parameters, and completes dynamic parameter adjustment of flight control commands. Based on the dynamic adjustment of flight control parameters, the S600 extracts the spatial distribution characteristics of abnormal areas during flight and the adaptation performance parameters of alternative information sources, dynamically updates the generation rules of the reflection risk distribution map, and realizes continuous identification and forward avoidance closed-loop control of aircraft-sensed interference.

2. The flight control method for a heavy-load, manned eVTOL aircraft according to claim 1, characterized in that, Step S200 includes: Read the reflection risk distribution map generated based on the cargo surface material type, geometry and loading angle parameters, and map it to the perception field of view coordinate system under the current attitude of the aircraft; Based on the current emission characteristic parameters of the sensor, a beam emission simulation spectrum is constructed to determine the scanning direction, energy coverage area, and scanning timing of the overlapping area of ​​the beam in each emission cycle. Based on the comparison between the beam simulation map and the reflection risk distribution map, beam masking operations are performed on the high reflection risk angular region. Echo signal data in the unshielded direction is extracted to obtain the effective echo area, eliminate high-reflection interference signals, and achieve accurate output of sensing data.

3. The flight control method for a heavy-load, manned eVTOL aircraft according to claim 1, characterized in that, Step S300 includes: Extract all echo signal data within the effective echo area during the current flight cycle and perform normalized mapping in the aircraft body coordinate system; The airspace echo reference map constructed based on historical flight mission data is called to build an expected model as a reference standard for echo consistency comparison; Perform an echo consistency comparison process to identify abnormal echo segments whose echo intensity or shape deviates significantly from the characteristics of the reference spectrum; The spatial location of the identified abnormal echo segments is marked, and the interference area caused by cargo reflection is identified by combining the sensor viewpoint, flight attitude and load orientation.

4. The flight control method for a heavy-load, manned eVTOL aircraft according to claim 1, characterized in that, Step S400 includes: The image features are acquired from the camera and combined with the aircraft's current position, attitude angle, and viewing angle parameters to project the image information onto a three-dimensional spatial coordinate system. Based on the position, velocity, and acceleration data provided by the inertial navigation system, the motion trajectory of the aircraft is constructed and the attitude and transformation matrix of the image frame in space are calculated. The spatial information generated by image features and inertial trajectories is fused with the effective echo region, and spatial matching, feature similarity verification and weighted processing are used to construct a three-dimensional perception fusion map; The accuracy of the 3D perception fusion image is optimized to improve the spatial continuity and resolvability of the perception data, ensuring the integrity and reliability of the perception information of the flight control system.

5. The flight control method for a heavy-load, manned eVTOL aircraft according to claim 1, characterized in that, Step S500 includes: Extract spatial regions marked as abnormal from the 3D perception fusion map and analyze the spatial attributes and source information provided by each data source; Based on the data stability, resolution, and synchronization accuracy of the image information and navigation trajectory, each is assigned a confidence level weight, and a weighted reconstruction operation is performed on the abnormal area to complete the local environment geometric completion. The reconstructed geometric feature information is input into the obstacle avoidance judgment and flight altitude recognition process, and the relevant parameters are updated based on the confidence threshold. Based on the updated obstacle avoidance results and altitude parameters, the control parameters in the flight control commands, such as target heading, hovering altitude, descent speed, and attitude angle correction, are dynamically adjusted.

6. The flight control method for a heavy-load, manned eVTOL aircraft according to claim 1, characterized in that, Step S600 includes: By retrospectively analyzing the spatial regions marked as anomalous during the flight cycle, their positional distribution, frequency of occurrence, and attitude relationship in the flight path, spatial features of the interference regions are extracted. Evaluate the compensation performance indicators of image information and navigation trajectory in each abnormal area to determine the adaptation capability level; Based on the spatial characteristics of abnormal regions and the adaptability of information sources, the spatial boundaries and weighting rules of the reflection risk distribution map are dynamically updated. The updated reflection risk distribution map is loaded into the perception control process as the input basis for subsequent occlusion control and perception compensation.

7. A flight control system for a heavy-load, manned eVTOL aircraft, used to implement the flight control method for a heavy-load, manned eVTOL aircraft as described in any one of claims 1-6, characterized in that, It includes a reflection risk modeling module, a dynamic occlusion module for the perceived field of view, an abnormal echo identification and marking module, a 3D environment perception fusion module, a flight control dynamic update module, and a perception closed-loop optimization module. The reflection risk modeling module acquires cargo parameter information carried by the aircraft and constructs a reflection risk distribution map covering the aircraft's perception range to mark spatial areas with high reflection interference risk. The dynamic masking module for sensing the field of view performs corner-domain masking on the lidar and millimeter-wave radar beams based on the reflection risk distribution map, blocking the echo signals in the corresponding directions and obtaining an effective echo area. The abnormal echo identification and marking module performs an environmental echo consistency comparison process on the echo signals within the valid echo area, identifies abnormal echo segments caused by the reflection characteristics of the cargo, and marks the spatial location of the abnormal segments. The 3D environment perception fusion module performs 3D environment compensation on the corner domain occlusion area for the spatial location of the marked abnormal echo segments, and generates a 3D perception fusion map covering the occlusion area and the effective echo area. The flight control dynamic update module, based on the generated 3D perception fusion map, reconstructs environmental information by combining the credibility level of images and navigation information in abnormal spatial areas, and updates obstacle avoidance judgment results and flight altitude recognition parameters to complete the dynamic parameter adjustment of flight control commands; The perception closed-loop optimization module, based on the dynamic adjustment of flight control parameters, extracts the spatial distribution characteristics of abnormal areas during flight and the adaptation performance parameters of alternative information sources, dynamically updates the generation rules of the reflection risk distribution map, and realizes continuous identification and forward avoidance closed-loop control of the aircraft's perceived interference.

Images