Automatic obstacle avoidance method and system for low-altitude aircraft

Abstract

The invention relates to the technical field of automatic obstacle avoidance, and provides an automatic obstacle avoidance method and system for a low-altitude aircraft. The method comprises the following steps: acquiring flight state information in real time; acquiring obstacle information through an environment sensing module; based on the flight state information and the obstacle information, calculating a relative position and a relative speed with the obstacle, and determining a potential collision risk area; if the potential collision risk exists, based on a preset obstacle avoidance strategy library, an obstacle avoidance strategy is retrieved in combination with a risk area, and an obstacle avoidance control instruction is generated; and transmitting the control instruction to a flight controller, and controlling the aircraft to execute automatic obstacle avoidance. The technical problems that an existing low-altitude aircraft is lagged in obstacle avoidance response in a complex environment and difficult to accurately avoid obstacles in real time are solved, and the technical effects that collision risk dynamic pre-judgment is achieved through multi-source information fusion and intelligent strategy matching, and the autonomous obstacle avoidance success rate and flight safety of the low-altitude aircraft are improved are achieved.

Description

Technical Field

[0001] This application relates to the field of automatic obstacle avoidance technology, specifically to automatic obstacle avoidance methods and systems for low-altitude aircraft. Background Technology

[0002] Currently, low-altitude aircraft (such as multi-rotor drones and vertical takeoff and landing fixed-wing aircraft) are widely used in logistics transportation, power line inspection, agricultural monitoring, and emergency rescue. However, in complex terrain, urban low-altitude areas, and high-speed flight environments, aircraft are prone to potential collision threats from various types of obstacles (such as buildings, utility poles, trees, and birds). Traditional obstacle avoidance technologies mostly rely on a single sensor (such as ultrasonic or visual sensors) for obstacle detection. Limited by sensing accuracy, field of view, and environmental adaptability, they struggle to achieve real-time identification and effective avoidance of dynamic obstacles. Furthermore, existing systems often embed obstacle avoidance strategies into flight control algorithms, lacking flexibility and intelligence. They cannot dynamically adjust response measures based on different flight states and obstacle characteristics, resulting in delayed response and insufficient obstacle avoidance accuracy in sudden risk situations, posing significant flight safety hazards. Summary of the Invention

[0003] This application provides an automatic obstacle avoidance method and system for low-altitude aircraft, aiming to solve the technical problem that existing low-altitude aircraft have a lagging obstacle avoidance response in complex environments and are unable to avoid obstacles accurately in real time.

[0004] The first aspect disclosed in this application provides an automatic obstacle avoidance method for a low-altitude aircraft. The method includes: acquiring flight status information of the low-altitude aircraft in real time, the flight status information including flight speed, flight altitude, flight attitude, and real-time position information; acquiring obstacle information of the surrounding environment of the low-altitude aircraft through an environmental perception module, the obstacle information including at least obstacle position and obstacle size; calculating the relative position and relative speed between the low-altitude aircraft and the obstacles based on the flight status information and obstacle information, and determining potential collision risk areas; if a potential collision risk is determined, retrieving and selecting an obstacle avoidance strategy based on a preset obstacle avoidance strategy library and the potential collision risk areas, and generating an obstacle avoidance control command; transmitting the obstacle avoidance control command to a flight controller to control the low-altitude aircraft to perform automatic obstacle avoidance actions.

[0005] Another aspect of this application discloses an automatic obstacle avoidance system for a low-altitude aircraft. The system includes: a flight status acquisition unit for real-time acquisition of flight status information of the low-altitude aircraft, including flight speed, flight altitude, flight attitude, and real-time position information; an environmental perception unit for acquiring obstacle information about the environment surrounding the low-altitude aircraft through an environmental perception module, including at least obstacle location and obstacle size; a collision risk determination unit for calculating the relative position and relative speed between the low-altitude aircraft and obstacles based on the flight status information and obstacle information, and determining potential collision risk areas; an obstacle avoidance strategy selection unit for selecting an obstacle avoidance strategy based on a preset obstacle avoidance strategy library and the potential collision risk areas, and generating obstacle avoidance control commands; and a flight control unit for transmitting the obstacle avoidance control commands to a flight controller to control the low-altitude aircraft to perform automatic obstacle avoidance actions.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0007] The aforementioned automatic obstacle avoidance method for low-altitude aircraft acquires real-time flight status information (such as speed, altitude, attitude, and position information) of the low-altitude aircraft, combines it with surrounding obstacle information (including the position and size of obstacles) collected by the environmental perception module, calculates the relative position and speed between the aircraft and the obstacles, and determines whether there is a potential collision risk. If a risk exists, the method intelligently selects an appropriate obstacle avoidance strategy based on a preset obstacle avoidance strategy library and the specific risk area, generates a corresponding obstacle avoidance control command, and sends the command to the flight controller, thereby realizing the autonomous obstacle avoidance operation of the aircraft and ensuring flight safety.

[0008] The above description is merely an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, specific embodiments of this application are given below. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying 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.

[0010] Figure 1 This is a flowchart illustrating an automatic obstacle avoidance method for a low-altitude aircraft in one embodiment.

[0011] Figure 2This is an architecture diagram of an automatic obstacle avoidance system for a low-altitude aircraft in one embodiment.

[0012] Explanation of reference numerals in the attached figures: Flight status acquisition unit 11, Environmental perception unit 12, Collision risk determination unit 13, Collision risk determination unit 14, Flight control unit 15. Detailed Implementation

[0013] This application provides an automatic obstacle avoidance method and system for low-altitude aircraft, which solves the technical problem that existing low-altitude aircraft have a lagging obstacle avoidance response in complex environments and are unable to avoid obstacles accurately in real time.

[0014] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0015] It should be noted that the terms “comprising” and “having”, and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to such process, method, product, or device.

[0016] Example 1, as Figure 1 As shown, this application provides an automatic obstacle avoidance method for low-altitude aircraft, the method comprising:

[0017] The flight status information of low-altitude aircraft is acquired in real time, including flight speed, flight altitude, flight attitude and real-time position information.

[0018] In this embodiment, the low-altitude aircraft continuously acquires its state parameters during flight through sensing components such as an inertial measurement unit (IMU), a global positioning system (GPS), and an altimeter. These parameters include the current flight speed (the distance the aircraft travels per unit time, used to determine its speed), flight altitude (the vertical distance of the aircraft relative to the ground, used to determine its flight level), flight attitude (typically including pitch, roll, and yaw angles, used to describe the aircraft's direction and attitude changes in space), and real-time location information (acquired via GPS or an inertial navigation system, used to accurately locate the aircraft's geographic coordinates in three-dimensional space). This information serves as the basic input for the low-altitude aircraft's perception and decision-making, providing dynamic data support for subsequent obstacle detection, collision assessment, and obstacle avoidance path planning.

[0019] The environmental perception module acquires obstacle information about the environment surrounding the low-altitude aircraft, including at least the location and size of the obstacles.

[0020] In one embodiment, an environmental perception module configured on a low-altitude aircraft is used to perceive and acquire obstacle information in the surrounding environment in real time. This module typically integrates multiple sensors, such as lidar, millimeter-wave radar, and visual cameras, to identify existing obstacles by scanning the space around the aircraft. The acquired obstacle information includes at least two key elements: the obstacle's position, representing its specific coordinates in space, used to determine its positional relationship relative to the aircraft; and the obstacle's dimensions, describing its length, width, and height in three-dimensional space, used to understand the obstacle's boundaries. This information provides the necessary environmental foundation for subsequent collision risk assessment and obstacle avoidance path calculation.

[0021] Furthermore, this application provides methods for acquiring obstacle information about the environment surrounding a low-altitude aircraft through an environmental perception module, including:

[0022] The environmental perception module is activated, which includes a radar sensor, a lidar sensor, and a vision sensor; obstacle identification is performed based on the environmental perception module to obtain multiple sets of obstacle identification information; the position information in the multiple sets of obstacle identification information is fused and corrected to obtain the obstacle position.

[0023] Preferably, firstly, the environmental perception module is activated. This module integrates multiple sensor types, including radar sensors (such as millimeter-wave radar), lidar sensors (such as LiDAR), and visual sensors (such as RGB cameras or depth cameras). These sensors operate simultaneously, acquiring raw perception data of obstacles around the aircraft from different angles and resolutions. Subsequently, obstacle recognition is performed based on the data acquired by these sensors, obtaining multiple sets of obstacle recognition information. Each set of obstacle recognition information corresponds to one sensor. Radar can quickly detect general outlines and moving obstacles, lidar provides detailed spatial point cloud distribution, and visual sensors assist in confirming object edges and categories. Next, the location information of the obstacles is extracted from the multiple sets of obstacle recognition information. Specifically, for obstacle recognition information collected by radar sensors, the straight-line distance from the aircraft to the obstacle is obtained by calculating the time difference between radar transmission and reception signals. The horizontal angular position of the obstacle relative to the aircraft is calculated using a beamforming algorithm of a multi-antenna array. Then, the distance and azimuth are projected onto the aircraft's coordinate system, and a rough radar position of the obstacle on the horizontal plane is obtained through calculation using a three-dimensional polar coordinate transformation formula. For obstacle recognition information acquired by LiDAR sensors, voxel downsampling or sparse point filtering is performed on the raw point cloud data to reduce computational burden. Then, clustering algorithms such as DBSCAN and Euclidean Cluster Extraction are used to separate individual obstacles, and the coarse LiDAR position is obtained by calculating the centroid coordinates of the clusters. For obstacle recognition information acquired by visual sensors, target detection algorithms such as YOLO and Faster R-CNN are applied to identify obstacles in the image and obtain their two-dimensional image coordinate center points. Then, combining image pixels with corresponding depth values, pixel back-projection is performed to map the image center points to three-dimensional spatial coordinates, thus obtaining the visual coarse position. Then, the inverse ratio of the measurement accuracy of each sensor's markings is used as the weight of each sensor's observation. The obtained coarse LiDAR position, coarse LiDAR position, and coarse visual position are weighted and fused to obtain the obstacle position. This obstacle position eliminates recognition bias caused by different viewing angles, resolution differences, or synchronization errors, providing accurate position input for subsequent risk analysis and obstacle avoidance decisions.

[0024] Furthermore, this application provides methods for acquiring obstacle information about the environment surrounding a low-altitude aircraft through an environmental perception module, including:

[0025] The multiple sets of obstacle recognition information are decomposed to obtain LiDAR obstacle recognition information; density clustering is used to segment the point cloud data in the LiDAR obstacle recognition information to extract independent obstacle point cloud clusters; the extreme point coordinates of the independent obstacle point cloud clusters in the three-dimensional coordinate system are calculated to generate obstacle bounding box size parameters; the three-dimensional size of the obstacle bounding box size parameters is calculated to obtain the obstacle size; the obstacle size is mapped to the obstacle position to construct the obstacle information.

[0026] Optionally, firstly, the obstacle identification information from multiple sensors is decomposed and processed to extract the obstacle identification data generated by the LiDAR sensor. This data mainly contains 3D point cloud information, i.e., a large number of spatial scattered points obtained by laser scanning. Each point has spatial coordinates (x, y, z), reflecting the distribution pattern of the obstacle surface in 3D space. Subsequently, a density clustering algorithm (such as DBSCAN) is used to process the LiDAR point cloud data. This algorithm can automatically identify the set of points belonging to the same entity based on the local density of each point in the point cloud, combined with a distance threshold and the minimum number of cluster points, and exclude outliers, thereby dividing the original continuous point cloud into several independent obstacle point cloud clusters, each corresponding to an actual obstacle. Then, for each independent obstacle point cloud cluster, its extreme point coordinates are calculated in a 3D coordinate system, i.e., the minimum and maximum values ​​of the point cloud cluster in the x, y, and z axes are extracted. These extreme points are used to construct a minimum bounding box surrounding the obstacle, forming the bounding box size parameters of the obstacle. Then, based on the three dimensions (length, width, and height) of the bounding box, three-dimensional dimensions are calculated. The length of the obstacle is obtained by subtracting the minimum value from the maximum value along the x-axis, the width by subtracting the minimum value along the y-axis, and the height by subtracting the minimum value along the z-axis. By summing these dimensional data, the size data of the obstacle in the actual physical space is obtained. Finally, the three-dimensional obstacle dimensions are mapped and integrated with the spatial position of the obstacle obtained through multi-sensor fusion. The position coordinates are combined with the bounding box dimensions to complete the structured representation of the obstacle in the environmental model. The output is complete obstacle information, including its position, size, and geometry in space, which serves as the input basis for flight obstacle avoidance judgment.

[0027] Based on the flight status information and obstacle information, the relative position and relative speed between the low-altitude aircraft and the obstacle are calculated to determine the potential collision risk area.

[0028] In one embodiment, after acquiring the current flight status information of the low-altitude aircraft and the information of surrounding obstacles, the position coordinates of the obstacles are transformed or mapped to a coordinate system unified with that of the aircraft to ensure the comparability of their spatial positions. Subsequently, based on the real-time position information of the low-altitude aircraft and the position of the obstacles, the straight-line distance between the obstacles and the low-altitude aircraft is calculated. This straight-line distance represents the relative positional relationship between the obstacles and the low-altitude aircraft. Then, based on the spatial point cloud data collected by the lidar sensor, the obstacle velocity is calculated. If the obstacle is stationary, its velocity is set to 0. The velocity of the low-altitude aircraft and the obstacle velocity are then decomposed and calculated to obtain their relative velocity. After obtaining the relative position and relative velocity, based on a preset safe distance threshold, it is predicted whether the aircraft may enter a dangerous approach state within a certain period of time. If the prediction result shows that the aircraft may approach the obstacle to a distance less than the safe distance within a certain period of time, a potential collision risk is determined, and a potential collision risk area is constructed based on the predicted trajectory, serving as the core basis for subsequent obstacle avoidance strategy selection and control command generation. This area can be considered a high-risk segment on the flight path. The system will intervene in advance to adjust the path or actively avoid obstacles, thereby improving the success rate of autonomous obstacle avoidance and flight safety of low-altitude aircraft.

[0029] Furthermore, this application provides a method for calculating the relative position and relative velocity between the low-altitude aircraft and obstacles based on the flight status information and obstacle information, and determining potential collision risk areas, including:

[0030] Obstacle information is mapped to the flight coordinate system, and the straight-line distance and relative speed between the obstacle and the low-altitude aircraft are calculated. Based on the preset safe distance threshold and relative speed, potential collision trajectories are dynamically predicted, and potential collision risk areas are determined.

[0031] Optionally, the obstacle position information acquired through the environmental perception module is transformed into a unified flight coordinate system used by the aircraft. This process typically employs coordinate transformation algorithms. For example, the extrinsic parameter matrix of the sensor installation location (i.e., the sensor's position and attitude relative to the aircraft body) is used to convert the obstacle's position data in the sensor coordinate system into coordinate data in the aircraft coordinate system. This ensures that the aircraft's own position and the obstacle's position are calculated within the same spatial reference frame. Subsequently, in the unified coordinate system, based on the real-time position information of the low-altitude aircraft and the obstacle's position, the straight-line distance between the obstacle and the low-altitude aircraft is calculated using Euclidean distance. This straight-line distance represents the relative positional relationship between the obstacle and the low-altitude aircraft. Simultaneously, the obstacle's velocity is calculated based on the spatial point cloud data collected by the lidar sensor. If the obstacle is stationary, its velocity is directly set to 0, and no velocity calculation is performed. Next, the velocity of the low-altitude aircraft and the obstacle are decomposed to obtain velocity components in the X, Y, and Z axes. The velocity difference between the obstacle and the aircraft is then subtracted from the velocity component of the obstacle in each direction to obtain a vector representing the relative velocity direction and components. Modulo operation is then used to calculate the relative velocity between the two. Then, based on a preset safe distance threshold (i.e., the minimum allowable distance set to ensure flight safety), the system predicts whether the low-altitude aircraft will enter this distance range within a future period. During the prediction process, the relative velocity is used to estimate the future trajectory of the aircraft, and it is determined whether the obstacle is also relatively close. If a distance less than the safe threshold exists within the preset time window, it indicates a potential collision risk. In this case, the system defines these risky trajectory points as potential collision risk areas, which serve as the core basis for subsequent obstacle avoidance strategy decisions to avoid entering high-risk areas and affecting the flight safety of the low-altitude aircraft.

[0032] Furthermore, this application provides that, before calculating the relative velocity, it also includes:

[0033] The spatial point cloud data of the obstacle is continuously collected at a set frequency using the lidar sensor to construct a temporal point cloud frame sequence of the obstacle; the spatial position of the temporal point cloud frame sequence of the obstacle is matched to obtain the initial position movement vector of the obstacle, and the aircraft motion compensation processing is performed through the flight state information to generate the actual position movement vector; the velocity vector is calculated from the actual position movement vector to obtain the speed of the obstacle.

[0034] Optionally, a lidar sensor is used to continuously collect spatial point cloud data of the obstacle at a set sampling frequency (e.g., 10 or 20 frames per second). Each frame of point cloud data represents the three-dimensional surface morphology of the obstacle at a certain moment. Point cloud frames collected at multiple consecutive time points are combined in chronological order to construct a temporal point cloud frame sequence of the obstacle, reflecting the spatial change process of the obstacle over time. Subsequently, spatial position matching processing is performed on these consecutive frames, that is, the point cloud registration algorithm (such as the ICP algorithm, Iterative Closest Point) is used to align adjacent frames of point cloud data to identify the positional changes of the obstacle in consecutive time frames. By comparing the point cloud positional changes at different times, the initial position movement vector of the obstacle can be calculated, that is, the direction and distance of the obstacle's movement within a certain time interval when the aircraft is stationary. If the initial position movement vector is 0 or close to 0 (the deviation from 0 is within the allowable range), it indicates that the obstacle is a static obstacle. In this case, the obstacle velocity is set to 0, and no further obstacle velocity calculation is performed. However, since the aircraft itself is also in motion, the motion vector calculated solely based on point cloud frame comparisons will contain deviations introduced by the aircraft's own motion. Therefore, it is necessary to incorporate flight state information (such as the aircraft's own velocity, attitude, and displacement data) to compensate for the obstacle's motion vector and eliminate the influence of the aircraft's motion. Specifically, acceleration, angular velocity, real-time position, and flight attitude information are extracted in real time from the low-altitude aircraft's flight state information to construct a complete aircraft motion transformation matrix. This matrix contains the aircraft's rotation information (represented by a rotation matrix) and displacement information (represented by a translation vector) within a certain time (Δt), used to describe the aircraft's rigid motion trajectory. Then, point cloud frame data from two consecutive time points are extracted from the lidar module: the source point cloud frame at time t and the target point cloud frame at time t+Δt. Because the aircraft is in motion during this period, the lidar's viewing angle also changes, resulting in the two point cloud frames having inconsistent spatial positions. To address this issue, point cloud registration techniques (such as ICP or NDT algorithms) are used to rigidly register two point cloud frames. Taking ICP as an example, the point cloud frame at time t is taken as the source frame, and the point cloud frame at time t+Δt is taken as the target frame. Through iterative optimization, the Euclidean distance between each point in the source frame and the closest point in the target frame is minimized, thereby solving for the optimal rigid transformation matrix. The inverse transformation of this matrix is ​​then applied to the point cloud data of the target frame to eliminate the coordinate offset caused by the aircraft motion.Subsequently, the obstacle point cloud region in the compensated target frame is paired with the corresponding obstacle region in the source frame to recover the obstacle's true spatial position change in the world coordinate system. In other words, the obstacle's actual displacement vector can be obtained by subtracting its position coordinate at time t from its position at time t+Δt after compensation. This accurately reflects the obstacle's actual position movement vector after the influence of aircraft motion is eliminated. Dividing this actual displacement vector by the time interval Δt yields the obstacle's velocity vector, which serves as a key input parameter for subsequent relative velocity calculations and collision risk assessment. This process ensures the accuracy of obstacle velocity estimation, effectively improving the system's ability to identify and predict dynamic targets, meeting the technical requirements for safe autonomous flight of aircraft in complex environments.

[0035] Furthermore, this application provides a method for dynamically predicting potential collision trajectories and determining collision risk areas based on a preset safe distance threshold and relative speed, including:

[0036] The minimum relative distance within a preset time window is calculated based on the relative speed; the minimum relative distance is compared with the preset safe distance threshold to generate a collision risk verification result, which includes whether there is a collision risk or not; if the collision risk verification result indicates that there is a collision risk, the temporal position sequence of the low-altitude aircraft is calculated according to the preset time window to construct the potential collision risk area.

[0037] Optionally, after obtaining the relative velocity between the low-altitude aircraft and the obstacle, the relative velocity is combined with the current position to predict the future motion trend of the aircraft and the obstacle along their future path within a preset time window (e.g., 2 seconds, 3 seconds, etc.). In this process, the following formula is used: The minimum distance between the two occurs at time point t. min Where ΔT is the prediction time window; It is a relative position vector; Let be the relative velocity vector. Then, substitute the calculated minimum distance at the time point into the formula: Calculate the minimum distance D at that moment. minThe calculated minimum distance is then compared with a preset safe distance threshold. This threshold is set based on parameters such as the size of the aircraft, its flight speed, control delay, and emergency braking capability, and represents the minimum avoidance distance for a low-altitude aircraft to maintain safe flight. When the minimum relative distance is less than or equal to this safety threshold, it indicates a potential collision hazard for the low-altitude aircraft within a preset time period. In this case, the generated collision risk verification result indicates a collision risk exists; otherwise, it indicates no collision risk exists. When a collision risk is determined to exist, the aircraft's current flight state (including speed, direction, and attitude) is further used to calculate its temporal position sequence within the entire time window. Specifically, within the preset time window [0, ΔT], each time point is substituted into the trajectory prediction formula: In this process, the future position and trajectory of a low-altitude aircraft relative to obstacles are predicted, where P is the predicted position vector at time t, and P0 is the position vector at the current time. The velocity vector is used. For each time point, the difference between the predicted position vectors of the low-altitude aircraft and the obstacle is calculated and moduloed to obtain the distance between the low-altitude aircraft and the obstacle. This distance and a preset safe distance threshold are then marked. All spatial ranges corresponding to times t that are less than or equal to the preset safe distance threshold are marked as potential collision risk areas. These potential collision risk areas will serve as key inputs for obstacle avoidance strategy selection, enabling accurate identification and early avoidance of impending risk areas, thereby improving the overall active obstacle avoidance capability of the system.

[0038] If a potential collision risk is identified, an obstacle avoidance strategy is retrieved and selected based on a preset obstacle avoidance strategy library and the potential collision risk area, and an obstacle avoidance control command is generated.

[0039] In one embodiment, once it is determined that a low-altitude aircraft faces a potential collision risk with an obstacle within a preset time window, the obstacle avoidance strategy decision-making process is immediately initiated. At this point, the identified potential collision risk area is used as the core input, combined with the aircraft's current state information (such as speed, attitude, controllable direction, remaining reaction time, etc.), to search a preset obstacle avoidance strategy library. This library contains various typical avoidance schemes, such as detour strategies (yaw to the left / right), emergency stop strategies (rapid deceleration and hovering), altitude adjustment strategies (ascending or descending to avoid obstacles), and optimal path replanning strategies. These strategies are typically categorized and set based on different obstacle shapes, approach angles, environmental constraints (such as narrow spaces, height-restricted areas), and the aircraft's controllability limits. The system performs joint matching based on the spatial characteristics of the current risk area and the aircraft's maneuverability parameters to determine the most suitable obstacle avoidance strategy for the current scenario. This matching process can employ rule-based logical judgment, priority ranking, fuzzy reasoning, or intelligent selection using machine learning models. Finally, the selected obstacle avoidance strategy is converted into executable obstacle avoidance control commands, which include specific flight parameter adjustment commands, such as yaw angle adjustment, flight speed change, and target trajectory correction path. These control commands are transmitted to the flight controller, which executes them directly, guiding the low-altitude aircraft to perform active obstacle avoidance maneuvers safely, effectively mitigating potential collision risks and ensuring mission continuity and flight safety.

[0040] Furthermore, this application provides a method for generating obstacle avoidance control instructions based on a preset obstacle avoidance strategy library, combined with the retrieval and selection of obstacle avoidance strategies in the potential collision risk area, including:

[0041] Based on the potential collision risk area and the flight status information of the low-altitude aircraft, the risk triggering direction and remaining reaction time are determined; the risk triggering direction and remaining reaction time are combined with the low-altitude aircraft's control capability parameters to perform a search of the preset obstacle avoidance strategy library and obtain the obstacle avoidance control command.

[0042] Preferably, after detecting a potential collision risk area, the spatial positional relationship between the risk area and the aircraft is analyzed based on the current flight status information of the aircraft (including velocity vector, attitude angle, heading direction, etc.). This determines the triggering direction in which the obstacle may collide with the aircraft. This direction can be described as directly in front, to the left front, to the right, above, or below, representing the relative direction in which the aircraft will encounter the obstacle under the current heading. Subsequently, using the relative velocity vector between the aircraft and the obstacle and the current relative distance, the time from now until the possible collision is calculated while maintaining the current speed. This is the remaining reaction time, which determines how long the system must complete obstacle avoidance decisions and flight adjustments, and is one of the core constraints for obstacle avoidance strategy selection. Then, the identified risk triggering direction and remaining reaction time are used as inputs and analyzed in conjunction with the aircraft's control capability parameters. These parameters include the aircraft's maximum pitch angle, maximum yaw rate, climb / descent speed, minimum turning radius, etc., representing the physical capability boundaries of the aircraft to safely complete a certain avoidance maneuver. After meeting the aforementioned constraints, the system enters the obstacle avoidance strategy library retrieval phase. In this phase, based on the current collision risk characteristics and aircraft status conditions, rule matching is used to filter and compare various strategies. For example, the strategy direction must be consistent; the remaining reaction time must be greater than or equal to the strategy execution time; and the action capability parameters must be less than or equal to the control capability parameters. Finally, through rule matching, multiple executable strategies are selected from the strategy library. Then, based on the priority score of each strategy (based on multiple factors such as energy consumption, safety, and route recovery), the optimal strategy is extracted as the obstacle avoidance strategy. Based on this, an obstacle avoidance control command containing specific instruction parameters is generated, such as changing the heading angle, adjusting speed, or executing trajectory redirection commands. This obstacle avoidance control command is sent to the flight controller to complete the aircraft's autonomous avoidance operation, ensuring a rapid obstacle avoidance response to sudden risks.

[0043] The obstacle avoidance control command is transmitted to the flight controller to control the low-altitude aircraft to perform automatic obstacle avoidance actions.

[0044] In one embodiment, after selecting an obstacle avoidance strategy and generating corresponding obstacle avoidance control commands, these commands are transmitted to the flight controller via an internal communication interface. The flight controller is the core execution unit of the aircraft, responsible for parsing and executing all flight-related action commands. Upon receiving these commands, the flight controller coordinates and controls multiple subsystems, including flight attitude, motor speed, and control surfaces, based on the current flight status, driving the aircraft to adjust its flight path in real time, thereby achieving proactive obstacle avoidance. The entire obstacle avoidance process requires no manual intervention; the system automatically identifies risks, selects strategies, and issues commands, which are ultimately executed by the flight control system. Through this closed-loop control mechanism, the aircraft can achieve autonomous obstacle avoidance flight in complex dynamic environments, effectively avoiding collisions and ensuring the continuity and safety of flight missions.

[0045] In summary, the embodiments of this application have at least the following technical effects:

[0046] This application embodiment first acquires the flight status information of the low-altitude aircraft in real time, including flight speed, flight altitude, flight attitude, and real-time position information. Then, it acquires obstacle information of the surrounding environment through an environmental perception module, including at least obstacle location and size. Next, based on the flight status information and obstacle information, it calculates the relative position and relative speed between the low-altitude aircraft and the obstacles to determine potential collision risk areas. Then, if a potential collision risk is determined, it retrieves and selects an obstacle avoidance strategy based on a preset obstacle avoidance strategy library and the potential collision risk areas, generating an obstacle avoidance control command. Finally, it transmits the obstacle avoidance control command to the flight controller to control the low-altitude aircraft to perform automatic obstacle avoidance actions. These technical effects collectively solve the technical problem of existing low-altitude aircraft having a lagging obstacle avoidance response in complex environments and difficulty in accurately avoiding obstacles in real time. It achieves the technical effect of dynamically predicting collision risks through multi-source information fusion and intelligent strategy matching, thereby improving the success rate of autonomous obstacle avoidance and flight safety of low-altitude aircraft.

[0047] Example 2, based on the same inventive concept as the automatic obstacle avoidance method for low-altitude aircraft in the aforementioned examples, such as... Figure 2As shown, this application provides an automatic obstacle avoidance system for low-altitude aircraft. The system includes: a flight status acquisition unit 11, which acquires flight status information of the low-altitude aircraft in real time, including flight speed, flight altitude, flight attitude, and real-time position information; an environment perception unit 12, which acquires obstacle information of the environment surrounding the low-altitude aircraft through an environment perception module, including at least obstacle position and obstacle size; a collision risk determination unit 13, which calculates the relative position and relative speed between the low-altitude aircraft and obstacles based on the flight status information and obstacle information, and determines potential collision risk areas; an obstacle avoidance strategy selection unit 14, which, if a potential collision risk is determined, retrieves and selects an obstacle avoidance strategy based on a preset obstacle avoidance strategy library and the potential collision risk areas, and generates an obstacle avoidance control command; and a flight control unit 15, which transmits the obstacle avoidance control command to the flight controller to control the low-altitude aircraft to perform automatic obstacle avoidance actions.

[0048] Furthermore, the environment sensing unit 12 is also used to perform the following method:

[0049] The environmental perception module is activated, which includes a radar sensor, a lidar sensor, and a vision sensor; obstacle identification is performed based on the environmental perception module to obtain multiple sets of obstacle identification information; the position information in the multiple sets of obstacle identification information is fused and corrected to obtain the obstacle position.

[0050] Furthermore, the environment sensing unit 12 is also used to perform the following method:

[0051] The multiple sets of obstacle recognition information are decomposed to obtain LiDAR obstacle recognition information; density clustering is used to segment the point cloud data in the LiDAR obstacle recognition information to extract independent obstacle point cloud clusters; the extreme point coordinates of the independent obstacle point cloud clusters in the three-dimensional coordinate system are calculated to generate obstacle bounding box size parameters; the three-dimensional size of the obstacle bounding box size parameters is calculated to obtain the obstacle size; the obstacle size is mapped to the obstacle position to construct the obstacle information.

[0052] Furthermore, the collision risk determination unit 13 is also used to perform the following method:

[0053] Obstacle information is mapped to the flight coordinate system, and the straight-line distance and relative speed between the obstacle and the low-altitude aircraft are calculated. Based on the preset safe distance threshold and relative speed, potential collision trajectories are dynamically predicted, and potential collision risk areas are determined.

[0054] Furthermore, the collision risk determination unit 13 is also used to perform the following method:

[0055] The spatial point cloud data of the obstacle is continuously collected at a set frequency using the lidar sensor to construct a temporal point cloud frame sequence of the obstacle; the spatial position of the temporal point cloud frame sequence of the obstacle is matched to obtain the initial position movement vector of the obstacle, and the aircraft motion compensation processing is performed through the flight state information to generate the actual position movement vector; the velocity vector is calculated from the actual position movement vector to obtain the speed of the obstacle.

[0056] Furthermore, the collision risk determination unit 13 is also used to perform the following method:

[0057] The minimum relative distance within a preset time window is calculated based on the relative speed; the minimum relative distance is compared with the preset safe distance threshold to generate a collision risk verification result, which includes whether there is a collision risk or not; if the collision risk verification result indicates that there is a collision risk, the temporal position sequence of the low-altitude aircraft is calculated according to the preset time window to construct the potential collision risk area.

[0058] Furthermore, the obstacle avoidance strategy selection unit 14 is also used to perform the following method:

[0059] Based on the potential collision risk area and the flight status information of the low-altitude aircraft, the risk triggering direction and remaining reaction time are determined; the risk triggering direction and remaining reaction time are combined with the low-altitude aircraft's control capability parameters to perform a search of the preset obstacle avoidance strategy library and obtain the obstacle avoidance control command.

[0060] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0061] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0062] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. An automatic obstacle avoidance method for low-altitude aircraft, characterized in that, include: Real-time acquisition of flight status information of low-altitude aircraft, including flight speed, flight altitude, flight attitude and real-time position information; The obstacle information of the environment surrounding the low-altitude aircraft is obtained through the environmental perception module. The obstacle information includes at least the location and size of the obstacle. Based on the flight status information and obstacle information, the relative position and relative speed between the low-altitude aircraft and the obstacle are calculated to determine the potential collision risk area. If a potential collision risk is determined, an obstacle avoidance control command is generated by searching and selecting an obstacle avoidance strategy based on a preset obstacle avoidance strategy library and the potential collision risk area. The obstacle avoidance control command is transmitted to the flight controller to control the low-altitude aircraft to perform automatic obstacle avoidance actions.

2. The automatic obstacle avoidance method for low-altitude aircraft as described in claim 1, characterized in that, Obstacle information about the environment surrounding the low-altitude aircraft is obtained through the environmental perception module, including: Activate the environmental perception module, which includes a radar sensor, a lidar sensor, and a vision sensor; Obstacle recognition is performed based on the environmental perception module to obtain multiple sets of obstacle recognition information; The location information in the multiple sets of obstacle identification information is fused and corrected to obtain the obstacle location.

3. The automatic obstacle avoidance method for low-altitude aircraft as described in claim 2, characterized in that, Obstacle information about the environment surrounding the low-altitude aircraft is obtained through the environmental perception module, including: The multiple sets of obstacle recognition information are decomposed to obtain lidar obstacle recognition information; Density clustering is used to segment the point cloud data in the lidar obstacle identification information into obstacle entities, and independent obstacle point cloud clusters are extracted. Calculate the extreme point coordinates of the independent obstacle point cloud cluster in the three-dimensional coordinate system to generate obstacle bounding box size parameters; The obstacle dimensions are obtained by performing three-dimensional dimension calculations on the obstacle bounding box size parameters; The obstacle information is constructed by mapping the obstacle size to the obstacle position.

4. The automatic obstacle avoidance method for low-altitude aircraft as described in claim 2, characterized in that, Based on the flight status information and obstacle information, the relative position and relative velocity between the low-altitude aircraft and the obstacle are calculated to determine potential collision risk areas, including: Map obstacle information to the flight coordinate system and calculate the straight-line distance and relative speed between the obstacle and the low-altitude aircraft; Based on preset safe distance thresholds and relative speeds, potential collision trajectories are dynamically predicted to determine potential collision risk areas.

5. The automatic obstacle avoidance method for low-altitude aircraft as described in claim 4, characterized in that, Before calculating relative velocity, the following is also included: The spatial point cloud data of the obstacle is continuously collected using the lidar sensor at a set frequency to construct a temporal point cloud frame sequence of the obstacle. Spatial position matching is performed on the temporal point cloud frame sequence of the obstacle to obtain the initial position movement vector of the obstacle, and the actual position movement vector is generated by performing aircraft motion compensation processing through the flight state information. The velocity vector of the obstacle is calculated by performing velocity vector calculation on the actual position movement vector.

6. The automatic obstacle avoidance method for low-altitude aircraft as described in claim 4, characterized in that, Based on preset safe distance thresholds and relative speeds, potential collision trajectories are dynamically predicted, and collision risk areas are identified, including: The minimum relative distance within a preset time window is calculated based on the relative velocity. The minimum relative distance is compared with the preset safe distance threshold to generate a collision risk verification result, which includes whether there is a collision risk or not. If the collision risk verification result indicates that there is a collision risk, the temporal position sequence of the low-altitude aircraft is calculated based on the preset time window to construct the potential collision risk area.

7. The automatic obstacle avoidance method for low-altitude aircraft as described in claim 1, characterized in that, Based on a pre-defined obstacle avoidance strategy library, and combined with the search and selection of obstacle avoidance strategies in the potential collision risk areas, obstacle avoidance control instructions are generated, including: Based on the potential collision risk area and the flight status information of the low-altitude aircraft, the risk triggering direction and remaining reaction time are determined. By combining the risk triggering direction, remaining reaction time, and low-altitude aircraft control capability parameters, a preset obstacle avoidance strategy library is searched to obtain the obstacle avoidance control command.

8. An automatic obstacle avoidance system for low-altitude aircraft, characterized in that, The system is used to execute the automatic obstacle avoidance method for a low-altitude aircraft according to any one of claims 1 to 7, comprising: Flight status acquisition unit: acquires the flight status information of the low-altitude aircraft in real time, including flight speed, flight altitude, flight attitude and real-time position information; Environmental perception unit: Obtains obstacle information of the environment surrounding the low-altitude aircraft through the environmental perception module. The obstacle information includes at least the obstacle's location and size. Collision risk determination unit: Based on the flight status information and obstacle information, calculate the relative position and relative speed between the low-altitude aircraft and the obstacle, and determine the potential collision risk area; Obstacle avoidance strategy selection unit: If a potential collision risk is determined, based on the preset obstacle avoidance strategy library, the unit retrieves and selects an obstacle avoidance strategy in conjunction with the potential collision risk area, and generates an obstacle avoidance control command. Flight control unit: Transmits the obstacle avoidance control commands to the flight controller to control the low-altitude aircraft to perform automatic obstacle avoidance actions.

Images