Urban low-altitude flight obstacle avoidance method in complex environment

Abstract

The application belongs to the technical field of flight control, and specifically discloses a city low-altitude flight obstacle avoidance method under complex environment, which comprises the following steps: collecting multiple types of data of the environment around a flight carrier through multiple types of sensors; constructing a static environment model and a dynamic obstacle motion model based on the multiple types of data; predicting a motion trajectory of a dynamic obstacle based on the dynamic obstacle motion model, and determining an obstacle avoidance priority of the dynamic obstacle based on a flight trajectory of the flight carrier and the motion trajectory of the dynamic obstacle; generating an obstacle avoidance path through a path planning algorithm based on the static environment model, the motion trajectory of the dynamic obstacle and the obstacle avoidance priority of the dynamic obstacle; and controlling the flight carrier to avoid obstacles based on the obstacle avoidance path. The method can improve the safety and reliability of low-altitude flight.

Description

Technical Field

[0001] This application belongs to the field of flight control technology, and more specifically, relates to a method for obstacle avoidance in urban low-altitude flight under complex environments. Background Technology

[0002] With the rapid development of the urban low-altitude economy, application scenarios such as drone logistics delivery, low-altitude tourism, and urban emergency rescue are constantly expanding, leading to a continuous increase in the operational demand for low-altitude flight vehicles. The urban low-altitude environment is characterized by dense building distribution, significant differences in elevation, complex electromagnetic signals, and the random appearance of dynamic obstacles, placing extremely high demands on the obstacle avoidance capabilities of flight vehicles.

[0003] Current low-altitude obstacle avoidance technologies primarily rely on single sensors (visual sensors, lidar) to collect environmental information and achieve obstacle avoidance through preset path planning or simple obstacle detection rules. However, in complex urban environments, single sensors are susceptible to environmental interference: visual sensors experience a significant decrease in detection accuracy under conditions such as rain, fog, haze, and insufficient nighttime lighting; lidar is prone to signal reflection anomalies when facing buildings made of special materials such as glass curtain walls and reflective metal surfaces, leading to missed or false detections of obstacles. Furthermore, existing obstacle avoidance algorithms are mostly designed for static environments and lack sufficient accuracy in predicting the trajectory of dynamic obstacles, making it difficult to cope with the sudden movement of dynamic obstacles in urban low-altitude environments and increasing the risk of collisions. Furthermore, there are numerous "visual blind spots" in urban low-altitude flight paths, such as narrow passages between buildings and shadowed areas created by tall buildings. Existing algorithms cannot cover these blind spots, resulting in a lack of reliable obstacle avoidance decision-making basis for flight vehicles when traversing such areas. As the number of low-altitude flight vehicles increases, the need for obstacle avoidance in multi-vehicle cooperative flight scenarios becomes even more prominent. Existing algorithms lack the ability to analyze the interactive behavior of multiple targets, making it difficult to achieve efficient cooperative obstacle avoidance and restricting the large-scale operation of urban low-altitude transportation systems. In summary, existing obstacle avoidance methods for low-altitude flight have low obstacle avoidance capabilities, resulting in lower safety and reliability for low-altitude flights. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this application aims to provide a method for obstacle avoidance in urban low-altitude flight under complex environments. This method addresses the problem of low safety and reliability in low-altitude flight due to reliance on a single sensor and insufficient accuracy in predicting the trajectory of dynamic obstacles.

[0005] To achieve the above objectives, in a first aspect, this application provides a method for obstacle avoidance during low-altitude urban flight in complex environments, comprising: Multiple types of data about the environment surrounding the flight vehicle are collected using various sensors. Based on the aforementioned multiple types of data, a static environment model and a dynamic obstacle motion model are constructed. Based on the dynamic obstacle motion model, the motion trajectory of the dynamic obstacle is predicted, and the obstacle avoidance priority of the dynamic obstacle is determined based on the flight trajectory of the flight vehicle and the motion trajectory of the dynamic obstacle. Based on the static environment model, the motion trajectory of the dynamic obstacle, and the obstacle avoidance priority of the dynamic obstacle, an obstacle avoidance path is generated through a path planning algorithm. Based on the obstacle avoidance path, the flight vehicle is controlled to avoid obstacles.

[0006] This application employs multiple types of sensors to collect various types of data about the environment surrounding the flight vehicle, including data on static and dynamic obstacles. This addresses the problem of insufficient detection accuracy of a single sensor in complex environments, improving the reliability and completeness of environmental data. Simultaneously, it constructs static environment models and dynamic obstacle motion models, predicts the motion trajectories of dynamic obstacles, and, combined with the real-time flight trajectory of the flight vehicle, prioritizes obstacle avoidance, providing a basis for subsequent obstacle avoidance decisions. This overcomes the shortcomings of existing technologies, such as insufficient accuracy in predicting the motion trajectory of dynamic obstacles, difficulty in handling sudden movements of dynamic obstacles in urban low-altitude environments, and the risk of collisions. Based on the static environment model, the motion trajectory of dynamic obstacles, and the obstacle avoidance priority of dynamic obstacles, an obstacle avoidance path is generated through a path planning algorithm, and the flight vehicle is controlled to avoid obstacles based on this path, thereby improving the safety and reliability of low-altitude flight.

[0007] According to the obstacle avoidance method for low-altitude urban flight in complex environments provided in this application, the method involves collecting various types of data about the environment surrounding the flight vehicle using multiple sensors, including: Three-dimensional point cloud data of the environment surrounding the flight vehicle are collected using lidar. The visual camera acquires color and depth images of the environment surrounding the flight vehicle. The speed and distance data of dynamic obstacles around the flight vehicle are collected using millimeter-wave radar. The real-time position and attitude information of the flight vehicle is collected through the BeiDou positioning module.

[0008] This application employs a multi-sensor acquisition system composed of lidar, visual camera, millimeter-wave radar, and BeiDou positioning module to collect comprehensive data on the environment surrounding the flight vehicle, solving the problem of insufficient detection accuracy of a single sensor in complex environments and improving the reliability and completeness of environmental data.

[0009] According to the obstacle avoidance method for low-altitude flight in complex environments provided in this application, the step of predicting the trajectory of dynamic obstacles based on the dynamic obstacle motion model includes: Based on the dynamic obstacle motion model, a linear prediction model is used to predict the short-term motion trajectory of a dynamic target moving at a constant speed, and a nonlinear prediction model is used to predict the medium- and long-term motion trajectory of a dynamic target moving at a non-uniform speed.

[0010] This application constructs a multi-model prediction system that combines linear and nonlinear methods, selects an appropriate prediction model for dynamic obstacles with different motion characteristics, and achieves accurate prediction of dynamic target trajectories.

[0011] According to the obstacle avoidance method for low-altitude urban flight in a complex environment provided in this application, the method for determining the obstacle avoidance priority of dynamic obstacles based on the flight trajectory of the flight vehicle and the motion trajectory of the dynamic obstacles includes: Based on the flight trajectory of the flight vehicle and the motion trajectory of the dynamic obstacle, calculate the minimum encounter distance and encounter time between the dynamic obstacle and the flight vehicle; Based on the minimum encounter distance and encounter time between the dynamic obstacle and the flight vehicle, a risk assessment model is established, and the risk level of the dynamic obstacle is determined based on the risk assessment model. Based on the risk level of dynamic obstacles, the obstacle avoidance priority of dynamic obstacles is determined.

[0012] This application combines a risk assessment model to prioritize obstacle avoidance, addressing the problems of low accuracy in predicting dynamic obstacle trajectories and delayed risk assessment in existing algorithms, and providing accurate risk basis for obstacle avoidance decisions.

[0013] According to the obstacle avoidance method for low-altitude urban flight in a complex environment provided in this application, the obstacle avoidance path is generated through a path planning algorithm based on the static environment model, the motion trajectory of the dynamic obstacle, and the obstacle avoidance priority of the dynamic obstacle, including: Based on the static environment model, the motion trajectory of the dynamic obstacle, and the obstacle avoidance priority of the dynamic obstacle, the improved A... The path planning algorithm generates the initial obstacle avoidance path; Based on the blind zone environmental data in the multiple types of data, the initial obstacle avoidance path is corrected to obtain the obstacle avoidance path.

[0014] This application improves A The path planning algorithm, combining urban geographic information system (GIS) data with supplementary information on blind spot environments, generates obstacle avoidance paths that adapt to static complex environments, thereby improving the obstacle avoidance flexibility and safety of flight vehicles in complex environments.

[0015] According to the obstacle avoidance method for low-altitude urban flight in a complex environment provided in this application, the step of controlling the flight vehicle to avoid obstacles based on the obstacle avoidance path includes: Based on the obstacle avoidance path, the flight vehicle is controlled to avoid obstacles. In areas with dense static obstacles, obstacle avoidance paths with high smoothness and small turning radius are selected first. In areas with high risk of dynamic obstacles, a rapid obstacle avoidance strategy is adopted to adjust the flight trajectory.

[0016] This application designs a multi-scenario adaptive decision-making mechanism to address the shortcomings of existing algorithms in terms of poor scenario adaptability, thereby improving the obstacle avoidance flexibility and safety of flight vehicles in complex environments.

[0017] According to the obstacle avoidance method for low-altitude flight in complex environments provided in this application, the method further includes: If there are multiple flight vehicles, the obstacle avoidance paths of each flight vehicle are optimized collaboratively to avoid path intersections.

[0018] This application optimizes the obstacle avoidance paths of each flight vehicle in a multi-vehicle cooperative flight scenario, avoiding path conflicts between multiple vehicles and improving the obstacle avoidance flexibility and safety of the flight vehicle in complex environments.

[0019] According to the obstacle avoidance method for low-altitude flight in complex environments provided in this application, the method further includes: If the position deviation of the obstacle exceeds the first preset threshold or the flight trajectory deviation of the flight vehicle exceeds the second preset threshold during the obstacle avoidance process, a new obstacle avoidance path will be generated.

[0020] Secondly, this application provides an urban low-altitude flight obstacle avoidance device for complex environments, comprising: The data acquisition module is used to collect various types of data about the environment surrounding the flight vehicle through multiple sensors. A construction module is used to construct a static environment model and a dynamic obstacle motion model based on the aforementioned multiple types of data; The prediction module is used to predict the trajectory of the dynamic obstacle based on the dynamic obstacle motion model, and to determine the obstacle avoidance priority of the dynamic obstacle based on the flight trajectory of the flight vehicle and the trajectory of the dynamic obstacle. The generation module is used to generate an obstacle avoidance path based on the static environment model, the motion trajectory of the dynamic obstacle, and the obstacle avoidance priority of the dynamic obstacle, using a path planning algorithm. The control module is used to control the flight vehicle to avoid obstacles based on the obstacle avoidance path.

[0021] Thirdly, this application provides an electronic device, comprising: at least one memory for storing a program; and at least one processor for executing the program stored in the memory. When the program stored in the memory is executed, the processor is used to execute the urban low-altitude flight obstacle avoidance method in a complex environment as described in the first aspect or any possible implementation of the first aspect.

[0022] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when run on a processor, causes the processor to execute the urban low-altitude flight obstacle avoidance method in a complex environment as described in the first aspect or any possible implementation of the first aspect.

[0023] Fifthly, this application provides a computer program product that, when run on a processor, causes the processor to execute the urban low-altitude flight obstacle avoidance method in a complex environment as described in the first aspect or any possible implementation of the first aspect.

[0024] It is understood that the beneficial effects of the second to fifth aspects mentioned above can be found in the relevant descriptions in the first aspect mentioned above, and will not be repeated here.

[0025] Overall, the technical solutions conceived in this application have the following beneficial effects compared with the prior art: (1) This application uses multiple types of sensors to collect multiple types of data about the environment around the flight vehicle, including data on static obstacles and dynamic obstacles, to solve the problem of insufficient detection accuracy of a single sensor in complex environments, improve the reliability and integrity of environmental data, and construct a static environment model and a dynamic obstacle motion model, and predict the motion trajectory of dynamic obstacles. Combined with the real-time flight trajectory of the flight vehicle, obstacle avoidance priority is divided to provide a basis for subsequent obstacle avoidance decisions. This overcomes the shortcomings of existing technologies in predicting the motion trajectory of dynamic obstacles, making it difficult to cope with the sudden movement of dynamic obstacles in urban low-altitude environments, and easily causing collision risks. Based on the static environment model, the motion trajectory of dynamic obstacles and the obstacle avoidance priority of dynamic obstacles, an obstacle avoidance path is generated through a path planning algorithm, and the flight vehicle is controlled to avoid obstacles based on the obstacle avoidance path, which can improve the safety and reliability of low-altitude flight.

[0026] (2) This application constructs a multi-model prediction system that combines linear and nonlinear methods, selects an appropriate prediction model for dynamic obstacles with different motion characteristics, and realizes accurate prediction of dynamic target trajectories.

[0027] (3) Improvement A of this application The path planning algorithm, combined with urban GIS data and supplementary information on blind spot environments, generates obstacle avoidance paths that are adapted to static and complex environments, thereby improving the obstacle avoidance flexibility and safety of flight vehicles in complex environments. Attached Figure Description

[0028] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 This is a flowchart illustrating the urban low-altitude flight obstacle avoidance method in complex environments provided in the embodiments of this application; Figure 2 This is a schematic diagram of the data acquisition process provided in an embodiment of this application; Figure 3 This is a schematic diagram of the structure of the urban low-altitude flight obstacle avoidance device in a complex environment provided in the embodiments of this application; Figure 4 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0031] In this article, the term "and / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. The symbol " / " in this article indicates that the related objects are in an "or" relationship; for example, A / B means A or B.

[0032] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design. Specifically, the use of the terms "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0033] In the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more, for example, multiple processing units means two or more processing units, multiple elements means two or more elements, etc.

[0034] Next, combined Figures 1-2 This application describes the obstacle avoidance method for low-altitude urban flight in complex environments provided in its embodiments.

[0035] Figure 1 This is a flowchart illustrating the obstacle avoidance method for low-altitude urban flight in complex environments provided in this application embodiment, as shown below. Figure 1 As shown, the method includes the following steps: Step S1: Collect various types of data about the environment surrounding the flight vehicle using multiple types of sensors; Alternatively, the various types of sensors can be radar, image acquisition devices, positioning devices, etc.

[0036] Alternatively, the data can be environmental image data, obstacle speed and distance data, flight vehicle positioning data, etc.

[0037] Optionally, after data acquisition, the acquired multi-source data is time-synchronized and spatially calibrated. Filtering algorithms, such as Kalman filtering, are used to remove noise from lidar point clouds, motion blur from visual images, and false target signals from millimeter-wave radar, generating a standardized set of environmental data.

[0038] In one embodiment of this application, the output data of each sensor is timestamped based on the pulse per second (PPS) signal of the Beidou positioning module, and the time deviation of each sensor data is controlled within 1ms to achieve time synchronization.

[0039] In one embodiment of this application, a hand-eye calibration method is used to establish the transformation relationship between each sensor and the coordinate system of the flight vehicle. The transformation matrix between the lidar and the coordinate system is obtained through a checkerboard calibration board (accuracy 0.01mm). The intrinsic parameters (focal length, principal point coordinates) and extrinsic parameters (rotation matrix, translation vector) of the visual camera are solved by Zhang's calibration method. The installation angle (horizontal deflection angle, vertical deflection angle) of the millimeter-wave radar is corrected by on-site measurement to ensure that the multi-source data are aligned in the same spatial coordinate system and the spatial calibration error is less than 0.1m, so as to achieve spatial calibration.

[0040] In one embodiment of this application, a statistical filtering algorithm (with the number of neighborhood points set to 15 and the standard deviation multiple set to 2) is used to remove isolated noise points in the point cloud; for false point clouds generated by special material surfaces such as glass curtain walls and reflective metal surfaces, the image texture features collected by the visual camera are combined to remove point cloud areas with continuous grayscale value changes of less than 10, and retain the real obstacle point cloud.

[0041] In one embodiment of this application, Gaussian filtering (convolution kernel size 5×5, standard deviation 1.2) is used to remove Gaussian noise from the image; the Retinex algorithm is used to enhance the image contrast in low-light environments and improve the clarity of the outline of dynamic targets; for motion-blurred images, a motion estimation-based deblurring algorithm (10 iterations, blur kernel size 9×9) is used to restore image details.

[0042] In one embodiment of this application, a Kalman filter algorithm is used to smooth the target range and velocity data of the millimeter-wave radar. The state equation is set as a uniform motion model, and the observation equation is set as the radar measurement value. The process noise covariance matrix Q and the observation noise covariance matrix R are determined through field tests. Remove radar false alarm targets (targets that are not detected for 3 consecutive frames are marked as false alarms and deleted).

[0043] A multi-sensor combination scheme is adopted, and the accurate fusion of multi-source data is achieved through time synchronization and spatial calibration technology. Targeted filtering algorithms are designed to remove noise and interference signals from various sensors, solve the problem of insufficient detection accuracy of single sensors in complex environments, and improve the reliability and integrity of environmental data.

[0044] Step S2: Based on multiple types of data, construct a static environment model and a dynamic obstacle motion model; Optionally, based on the preprocessed lidar point cloud data, a region growing algorithm can be used to segment the ground and non-ground areas, perform cluster analysis on the non-ground areas, extract the three-dimensional contour features of static obstacles such as buildings, utility poles, and bridges, construct a low-altitude static environment model of the city, and combine the building coordinate data in GIS to correct the model to ensure the accuracy of the position and contour of static obstacles.

[0045] Optionally, by utilizing image information and speed and distance data of dynamic obstacles, the YOLO target detection algorithm is used to identify dynamic obstacles (birds, other flying vehicles, floating objects), extract the shape features, speed and heading angle information of dynamic targets, establish a dynamic target database, and realize real-time tracking of dynamic obstacles.

[0046] In one embodiment of this application, a static environment model and a dynamic obstacle database are constructed using a core urban area of ​​a city as the experimental scenario: Static environment modeling: Import 1:2000 urban GIS data for this area, extract coordinates and height information of static obstacles such as buildings, utility poles, bridges, and trees; combine point cloud data collected by LiDAR, use a region growing algorithm (growth threshold set to 0.3m) to segment the ground and non-ground areas, perform cluster analysis on the non-ground areas (cluster radius 0.5m), generate a 3D contour model of static obstacles, and correct the building height error in the GIS data (error controlled within ±0.5m). Dynamic obstacle database construction: Using image information and speed and distance data of dynamic obstacles, the YOLOv8 target detection algorithm (confidence threshold set to 0.7) is used to identify dynamic targets, which are classified and labeled as "birds", "drones", and "floating objects". Typical motion parameters of each type of dynamic target are recorded (birds: speed 0~15m / s, heading angle change frequency 0.5~2 times / second; drones: speed 0~8m / s, heading angle change frequency 0.1~0.5 times / second; floating objects: speed 0~3m / s, heading angle change random) to establish a dynamic obstacle motion feature database.

[0047] Step S3: Based on the dynamic obstacle motion model, predict the motion trajectory of the dynamic obstacle, and determine the obstacle avoidance priority of the dynamic obstacle based on the flight trajectory of the flight vehicle and the motion trajectory of the dynamic obstacle. Optionally, for dynamic obstacles, the trajectory is predicted based on their historical motion data, and the obstacle avoidance priority is determined by combining the real-time flight trajectory of the flight vehicle, so as to provide a basis for subsequent obstacle avoidance decisions.

[0048] Step S4: Based on the static environment model, the motion trajectory of the dynamic obstacle, and the obstacle avoidance priority of the dynamic obstacle, an obstacle avoidance path is generated through a path planning algorithm. Step S5: Based on the obstacle avoidance path, control the flight vehicle to avoid obstacles.

[0049] Optionally, the planned obstacle avoidance path is converted into control commands for the flight vehicle (pitch angle, roll angle, throttle), and the flight control system drives the actuators to achieve the obstacle avoidance maneuver. The specific steps are as follows: 1a. Path discretization: Discretize the obstacle avoidance path according to the time step. Discretized into a sequence of path points ( ; 2a. Control command calculation: Calculate the control commands (pitch angle) of the flight vehicle, taking each waypoint as the target position. Roll angle throttle size Among them, pitch angle Determined by the rate of change of elevation of the path points The roll angle φ is determined by the rate of change of the path points in the horizontal direction. The throttle size T is adjusted according to the flight speed deviation (if the actual speed is less than the target speed, the throttle is increased; otherwise, the throttle is decreased), and the output frequency of the control command is consistent with the receiving frequency of the flight control system (50Hz). 3a. Command issuance: Control commands are issued to the actuators (motors, servos) of the flight vehicle via the CAN bus, ensuring that the command transmission delay is ≤10ms and the actuator response delay is ≤20ms.

[0050] The obstacle avoidance method for low-altitude flight in complex environments provided in this application employs multiple types of sensors to collect various types of data about the environment surrounding the flight vehicle, including data on static and dynamic obstacles. This addresses the problem of insufficient detection accuracy of a single sensor in complex environments, improving the reliability and completeness of environmental data. Simultaneously, it constructs static environment models and dynamic obstacle motion models, predicts the motion trajectories of dynamic obstacles, and, combined with the real-time flight trajectory of the flight vehicle, prioritizes obstacle avoidance, providing a basis for subsequent obstacle avoidance decisions. This overcomes the shortcomings of existing technologies, such as insufficient accuracy in predicting the motion trajectory of dynamic obstacles, difficulty in handling sudden movements of dynamic obstacles in urban low-altitude environments, and the risk of collisions. Based on the static environment model, the motion trajectory of dynamic obstacles, and the obstacle avoidance priority of dynamic obstacles, an obstacle avoidance path is generated through a path planning algorithm, and the flight vehicle is controlled to avoid obstacles based on the path, thereby improving the safety and reliability of low-altitude flight.

[0051] In some embodiments, step S1 specifically includes: Step S11: Collect three-dimensional point cloud data of the environment around the flight vehicle using lidar; Step S12: Acquire color and depth image information of the environment surrounding the flight vehicle using a visual camera; Step S13: Collect speed and distance data of dynamic obstacles around the flight vehicle using millimeter-wave radar; Step S14: Collect the real-time position and attitude information of the flight vehicle through the BeiDou positioning module.

[0052] Figure 2 This is a schematic diagram of the data acquisition process provided in the embodiments of this application, such as... Figure 2 As shown, in one embodiment of this application, a multi-sensor system consisting of a lidar, a vision camera, a millimeter-wave radar, and a BeiDou positioning module is used for data acquisition. The parameters of each device are set as follows: LiDAR: A 16-line LiDAR is selected, with a horizontal field of view of 360°, a vertical field of view of -15°~90°, a ranging range of 0.5m~200m, a point cloud density of 120,000 points / second, and a data output frequency of 10Hz, used to acquire high-precision 3D environmental point cloud data. Visual camera: It adopts a binocular visual camera with a resolution of 1920×1080, a frame rate of 30fps, a field of view of 60°, and is equipped with an infrared supplementary light module (supplementary light distance 0~50m) to acquire color images and depth images, and is suitable for night and low light environments. Millimeter-wave radar: operating frequency band 24GHz, ranging range 0.1m~150m, speed range -30m / s~30m / s, angular resolution 5°, data output frequency 20Hz, used to capture speed and distance information of dynamic obstacles, and the anti-electromagnetic interference capability level reaches IP67. Beidou positioning module: positioning accuracy 1m (single point positioning), 0.1m (differential positioning), update frequency 10Hz, and outputs the attitude angles (roll angle, pitch angle, yaw angle) of the flight vehicle, with an attitude measurement accuracy of 0.1°, providing real-time position and attitude reference for the flight vehicle.

[0053] In some embodiments, predicting the trajectory of a dynamic obstacle based on a dynamic obstacle motion model in step S3 specifically includes: Based on the dynamic obstacle motion model, a linear prediction model is used to predict the short-term motion trajectory of a dynamic target moving at a constant speed, while a nonlinear prediction model is used to predict the medium- and long-term motion trajectory of a dynamic target moving at a non-uniform speed.

[0054] Optionally, for dynamic obstacles, a multi-model fusion prediction framework is constructed based on their historical motion data: a linear prediction model, such as a linear regression model, is used to predict the short-term trajectory of dynamic targets moving at a constant speed, such as drones flying at a constant speed; a nonlinear prediction model, such as a Long Short-Term Memory (LSTM) network, is used to predict the medium- and long-term trajectory of dynamic targets moving at varying speeds and directions, such as birds.

[0055] In one embodiment of this application, the linear prediction model is applicable to a target moving at a constant speed. The model input is the position and velocity data of the target in the past 5 frames (time interval 0.1s). The target's motion trajectory is fitted by the least squares method to predict its position in the next 10s (prediction step size 0.1s).

[0056] In one embodiment of this application, the LSTM nonlinear prediction model is suitable for targets moving at varying speeds and directions. The model structure is a 3-layer LSTM (input layer dimension 6, hidden layer number of neurons 64, output layer dimension 3), the activation function is ReLU, the optimizer is Adam (learning rate 0.001, batch size 32), the training dataset is 1000 sets of bird motion data (including sequences of position, speed, and heading angle changing over time), the model is trained for 500 iterations, and the loss function (mean squared error) converges to below 0.01; during prediction, the target's motion data of the past 10 frames are input, and the position prediction result within the next 30 seconds is output (prediction step size 0.1s).

[0057] Optionally, the velocity change rate of the dynamic target over the past 5 frames can be calculated (velocity change rate = |current velocity - previous frame velocity| / time interval). If the velocity change rate is ≤0.5m / s², a linear prediction model is selected; if the velocity change rate is >0.5m / s², an LSTM nonlinear prediction model is selected to achieve adaptive prediction of the dynamic target trajectory.

[0058] In some embodiments, determining the obstacle avoidance priority of the dynamic obstacle based on the flight trajectory of the flight vehicle and the motion trajectory of the dynamic obstacle in step S3 specifically includes: Based on the flight trajectory of the flight vehicle and the motion trajectory of the dynamic obstacle, calculate the minimum encounter distance and encounter time between the dynamic obstacle and the flight vehicle; A risk assessment model is established based on the minimum encounter distance and encounter time between dynamic obstacles and flight vehicles, and the risk level of dynamic obstacles is determined based on the risk assessment model. Based on the risk level of dynamic obstacles, the obstacle avoidance priority of dynamic obstacles is determined.

[0059] Specifically, based on the predicted trajectory of the dynamic target and real-time data of the flight vehicle, the steps for performing a risk assessment are as follows: 1b. Obtain the real-time location of the flight vehicle. ,speed The flight path (planned position within the next 30 seconds) and the predicted trajectory of a dynamic target (position sequence within the next 30 seconds) ; 2b. For each time step t, calculate the spatial distance d between the flight vehicle and the dynamic target d = √[(x0 + v0x×t - xt)² + (y0 + v0y×t - yt)² + (z0 + v0z×t - zt)²], and record the minimum encounter distance d_min and the corresponding encounter time t_min; 3b. Set risk level thresholds (low risk: d_min≥50m; medium risk: 20m≤d_min<50m; high risk: d_min<20m), and determine the risk level based on the range of d_min; at the same time, set the response priority based on the encounter time t_min (t_min<5s: emergency response; 5s≤t_min<10s: normal response; t_min≥10s: delayed response), and finally output the evaluation result of "risk level - response priority" to provide a basis for obstacle avoidance decision-making.

[0060] Obstacle avoidance priorities are assigned based on risk level (low risk, medium risk, high risk), with higher risk levels having higher priority, thus providing a basis for subsequent obstacle avoidance decisions.

[0061] In some embodiments, step S4 specifically includes: Step S41, based on the static environment model, the motion trajectory of dynamic obstacles, and the obstacle avoidance priority of dynamic obstacles, through the improved A The path planning algorithm generates the initial obstacle avoidance path; Step S42: Based on the blind zone environmental data in the multi-class data, the initial obstacle avoidance path is corrected to obtain the obstacle avoidance path.

[0062] Based on the static environment model, dynamic obstacle prediction trajectory, and risk assessment results, an improved A method is adopted. The path planning algorithm generates an initial obstacle avoidance path, and the specific steps are as follows: 1c. Map Gridding: The flight area is divided into a 1m×1m×1m three-dimensional grid, and each grid is marked as "passable" (no static obstacles and no high-risk dynamic targets) or "impassable" (static obstacles or high-risk dynamic targets exist). 2c. Cost function design: Improved A The algorithm's cost function is f(n) = g(n) + ω × h(n), where g(n) is the actual cost from the starting point to the current node n (Euclidean distance), and h(n) is the heuristic function from the current node n to the destination (calculated using Manhattan distance). (the coordinates of the endpoint) Weighting coefficients (static dense region) Dynamic high-risk areas (Adaptively adjusted based on risk assessment results) 3c. Path Search: Initialize the OPEN list (stores nodes to be explored) and the CLOSE list (stores explored nodes), and add the starting point to the OPEN list. Each time, select the node with the smallest f(n) from the OPEN list as the current node, expand its 6 adjacent nodes (front, back, left, right, up, down). If the adjacent node is "passable" and is not in the CLOSE list, calculate its f(n) value and add it to the OPEN list. Repeat the above process until the endpoint is added to the CLOSE list. Backtrack the CLOSE list to generate the initial obstacle avoidance path. The path search time is ≤1s, which meets the real-time requirements.

[0063] To address urban low-altitude "visual blind spots" (building gaps, shadowed areas), penetration data from both lidar and millimeter-wave radar are integrated to supplement environmental information in these blind spots and correct the initial path. The specific steps are as follows: 1d. Blind spot identification: Identify visual blind spots by analyzing the grayscale distribution of visual camera images (grayscale value < 50 and continuous area > 100 pixels); combine the density distribution of LiDAR point clouds (areas with point cloud density < 10 points / m³) to confirm the blind spot range, and record the coordinates and size of the blind spot. 2d. Supplementing blind zone information: Using millimeter-wave radar penetration data (millimeter waves can penetrate obstacles such as glass and fog), the location information (distance and angle) of obstacles inside the blind zone is obtained; combined with the building structure data (building spacing and height) in the static environment model, the passable space inside the blind zone is inferred. 3d. Path correction: For road segments that cross blind spots in the initial path, adjust the coordinates of path nodes based on the supplemented blind spot obstacle information (avoid obstacles inside the blind spot and ensure that the distance between the path and obstacles is ≥2m) to generate a corrected obstacle avoidance path. The corrected path meets the blind spot passage safety distance requirements.

[0064] In some embodiments, step S5 specifically includes: Based on the obstacle avoidance path, the flight vehicle is controlled to avoid obstacles. In areas with dense static obstacles, obstacle avoidance paths with high smoothness and small turning radius are selected first. In areas with high risk of dynamic obstacles, a rapid obstacle avoidance strategy is adopted to adjust the flight trajectory.

[0065] Optionally, in areas with dense static obstacles (building spacing <20m): prioritize planning strategies with high path smoothness, use B-spline curves to smooth the initial path (the number of control points is 1 / 3 of the number of path nodes, and the curve order is 3), so that the turning radius of the path is ≥5m (meeting the maneuverability requirements of the UAV), and reduce the number of turns of the flight vehicle (≤2 turns per 100m path).

[0066] Optionally, in high-risk areas of dynamic obstacles (d_min<20m and t_min<5s): execute an emergency obstacle avoidance strategy, calculate the emergency avoidance direction (the direction away from the predicted trajectory of the dynamic target), adjust the pitch or roll angle of the flight vehicle (adjustment range ≤10° / s), achieve emergency climb (climb speed ≥2m / s) or lateral avoidance (lateral speed ≥1m / s), and ensure that the distance to the dynamic target is ≥20m before t_min.

[0067] In some embodiments, the method further includes: If there are multiple flight vehicles, the obstacle avoidance paths of each flight vehicle are optimized collaboratively to avoid path intersections.

[0068] Optionally, in a multi-carrier collaborative area (where ≥2 other flying vehicles exist simultaneously): the obstacle avoidance intentions (path planning results, speed, heading angle) of other flying vehicles are obtained through vehicle-to-everything (V2X) technology. A multi-carrier collaborative optimization objective function is established (minimizing the path length and number of conflicts of all vehicles). A distributed optimization algorithm (alternating direction multiplier method, 20 iterations) is used to adjust the paths of each vehicle to avoid path intersection conflicts (minimum distance between vehicles ≥10m), thereby achieving collaborative obstacle avoidance.

[0069] In some embodiments, the method further includes: If the position deviation of the obstacle exceeds the first preset threshold or the flight trajectory deviation of the flight vehicle exceeds the second preset threshold during the obstacle avoidance process, a new obstacle avoidance path will be generated.

[0070] During obstacle avoidance, sensor feedback data is collected in real time to monitor the deviation between the obstacle position and the flight trajectory. If the deviation exceeds the preset threshold, the dynamic obstacle detection, trajectory prediction and path planning steps are re-executed to iteratively adjust the obstacle avoidance command and ensure that the flight vehicle is always in a safe flight state.

[0071] In one embodiment of this application, during obstacle avoidance, real-time adjustments are performed based on sensor feedback data, and the specific steps are as follows: 1e. Deviation Monitoring: Real-time acquisition of the actual position of the flight vehicle (obtained via the BeiDou positioning module) and the target position on the obstacle avoidance path, and calculation of position deviation. ; 2e. Adjust the trigger: If Δd>0.5m, the trigger path is replanned; at the same time, the deviation between the actual position of the dynamic obstacle and the predicted trajectory is detected. If the deviation>2m, the dynamic target detection, trajectory prediction and risk assessment steps are re-executed. 3e. Iterative optimization: Based on the updated environmental information (obstacle positions and trajectories after deviation correction), the obstacle avoidance path planning algorithm is rerun to generate new obstacle avoidance paths and control commands, replacing the original commands; the iteration cycle is set to 0.5s to ensure that the flight vehicle always tracks the corrected obstacle avoidance path, and the position deviation is controlled within 0.5m, realizing real-time obstacle avoidance in dynamic environments.

[0072] The following describes the urban low-altitude flight obstacle avoidance device in complex environments provided in this application. The urban low-altitude flight obstacle avoidance device in complex environments described below can be referred to in correspondence with the urban low-altitude flight obstacle avoidance method in complex environments described above.

[0073] Figure 3 This is a schematic diagram of the structure of an urban low-altitude flight obstacle avoidance device in a complex environment, as provided in an embodiment of this application. Figure 3 As shown, the device 300 includes: The data acquisition module 310 is used to acquire various types of data about the environment surrounding the flight vehicle through multiple types of sensors. Module 320 is used to build static environment models and dynamic obstacle motion models based on multiple types of data; The prediction module 330 is used to predict the motion trajectory of dynamic obstacles based on the dynamic obstacle motion model, and to determine the obstacle avoidance priority of dynamic obstacles based on the flight trajectory of the flight vehicle and the motion trajectory of the dynamic obstacles. The generation module 340 is used to generate an obstacle avoidance path based on a static environment model, the motion trajectory of dynamic obstacles, and the obstacle avoidance priority of dynamic obstacles, using a path planning algorithm. The control module 350 is used to control the flight vehicle to avoid obstacles based on the obstacle avoidance path.

[0074] It should be understood that the above-described device is used to execute the methods in the above embodiments. The implementation principle and technical effect of the corresponding program modules in the device are similar to those described in the above methods. The working process of the device can be referred to the corresponding process in the above methods, and will not be repeated here.

[0075] Based on the methods in the above embodiments, Figure 4 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 4As shown in the illustration, this application provides an electronic device that may include a processor 410, a communication interface 420, a memory 430, and a communication bus 440. The processor 410, communication interface 420, and memory 430 communicate with each other via the communication bus 440. The processor 410 can call logical instructions stored in the memory 430 to execute the urban low-altitude flight obstacle avoidance method in complex environments described in the above embodiment.

[0076] Furthermore, the logical instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the urban low-altitude flight obstacle avoidance method in complex environments described in the various embodiments of this application.

[0077] Based on the methods in the above embodiments, this application provides a computer-readable storage medium storing a computer program. When the computer program runs on a processor, it causes the processor to execute the urban low-altitude flight obstacle avoidance method in complex environments as described in the above embodiments.

[0078] Based on the methods in the above embodiments, this application provides a computer program product that, when running on a processor, causes the processor to execute the urban low-altitude flight obstacle avoidance method in complex environments as described in the above embodiments.

[0079] It is understood that the processor in the embodiments of this application can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. A general-purpose processor can be a microprocessor or any conventional processor.

[0080] The method steps in this application embodiment can be implemented in hardware or by a processor executing software instructions. The software instructions can consist of corresponding software modules, which can be stored in random access memory (RAM), flash memory, read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), registers, hard disks, portable hard disks, CD-ROMs, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor, enabling the processor to read information from and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and the storage medium can reside in an ASIC.

[0081] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially as a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted through the computer-readable storage medium. The computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0082] It is understood that the various numerical designations used in the embodiments of this application are merely for the convenience of description and are not intended to limit the scope of the embodiments of this application.

[0083] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for obstacle avoidance in urban low-altitude flight under complex conditions, characterized in that, include: Multiple types of data about the environment surrounding the flight vehicle are collected using various sensors. Based on the aforementioned multiple types of data, a static environment model and a dynamic obstacle motion model are constructed. Based on the dynamic obstacle motion model, the motion trajectory of the dynamic obstacle is predicted, and the obstacle avoidance priority of the dynamic obstacle is determined based on the flight trajectory of the flight vehicle and the motion trajectory of the dynamic obstacle. Based on the static environment model, the motion trajectory of the dynamic obstacle, and the obstacle avoidance priority of the dynamic obstacle, an obstacle avoidance path is generated through a path planning algorithm. Based on the obstacle avoidance path, the flight vehicle is controlled to avoid obstacles.

2. The obstacle avoidance method for low-altitude flight in complex environments according to claim 1, characterized in that, The collection of various types of data about the environment surrounding the flight vehicle through multiple sensors includes: Three-dimensional point cloud data of the environment surrounding the flight vehicle are collected using lidar. The visual camera acquires color and depth images of the environment surrounding the flight vehicle. The speed and distance data of dynamic obstacles around the flight vehicle are collected using millimeter-wave radar. The real-time position and attitude information of the flight vehicle is collected through the BeiDou positioning module.

3. The obstacle avoidance method for low-altitude flight in complex environments according to claim 1, characterized in that, The prediction of the motion trajectory of the dynamic obstacle based on the dynamic obstacle motion model includes: Based on the dynamic obstacle motion model, a linear prediction model is used to predict the short-term motion trajectory of a dynamic target moving at a constant speed, and a nonlinear prediction model is used to predict the medium- and long-term motion trajectory of a dynamic target moving at a non-uniform speed.

4. The obstacle avoidance method for low-altitude flight in complex environments according to claim 1, characterized in that, The determination of obstacle avoidance priority for dynamic obstacles based on the flight trajectory of the flight vehicle and the motion trajectory of the dynamic obstacle includes: Based on the flight trajectory of the flight vehicle and the motion trajectory of the dynamic obstacle, calculate the minimum encounter distance and encounter time between the dynamic obstacle and the flight vehicle; Based on the minimum encounter distance and encounter time between the dynamic obstacle and the flight vehicle, a risk assessment model is established, and the risk level of the dynamic obstacle is determined based on the risk assessment model. Based on the risk level of dynamic obstacles, the obstacle avoidance priority of dynamic obstacles is determined.

5. The obstacle avoidance method for low-altitude flight in complex environments according to claim 1, characterized in that, The step of generating an obstacle avoidance path based on the static environment model, the motion trajectory of the dynamic obstacle, and the obstacle avoidance priority of the dynamic obstacle through a path planning algorithm includes: Based on the static environment model, the motion trajectory of the dynamic obstacle, and the obstacle avoidance priority of the dynamic obstacle, the improved A... The path planning algorithm generates the initial obstacle avoidance path; Based on the blind zone environmental data in the multiple types of data, the initial obstacle avoidance path is corrected to obtain the obstacle avoidance path.

6. The obstacle avoidance method for low-altitude flight in complex environments according to claim 1, characterized in that, The control of the flight vehicle to avoid obstacles based on the obstacle avoidance path includes: Based on the obstacle avoidance path, the flight vehicle is controlled to avoid obstacles. In areas with dense static obstacles, obstacle avoidance paths with high smoothness and small turning radius are selected first. In areas with high risk of dynamic obstacles, a rapid obstacle avoidance strategy is adopted to adjust the flight trajectory.

7. The obstacle avoidance method for low-altitude urban flight in complex environments according to claim 1, characterized in that, The method further includes: If there are multiple flight vehicles, the obstacle avoidance paths of each flight vehicle are optimized collaboratively to avoid path intersections.

8. The obstacle avoidance method for low-altitude flight in complex environments according to claim 1, characterized in that, The method further includes: If the position deviation of the obstacle exceeds the first preset threshold or the flight trajectory deviation of the flight vehicle exceeds the second preset threshold during the obstacle avoidance process, a new obstacle avoidance path will be generated.

9. A low-altitude obstacle avoidance device for urban flight in complex environments, characterized in that, include: The data acquisition module is used to collect various types of data about the environment surrounding the flight vehicle through multiple sensors. A construction module is used to construct a static environment model and a dynamic obstacle motion model based on the aforementioned multiple types of data; The prediction module is used to predict the trajectory of the dynamic obstacle based on the dynamic obstacle motion model, and to determine the obstacle avoidance priority of the dynamic obstacle based on the flight trajectory of the flight vehicle and the trajectory of the dynamic obstacle. The generation module is used to generate an obstacle avoidance path based on the static environment model, the motion trajectory of the dynamic obstacle, and the obstacle avoidance priority of the dynamic obstacle, using a path planning algorithm. The control module is used to control the flight vehicle to avoid obstacles based on the obstacle avoidance path.

10. An electronic device, characterized in that, include: At least one memory for storing computer programs; At least one processor is configured to execute a program stored in the memory, wherein when the program stored in the memory is executed, the processor is configured to perform an urban low-altitude flight obstacle avoidance method in a complex environment as described in any one of claims 1-8.

Images