Helicopter landing auxiliary system based on fusion of laser radar and inertial navigation system

Abstract

A system for assisting helicopter landing based on laser radar and inertial navigation comprises a data and control interface module used for receiving and processing laser radar point cloud data and inertial navigation data, sending the data to a terrain evaluation module after coordinate conversion is completed, analyzing a user control instruction and feeding back a terrain evaluation result; the terrain evaluation module is used for screening out an optimal landing area through grid division and height difference and gradient calculation according to the received coordinate data and the control instruction; and the image display module is used for displaying visual information of the laser radar point cloud and the terrain evaluation result in real time. According to the method, the three-dimensional terrain model under the UTM coordinate system can be generated in real time, the sliding window traversal and multi-weight scoring algorithm is combined, the optimal landing area is dynamically screened, the problems of low environmental perception precision, decision lag and insufficient adaptability in the prior art are effectively solved, and the landing safety and operation efficiency of a helicopter in a complex environment are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the fields of avionics systems and autonomous navigation technologies, and particularly relates to a helicopter landing assistance system based on the fusion of lidar (LiDAR) and inertial navigation system (INS), which is particularly suitable for real-time terrain assessment and safe landing decision-making of helicopters (including manned and unmanned models) in complex terrain environments. Background Art

[0002] Due to its vertical takeoff and landing capabilities, helicopters are widely used in fields such as rescue, military, and transportation in remote areas. However, in unstructured environments (such as mountains, jungles, or disaster sites), the safe landing of helicopters highly depends on the visual judgment of operators, which has the following limitations:

[0003] 1. Limited environmental perception: Visual judgment is easily affected by light, weather, and visual range, making it difficult to accurately evaluate the flatness, slope, and obstacle distribution of long-distance or complex terrains.

[0004] 2. Lack of real-time performance: Manual terrain analysis takes a long time and cannot meet the requirements of rapid decision-making in emergency tasks.

[0005] 3. Low degree of automation: Existing unmanned aircraft landing systems mostly rely on GPS or visual sensors. GPS is vulnerable to signal interference in environments such as canyons and dense forests, while the performance of visual sensors significantly degrades under low-light or foggy conditions.

[0006] In recent years, lidar technology has been gradually applied to the fields of autonomous driving and unmanned aircraft navigation due to its high-precision three-dimensional modeling ability. However, existing solutions do not fully integrate lidar point clouds and high-frequency attitude data of inertial navigation, resulting in limited real-time performance and accuracy of terrain assessment. Most systems only focus on obstacle detection or slope calculation, lacking a multi-index scoring mechanism that comprehensively considers height difference, slope, and regional continuity. Moreover, existing methods do not optimize the assessment algorithm for the takeoff and landing characteristics of helicopters (such as the influence of rotor downwash), making it difficult to meet the safety requirements in complex environments. Summary of the Invention

[0007] To overcome the deficiencies of the above-mentioned prior art, the present invention proposes a helicopter landing assistance system based on the fusion of lidar and inertial navigation through the deep fusion of lidar and inertial navigation. It can generate a three-dimensional terrain model in the UTM coordinate system in real time, and combine a sliding window traversal and a multi-weight scoring algorithm to dynamically screen the optimal landing area, effectively solving the problems of low environmental perception accuracy, decision-making lag, and insufficient adaptability in the prior art, and significantly improving the landing safety and operation efficiency of helicopters in complex environments.

[0008] The technical solution of the present invention is as follows:

[0009] A method for calibrating the reference plane of a lidar for calibrating an airborne pod of an unmanned aerial vehicle

[0010] A helicopter landing assistance system based on the fusion of a lidar and an inertial navigation system, characterized by comprising:

[0011] A data and control interface module, configured to receive lidar point cloud data and inertial navigation data, generate UTM three-dimensional coordinate data after completing multi-level coordinate transformation, and parse control commands input by a user and a terrain evaluation result fed back;

[0012] A terrain evaluation module, configured to divide grids based on the UTM three-dimensional coordinate data and control commands, calculate the elevation difference Δh and slope slp of each grid, and screen for the optimal landing grid by combining sliding window traversal and a multi-weight scoring algorithm;

[0013] An image display module, configured to visually display in real time the three-dimensional terrain point cloud in the UTM coordinate system, and mark the position and evaluation information of the landable area.

[0014] The coordinate transformation process of the data and control interface module includes:

[0015] Sequentially transform the lidar point cloud data from the lidar coordinate system to the NED coordinate system, the ECEF coordinate system, the WGS84 coordinate system, and the UTM three-dimensional coordinate system;

[0016] The Z-axis direction and value of the UTM three-dimensional coordinates are consistent with the height of the WGS84 coordinate system.

[0017] The terrain evaluation module performs the following steps:

[0018] (a) Centering on the pre-landing point, divide a square evaluation area with a side length of d to generate grids with a side length of d;

[0019] (b) According to the UTM coordinate point cloud data, calculate the elevation difference Δh and slope slp of each grid, and determine the grid level through a threshold;

[0020] (c) Adopt sliding window traversal to evaluate the area, and screen for continuous grids that meet the elevation difference threshold Hmax and slope threshold Smax;

[0021] (d) Determine the optimal landing grid, and the formula is as follows:

[0022] Score = W slop *(1 - slop / S max ) + W height *(1 - Δh / H max )

[0023] In the formula, W slop 、W heightThey are the weight of the slope and the weight of the elevation difference respectively.

[0024] The calculation method of the slope slp is as follows:

[0025] Perform least squares plane fitting on the point cloud coordinates within the grid to obtain the normal vector of the reference plane;

[0026] Calculate the slope value through the angle between the normal vector and the vertical direction.

[0027] The size of the sliding window is a D×D grid, where D = nd, n is a positive integer, and the sliding step of the window is the side length of a single grid.

[0028] The weight coefficients in the scoring formula satisfy W slop +W height = 1, and W slop 、W height are preset values.

[0029] The detection distance of the lidar ≥ 500 meters, and the response time of the terrain evaluation module ≤ 1 second.

[0030] The image display module marks the terrain elevation through color mapping, marks the "landable grid" with a green polygon, and marks the "hazardous grid" with a red area.

[0031] The system is applicable to manned and unmanned helicopters and does not rely on GPS signals.

[0032] The inertial navigation system and the lidar share the same coordinate system reference through a rigid connection to eliminate relative displacement errors.

[0033] Compared with the prior art, the beneficial effects of the present invention are

[0034] 1. High-precision terrain modeling: By fusing lidar and inertial navigation data, the limitations of a single sensor are eliminated, the terrain resolution reaches the centimeter level, and the detection distance exceeds 500 meters.

[0035] 2. Real-time decision-making ability: The fully automated processing flow can complete terrain scanning, evaluation, and result feedback within 1 second, significantly shortening the decision-making time in emergency tasks.

[0036] 3. Multidimensional evaluation and optimization: Comprehensive indicators such as elevation difference, slope, and regional continuity are integrated to avoid missed detections or misjudgments caused by a single parameter in traditional methods and improve the adaptability to complex environments.

[0037] 4. Wide applicability: It supports manned / unmanned helicopters, is compatible with various takeoff and landing scenarios (mountains, jungles, disaster sites, etc.), and is not restricted by GPS signals or lighting conditions. Description of the Drawings

[0038] Figure 1 It is a schematic diagram of a helicopter landing assistance system based on the fusion of lidar and inertial navigation system;

[0039] Figure 2 It is a schematic diagram of grid division within the landing site area;

[0040] Figure 3 It is a schematic diagram of the terrain assessment display image. Specific implementation manners

[0041] The technical solutions of the present invention will be described in detail below in conjunction with the accompanying drawings and embodiments, but the protection scope of the present invention should not be limited thereby.

[0042] Please first refer to Figure 1 , Figure 1 which is a schematic diagram of a helicopter landing assistance system based on the fusion of lidar and inertial navigation system. As shown in the figure, a system for assisting a helicopter to land based on lidar and inertial navigation includes a data and control interface module, a terrain assessment module, and an image display module.

[0043] The data and control interface module is responsible for receiving and processing lidar data and inertial navigation data, sending the processed data to the terrain assessment module, and sending the received user control instructions to the terrain assessment module, and sending the feedback information sent by the terrain assessment module to the user. These three modules work together to receive and process data from LiDAR and INS, conduct terrain assessment, and present the results to the user in a visual manner.

[0044] The data and control interface module has data processing functions, control instruction processing functions, and information feedback functions.

[0045] Among them, the data processing function: receive LiDAR point cloud data and INS data, and after converting the LiDAR point cloud data from the LiDAR coordinate system to the UTM three-dimensional coordinates through the Euler matrix, send it to the terrain assessment module. The specific processing steps are as follows:

[0046] Step D01: Receive lidar point cloud data and inertial navigation data;

[0047] Step D02: Transform the lidar point cloud data from the lidar coordinate system to NED coordinate data through the Euler matrix transformation;

[0048] Step D03: Transform from NED coordinate data to ECEF coordinate data;

[0049] Step D04: Transform from ECEF coordinate data to WGS84 coordinate data;

[0050] Step D05: Transform from WGS84 coordinate data to UTM three-dimensional coordinate data. For the added Z-axis in the UTM three-dimensional coordinate, its direction and value are consistent with the altitude in the WGS84 coordinate.

[0051] Step D06: Send the UTM three-dimensional coordinate data to the terrain evaluation module.

[0052] Control instruction processing function: Receive and parse user control instructions, such as the landing site range, pre-landing point location, grid size, sliding window size, etc., and send them to the terrain evaluation module.

[0053] Information feedback function: Receive the feedback information from the terrain evaluation module and send it to the user.

[0054] The terrain evaluation module is responsible for performing terrain evaluation on the received lidar point cloud information according to user requirements, using terrain evaluation algorithms to calculate the area most suitable for helicopter landing, and sending the evaluation results to the data and control interface module and the image display module to visually display the location information of the most suitable landing area. The specific steps are as follows:

[0055] Step E01: Receive UTM coordinate point cloud and control instruction information from the data and control interface module, and parse parameters such as the landing site range, pre-landing point location information E0, grid size d within the evaluation area, grid height difference threshold Hmax, slope threshold Smax, sliding window size D (D = nd, where n is a positive integer), etc. in the control instruction;

[0056] Step E02: Take the pre-landing point as the evaluation center E0, and divide a square evaluation area with the east-west and north-south sides. This area is further divided into several square grid areas with a side length of grid size d. The grid level of each grid is initialized to "blank level", and the point cloud count is initialized to 0;

[0057] Step E03: Fill the UTM coordinate point cloud into different grids according to the x and y coordinates, calculate the height difference Δh and slope slp between the highest and lowest points of elevation in each grid, update the point cloud count within the grid, and update the slope fitting matrix M according to the point cloud coordinates. The M matrix is

[0058]

[0059] M x = M x + x, M y = M y + y, M z = M z + z

[0060] M xx = M xx + x2 , M yy = M yy + y 2 , M zz = M zz + z 2

[0061] M xy = M xy + x·y, M xz = M xz + x·z, M yz = M yz + y·z

[0062] where (x, y, z) are the point cloud coordinates;

[0063] Step E04: Determine the grid level LEVEL. When Δh > H max , it is determined as a "dangerous grid", otherwise it is determined as a "landable grid";

[0064] Step E05: After all UTM point clouds are filled, starting from the upper left corner of the evaluation area, set a sliding window. The sliding window fixedly contains D rows and D columns of grids. Evaluate the grids under the sliding window. The evaluation method is as follows:

[0065] First, the levels of all grids under the sliding window are not "dangerous grids". Otherwise, determine the level of the grid at the center of the sliding window as an "unlandable grid", and move one grid to the right or down until a sliding window where the levels of all grids are not "dangerous grids" is found. Then, evaluate the terrain of the sliding window, including height difference, oblique distance, height, and slope. Among them, the slope evaluation algorithm is as follows:

[0066]

[0067] where, is the normal vector of the reference plane, c is the number of point clouds, and slp is the slope evaluation value. If slp < S max , the level of the grid at the center of the sliding window is determined as a "landable grid", otherwise it is an "unlandable grid".

[0068] Step E06: Compare all grids with the level of "landable grid", select the grid with the highest comparison score, and send the information of this grid to the data and control interface module and the image display module. The grid comparison method is as follows:

[0069] Score = W slop *(1 - slop / S max ) + W height *(1 - Δh / H max ),

[0070] Among them, W slop and W height are the weights of slope and height difference respectively.

[0071] The image display module displays the lidar point cloud image in real time and marks the area information of terrain assessment at the same time.

[0072] Embodiment

[0073] Hardware Deployment and Integration

[0074] Lidar selection and installation:

[0075] A high-precision semi-solid lidar is adopted, with a horizontal field of view of 360°, a vertical field of view of 30°, a detection distance of ≥ 300 meters, and a point cloud density of ≥ 100,000 points / second. The lidar is installed at the bottom of the helicopter fuselage and fixed through a shock-absorbing bracket to ensure that the scanning plane is parallel to the ground, and the scanning range covers an area with a diameter of 500 meters directly below the helicopter.

[0076] Inertial navigation system configuration:

[0077] An integrated high-precision MEMS inertial navigation system (such as Honeywell HG4930) is included, which contains three-axis gyroscopes, accelerometers and magnetometers, with a data output frequency of ≥ 100Hz, an attitude angle accuracy of ≤ 0.1°, and a position accuracy of ≤ 1 meter (short-term). The inertial navigation and the lidar share the same coordinate system reference through a rigid connection to eliminate relative displacement errors.

[0078] Data processing unit:

[0079] An embedded computer (such as NVIDIA Jetson AGX Xavier) is adopted, equipped with a Linux real-time operating system and CUDA cores to accelerate point cloud processing and coordinate transformation calculations.

[0080] Software Implementation and Parameter Configuration

[0081] Data and control interface module:

[0082] 1. Coordinate transformation process:

[0083] Lidar coordinate system → NED coordinate system: realized through the Euler angle (pitch angle θ, roll angle yaw angle ψ) transformation matrix;

[0084] NED coordinate system → ECEF coordinate system: translation and rotation transformations are carried out based on the current position of the helicopter (longitude λ, latitude φ, altitude h).

[0085] ECEF coordinate system → WGS84 coordinate system: transformed using the standard ellipsoid model;

[0086] WGS84 coordinate system → UTM three-dimensional coordinate system: The plane coordinates (x, y) are converted through the UTM projection algorithm, and the Z-axis height directly inherits the WGS84 elevation value.

[0087] 2. Control instruction parsing: The user inputs parameters through the touch screen, including a landing site radius of 300 meters, a grid size d = 2 meters, a sliding window size D = 30 meters, a height difference threshold Hmax = 0.5 meters, a slope threshold Smax = 2°, and weight coefficients W slop = 0.6, W height = 0.4.

[0088] Terrain evaluation module:

[0089] 1. Grid division and filling:

[0090] With the pre-landing point E0 as the center, a 300×300-meter square evaluation area (corresponding to the UTM coordinate range) is generated and divided into 2×2-meter grids. Each grid records the number of point clouds, the maximum / minimum elevation values, and the slope fitting matrix.

[0091] 2. Slope calculation:

[0092] 2.1 Perform least squares plane fitting on all point cloud coordinates (xi, yi, zi) within the grid to obtain the reference plane equation ax + by + cz + d = 0;

[0093] 2.2 Calculate the normal vector of the reference plane;

[0094] 2.3 Calculate the slope value slp.

[0095] 3. Sliding window traversal and scoring:

[0096] The sliding window size is 15×15 grids (D = 30 meters), and it slides step by step for each grid. If Δh ≤ 0.5 meters and slp ≤ 2° for all grids within the sliding window, then calculate the score of the central grid Score = 0.6×(1 - slop / 2) + 0.4×(1 - Δh / 0.5), and select the grid with the highest score as the optimal landing point.

[0097] Image display module:

[0098] In the cockpit display or the UAV ground station, with the UTM coordinate system as the reference, the three-dimensional point cloud terrain is rendered in real time (color mapping the elevation value), and the "landable grids" are marked with green polygons, and the "dangerous grids" (Δh > 0.5 meters or slp > 2°) are marked with red areas. At the same time, the coordinates (x, y, z) of the optimal landing point and the comprehensive score are displayed.

[0099] 3. Implementation verification and test results

[0100] Simulation environment test:

[0101] Simulate mountain, jungle and urban ruins scenes in the digital terrain model (DTM). The average system response time is 0.8 seconds, and the terrain evaluation accuracy rate ≥ 95% (compared with the results of manual annotation).

[0102] Actual flight test:

[0103] Deploy this system on a certain type of unmanned helicopter (payload 50 kg), and perform 10 autonomous landing tasks in hilly areas. Successfully screen out flat areas (slope ≤ 0.9°, height difference ≤ 0.3 m), and the average deviation of the landing position < 0.5 m.

[0104] Extreme condition test:

[0105] In a canyon environment with GPS signal shielding, the system relies on pure lidar and inertial navigation data and can still stably output terrain evaluation results, verifying the anti-interference ability.

[0106] Through the description of the above embodiments, those skilled in the art can clearly understand the specific implementation manners of a method for calibrating the lidar reference plane applicable to an unmanned aerial vehicle airborne pod provided by the present invention, and can make appropriate modifications and changes according to needs. These modifications and changes should all fall within the protection scope of the present invention.

Claims

1. A helicopter landing assistance system based on the fusion of lidar and inertial navigation system, characterized in that, Including: A data and control interface module, which is used to receive lidar point cloud data and inertial navigation data, generate UTM three-dimensional coordinate data after completing multi-level coordinate transformation, and parse the control instructions input by the user and the feedback terrain evaluation results; A terrain evaluation module, which is used to divide grids based on the UTM three-dimensional coordinate data and control instructions, calculate the height difference Δh and slope slp of each grid, and screen the optimal landing grid by combining sliding window traversal and multi-weight scoring algorithm; An image display module, which is used to visually display the three-dimensional terrain point cloud in the UTM coordinate system in real time, and mark the position and evaluation information of the landable area.

2. The system according to claim 1, characterized in that, The coordinate transformation process of the data and control interface module includes: Sequentially transform the lidar point cloud data from the lidar coordinate system to the NED coordinate system, ECEF coordinate system, WGS84 coordinate system, and UTM three-dimensional coordinate system; The direction and value of the Z axis of the UTM three-dimensional coordinate are consistent with the height of the WGS84 coordinate system.

3. The system according to claim 1, wherein The terrain evaluation module performs the following steps: (a) Taking the pre-landing point as the center, divide a square evaluation area with a side length of d to generate grids with a side length of d; (b) According to the UTM coordinate point cloud data, calculate the height difference Δh and slope slp of each grid, and determine the grid level through a threshold; (c) Use a sliding window to traverse the evaluation area to screen continuous grids that meet the height difference threshold Hmax and slope threshold Smax; (d) Determine the optimal landing grid, and the formula is as follows: Score=W slop *(1 - slop / S max ) + W height *(1 - Δh / H max ) Where, W slop and W height are the weights of slope and elevation difference respectively.

4. The system according to claim 3, wherein The calculation method of the slope slp is: Perform least squares plane fitting on the point cloud coordinates in the grid to obtain the normal vector of the reference plane; Calculate the slope value through the angle between the normal vector and the vertical direction.

5. The system according to claim 3, wherein The size of the sliding window is a D×D grid, where D = nd and n is a positive integer, and the sliding window moving step is the side length of a single grid.

6. The system according to claim 3, characterized in that The weight coefficients in the scoring formula satisfy W slop +W height = 1, and W slop , W height are preset values.

7. The system according to claim 1, wherein The detection distance of the lidar ≥ 500 meters, and the response time of the terrain evaluation module ≤ 1 second.

Images