Autonomous navigation system and method of unmanned helicopter for denial environment

Abstract

The invention discloses a denial environment-oriented autonomous navigation system and method for an unmanned helicopter, and the system employs system architecture design, sensor configuration, a high-precision inertial navigation combination mode and application design for each functional stage. And the unmanned helicopter is ensured to have the capabilities of taking off, landing and executing flight tasks for a long time under the environment of complete rejection of satellites and links.

Description

Technical Field

[0001] The present invention relates to an autonomous navigation system for an unmanned helicopter, and in particular to an autonomous navigation system and method for an unmanned helicopter in a denied environment. Background Art

[0002] Currently, most unmanned helicopters are equipped with anti-jamming satellite navigation and use a multi-element method to improve the anti-jamming ability of satellite navigation equipment. Nowadays, the problem of electronic countermeasures is becoming increasingly acute, and the demand for the anti-jamming ability of the system is increasing. However, unmanned helicopters are often restricted by cost and weight, and can only support the design of anti-jamming satellite navigation to a certain extent. Since the implementation structure of a high-power narrowband interference transmitter is relatively simple, several narrowband interference signals can cover a certain frequency band. In addition, the wireless signals in the public environment are becoming more and more complex, and the transmission frequencies of various wireless signals may fall within the frequency band of the satellite navigation signal, causing interference to the satellite navigation signal. Therefore, unmanned helicopters still face the situation of satellite navigation failure.

[0003] Simultaneous Localization and Mapping (SLAM) uses measurement information of vision or laser point cloud to obtain environmental information, enabling a robot to complete localization, mapping, and path planning in an unknown environment. However, in the field of unmanned helicopters, when flying at an altitude of hundreds of meters to kilometers, there is often no object information to refer to, and environmental information cannot be constructed. Currently, most unmanned helicopters are equipped with inertial navigation as the main navigation device. However, the pure inertial navigation capabilities of inertial navigation devices vary. If the natural frequency of the shock-absorbing device is too low, it will lead to a delay in the measurement of attitude and angular velocity. Therefore, most fiber optic inertial navigation and micro-electromechanical inertial navigation cannot cope with the low-frequency vibration characteristics of unmanned helicopters, resulting in poor pure inertial navigation capabilities, and the auxiliary positioning devices configured with them lack redundancy and do not have the ability to fly and perform tasks for a long time in a denied environment. Summary of the Invention

[0004] Object of the Invention: Aiming at the above problems, the present invention proposes an autonomous navigation system and method for an unmanned helicopter in a denied environment, which solves the problem of autonomous navigation of an unmanned helicopter throughout the entire mission process from takeoff to landing in a completely denied environment of satellite navigation and link navigation.

[0005] Technical Solution: The technical solution adopted by the present invention is an autonomous navigation system for an unmanned helicopter in a denied environment, including a flight control computer, an intelligent computer, and a laser inertial navigation.

[0006] The laser inertial navigation system includes a pure laser inertial navigation module and a combined navigation unit. The combined navigation unit is electrically connected to the basic auxiliary navigation equipment. The combined navigation unit makes a judgment based on the acquisition information of the basic auxiliary navigation equipment and outputs a combined option. The basic auxiliary navigation equipment includes: high-precision satellite navigation, anti-jamming satellite navigation, direction-finding and ranging link, atmosphere computer, and radio altimeter. The laser inertial navigation system obtains the results of terrain matching and scene matching output by the intelligent computer and corrects the pure inertial position based on the results of terrain matching and scene matching.

[0007] The intelligent computer is electrically connected to the pod, lidar, and radio altimeter and is used for: obtaining the data of the pod, lidar, and radio altimeter; realizing mapping and positioning based on the point cloud measurement data of the lidar and the visual data collected by the pod; realizing scene matching and target guidance by matching the visual data collected by the pod with the map; realizing terrain matching based on the altitude information provided by the radio altimeter; realizing positioning and navigation based on the pose information of the inertial navigation provided by the radio altimeter; detecting obstacles and re-planning the real-time flight path during flight based on the point cloud measurement data of the lidar.

[0008] The flight control computer obtains the navigation results of the laser inertial navigation system and the intelligent computer for guidance and control, and sends a combined navigation data source selection command to the laser inertial navigation system.

[0009] The combined navigation unit makes a judgment based on the acquisition information of the basic auxiliary navigation equipment and outputs a combined option, including the following steps:

[0010] (1) Judge whether the high-precision satellite navigation data is valid. If it is valid, output the high-precision satellite navigation option; otherwise, go to the next step.

[0011] (2) Judge whether the anti-jamming satellite navigation data is valid. If it is valid, output the anti-jamming satellite navigation option; otherwise, go to the next step.

[0012] (3) Judge whether the direction-finding and ranging link data is valid. If it is valid, further judge whether the atmosphere computer data is valid; otherwise, go to the next step. If the atmosphere computer data is valid, output the combined navigation option of the direction-finding and ranging link and the atmosphere computer; if the atmosphere computer data is invalid, further judge whether the radio altimeter data is valid. If the radio altimeter data is valid, output the combined navigation option of the direction-finding and ranging link and the radio altimeter; if the radio altimeter is invalid, output the direction-finding and ranging link navigation option.

[0013] (4) Determine whether the terrain matching data is valid. If it is valid, further determine whether the air data computer data is valid; otherwise, proceed to the next step. If the air data computer data is valid, output the combined navigation option of terrain matching and air data computer. If the air data computer data is invalid, further determine whether the radio altimeter data is valid. If the radio altimeter data is valid, output the combined navigation option of terrain matching and radio altimeter. If the radio altimeter is invalid, output the terrain matching navigation option;

[0014] (5) Determine whether the scene matching data is valid. If it is valid, further determine whether the air data computer data is valid; otherwise, proceed to the next step. If the air data computer data is valid, output the combined navigation option of scene matching and air data computer. If the air data computer data is invalid, further determine whether the radio altimeter data is valid. If the radio altimeter data is valid, output the combined navigation option of scene matching and radio altimeter. If the radio altimeter is invalid, output the scene matching navigation option;

[0015] (6) Output the pure inertial navigation option.

[0016] The flight control computer commands the inertial navigation to select one of "high-precision satellite navigation", "anti-jamming satellite navigation", "link direction finding and ranging", "terrain matching", and "scene matching" through an interactive command for horizontal position combined navigation or enter the pure inertial navigation state; the flight control computer commands the inertial navigation to select one of "high-precision satellite navigation", "anti-jamming satellite navigation", "air data computer", and "radio altimeter" through an interactive command for altitude channel combined navigation;

[0017] When the data of high-precision satellite navigation, anti-jamming satellite navigation, direction finding and ranging link, and intelligent computer all fail, the laser inertial navigation automatically enters the pure inertial navigation, and the position accuracy of the pure inertial navigation reaches at least 1 nautical mile per hour.

[0018] The present invention proposes an autonomous navigation method applied to the autonomous navigation system of the unmanned helicopter for the denied environment, including:

[0019] (1) In the takeoff stage, perform self-alignment through the laser inertial navigation, and use the visual data of the pod, lidar, and inertial navigation odometer information for fusion to complete mapping and positioning;

[0020] (2) In the flight-navigation stage, perform terrain matching through the radio altimeter and elevation map, and perform scene matching through the visual data of the pod and high-precision map. The results of terrain matching and scene matching are used to correct the results of pure inertial navigation to achieve long-time flight operations;

[0021] (3) In the flight-obstacle avoidance stage, perform obstacle scanning through the lidar, and ensure flight safety through real-time route replanning;

[0022] (4) During the return flight phase, the landing point is detected using the visual data of the pod and target guidance is performed to fly the unmanned helicopter above the landing point, and hovering or dive height reduction is carried out in combination with the radio altimeter.

[0023] (5) During the landing phase, after hovering above the target point, the visual component of the pod is used to identify the relative position of the target. The height and position are adjusted through the radio altimeter and the pure inertial navigation information of the laser inertial navigation, and mapping and positioning are carried out in combination with the lidar to achieve precise landing.

[0024] The strategy design for terrain matching includes:

[0025] (1) Preprocess the elevation data collected by the radio altimeter, including filtering, denoising, and splicing, to improve the data quality. Using digital elevation model technology, the elevation data is transformed into a finite sequence of three-dimensional vectors to generate high-precision terrain data.

[0026] (2) Extract feature points from the real-time terrain data and the terrain elevation database. Adopt a feature point-based matching algorithm to match the feature points in the real-time terrain data with the feature points in the terrain elevation database. By calculating the similarity between the feature points, the best matching points are searched.

[0027] (3) According to the matching results, calculate the current position and attitude of the unmanned helicopter. Use filtering to smooth the position and attitude, and then dynamically adjust the position and attitude according to the matching error and the motion state of the unmanned helicopter to improve the matching accuracy.

[0028] The strategy design for scene matching includes:

[0029] (1) Relative position estimation uses frame-to-frame relative motion estimation based on the optimization of visual reprojection error to obtain the position of the real-time image relative to the initial reference frame.

[0030] (2) Absolute position estimation is performed by matching the real-time image collected by the unmanned helicopter's on-board pod with a pre-stored local reference database with geographical location information, and calculating the absolute geographical location corresponding to the real-time image.

[0031] (3) The real-time image acquisition of the pod uses infrared images that can work all day and night.

[0032] (4) The position information is fused to construct a factor graph, adding absolute position constraints to the frame-to-frame relative position to obtain the fused and optimized absolute position, thereby implementing the correction of the inertial navigation data.

[0033] The target guidance includes: the visual component of the pod collects images when the unmanned helicopter reaches near the target point and transmits them to the data processing module; the data processing module includes an intelligent computer vision image processing unit, which processes the collected image data to calculate the pose of the cooperation object mark in the pod coordinate system.

[0034] Pure inertial navigation includes:

[0035] (1) During the initialization process, create an initial reference frame, perform coordinate system creation and coordinate alignment, and estimate sensor biases;

[0036] (2) Geometric tracking receives visual images and point cloud data, performs feature extraction, moving object recognition, and motion feature filtering, then performs feature matching, and conducts geometric consistency verification on the matching results to ensure the geometric consistency of the matched features;

[0037] (3) Combine the inertial navigation pre-integration technology to achieve the non-linear fusion of vision-lidar-high-precision pure inertia. By constructing a factor graph, use the global optimization method to accurately calculate the trajectory and attitude of the unmanned helicopter, and at the same time optimize the visual and point cloud feature points to realize the construction of a three-dimensional point cloud map in the scene.

[0038] The pure inertial navigation and control flight process includes the following steps:

[0039] (1) The laser inertial navigation forwards the positioning information of the high-precision satellite navigation and anti-jamming satellite navigation to the flight control computer, so that the flight control computer obtains the current satellite navigation positioning status information;

[0040] (2) The flight control computer commands the laser inertial navigation to enter the integrated navigation state normally when selecting anti-jamming satellite navigation, and enter the pure inertial navigation state when selecting high-precision satellite navigation;

[0041] (3) The flight control computer always uses the navigation result of the laser inertial navigation for navigation and control flight;

[0042] (4) After the aircraft takes off, the flight control computer sets the laser inertial navigation to pure inertia through instructions to perform hovering climb, acceleration, cruise, and route flight actions; before landing, the flight control computer sets the laser inertial navigation to integrated navigation through instructions to complete the landing;

[0043] (5) During the flight, the laser inertial navigation goes through three stages: entering the pure inertial state; exiting the pure inertial state; restoring the integrated navigation state;

[0044] (6) The flight control computer obtains the flight position deviation in real time by comparing the deviation between pure inertial navigation and satellite navigation.

[0045] Advantages: Compared with the prior art, the present invention has the following advantages. The present invention designs a complementary autonomous navigation system: terrain matching, scene matching, and pure inertial navigation, all of which can achieve autonomous navigation without relying on external facilities. The error of pure inertial navigation accumulates over time, but it can track the movement of the vehicle well in a short time. Once the terrain matching and scene matching are successful, a high position estimation accuracy can be obtained, but their disadvantage is low real-time performance. Pure inertial navigation and terrain matching, scene matching have good complementarity, which can improve the navigation accuracy of the system and achieve flight in a long-term denied environment. In the present invention, the intelligent computer receives the point cloud measurement data of the lidar to detect obstacles and re-plan the real-time flight path during flight, and performs mapping and positioning during the takeoff and landing phases.

[0046] The present invention designs a mapping and positioning scheme for the takeoff phase of an unmanned helicopter, which has the characteristics of stability, robustness, strong environmental adaptability, and high precision: using the heterogeneous data of "vision-lidar-high-precision pure inertial navigation" for fusion, associating the mapping result with the object information of the takeoff site to complete the closed-loop detection, realizing more stable and robust mapping and positioning, and having better environmental adaptability and precision compared with traditional vision mapping.

[0047] The present invention designs a mapping and positioning scheme for the landing phase of an unmanned helicopter, which has the characteristics of redundancy and high precision: after the unmanned helicopter hovers over the landing point, the pod is placed in the directly downward viewing angle to identify and position the characteristic objects of the landing point, and uses "lidar-pure inertial navigation" to map and position the landing site environment, forming a redundant system with visual guidance to achieve precise landing. The present invention solves the problem of autonomous navigation of an unmanned helicopter in the full mission process from takeoff to landing in a completely denied environment of satellite navigation and link navigation. It is applicable to improving the ability of an unmanned helicopter to perform tasks under countermeasures such as electromagnetic information interference, electromagnetic spectrum suppression, and electromagnetic information deception in satellite navigation and link navigation. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is the architecture diagram of the autonomous navigation system of the unmanned helicopter according to the present invention;

[0049] Figure 2 It is the flow chart for judging the high-precision inertial navigation combination method of the horizontal position according to the present invention;

[0050] Figure 3 It is the flow chart for judging the high-precision inertial navigation combination method according to the present invention;

[0051] Figure 4 It is the working flow chart of each task phase of the autonomous navigation system according to the present invention;

[0052] Figure 5Topographic matching framework diagram of the present invention;

[0053] Figure 6 Scene matching framework diagram of the present invention;

[0054] Figure 7 Target guidance framework diagram of the present invention;

[0055] Figure 8 Mapping and positioning framework diagram of the present invention. Detailed implementation mode

[0056] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0057] Embodiment 1

[0058] The autonomous navigation system of the unmanned helicopter according to the present invention for a denied environment, the system architecture diagram is as Figure 1 shown, including: a flight control computer, an intelligent computer, and a laser inertial navigation. The three computing units communicate with each other pairwise for information sharing. The laser inertial navigation includes laser pure inertial navigation and integrated navigation. The laser inertial navigation obtains the information of the basic auxiliary navigation equipment for integrated navigation. The basic auxiliary navigation equipment includes: high-precision satellite navigation, anti-jamming satellite navigation, direction finding and ranging link, atmosphere computer, radio altimeter, pod, and lidar equipment. The laser inertial navigation can also obtain the results of terrain matching, scene matching, etc. output by the intelligent computer for pure inertial position correction. The ground operator interacts with the flight control computer through the direction finding and ranging link and can obtain the state of the current navigation system.

[0059] The flight control computer obtains the navigation results of the current laser inertial navigation and the intelligent computer for guidance and control, and can send a combined navigation data source selection command to the laser inertial navigation, commanding it to select the data source at the required horizontal position and altitude for combined navigation; the laser inertial navigation has the ability of high-precision pure inertial navigation, and forms a basic combined navigation function with the externally connected high-precision satellite navigation, anti-jamming satellite navigation, direction finding and ranging link, atmosphere computer, and radio altimeter. The laser inertial navigation accesses the navigation information such as scene matching and terrain matching of the intelligent computer to correct the errors of pure inertial navigation; the intelligent computer is externally connected to a pod, lidar, and radio altimeter, and the radio altimeter can also obtain the pose information of the inertial navigation to realize functions such as scene matching, terrain matching, target guidance, mapping and positioning, and obstacle avoidance. The pose information of the inertial navigation is output from the inertial navigation to the intelligent computer. In the pure inertial state, the inertial navigation will have position drift and can only provide rough pose information. This rough pose information is provided to the intelligent computer for optimization through means such as vision or laser to obtain accurate pose. When the intelligent computer performs functions such as target guidance, mapping and positioning, obstacle avoidance planning, and scene matching, it needs to use the rough pose estimation of the inertial navigation. Obtaining the pose of the inertial navigation can help the computer calculate accurate results more quickly.

[0060] Terrain matching navigation uses the terrain height contour as the matching feature. Usually, a radio altimeter is used to measure the height data along the flight path and compare it with the regional terrain data on the pre-acquired flight path. Scene matching navigation uses regional ground objects as the feature. A visual sensor is used to obtain the regional image near the target area and match it with the reference image stored on the aircraft.

[0061] Currently, most equipment-level unmanned helicopters use high-precision pure inertial navigation devices to improve their autonomous navigation capabilities: the inertial navigation includes an accelerometer and a gyroscope. By measuring the acceleration and angular velocity of the carrier and performing integral operations to obtain the speed and position, it does not rely on external information during operation, does not radiate energy to the outside, and is not easily interfered with. It is an autonomous navigation system. Pure inertial navigation can track the rapid movement of the carrier well in a short time, but its error accumulates over time and cannot guarantee long-term flight accuracy. Moreover, high-precision inertial navigation often has high costs, large volumes, and large weights. The laser inertial navigation of the present invention is equipped on the unmanned helicopter. Since the vibration frequencies of the unmanned helicopter mainly come from the main rotor, tail rotor, transmission system, and power system; among them, the vibration frequencies generated by the rotor and transmission are low and the amplitudes are large. The inertial navigation system needs to have good vibration adaptability and cannot use relatively soft shock absorbers for adaptation because the natural frequency of the shock absorber is too low, which will cause the inertial navigation to lose the measurement accuracy and real-time performance of the attitude and angular velocity. Through a large number of tests and adaptation work on various types of inertial navigation such as micromachined, fiber optic, and laser, it is concluded that the laser inertial navigation has good vibration adaptability for the unmanned helicopter. Therefore, the autonomous navigation system of the unmanned helicopter proposed in the present invention is equipped with laser inertial navigation and gives the core performance indicators required.

[0062] The horizontal position information data used in the integrated navigation of laser inertial navigation are, in sequence, high-precision satellite navigation data, anti-jamming satellite navigation data, direction finding and ranging link data, terrain matching data, and scene matching data; the judgment process for the high-precision inertial navigation combination method in the horizontal position is as Figure 2 shown. The altitude position information data used in the integrated navigation of laser inertial navigation are, in sequence, high-precision satellite navigation data, anti-jamming satellite navigation data, direction finding and ranging link data, and atmospheric computer. The flowchart for judging the combination method of laser inertial navigation is as Figure 3 shown.

[0063] The flight control computer can command the inertial navigation to select one of "high-precision satellite navigation", "anti-jamming satellite navigation", "link direction finding and ranging", "terrain matching", and "scene matching" through interactive instructions for horizontal position integrated navigation or enter the pure inertial navigation state.

[0064] The altitude information data used in the integrated navigation of laser inertial navigation are, in sequence, high-precision satellite navigation, anti-jamming satellite navigation, atmospheric computer, and radio altimeter. Among them, the atmospheric computer can normally collect effective data in a complex electromagnetic environment, which can ensure that the altitude solution of high-precision inertial navigation is in the combined operation, as Figure 3 shown.

[0065] The flight control computer can command the inertial navigation to select one of "high-precision satellite navigation", "anti-jamming satellite navigation", "atmospheric computer", and "radio altimeter" through interactive instructions for altitude channel integrated navigation.

[0066] When the data of devices such as high-precision satellite navigation, anti-jamming satellite navigation, direction finding and ranging link, and intelligent computer all fail, the laser inertial navigation enters the pure inertial navigation state. The laser inertial navigation can adapt to the low-frequency and high-frequency vibrations of the unmanned helicopter platform, and the position accuracy of the pure inertial navigation needs to reach 1 nautical mile per hour.

[0067] Equipped with a high-performance intelligent computer, a radio altimeter, an optoelectronic pod, and a lidar, it performs functions such as terrain matching positioning, scene matching positioning, terminal target guidance, takeoff and landing site mapping and positioning, etc., to achieve the navigation task throughout the process from takeoff, mission flight to landing.

[0068] Embodiment 2

[0069] Based on the autonomous navigation system described in Embodiment 1, the present invention proposes an autonomous navigation method for an unmanned helicopter facing a denied environment.

[0070] The present invention designs the working forms of the autonomous navigation system in each mission stage in a completely denied environment of satellites and links, as Figure 4 shown, including the following content:

[0071] (1) Takeoff phase: Use laser inertial navigation for self-alignment, and fuse the visual components of the pod, lidar, and inertial odometer information to complete mapping and positioning;

[0072] (2) Flight phase - Navigation: Use a radio altimeter for terrain matching with an elevation map, and use the pod for scene matching with a high-precision map. The results of terrain matching and scene matching are used to correct the results of pure inertial navigation to achieve long-duration flight operations;

[0073] Terrain matching navigation can obtain a series of terrain elevations of the true flight path through a radio altimeter, perform correlation analysis on the measured data and the stored digital map, determine the grid position corresponding to the aircraft's flight path, and achieve position estimation. The defect of existing terrain matching navigation is that it can only handle the flight terrain in mountainous and hilly areas and has greater restrictions on flight altitude. There are more restrictions on the application environment and flight tasks. Without the starting information provided by high-precision pure inertial navigation as a basis, it is extremely easy to cause matching failures. The present invention can improve this situation.

[0074] Scene matching stores the high-definition map scene information within the predetermined flight range on the airborne device. During flight, the scene along the flight path is acquired through the airborne imaging device and compared with the pre-stored data to determine the position of the aircraft. The defect of existing scene matching is that it cannot meet the all-weather flight requirements such as clouds, dust, etc., cannot meet the all-region flight requirements such as deserts, oceans, etc., and without the starting information provided by high-precision pure inertial navigation as a basis, it will cause a large search range, long search time, and matching failures. The present invention can improve this situation.

[0075] (3) Flight phase - Obstacle avoidance: Use lidar for obstacle scanning and ensure flight safety through real-time route replanning;

[0076] (4) Return flight phase: Use the pod to detect the landing point and perform target guidance to fly the unmanned helicopter above the landing point, and combine with the radio altimeter for hovering or diving to reduce altitude;

[0077] Target guidance is equipped with visual guidance equipment, identifies the established marker as a reference target, and completes pose calculation and control through image processing technology and corresponding visual measurement algorithms. The disadvantage of existing target guidance is that the operating range is short and the positioning accuracy is greatly affected by weather. The present invention can overcome this defect.

[0078] (5) Landing phase: After hovering above the target point, use the pod to identify the relative position of the target, adjust the height and position through the radio altimeter and laser inertial pure inertial navigation information, and combine with lidar for mapping and positioning to achieve precise landing.

[0079] The strategy design adopted for terrain matching is specifically manifested as:

[0080] (1) Preprocess the elevation data collected by the radio altimeter, including filtering, denoising, and splicing, to improve the data quality. Then, use the digital elevation model (DEM) technology to convert the elevation data into a three-dimensional vector finite sequence and generate high-precision terrain data;

[0081] (2) Extract feature points from the real-time terrain data and the terrain elevation database. Adopt a feature point-based matching algorithm to match the feature points in the real-time terrain data with those in the terrain elevation database. By calculating the similarity between the feature points, search for the best matching points;

[0082] (3) According to the matching results, calculate the current position and attitude of the unmanned helicopter. Use filtering to smooth the position and attitude, and then dynamically adjust the position and attitude according to the matching error and the motion state of the unmanned helicopter to improve the matching accuracy. The overall block diagram of terrain matching is as shown in Figure 5 shown.

[0083] Scene matching adopts the fusion method of infrared / visible light heterogeneous scene matching technology and inter-frame motion estimation technology of the pod to achieve high-precision and high-frequency autonomous navigation information output, including relative position estimation, absolute position estimation, and position information fusion. The strategy design adopted by scene matching is specifically manifested as follows:

[0084] (1) Relative position estimation adopts inter-frame relative motion estimation based on the optimization of visual reprojection error to obtain the position of the real-time image relative to the initial reference frame;

[0085] (2) Absolute position estimation is achieved by matching the real-time image collected by the unmanned helicopter's airborne pod with the pre-stored local reference database with geographical location information, and calculating the absolute geographical location corresponding to the real-time image;

[0086] (3) To cope with the influence of light, cloud and fog occlusion, etc., the real-time image acquisition of the pod can use infrared images that can work all day long;

[0087] (4) Position information fusion constructs a factor graph, adds absolute position constraints to the inter-frame relative position, and obtains the fused and optimized absolute position, so as to correct the inertial navigation data. The overall block diagram of scene matching is as shown in Figure 6 shown.

[0088] Target guidance consists of three parts: a cooperative object marker, an image acquisition module, and a data processing module, as shown in Figure 7 shown.

[0089] (1) The cooperative object marker is designed manually and is used to guide the unmanned helicopter to perform tasks.

[0090] (2) The image acquisition module consists of an airborne pod, which is responsible for acquiring images when the unmanned helicopter reaches near the target point and transmitting them to the data processing module.

[0091] (3) The data processing module consists of an intelligent computer vision image processing unit, which processes the acquired image data to calculate and cooperate the pose of the cooperative object marker in the pod coordinate system.

[0092] Based on the SLAM technology of vision - lidar - high - precision pure inertial navigation, more stable and robust mapping and positioning are realized, as Figure 8 shown.

[0093] Vision - lidar - high - precision pure inertial navigation includes initialization and geometric tracking:

[0094] (1) During the initialization process, an initial reference frame is created, the coordinate system is created and aligned, and the sensor bias is estimated.

[0095] (2) The geometric tracking receives visual images and point cloud data, performs feature extraction, moving object recognition, and motion feature filtering, then performs feature matching, and conducts geometric consistency verification on the matching results to ensure the geometric consistency of the matched features in the geometric space.

[0096] (3) Finally, combined with the inertial navigation pre - integration technology, the non - linear fusion of vision - lidar - high - precision pure inertia is realized. By constructing a factor graph and using the global optimization method, the trajectory and attitude of the unmanned helicopter are accurately calculated, and at the same time, the visual and point cloud feature points are optimized to realize the construction of the three - dimensional point cloud map in the scene.

[0097] Example of the implementation of laser - inertial pure inertial navigation: In order to adapt to the complex electromagnetic environment, ensure the safety of the unmanned helicopter, and improve the effectiveness of mission execution, the unmanned helicopter needs to perform tasks in a denied environment and meet the index requirement of pure inertia of 1 nautical mile per hour. The process of this pure inertial navigation and control flight is as follows:

[0098] (1) The flight control computer and the laser inertial navigation are disconnected from the intelligent computer to ensure that the system does not use the navigation results of scene matching and terrain matching of the intelligent computer.

[0099] (2) The laser inertial navigation only accesses one anti - interference satellite navigation device, and one high - precision satellite navigation, and blocks the positioning antenna of the high - precision satellite navigation so that it does not position.

[0100] (3) The laser inertial navigation forwards the positioning information of the high - precision satellite navigation and the anti - interference satellite navigation to the flight control computer, enabling the flight control computer to obtain the current satellite navigation positioning status information.

[0101] (4) When the flight control computer commands the laser inertial navigation to select anti-jamming satellite navigation, it can normally enter the integrated navigation state. When it commands the laser inertial navigation to select high-precision satellite navigation, it enters the pure inertial navigation state;

[0102] (5) The flight control computer always uses the navigation results of the laser inertial navigation (including pure inertia) for navigation and flight control;

[0103] (6) After the aircraft takes off, the flight control computer immediately sets the laser inertial navigation to pure inertia through commands to perform actions such as hovering climb, acceleration, cruise, and route flight. Before landing, the flight control computer sets the laser inertial navigation to integrated navigation through commands to complete the landing;

[0104] (7) During the flight, the laser inertial navigation mainly goes through three stages: it enters the pure inertial state at the 1364th second; it exits the pure inertial state at the 6605th second; it resumes the integrated navigation state at the 6621st second, maintaining the pure inertial state for a total of 5241 seconds, approximately 1 hour and 30 minutes.

[0105] (8) The flight control computer obtains the flight position deviation in real time by comparing the deviation between the pure inertial navigation and the satellite navigation. The maximum deviation during this flight is 1454 meters, meeting the index requirements.

Claims

1. An autonomous navigation system for an unmanned helicopter in a denied environment, characterized by: Including flight control computer, intelligent computer and laser inertial navigation; The laser inertial navigation comprises a laser pure inertial navigation module and a combined navigation unit, the combined navigation unit is electrically connected to a basic auxiliary navigation device, and the combined navigation unit makes a judgment based on the collected information of the basic auxiliary navigation device and outputs a combined option; the basic auxiliary navigation device comprises: a high-precision satellite navigation, an anti-interference satellite navigation, a direction-finding and ranging link, an atmospheric machine, and a radio altimeter; the laser inertial navigation obtains the results of terrain matching and scene matching output by an intelligent computer, and performs a pure inertial position correction based on the results of terrain matching and scene matching; The intelligent computer is electrically connected to the pod, the laser radar, and the radio altimeter, and is used to: obtain data from the pod, the laser radar, and the radio altimeter; realize mapping and positioning according to the point cloud measurement data of the laser radar and the visual data collected by the pod; realize scene matching and target guidance by matching the visual data collected by the pod with the map; realize terrain matching according to the altitude information provided by the radio altimeter; realize positioning and navigation according to the inertial navigation posture information provided by the radio altimeter; and perform obstacle detection and real-time route replanning during flight according to the point cloud measurement data of the laser radar; The flight control computer obtains the navigation results of the laser inertial navigation and the intelligent computer for guidance and control, and sends a combined navigation data source selection command to the laser inertial navigation.

2. The autonomous navigation system for an unmanned helicopter in a denied environment according to claim 1, characterized in that: The combined navigation unit makes a judgment based on the collected information of the basic auxiliary navigation device and outputs the combined options, including the following steps: (1) Determine whether the high-precision satellite navigation data is valid. If so, output the high-precision satellite navigation option; otherwise, proceed to the next step; (2) judging whether the anti-interference satellite navigation data is valid, and if so, outputting the anti-interference satellite navigation option, otherwise proceeding to the next step; (3) Determine whether the direction finding and ranging link data is valid. If so, further determine whether the atmospheric aircraft data is valid, otherwise proceed to the next step; if the atmospheric aircraft data is valid, output the combined navigation option of the direction finding and ranging link and the atmospheric aircraft; if the atmospheric aircraft data is invalid, further determine whether the radio altimeter data is valid. If the radio altimeter data is valid, output the combined navigation option of the direction finding and ranging link and the radio altimeter; if the radio altimeter is invalid, output the navigation option of the direction finding and ranging link; (4) Determine whether the terrain matching data is valid. If so, further determine whether the atmospheric aircraft data is valid, otherwise proceed to the next step; if the atmospheric aircraft data is valid, output a combined navigation option of terrain matching and atmospheric aircraft; if the atmospheric aircraft data is invalid, further determine whether the radio altimeter data is valid. If the radio altimeter data is valid, output a combined navigation option of terrain matching and radio altimeter; if the radio altimeter is invalid, output a terrain matching navigation option; (5) Determine whether the scene matching data is valid. If so, further determine whether the atmospheric aircraft data is valid, otherwise proceed to the next step; if the atmospheric aircraft data is valid, output a combined navigation option of scene matching and atmospheric aircraft; if the atmospheric aircraft data is invalid, further determine whether the radio altimeter data is valid. If the radio altimeter data is valid, output a combined navigation option of scene matching and radio altimeter; if the radio altimeter is invalid, output a scene matching navigation option; (6) Output pure inertial navigation option.

3. The autonomous navigation system for an unmanned helicopter in a denied environment according to claim 1, characterized in that: The flight control computer commands the inertial navigation system to select one of the following options: "high-precision satellite navigation", "anti-interference satellite navigation", "link direction finding and ranging", "terrain matching" and "scene matching" to perform horizontal position integrated navigation or enter the pure inertial navigation state through interactive commands; the flight control computer commands the inertial navigation system to select one of the following options: "high-precision satellite navigation", "anti-interference satellite navigation", "atmospheric machine" and "radio altimeter" to perform altitude channel integrated navigation through interactive commands; When the data from high-precision satellite navigation, anti-interference satellite navigation, direction-finding and ranging links, and intelligent computers all fail, the laser inertial navigation automatically enters pure inertial navigation, and the pure inertial navigation position accuracy reaches at least 1 nautical mile / hour.

4. An autonomous navigation method for an autonomous navigation system for an unmanned helicopter facing a denied environment as claimed in claim 1, characterized in that: include: (1) During the takeoff phase, the laser inertial navigation system is used for self-alignment, and the pod’s visual data is integrated with the laser radar and inertial navigation odometer information to complete mapping and positioning. (2) During the flight-navigation phase, terrain matching is performed using the radio altimeter and the elevation map, and scene matching is performed using the pod's visual data and the high-precision map. The results of terrain matching and scene matching are used to correct the results of pure inertial navigation to achieve long-term flight operations; (3) During the flight-obstacle avoidance phase, obstacle scanning is performed using lidar, and real-time route replanning is performed to ensure flight safety; (4) During the return flight phase, the pod’s visual data is used to find the landing point and provide target guidance to enable the unmanned helicopter to fly over the landing point, and the unmanned helicopter is then hovered or dived to lower its altitude in conjunction with the radio altimeter; (5) During the landing phase, after flying to and hovering above the target point, the pod’s visual components are used to identify the target’s relative position, and the altitude and position are adjusted through the radio altimeter and pure inertial navigation information of the laser inertial navigation system. The laser radar is used for mapping and positioning to achieve precise landing.

5. The autonomous navigation method for an unmanned helicopter in a denied environment according to claim 4, characterized in that: The strategy design adopted for terrain matching includes: (1) Preprocess the elevation data collected by the radio altimeter, including filtering, denoising, and splicing, to improve data quality. Use digital elevation model technology to convert the elevation data into a three-dimensional vector finite sequence to generate high-precision terrain data. (2) Extracting feature points from the real-time terrain data and the terrain elevation database, using a matching algorithm based on feature points to match the feature points in the real-time terrain data with the feature points in the terrain elevation database, and searching for the best matching point by calculating the similarity between the feature points; (3) Based on the matching results, the current position and attitude of the unmanned helicopter are calculated, and the position and attitude are smoothed using filtering. Then, the position and attitude are dynamically adjusted according to the matching error and the motion state of the unmanned helicopter to improve the matching accuracy.

6. The autonomous navigation method for an unmanned helicopter in a denied environment according to claim 4, characterized in that: The strategy design adopted by scene matching includes: (1) Relative position estimation uses inter-frame relative motion estimation based on visual reprojection error optimization to obtain the position of the real-time image relative to the initial reference frame; (2) Absolute position estimation: The real-time image collected by the pod on the unmanned helicopter is matched with a pre-stored local reference database with geographic location information to calculate the absolute geographic location corresponding to the real-time image; (3) The pod’s real-time image acquisition uses infrared images that can work around the clock; (4) Position information fusion constructs a factor graph, adds absolute position constraints to the relative position between frames, and obtains the absolute position after fusion optimization, thereby implementing the correction of inertial navigation data.

7. The autonomous navigation method for an unmanned helicopter in a denied environment according to claim 4, characterized in that: Target guidance includes: the visual component of the pod collects images and transmits them to the data processing module when the unmanned helicopter arrives near the target point; the data processing module includes an intelligent computer vision image processing unit, which processes the collected image data to calculate the position and posture of the collaborative object mark in the pod coordinate system.

8. The autonomous navigation method for an unmanned helicopter in a denied environment according to claim 4, characterized in that: Pure inertial navigation includes: (1) During the initialization process, an initial reference frame is created, the coordinate system is created and aligned, and the sensor bias is estimated; (2) Geometric tracking receives visual images and point cloud data and performs feature extraction, moving object recognition, and motion feature filtering, and then performs feature matching. The matching results are subjected to geometric consistency checks to ensure the consistency of the matched features in the geometric space. (3) Combined with inertial navigation pre-integration technology, the nonlinear fusion of vision, lidar and high-precision pure inertial is realized. By constructing a factor graph and using a global optimization method, the trajectory and posture of the unmanned helicopter are accurately calculated. At the same time, the visual and point cloud feature points are optimized to realize the construction of a three-dimensional point cloud map in the scene.

9. The autonomous navigation method for an unmanned helicopter in a denied environment according to claim 4, characterized in that: The pure inertial navigation and control flight process includes the following steps: (1) The laser inertial navigation system forwards the positioning information of high-precision satellite navigation and anti-interference satellite navigation to the flight control computer, so that the flight control computer can obtain the status information of the current satellite navigation positioning; (2) When the flight control computer commands the laser inertial navigation system to select anti-interference satellite navigation, it normally enters the integrated navigation state; when the flight control computer commands the laser inertial navigation system to select high-precision satellite navigation, it enters the pure inertial navigation state; (3) The flight control computer always uses the laser inertial navigation results to navigate and control the flight; (4) After the aircraft takes off, the flight control computer sets the laser inertial navigation to pure inertial through instructions, and performs hovering, climbing, acceleration, cruising, and route flight actions; before landing, the flight control computer sets the laser inertial navigation to integrated navigation through instructions to complete the landing; (5) During the flight, the laser inertial navigation system goes through three stages: entering the pure inertial state; exiting the pure inertial state; and restoring the integrated navigation state; (6) The flight control computer obtains the flight position deviation in real time by comparing the deviations between pure inertial navigation and satellite navigation.

Images