Arresting hook test system and method for fixed-wing UAV offshore recovery

Abstract

The present invention relates to an arresting hook test system and method for recovering fixed-wing unmanned aerial vehicles at sea, wherein a simulation test vehicle is movably arranged in a guide frame, and recovery rope height adjustment components with lifting sliders are arranged on both sides of the buffer frame, and recovery rope steering components are arranged on the sliders, and the recovery rope steering components are arranged with tension sensors, and the bottom of each recovery rope height adjustment component is provided with a recovery rope retracting component, and the two ends of the recovery rope are respectively wound around the tension sensors on the recovery rope steering components on the corresponding sides and then wound in the recovery rope retracting components on the corresponding sides, and an arresting hook is arranged on the lower side of the simulation test vehicle to cooperate with the recovery rope, and a parameter capture module is arranged on the simulation test vehicle, and a pan-tilt camera is arranged on any recovery rope height adjustment component. The present invention has low testing cost and can conduct in-depth analysis of the moment when the arresting hook is in action, and can analyze and estimate the optimal tension of the recovery rope and the optimal parameters of the arresting hook through experimental testing.

Description

Technical Field

[0001] The invention relates to an arresting hook testing system and method for offshore recovery of fixed-wing unmanned aerial vehicles. Background Art

[0002] Unmanned aerial vehicles and unmanned ships play an extremely important role in the fields of marine environmental monitoring, marine surveying and mapping, and maritime patrol and rescue. In application scenarios such as ocean environmental monitoring, it is difficult to conduct large-scale marine environmental coverage monitoring in a short period of time due to the limitations of the ship's movement speed and maneuverability. Fixed-wing drones have the advantages of fast speed, strong maneuverability, large range of activities, and long endurance. They can quickly conduct large-scale coverage environmental monitoring, especially when multiple drones work together at the same time, which can greatly improve the operating efficiency. In addition, fixed-wing drone technology is relatively mature, with low development costs, and small fixed-wing drones can be stored in boxes, which have the advantages of small space occupation and have been widely used. If unmanned ships and small fixed-wing drones can be combined, the unmanned ship will serve as the mother ship of the drone, providing energy supply, unified planning and placement and storage. The drone will serve as a large-scale fast and efficient search operation tool, which will greatly expand the application of drone-unmanned ship sub-platforms in the marine environment and have great research value.

[0003] There are many types of landing and recovery methods for shipborne fixed-wing UAVs. The common landing and recovery methods currently include: runway arrest landing, net recovery, parachute / wing parachute recovery, rope hook recovery, sky hook recovery, over-stall recovery, smart drop landing, wind vane recovery, swing-type hanger recovery, etc. However, no matter which recovery mechanism is used, there are certain risks, especially in the testing stage when the technology is not yet mature. During the testing stage, how to safely and effectively conduct ground tests on the rope hook arrest method for the recovery of fixed-wing UAVs at sea is a problem that needs to be considered. Summary of the invention

[0004] The purpose of the present invention is to provide an arresting hook testing system and method for offshore recovery of fixed-wing UAVs, which has low testing cost and can perform in-depth analysis of the moment when the arresting hook is in action. Through a large number of experimental tests, the optimal tension of the recovery rope and the optimal parameters of the arresting hook can be analyzed and estimated, providing a basis for the design of related mechanisms.

[0005] The objective of the present invention is achieved through the following technical solutions:

[0006] The invention discloses an arresting hook test system for offshore recovery of fixed-wing unmanned aerial vehicles, comprising a guide frame, a simulation test vehicle, a recovery rope height adjustment component, a buffer frame, a recovery rope, a recovery rope steering component and a recovery rope retracting and releasing component, wherein the simulation test vehicle is movably arranged in the guide frame, the buffer frame is arranged on one side of the guide frame, and recovery rope height adjustment components are arranged on both sides of the buffer frame, the recovery rope height adjustment component is provided with a liftable slider, and a recovery rope steering component is arranged on the slider, the recovery rope steering component is provided with a tension sensor, a recovery rope retracting and releasing component is arranged at the bottom end of each recovery rope height adjustment component, and the two ends of the recovery rope are respectively wound around the tension sensor on the recovery rope steering component on the corresponding side and then wound in the recovery rope retracting and releasing component on the corresponding side, an arresting hook is arranged on the lower side of the simulation test vehicle to cooperate with the recovery rope, a parameter capture module is arranged on the simulation test vehicle, and a pan / tilt camera is arranged on any recovery rope height adjustment component.

[0007] A top plate is provided at the upper end of the guide frame, the simulation test vehicle includes a vehicle body, and the upper side of the vehicle body is slidably connected to the top plate, elastic cylinders, connecting rod arm assemblies and wheels are provided on both sides of the vehicle body, wherein a wheel seat is provided at the end of the wheel axle, one end of the connecting rod arm assembly is hinged to the vehicle body, and the other end is hinged to the wheel seat on the corresponding side, the lower end of the elastic cylinder is hinged to the lower side of the vehicle body, and the upper end is hinged to the connecting rod arm assembly on the corresponding side.

[0008] A guide rail is arranged on the lower side of the top plate, a guide suspension rod is arranged on the upper side of the vehicle body, and a guide pulley is arranged on the upper end of the guide suspension rod to cooperate with the guide rail on the lower side of the top plate.

[0009] The link arm assembly includes a first link arm and a second link arm, wherein one end of the first link arm is hinged to the upper side of the corresponding side wheel seat, and the other end is hinged to the upper side of the vehicle body, one end of the second link arm is hinged to the lower side of the corresponding side wheel seat, and the other end is hinged to the lower side of the vehicle body, and the upper end of the elastic cylinder is hinged to the first link arm.

[0010] The recovery rope height adjustment assembly includes a support rod and a belt, a pulley and an adjustment drive device arranged on the support rod, wherein pulleys are provided at the upper and lower ends of the support rod, and the two ends of the belt are respectively passed around the pulleys at the corresponding ends, and the pulley at either end is driven to rotate by the adjustment drive device, and the slider is fixedly connected to one side of the belt.

[0011] The recovery rope steering assembly includes a first guide element, a tension sensor, a first tensioning clamp, a second tensioning clamp and a second guide element. The recovery rope is guided input via the first guide element, bypasses the tension sensor, passes between the first tensioning clamp and the second tensioning clamp, and is guided output via the second guide element.

[0012] The recovery rope steering assembly is provided with a spring shaft and a spring, and the spring shaft passes through the first tensioning clamp and the second tensioning clamp, and the spring shaft is provided with a spring and a spring pressure plate at one end close to the first tensioning clamp, and one end of the spring is against the spring pressure plate, and the other end is against the first tensioning clamp.

[0013] The recovery rope retracting assembly includes a mounting frame, a winding wheel and a retracting drive device arranged on the mounting frame. A guide ring is provided at the upper end of the mounting frame. The recovery rope passes through the guide ring and is wound around the winding wheel. The winding wheel is driven to rotate by the retracting drive device.

[0014] A method for testing an arresting hook system for recovering fixed-wing UAVs at sea:

[0015] The steps include:

[0016] Step 1: Build a neural network;

[0017] Step 2: Obtain multiple recovery test results through tests under multiple different input conditions;

[0018] Step 3: Use the input conditions and test results of step 2 as input and output parameters to train the neural network, and use the descent gradient method to iteratively solve the weight value and bias of the neural network. The formula is as follows:

[0019]

[0020] In the above formula, m is the number of points in the test result data set, y is the output result value of the test, h is the value predicted by the neural network, and J(w, b) is a function of the weight value w and the bias b. By solving the optimal solution of J(w, b), the parameters in the neural network are determined;

[0021] Step 4: Design the arresting hook (201) according to the optimal solution parameters.

[0022] In step 1, a 4-layer neural network is constructed using the data generated in the simulated recycling process, wherein the neural network includes 23 neurons, including 7 neurons in the input layer and 5 neurons in the output layer;

[0023] The 7 neurons in the input layer represent the UAV’s impact speed, material diameter, arresting hook material, arresting hook length, arresting hook end joint elasticity, arresting hook end joint damping, and the UAV’s impact height relative to the recovery rope.

[0024] The five neurons in the output layer represent the five initial tension states of the recovery rope: state A (very tight), state B (tight), state C (normal), state D (loose), and state E (very loose);

[0025] Using an S-type neural network, the output of each neuron is:

[0026]

[0027] In the above formula, x j represents the output value of the jth neuron in the previous layer, w j represents the weight value of the jth neuron in the previous layer, and b represents the bias of this neural network.

[0028] The advantages and positive effects of the present invention are:

[0029] The present invention can greatly reduce the risks brought by real deck tests at sea, while controlling the testing costs and minimizing the economic losses caused by test failures. It can also assist testers in conducting in-depth analysis of the moment when the arresting hook is in action from multiple aspects such as data and images. Through a large number of experimental tests, the optimal tension of the recovery rope and the optimal parameters of the arresting hook can be estimated, providing a basis for designing an arresting mechanism for the recovery of fixed-wing UAVs at sea. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 is a schematic diagram of the structure of the system of the present invention,

[0031] Figure 2 for Figure 1 A side view of the system of the present invention,

[0032] Figure 3 for Figure 1 The front view of the simulated test vehicle and the guide frame,

[0033] Figure 4 for Figure 3 A magnified schematic diagram of the simulated test vehicle in the figure.

[0034] Figure 5 for Figure 1 Schematic diagram of the recovery rope height adjustment component structure,

[0035] Figure 6 for Figure 1 Schematic diagram of the recovery rope steering assembly structure,

[0036] Figure 7 for Figure 1 Schematic diagram of the recovery rope retracting and releasing component structure,

[0037] Figure 8 for Figure 1 Schematic diagram of the arresting hook structure on the lower side of the simulation test vehicle.

[0038] Fig. 9 Schematic diagram of a neural network constructed by the method of the present invention.

[0039] Among them, 1 is a guide frame, 101 is a top plate, 102 is a side support beam, 2 is a simulation test vehicle, 201 is an arresting hook, 2011 is a steering gear, 202 is a vehicle body, 2021 is a cylinder mounting seat, 203 is a guide boom, 204 is a guide pulley, 205 is an elastic cylinder, 206 is a connecting rod arm assembly, 2061 is a first connecting rod arm, 2062 is a second connecting rod arm, 207 is a wheel, 2071 is a wheel seat, 3 is a recovery rope height adjustment assembly, 301 is an adjustment drive device, 302 is a slider, 303 is a belt, 304 is a support rod, 4 is a buffer frame, 4 01 is a buffer pad, 402 is a side baffle, 5 is a recovery rope, 6 is a recovery rope steering assembly, 601 is a first guide element, 602 is a second tensioning clamp, 603 is a tension sensor, 604 is a spring, 605 is a first tensioning clamp, 606 is a second guide element, 607 is a spring shaft, 7 is a pan / tilt camera, 8 is a recovery rope retracting assembly, 801 is a mounting frame, 8011 is a base frame, 8012 is a stand, 8013 is a guide frame, 802 is a winding wheel, 803 is a wire buckle, 804 is a baffle, 805 is a retracting drive device, and 806 is a guide ring. DETAILED DESCRIPTION

[0040] The present invention will be further described below in conjunction with the accompanying drawings.

[0041] like Figures 1 to 9 As shown, the present invention comprises a guide frame 1, a simulation test vehicle 2, a recovery rope height adjustment assembly 3, a buffer frame 4, a recovery rope 5, a recovery rope steering assembly 6, a pan / tilt camera 7 and a recovery rope retracting assembly 8, wherein Figure 2 As shown, the simulation test vehicle 2 is movably arranged in the guide frame 1, and the buffer frame 4 is arranged on one side of the guide frame 1, and as shown in FIG. Figure 1 As shown, both sides of the buffer frame 4 are provided with a recovery rope height adjustment component 3, such as Figure 5 As shown, the recovery rope height adjustment assembly 3 is provided with a slider 302 that can be lifted and lowered, and the slider 302 is provided with a recovery rope steering assembly 6, as shown in FIG. Figure 6 As shown, the recovery rope steering assembly 6 is provided with a tension sensor 603. Figure 1 As shown, each recovery rope height adjustment assembly 3 is provided with a recovery rope retracting assembly 8 at the bottom end, and the two ends of the recovery rope 5 are respectively wound around the tension sensor 603 on the recovery rope steering assembly 6 on the corresponding side and then wound in the recovery rope retracting assembly 8 on the corresponding side. While the tension sensor 603 guides the recovery rope 5 to turn, it can also obtain the tension data of the recovery rope 5 during the whole arresting test. The tension sensor 603 is a well-known technology in the art and is a commercially available product, such as Figures 1-2As shown, an arresting hook 201 is provided on the lower side of the simulation test vehicle 2 to cooperate with the recovery rope 5. In addition, a parameter capture module is provided on the simulation test vehicle 2 to obtain the posture information of the simulation test vehicle 2 in real time. A pan-tilt camera 7 is provided on any recovery rope height adjustment component 3 to shoot the instantaneous shape of the arresting hook 201 hooking the recovery rope 5. The data obtained by the parameter capture module, the tension sensor 603 and the pan-tilt camera 7 are analyzed by the analysis module of the system of the present invention to provide a basis for the design of the arresting hook 201 and related mechanisms.

[0042] like Figures 3-4 As shown, the guide frame body 1 includes a top plate 101 and side beams 102 supporting both sides of the top plate 101, a guide rail is provided on the lower side of the top plate 101, and the simulation test vehicle 2 includes a vehicle body 202, an elastic cylinder 205, a connecting rod arm assembly 206 and a wheel 207, wherein a guide hanger 203 is provided on the upper side of the vehicle body 202, and a guide pulley 204 is provided on the upper end of the guide hanger 203 to cooperate with the guide rail on the lower side of the top plate 101. In this embodiment, the guide pulley 204 adopts a high-speed copper alloy bearing to be fully embedded in the guide rail to realize hanging sliding, and the guide rail has a downward inclination angle of 5°, which helps the simulation test vehicle 2 to accelerate to 16 m / s. Both sides of the vehicle body 202 are provided with elastic cylinders 205, connecting rod arm assemblies 206 and wheels 207, wherein the wheel 207 is provided with a wheel seat 2071 at the end of the wheel axle, and a driving motor is provided on the wheel seat 2071 to drive the wheel 207 to rotate along the lower side of the top 101, thereby realizing the movement of the simulated test vehicle 2, one end of the connecting rod arm assembly 206 is hinged to the vehicle body 202, and the other end is hinged to the wheel seat 2071 on the corresponding side, the lower end of the elastic cylinder 205 is hinged to the lower side of the vehicle body 202, and the upper end is hinged to the connecting rod arm assembly 206 on the corresponding side. In this embodiment, the wheels 207 are made of rubber tires, and their grip becomes stronger as the vehicle speed increases. The elastic cylinder 205 and the connecting rod arm assembly 206 can effectively avoid the interference caused by the unevenness of the lower side of the top plate 101 or the guide rail, and can appropriately adjust the posture of the vehicle body. In this embodiment, the simulation test vehicle 2 has a length, width and height of 549 mm × 368 mm × 232 mm, a wheelbase of 329 mm, a weight of 4 kg, is driven by a 2400 kV brushless motor, and has a maximum speed of 90 km / h. The simulation test vehicle 2 is used to simulate the high-speed recovery and movement process of a fixed-wing UAV.

[0043] like Figure 4As shown, in this embodiment, the link arm assembly 206 includes a first link arm 2061 and a second link arm 2062, and the first link arm 2061 and the second link arm 2062 are both in the shape of a triangular bracket, wherein one end of the first link arm 2061 is hinged to the upper side of the corresponding side wheel seat 2071, and the other end is hinged to the upper side of the vehicle body 202, and one end of the second link arm 2062 is hinged to the lower side of the corresponding side wheel seat 2071, and the other end is hinged to the lower side of the vehicle body 202, and a cylinder mounting seat 2021 is provided on the lower side of the vehicle body 202, the lower end of the elastic cylinder 205 is hinged to the cylinder mounting seat 2021, and the upper end of the elastic cylinder 205 is hinged to the first link arm 2061.

[0044] like Figure 5 As shown, the recovery rope height adjustment assembly 3 includes a support rod 304 and a belt 303, a pulley and an adjustment drive device 301 arranged on the support rod 304, wherein pulleys are arranged at both ends of the support rod 304, and both ends of the belt 303 are respectively passed around the pulleys at the corresponding ends, and the pulley at either end is driven to rotate by the adjustment drive device 301, thereby realizing the movement of the belt 303, and the slider 302 is fixedly connected to one side of the belt 303. In this embodiment, the adjustment drive device 301 can adopt a servo reduction motor. Figure 1 As shown, the slider 302 moves up and down to adjust the height of the recovery rope steering assembly 6 and also adjusts the height of the recovery rope 5.

[0045] like Figure 6As shown, the recovery rope steering assembly 6 includes a base and a first guide element 601, a tension sensor 603, a first tensioning clamp 605, a second tensioning clamp 602, a spring shaft 607, a spring 604, and a second guide element 606 arranged on the base, wherein the recovery rope 5 is guided by the first guide element 601 to enter between the first tensioning clamp 605 and the second tensioning clamp 602 after bypassing the tension sensor 603, the spring shaft 607 passes through the first tensioning clamp 605 and the second tensioning clamp 602, and the spring shaft 607 is provided with a spring 604 and a spring pressure plate at one end close to the first tensioning clamp 605 , and the other end passes through the stopper block provided on the base and is threadedly connected with a nut. One end of the spring 604 abuts against the spring pressure plate, and the other end abuts against the first tensioning clamping plate 605. The spring 604 provides elastic force to make the first tensioning clamping plate 605 and the second tensioning clamping plate 602 cooperate to clamp the recovery rope 5 to achieve pre-tightening. By screwing the spring shaft 607, the spring shaft 607 can move and adjust the compression amount of the spring 604, that is, adjust the pre-tightening force of the two tensioning clamping plates. The recovery rope 5 is output between the first tensioning clamping plate 605 and the second tensioning clamping plate 602 and then guided and output through the second guide element 606. The base is mounted on the slider 302 of the recovery rope height adjustment assembly 3 on the corresponding side. In addition, in this embodiment, the first guide element 601 and the second guide element 606 can adopt structures such as guide copper sheets, which is a well-known technology in the art.

[0046] like Figure 7 As shown, the recovery rope retracting assembly 8 includes a mounting frame 801, a winding wheel 802 and a retracting drive device 805 arranged on the mounting frame 801, a guide ring 806 is provided at the upper end of the mounting frame 801, the recovery rope 5 passes through the guide ring 806 and is wound around the winding wheel 802, the winding wheel 802 is provided with a wire buckle 803 for fixing the end of the recovery rope 5, the winding wheel 802 is driven to rotate by the retracting drive device 805 to realize the retraction and release of the recovery rope 5, and a baffle 804 is provided on the side of the winding wheel 802 away from the retracting drive device 805 to prevent the recovery rope 5 from escaping. In this embodiment, the retracting drive device 805 adopts a reduction servo motor.

[0047] like Figure 7 As shown, in this embodiment, the mounting frame 801 includes a base frame 8011, a vertical frame 8012 and a guide frame 8013, wherein the vertical frame 8012 is vertically arranged on the base frame 8011, a guide frame 8013 is arranged on the upper end of the vertical frame 8012, the winding wheel 802 and the retracting and releasing driving device 805 are both installed on the vertical frame 8012, and the guide ring 806 is installed on the guide frame 8013.

[0048] like Figure 1As shown, the buffer frame 4 is provided with a buffer pad 401 made of sponge, and side guards 402 are provided on both sides of the buffer frame 4. When the simulation test vehicle 2 fails to be recovered and flies out of the guide rail on the lower side of the top plate 101, it can fall on the buffer pad 401, thereby effectively protecting the simulation test vehicle 2.

[0049] like Figure 8 As shown, in this embodiment, a steering gear 2011 is provided at the lower side of the simulation test vehicle 2, and the arresting hook 201 is driven by the steering gear 2011 to rotate and adjust the angle so that it can be hooked on the recovery rope 5. The steering gear 2011 is a well-known technology in the art and is a commercially available product.

[0050] The parameter capture module on the simulation test vehicle 2 includes an SBG-IMU module and an accelerometer. In addition, a differential GPS positioning device is installed to obtain positioning information with an accuracy of 1 mm and a collection frequency of 50 Hz. At the same time, the module can obtain the attitude angle of the simulation test vehicle 2 and store it in the vehicle SD card. The SBG-IMU module, accelerometer, and differential GPS positioning device are all well-known technologies in the art. In addition, Figure 1 As shown, in this embodiment, the PTZ camera 7 uses a 32x slow-motion camera with a maximum capture speed of 960 frames per second, which can capture the moment when the arresting hook 201 hooks the recovery rope 5, and the shape changes of the arresting hook 201 and the recovery rope 5. The PTZ camera 7 is a commercially available product.

[0051] The working principle of the present invention is:

[0052] When the system of the present invention is working, the simulation test vehicle 2 moves along the lower side of the top 101 of the guide frame body 1 to simulate the recovery movement process of the UAV, and the two ends of the recovery rope 5 are respectively passed around the recovery rope steering assembly 6 on the corresponding side and then wound on the winding wheel 802 of the recovery rope retracting assembly 8 on the corresponding side, and the winding wheel 802 is driven to rotate by the retracting drive device 805 to realize the retracting and releasing of the recovery rope 5, thereby realizing the tensioning or loosening of the recovery rope 5, and the recovery rope steering assembly 6 is arranged on the slider 302 on the recovery rope height assembly 3 on the corresponding side, and the height is adjusted by the slider 302. The tension sensor 603 on the recovery rope steering assembly 6 guides the recovery rope 5 to turn while also monitoring the tension of the recovery rope 5 in real time. The simulation test vehicle 2 is provided with a parameter capture module to obtain the posture information of the simulation test vehicle 2 in real time. The pan-tilt camera 7 is used to capture the moment when the arresting hook 201 hooks the recovery rope 5, and the shape changes of the arresting hook 201 and the recovery rope 5. The parameter capture module, the tension sensor 603 and the pan-tilt camera 7 obtain data and transmit it to the system analysis module for analysis, and the arresting hook 201 and related mechanism structures are designed according to the analysis results.

[0053] The method of the present invention comprises the following steps:

[0054] Step 1: Build a neural network.

[0055] like Fig. 9 As shown, the present invention utilizes the simulation test vehicle 2 to simulate the high-speed recovery of the fixed-wing UAV, and the system analysis module utilizes the data generated during the recovery process to construct a 4-layer neural network, which includes 23 neurons, including 7 neurons in the input layer and 5 neurons in the output layer.

[0056] The 7 neurons in the input layer represent the UAV's incident speed, material diameter, arresting hook material, arresting hook length, arresting hook end joint elasticity, arresting hook end joint damping, and the UAV's incident height relative to the recovery rope 5. In this embodiment, the specific acquisition process of the above neurons is as follows: the UAV speed is obtained by the IMU device in the parameter capture module; the material diameter is the diameter of the arresting hook 201; the hook material uses discrete points: aluminum alloy (1), steel (2), and titanium alloy (3); the hook length is the total length of the arresting hook 201; the joint elasticity is determined by the different elastic properties of the coil spring connecting the end of the arresting hook 201 and the servo 2011; the joint damping is determined by the properties of the damping oil used by the arresting hook 201; the incident height is confirmed by the height of the slider 302 in the recovery rope height adjustment component 3.

[0057] The five neurons in the output layer represent the five initial tension states of the recovery rope 5: state A (very tight), state B (tight), state C (normal), state D (loose), and state E (very loose). The present invention adopts an S-type neural network, that is, the output of each neuron is:

[0058]

[0059] In the above formula (1), x j represents the output value of the jth neuron in the previous layer, w j represents the weight value of the jth neuron in the previous layer, and b represents the bias of this neural network.

[0060] Step 2: Obtain a variety of recovery test results through tests under a variety of different input conditions.

[0061] Step 3: Use the input conditions and test results of step 2 as input and output parameters to train the neural network, and use the descent gradient method to iteratively solve the weight value and bias of the neural network. The formula is as follows:

[0062]

[0063] In the above formula (2), m is the number of points in the test result data set, y is the output result value of the test, and h is the value predicted by our neural network. Therefore, J(w, b) is a function of the weight value w and the bias b. By solving the optimal solution of J(w, b), the parameters in the neural network are determined.

[0064] Step 4: Design the arresting hook 201 according to the optimal solution parameters obtained in step 3.

[0065] An example is listed below for further explanation.

[0066] Step 1: Build Fig. 9 The neural network shown.

[0067] Step 2: Conduct tests under various input conditions and obtain different recovery test results, where the input variables are shown in Table 1 below, including:

[0068]

[0069] Table 1

[0070] The final test statistical results under different input conditions are shown in Table 2 below:

[0071]

[0072] Table 2

[0073] Step 3: Train the neural network based on the input conditions and test results of step 2 as input and output parameters, and calculate the parameters corresponding to all neurons. Parameter list 3 is as follows:

[0074] serial number Input Layer Hidden layer 1 Hidden Layer 2 Output Layer 1 w = 0.24, b = 0.026 w = 0.04, b = 0.13 w=0.02, b=0.02 w = 0.03, b = 0.16 2 w = 0.12, b = 0.01 w=0.04,b=0.05 w = 0.01, b = 0.07 w=0.012,b=0.17 3 w=-0.03,b=0.18 w=-0.18,b=0.17 w = -0.1, b = 0.17 w=-0.11,b=0.06 4 w = -0.2, b = 0.07 w=-0.12,b=0.28 w = -0.12, b = 0.13 w=-0.07,b=0.076 5 w = 0.02, b = 0.04 w = 0.02, b = 0.20 w=0.06,b=0.09 w = 0.03, b = 0.02 6 w=0.03,b=0.09 w = 0.07, b = 0.14 7 w = -0.06, b = 0.13

[0075] Table 3

[0076] According to the above parameters and formula (2), the weight value and bias J(w, b) of the neural network are iteratively solved by the descent gradient method, and the optimal solution of J(w, b) is solved. The optimal solution obtained is shown in Table 4 below.

[0077] serial number Input Layer Hidden layer 1 Hidden Layer 2 Output Layer 1 w = 0.24, b = 0.026 w=0.04.b=0.13 w=0.02, b=0.02 w=0.03.b=0.16 2 w=0.12.b=0.01 w=0.04.b=0.05 w = 0.01, b = 0.07 w=0.012,b=0.17 3 w=-0.03,b=0.18 w=-0.18,b=0.17 w=-0.1.b=0.17 w=-0.11,b=0.06

[0078] Table 4

[0079] Step 4: Design the arresting hook 201 according to Table 4 above. The optimal parameters of the arresting hook 201 are: hook diameter 1.5mm, incident speed 12m / s, hook incident height 23cm, coil spring elasticity 0.3Nm / °, and a certain amount of damping oil.

Claims

1. An arresting hook test system for offshore recovery of fixed-wing UAVs, characterized by: The invention comprises a guide frame (1), a simulation test vehicle (2), a recovery rope height adjustment assembly (3), a buffer frame (4), a recovery rope (5), a recovery rope steering assembly (6) and a recovery rope retracting assembly (8), wherein the simulation test vehicle (2) is movably arranged in the guide frame (1), the buffer frame (4) is arranged on one side of the guide frame (1), and the recovery rope height adjustment assembly (3) is arranged on both sides of the buffer frame (4), the recovery rope height adjustment assembly (3) is provided with a liftable slider (302), and the slider (302) is provided with a recovery rope steering assembly (6) The recovery rope steering assembly (6) is provided with a tension sensor (603), and each recovery rope height adjustment assembly (3) is provided with a recovery rope retracting assembly (8) at the bottom end, and the two ends of the recovery rope (5) are respectively wound around the tension sensor (603) on the recovery rope steering assembly (6) on the corresponding side and then wound in the recovery rope retracting assembly (8) on the corresponding side. The lower side of the simulation test vehicle (2) is provided with an arresting hook (201) to cooperate with the recovery rope (5), and the simulation test vehicle (2) is provided with a parameter capture module, and any recovery rope height adjustment assembly (3) is provided with a pan-tilt camera (7).

2. The arresting hook test system for offshore recovery of fixed-wing UAV according to claim 1 is characterized by: The guide frame (1) is provided with a top plate (101) at the upper end, the simulation test vehicle (2) comprises a vehicle body (202), and the upper side of the vehicle body (202) is slidably connected to the top plate (101), and both sides of the vehicle body (202) are provided with elastic cylinders (205), connecting rod arm assemblies (206) and wheels (207), wherein the ends of the wheel axles of the wheels (207) are provided with wheel seats (2071), one end of the connecting rod arm assembly (206) is hinged to the vehicle body (202), and the other end is hinged to the wheel seat (2071) on the corresponding side, and the lower end of the elastic cylinder (205) is hinged to the lower side of the vehicle body (202), and the upper end is hinged to the connecting rod arm assembly (206) on the corresponding side.

3. The arresting hook test system for recovering fixed-wing UAV at sea according to claim 2 is characterized in that: A guide rail is provided on the lower side of the top plate (101), a guide suspension rod (203) is provided on the upper side of the vehicle body (202), and a guide pulley (204) is provided on the upper end of the guide suspension rod (203) to cooperate with the guide rail on the lower side of the top plate (101).

4. The arresting hook test system for offshore recovery of fixed-wing UAV according to claim 2 is characterized by: The link arm assembly (206) comprises a first link arm (2061) and a second link arm (2062), wherein one end of the first link arm (2061) is hinged to the upper side of the corresponding side wheel seat (2071), and the other end is hinged to the upper side of the vehicle body (202); one end of the second link arm (2062) is hinged to the lower side of the corresponding side wheel seat (2071), and the other end is hinged to the lower side of the vehicle body (202); the upper end of the elastic cylinder (205) is hinged to the first link arm (2061).

5. The arresting hook test system for recovering fixed-wing UAV at sea according to claim 1 is characterized in that: The recovery rope height adjustment assembly (3) comprises a support rod (304), a belt (303) arranged on the support rod (304), a pulley and an adjustment drive device (301), wherein pulleys are arranged at both the upper and lower ends of the support rod (304), the two ends of the belt (303) respectively pass around the pulleys at the corresponding ends, and the pulley at either end is driven to rotate by the adjustment drive device (301), and the slider (302) is fixedly connected to one side of the belt (303).

6. The arresting hook test system for recovering fixed-wing UAV at sea according to claim 1 is characterized by: The recovery rope steering assembly (6) comprises a first guide element (601), a tension sensor (603), a first tensioning clamp (605), a second tensioning clamp (602) and a second guide element (606). The recovery rope (5) is guided to input via the first guide element (601), bypasses the tension sensor (603), passes between the first tensioning clamp (605) and the second tensioning clamp (602), and is guided to output via the second guide element (606).

7. The arresting hook test system for recovering fixed-wing UAV at sea according to claim 6 is characterized by: The recovery rope steering assembly (6) is provided with a spring shaft (607) and a spring (604), and the spring shaft (607) passes through the first tensioning clamp (605) and the second tensioning clamp (602), and the spring shaft (607) is provided with a spring (604) and a spring pressure plate at one end close to the first tensioning clamp (605), and one end of the spring (604) is against the spring pressure plate, and the other end is against the first tensioning clamp (605).

8. The arresting hook test system for recovering fixed-wing UAV at sea according to claim 1 is characterized by: The recovery rope retracting assembly (8) comprises a mounting frame (801), a winding wheel (802) and a retracting drive device (805) arranged on the mounting frame (801), a guide ring (806) being arranged at the upper end of the mounting frame (801), the recovery rope (5) passing through the guide ring (806) and then being wound around the winding wheel (802), and the winding wheel (802) is driven to rotate by the retracting drive device (805).

9. A method for testing an arresting hook system for recovering a fixed-wing UAV at sea according to claim 1, characterized in that: The steps include: Step 1: Build a neural network; Step 2: Obtain multiple recovery test results through tests under multiple different input conditions; Step 3: Use the input conditions and test results of step 2 as input and output parameters to train the neural network, and use the descent gradient method to iteratively solve the weight value and bias of the neural network. The formula is as follows: In the above formula, m is the number of points in the test result data set, y is the output result value of the test, h is the value predicted by the neural network, and J(w, b) is a function of the weight value w and the bias b. By solving the optimal solution of J(w, b), the parameters in the neural network are determined; Step 4: Design the arresting hook (201) according to the optimal solution parameters.

10. The method for testing the arresting hook system for recovering fixed-wing UAV at sea according to claim 9, characterized in that: In step 1, a 4-layer neural network is constructed using the data generated in the simulated recycling process, wherein the neural network includes 23 neurons, including 7 neurons in the input layer and 5 neurons in the output layer; The 7 neurons in the input layer represent the UAV’s impact velocity, material diameter, arresting hook material, arresting hook length, arresting hook end joint elasticity, arresting hook end joint damping, and the UAV’s impact height relative to the recovery rope (5). The five neurons in the output layer represent the five initial tension states of the recovery rope (5): state A (very tight), state B (tight), state C (normal), state D (loose), and state E (very loose); Using an S-type neural network, the output of each neuron is: In the above formula, x j represents the output value of the jth neuron in the previous layer, w j represents the weight value of the jth neuron in the previous layer, and b represents the bias of this neural network.

Images