A multi-redundant flight control method and system for unmanned aerial vehicles

Abstract

The present invention provides a multi-redundant flight control method and system for an unmanned aerial vehicle (UAV), relating to the field of control or regulation systems for non-electrical variables. The method includes: a flight control arbitration master control device uses an arbitration algorithm to use the data of one of the flight control modules as actual output data based on the electronic control and servo data of all connected flight control modules, and uses a hot-swap detection I / O port to detect whether the master control side interface and the module side interface are connected; if connected, power is supplied to the corresponding flight control module, data is sent to the flight control arbitration master control, and flight control status information is updated; if the connection is disconnected, the corresponding flight control module is powered off and the flight control status information is updated. In the present invention, flight control modules can be dynamically connected or removed during UAV operation without the need for a system restart, significantly reducing maintenance downtime. After a flight control module is connected, power is automatically supplied and the module is included in the arbitration process. After a flight control module is disconnected, power is immediately disconnected and the status is updated, ensuring that the system always operates based on the currently available set of flight control modules.

Description

Technical Field

[0001] The present invention relates to the field of control or regulation systems for non-electrical variables, and in particular to a multi-redundant flight control method and system for unmanned aerial vehicles. Background Art

[0002] Unmanned aerial vehicles (UAVs), also known as drones, are unmanned aircraft controlled by radio remote control and self-contained program control devices, or operated completely or intermittently autonomously by an onboard computer. Currently, with the development of science and technology, UAV technology is becoming increasingly mature, and UAVs are widely used due to their high speed and flexible operation.

[0003] A drone's flight control system is the core control brain, responsible for controlling the entire flight process. A failure can lead to a crash or other serious consequences. Therefore, redundant backups are essential. Conventional backups involve dual-redundant hot backups, but this backup mechanism suffers from slow speeds and limited stability. Summary of the Invention

[0004] The purpose of the present invention is to provide a multi-redundant flight control method and system for an unmanned aerial vehicle, which aims to overcome the above-mentioned problems existing in the prior art.

[0005] To achieve the purpose, the present invention provides the following technical solutions:

[0006] A multi-redundant flight control method and system for an unmanned aerial vehicle (UAV) includes a flight control arbitration master control device and multiple flight control modules. The flight control arbitration master control device includes a power control module and multiple master control-side interfaces. The power control module is connected to each flight control module, and the flight control modules include module-side interfaces that cooperate with the master control-side interfaces. The method includes the following steps:

[0007] Step S100: The flight control arbitration master control device uses an arbitration algorithm based on the ESC and servo data of all currently connected flight control modules to select the data of one of the flight control modules as the actual output data;

[0008] Step S200: The flight control arbitration master control device detects whether the master control side interface and the module side interface are connected through hot plug detection I / O port;

[0009] Step S300: If the master control side interface is connected to the module side interface, the power control module supplies power to the corresponding flight control module; after the flight control module is powered on, it sends data to the flight control arbitration master control via the CAN bus, and the flight control arbitration master control device updates the flight control status information; then jump to step S100;

[0010] Step S400: If the connection between the master control side interface and the module side interface is disconnected, the power control module cuts off the power to the corresponding flight control module, and the flight control arbitration master control device updates the flight control status information; then jumps to step S100.

[0011] Furthermore, the hot-swap detection I / O port includes a first pin and a second pin, the first pin is fixedly grounded, and the second pin is connected to a preset voltage through a weak pull-up resistor and maintained in a high level state; the two pins of the module-side interface corresponding to the first pin and the second pin are connected to each other;

[0012] The above step S300 specifically includes: when the module side interface is fully inserted into the master control side interface, the module side interface short-circuits the second pin and the first pin, and at this time the arbitration master control device detects that the second pin is at a low level;

[0013] The above step S400 specifically includes: when the module side interface is not plugged into the master control side interface, the second pin is disconnected from the first pin, and the arbitration master control device detects that the second pin is high level through the hot plug detection I / O port.

[0014] Furthermore, the above step S100 specifically includes the following steps:

[0015] Step S110: The flight control arbitration master control device obtains the real-time ESC values ​​and real-time servo values ​​of each flight control module in real time during the current sampling period, and obtains historical data of each flight control module, including the confidence level of the ESC values ​​and servo values ​​of the flight control module over the past period of time, and the connection duration between the flight control module and the flight control arbitration master control device;

[0016] Step S120: The neural network model outputs the weight value of the corresponding flight control module according to the input confidence and connection duration.

[0017] Step S130: Calculate the weighted average value E1 of the electric control according to the weight values ​​and real-time electric control values ​​of all currently connected flight control modules. avg ;

[0018] Step S140: Calculate the weighted average value R1 of the servo according to the weight values ​​of all currently connected flight control modules and the real-time servo values. avg ;

[0019] Step S150: Set the ESC error threshold to ±ΔE1 and the servo error threshold to ±ΔR1; determine the actual ESC error value In (E1 i ) = 100 × (E1 i -E1 avg ) / E1 avg , the actual error value of the servo In (R1 i ) = 100 × (R1 i-△R1) / △R1; select the actual error value of the ESC In (E1 i ) falls within the range of ±ΔE1, and the actual error value of the servo In (R1 i ) The flight control modules that fall within the range of ±ΔR1 are selected and their data are used as preliminary candidate data; where i represents the i-th flight control module, i=1, 2, 3, ..., M;

[0020] Step S160: Perform a second round of weighted average calculation on the preliminary candidate data, and then perform a second round of actual error calculation, and calculate the actual error value In (E1) of the second round. i ) absolute value and the actual error value of the servo In (R1 i ) The data of the flight control module with the smallest sum of absolute values ​​is used as the actual output data.

[0021] Furthermore, the confidence calculation steps are as follows:

[0022] Step S111: Set the time window length to ΔT and the sampling interval to Δt, then the number of data points in the window is N=T / Δt;

[0023] Step S112: For the i-th flight control module, calculate the deviation between the real-time data and the historical mean:

[0024] ;

[0025] in, is the ESC + servo vector of the kth sampling; , is the historical mean and standard deviation of the previous k-1 times;

[0026] Step S113: Update the abnormality count:

[0027] ;

[0028] Among them, α∈(0,1) is the forgetting factor, θ is the abnormal threshold, and I(·) is the indicator function;

[0029] Step S114: Calculate the basic confidence:

[0030] ;

[0031] Step S115: Correct the basic confidence and calculate the confidence:

[0032] ;

[0033] Among them, β∈(0,1) is the control time gain amplitude, is the connection duration, λ is the time gain coefficient (controls the correction speed, 0.1 is recommended).

[0034] Furthermore, the input layer of the above neural network model has two input nodes, namely confidence , connection duration The hidden layer has four nodes, the activation function uses LeakyReLU (α=0.1), and the output layer has one output node, using the Softmax function.

[0035] The mathematical expression of the above neural network model for the i-th flight control module is as follows:

[0036] ;

[0037] in, is the first layer weight, is the first layer bias, is the second layer weight, is the second layer bias;

[0038] Weight normalization:

[0039] ,make sure .

[0040] Furthermore, calculate the weighted average value of the electrical adjustment:

[0041] ;

[0042] in, is the weight value of the i-th flight control module, is the real-time electrical adjustment value of the i-th flight control module;

[0043] Calculate the weighted average value of the servos:

[0044] ;

[0045] in, is the weight value of the i-th flight control module, is the real-time servo value of the i-th flight control module.

[0046] The present invention also discloses a multi-redundant flight control system for an unmanned aerial vehicle, which is used to execute any of the above-mentioned flight control methods. The system includes a flight control arbitration master control device and multiple flight control modules. The flight control arbitration master control device is provided with a power control module and multiple master control side interfaces, and the master control side interfaces have hot plug detection I / O ports; the flight control modules are provided with module side interfaces that cooperate with the master control side interfaces;

[0047] The above-mentioned hot plug detection I / O port is used to detect whether the master control side interface and the module side interface are connected, and the above-mentioned power control module is used to power on or off the corresponding flight control module according to the detection result of the hot plug detection I / O port; the above-mentioned flight control arbitration master control device uses the arbitration algorithm based on the electronic control and servo data of all currently connected flight control modules, and uses the data of one of the flight control modules as the actual output data.

[0048] Furthermore, the hot plug detection I / O port includes a first pin and a second pin, wherein the first pin is fixedly grounded, and the second pin is connected to a preset voltage via a weak pull-up resistor to keep the second pin at a high level;

[0049] The two pins corresponding to the first and second pins of the module-side interface are connected to each other; when the module-side interface is fully inserted into the master-control-side interface, the second pin is connected to the first pin, thereby causing the electrical level of the second pin to become low; when the control arbitration master control device detects through the hot-swap detection I / O port that the electrical level of the second pin becomes low, the power control module supplies power to the flight control module corresponding to the master-control-side interface; after the flight control module is powered on, it sends data to the flight control arbitration master control through the CAN bus, and the flight control arbitration master control device updates the flight control status information; when the control arbitration master control device detects through the hot-swap detection I / O port that the electrical level of the second pin becomes high, the power control module cuts off the power supply to the corresponding master-control-side interface.

[0050] Furthermore, the above-mentioned master control side interface is a PCIE interface with an A / B double-sided pin layout, the first pin is PRSNT1#, defined on the 12th pin of the A side; the second pin is PRSNT2#, defined on the 12th pin of the B side.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] First, the present invention achieves full automation of the "live plug-in / out - state synchronization - arbitration adaptation" process through the physical design of the module-side interface and the master control-side interface, combined with the on-demand power supply logic of the power control module. The flight control module can be dynamically connected or removed during drone operation without requiring a system restart, significantly reducing maintenance downtime. Once connected, the flight control module automatically powers on and participates in the arbitration process. Once disconnected, the module immediately powers off and updates its status, ensuring that the system always operates based on the currently available set of flight control modules, preventing the failure of a single flight control module from causing overall system failure.

[0053] Secondly, in the present invention, the arbitration algorithm integrates the ESC and servo data of multiple flight control modules, and effectively suppresses the impact of abnormalities or failures of a single flight control module through weighted averaging, threshold screening and other processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1The figure is a simplified flowchart of a multi-redundant flight control method in the present invention.

[0055] Figure 2 This is a simplified structural block diagram of a multi-redundant flight control system in the present invention.

[0056] Figure 3 This is a schematic diagram of a hot-swap detection structure for a master control side interface and a module side interface in the present invention. DETAILED DESCRIPTION

[0057] The specific embodiments of the present invention are described below with reference to the accompanying drawings. In order to fully understand the present invention, many details are described below, but for those skilled in the art, the present invention can be implemented without these details.

[0058] like Figures 1 to 3 As shown, a multi-redundant flight control method and system for an unmanned aerial vehicle (UAV) is provided. The multi-redundant flight control system for the UAV includes a flight control arbitration master control device and multiple flight control modules. The flight control arbitration master control device is provided with a power control module and multiple master control side interfaces, and the master control side interfaces have hot plug detection I / O ports. The power control module is connected to each flight control module, and the flight control module is provided with a module side interface that cooperates with the master control side interface. The method includes the following steps:

[0059] Step S100: The flight control arbitration master control device uses an arbitration algorithm based on the ESC and servo data of all currently connected flight control modules to take the data of one of the flight control modules as the actual output data.

[0060] Step S200: The flight control arbitration master control device detects whether the master control side interface and the module side interface are connected by hot plugging the I / O port.

[0061] Step S300: If the master control side interface is connected to the module side interface, the power control module supplies power to the corresponding flight control module; after the flight control module is powered on, it sends data to the flight control arbitration master control through the CAN bus, and the flight control arbitration master control device updates the flight control status information; then jump to step S100.

[0062] Step S400: If the connection between the master control side interface and the module side interface is disconnected, the power control module cuts off the power to the corresponding flight control module, and the flight control arbitration master control device updates the flight control status information; then jumps to step S100.

[0063] The present invention realizes the full process automation of "live plug-in-and-remove-state synchronization-arbitration adaptation" through the physical design of the module-side interface and the master control-side interface, combined with the on-demand power supply logic of the power control module. Specifically, the flight control module can be dynamically connected or removed during the operation of the UAV without the need for a system restart, significantly reducing maintenance downtime. After the flight control module is connected, it is automatically powered on and joins the arbitration process. After the flight control module is disconnected, it is immediately powered off and the status is updated to ensure that the system always operates based on the currently available set of flight control modules, avoiding the overall failure caused by the failure of a single flight control module. The arbitration algorithm integrates the electronic control and servo data of multiple flight control modules, and effectively suppresses the impact of abnormalities or failures of a single flight control module through weighted averaging, threshold screening and other processing.

[0064] like Figure 3 As shown, in a specific embodiment, the hot plug detection I / O port includes a first pin and a second pin, the first pin is fixedly grounded, and the second pin is connected to a preset voltage through a weak pull-up resistor and maintained in a high level state; the two pins (i.e., pin 1 and pin 2) corresponding to the first pin and the second pin of the module side interface are connected to each other;

[0065] In a specific embodiment, step S300 is specifically: when the module side interface is fully inserted into the master control side interface, the module side interface short-circuits the second pin and the first pin, and at this time the flight control arbitration master control device detects that the second pin is at a low level.

[0066] In a specific embodiment, step S400 is specifically: when the module side interface is not inserted into the master control side interface, the second pin and the first pin are disconnected from each other. At this time, the flight control arbitration master control device detects that the second pin is high level through the hot plug detection I / O port.

[0067] In a specific embodiment, when the operator manually inserts the module-side interface into the master-side interface, step S300 is executed; when the operator manually unplugs the module-side interface from the master-side interface, step S400 is executed.

[0068] The flight control arbitration master control device updates flight control status information as mentioned in steps 300 and 400. Specifically, the flight control arbitration master control, as the core management unit in the multi-flight control system, establishes a real-time data channel through a communication protocol to continuously obtain configuration parameters (such as hardware configuration information and software version information) and flight data (such as sensor calibration parameters, preset flight paths, and dynamic flight status data) from each currently connected flight control module. Furthermore, the flight control arbitration master control device creates a three-dimensional digital twin model containing static features and dynamic trajectories for each flight control module through a combination of memory mapping and persistent storage, managing multiple flight control modules simultaneously. When the flight control arbitration master control detects that the flight control module corresponding to the actual output data (referred to as the original flight control module) has an abnormal operating status (such as a control response delay exceeding a threshold or abnormal sensor data), the flight control arbitration master control system sends an alert to the ground station. The operator replaces the original flight control module by plugging and unplugging the module-side interface and the master control-side interface. After replacing the new flight control module, the flight control arbitration master control synchronizes the complete configuration set of the original flight control module (including coordinate system settings, PID parameters, waypoint sequence and other key information) to the new flight control module, realizing hot-swap replacement in software.

[0069] like Figures 1 to 3 As shown, in a specific embodiment, step S100 specifically includes the following steps:

[0070] Step S110: The flight control arbitration master control device obtains the real-time ESC values ​​and real-time servo values ​​of each flight control module in real time during the current sampling period, and obtains historical data for each flight control module, including the confidence level of the ESC values ​​and servo values ​​of the flight control module over the past period, as well as the connection duration between the flight control module and the flight control arbitration master control device. The "connection duration" refers to the duration that the flight control module is currently connected to the flight control arbitration master control device, i.e., the current connection duration. If power is restored after a flight control module is powered off, its connection duration must be accumulated again. The "past period" can refer to the entire current connection duration or a preset duration within the current connection duration. For example, if the current connection duration is accumulated for 1500 sampling periods, the 1000 sampling periods prior to the current time point are set as the "past period." In this case, the confidence level of the ESC values ​​and servo values ​​over the past period is calculated by obtaining the ESC values ​​and servo values ​​of the flight control module for the 1000 sampling periods prior to the current time point and performing confidence calculations.

[0071] In this step, real-time (i.e., real-time data sampling) and persistence (i.e., confidence and connection duration) are combined to form a hybrid decision-making basis of "immediate status + long-term reliability", which is better than the traditional method that relies only on real-time data.

[0072] Step S120: The neural network model outputs a weight value of the corresponding flight control module according to the input confidence and connection duration.

[0073] In this step, the neural network model can capture the complex coupling relationship between confidence and connection duration. The weight value changes nonlinearly with the input features, can adapt to complex scenarios, and can dynamically adapt to changes in the characteristics of different flight control modules (such as confidence attenuation caused by aging of the flight control module).

[0074] Step S130: Calculate the weighted average value E1 of the electric control according to the weight values ​​and real-time electric control values ​​of all currently connected flight control modules. avg .

[0075] Step S140: Calculate the weighted average value R1 of the servo according to the weight values ​​of all currently connected flight control modules and the real-time servo values. avg .

[0076] Step S150: Set the ESC error threshold to ±ΔE1 and the servo error threshold to ±ΔR1; determine the actual ESC error value In (E1 i ) = 100 × (E1 i -E1avg) / E1 avg , the actual error value of the servo In (R1 i ) = 100 × (R1 i -△R1) / △R1; select the actual error value of the ESC In (E1 i ) falls within the range of the ESC error threshold ±ΔE1, and the actual servo error value In (R1 i ) The flight control modules that fall within the range of the servo error threshold ±ΔR1 are selected and their data are used as preliminary candidate data; where i represents the i-th flight control module, i=1, 2, 3, …, M.

[0077] Step S160: Perform a second round of weighted average calculation on the preliminary candidate data, and then perform a second round of actual error calculation, and calculate the actual error value In (E1) of the second round. i ) absolute value and the actual error value of the servo In (R1 i ) The data of the flight control module with the smallest sum of absolute values ​​is used as the actual output data.

[0078] The neural network weight calculation in step S120 provides pre-screening optimization for the threshold screening in step S150, reducing the amount of data to be processed. The second round of screening in step S160, based on the preliminary candidate dataset from step S150, further utilizes this weight information to improve output accuracy, forming a progressive decision chain from "coarse-grained weights to fine-grained errors."

[0079] like Figures 1 to 3As shown, in a specific embodiment, the confidence calculation step of step S110 is as follows:

[0080] Step S111: Set the time window length to ΔT and the sampling interval to Δt, then the number of data points in the window is N=T / Δt.

[0081] Step S112: For the i-th flight control module, calculate the deviation between the real-time data and the historical mean:

[0082] ;

[0083] in, is the ESC+servo vector of the kth sampling; , is the historical mean and standard deviation of the previous k-1 times.

[0084] Step S113: Update the abnormality count:

[0085] ;

[0086] Among them, α∈(0,1) is the forgetting factor, and its preferred value is 0.95; θ is the abnormality threshold, and its preferred value is 3 according to the 3σ principle in statistics; I(·) is the indicator function.

[0087] Step S114: Calculate the basic confidence:

[0088] ;

[0089] Step S115: Correct the basic confidence and finally calculate the confidence:

[0090] ;

[0091] Among them, β∈(0,1) is the control time gain amplitude; is the connection duration; λ is the time gain coefficient, which is used to control the correction speed, and the preferred value is 0.1).

[0092] In the above formula, for the newly connected flight module, that is, :

[0093] , retain the basic confidence;

[0094] For long-term connected flight modules, i.e. :

[0095] , the time gain is significant.

[0096] In a specific embodiment, the neural network model of step S120 includes:

[0097] The input layer has two input nodes, namely confidence , connection duration .

[0098] The hidden layer has four nodes, and the activation function uses LeakyReLU (α=0.1).

[0099] The output layer has one output node and uses the Softmax function.

[0100] The mathematical expression of the neural network model for the i-th flight control module is as follows:

[0101] ;

[0102] in, is the first layer weight, is the first layer bias, is the second layer weight, is the second layer bias;

[0103] Then normalize the weights output by the neural network model:

[0104] ,make sure .

[0105] In a specific embodiment, the weighted average value E1avg of the electrical adjustment in step S130 is calculated as follows:

[0106] ;

[0107] in, is the weight value of the i-th flight control module, is the real-time electrical regulation value of the i-th flight control module.

[0108] The weighted average value R1 of the servo in step S240 avg , the calculation formula is as follows:

[0109] ;

[0110] in, is the weight value of the i-th flight control module, is the real-time servo value of the i-th flight control module.

[0111] like Figure 1-Figure 3 As shown, a multi-redundant flight control system for an unmanned aerial vehicle is used to execute the above-mentioned flight control method. The system includes a flight control arbitration master control device and multiple flight control modules. Of course, the multi-redundant flight control system of the unmanned aerial vehicle also includes other conventional components, which will not be described in detail here.

[0112] Among them, the flight control arbitration master control device is equipped with a power control module and multiple master control side interfaces, and the master control side interface has a hot plug detection I / O port.

[0113] The flight control module is equipped with a module-side interface that matches the master control interface. Multiple flight control modules are connected to the flight control arbitration master control device in parallel through the connection between the module-side interface and the master control interface.

[0114] The hot-swap detection I / O port is used to detect whether the master control side interface and the module side interface are connected.

[0115] The power control module is used to power on or off the corresponding flight control module according to the detection result of the hot plug detection I / O port.

[0116] The flight control arbitration master control device uses the ESC and servo data from all currently connected flight control modules, employing an arbitration algorithm to determine the output data from one flight control module. This output data is then converted into a drive signal via the actuator interface card, enabling real-time control of the ESC speed and rudder deflection, achieving closed-loop tracking of the drone's flight attitude.

[0117] Specifically, the hot plug detection I / O port includes a first pin and a second pin, wherein the first pin is fixedly grounded, and the second pin is connected to a preset voltage through a weak pull-up resistor to keep the second pin at a high level;

[0118] The two pins of the module-side interface corresponding to the first pin and the second pin are connected to each other; when the module-side interface is fully inserted into the master-control-side interface, the second pin is connected to the first pin, thereby causing the electrical level of the second pin to become low; when the flight control arbitration master control device detects that the electrical level of the second pin becomes low, the power control module supplies power to the flight control module corresponding to the master-control-side interface; after the flight control module is powered on, it sends data to the flight control arbitration master control via the CAN bus, and the flight control arbitration master control device updates the flight control status information; when the flight control arbitration master control device detects that the electrical level of the second pin becomes high, the power control module cuts off the power supply to the corresponding master-control-side interface.

[0119] Preferably, the master control interface is a PCIE interface with an A / B double-sided pin layout. The first pin is PRSNT1#, defined on the A side, pin 12. The second pin is PRSNT2#, defined on the B side, pin 12. Of course, the master control interface is not limited to a PCIE interface and can include network ports, USB ports, MIL-DTL-38999 ports, and other interfaces, simply by leveraging PCIE's hot-swap pin detection technology.

[0120] The above is only a specific implementation of the present invention, but the design concept of the present invention is not limited to this. Any non-substantial changes to the present invention using this concept shall be deemed as an infringement of the protection scope of the present invention.

Claims

1. A multi-redundant flight control method for an unmanned aerial vehicle, characterized by: The multi-redundant flight control system of the UAV includes a flight control arbitration master control device and multiple flight control modules. The flight control arbitration master control device is equipped with a power control module and multiple master control side interfaces, and the master control side interface has a hot plug detection I / O port. The power control module is connected to each flight control module, and the flight control module is equipped with a module side interface that cooperates with the master control side interface. The method comprises the following steps: Step S100: The flight control arbitration master control device uses an arbitration algorithm based on the ESC and servo data of all currently connected flight control modules to select the data of one of the flight control modules as the actual output data. Step S100 specifically includes the following steps: Step S110: The flight control arbitration master control device obtains the real-time ESC values ​​and real-time servo values ​​of each flight control module in real time during the current sampling period, and obtains historical data of each flight control module, including the confidence level of the ESC values ​​and servo values ​​of the flight control module over the past period of time, and the connection duration between the flight control module and the flight control arbitration master control device; Step S120: The neural network model outputs the weight value of the corresponding flight control module according to the input confidence and connection duration. Step S130: Calculate the weighted average value E1 of the electric control according to the weight values ​​and real-time electric control values ​​of all currently connected flight control modules. avg ; Step S140: Calculate the weighted average value R1 of the servo according to the weight values ​​of all currently connected flight control modules and the real-time servo values. avg ; Step S150: Set the ESC error threshold to ±ΔE1 and the servo error threshold to ±ΔR1; determine the actual ESC error value In (E1 i ) = 100 × (E1 i -E1 avg ) / E1 avg , the actual error value of the servo In (R1 i ) = 100 × (R1 i -△R1) / △R1; select the actual error value of the ESC In (E1 i ) falls within the range of ±ΔE1, and the actual error value of the servo In (R1 i ) The flight control modules that fall within the range of ±ΔR1 are selected and their data are used as preliminary candidate data; where i represents the i-th flight control module, i=1, 2, 3, ..., M; Step S160: Perform a second round of weighted average calculation on the preliminary candidate data, and then perform a second round of actual error calculation, and calculate the actual error value In (E1) of the second round. i ) absolute value and the actual error value of the servo In (R1 i ) The data of the flight control module with the smallest sum of absolute values ​​is used as the actual output data; Step S200: The flight control arbitration master control device detects whether the master control side interface and the module side interface are connected through hot plug detection I / O port; Step S300: If the master control side interface is connected to the module side interface, the power control module supplies power to the corresponding flight control module; after the flight control module is powered on, it sends data to the flight control arbitration master control via the CAN bus, and the flight control arbitration master control device updates the flight control status information; then jump to step S100; Step S400: If the connection between the master control side interface and the module side interface is disconnected, the power control module cuts off the power to the corresponding flight control module, and the flight control arbitration master control device updates the flight control status information; then jumps to step S100.

2. The multi-redundant flight control method for an unmanned aerial vehicle according to claim 1, characterized in that: The hot plug detection I / O port includes a first pin and a second pin, the first pin is fixedly grounded, and the second pin is connected to a preset voltage through a weak pull-up resistor and maintained in a high level state; the two pins of the module side interface corresponding to the first pin and the second pin are connected to each other; The step S300 specifically includes: when the module-side interface is fully inserted into the master-side interface, the module-side interface short-circuits the second pin and the first pin, and at this time, the flight control arbitration master control device detects that the second pin is at a low level; The step S400 is specifically as follows: when the module side interface is not plugged into the master control side interface, the second pin is disconnected from the first pin, and at this time the flight control arbitration master control device detects that the second pin is at a high level through the hot plug detection I / O port.

3. The multi-redundant flight control method for an unmanned aerial vehicle according to claim 1, characterized in that: The steps for calculating confidence are as follows: Step S111: Set the time window length to ΔT and the sampling interval to Δt, then the number of data points in the window is N=T / Δt; Step S112: For the i-th flight control module, calculate the deviation between the real-time data and the historical mean: ； in, is the ESC + servo vector of the kth sampling; , is the historical mean and standard deviation of the previous k-1 times; Step S113: Update the abnormality count: ； Among them, α∈(0,1) is the forgetting factor, θ is the abnormal threshold, and I(·) is the indicator function; Step S114: Calculate the basic confidence: ； Step S115: Correct the basic confidence and calculate the confidence: ； Among them, β∈(0,1) is the control time gain amplitude, is the connection duration, and λ is the time gain coefficient.

4. The multi-redundant flight control method for an unmanned aerial vehicle according to claim 3, characterized in that: The neural network model of step S120 includes the following specific network architecture: The input layer has two input nodes, namely confidence , connection duration ; The hidden layer has four nodes and the activation function uses LeakyReLU; The output layer has one output node and uses the Softmax function; The mathematical expression of the neural network model for the i-th flight control module is as follows: ； in, is the weight value of the i-th flight control module, is the first layer weight, is the first layer bias, is the second layer weight, is the second layer bias; Weight normalization: ,make sure .

5. The multi-redundant flight control method for an unmanned aerial vehicle according to claim 4, characterized in that: Calculate the weighted average value of the ESC: ； in, is the real-time electrical adjustment value of the i-th flight control module; Calculate the weighted average value of the servo: ； in, is the real-time servo value of the i-th flight control module.

6. A multi-redundant flight control system for an unmanned aerial vehicle, characterized by: Used to execute the flight control method according to any one of claims 1 to 5, the system includes a flight control arbitration master control device and multiple flight control modules, the flight control arbitration master control device is provided with a power control module and multiple master control side interfaces, and the master control side interfaces have hot plug detection I / O ports; the flight control modules are provided with module side interfaces that cooperate with the master control side interfaces; The hot-swap detection I / O port is used to detect whether the master control side interface and the module side interface are connected, and the power control module is used to power on or off the corresponding flight control module according to the detection result of the hot-swap detection I / O port; the flight control arbitration master control device uses the arbitration algorithm based on the electronic control and servo data of all currently connected flight control modules, and uses the data of one of the flight control modules as the actual output data.

7. The multi-redundant flight control system for an unmanned aerial vehicle according to claim 6, characterized in that: The hot plug detection I / O port includes a first pin and a second pin, wherein the first pin is fixedly grounded, and the second pin is connected to a preset voltage via a weak pull-up resistor to keep the second pin at a high level; The module-side interface is connected to two pins corresponding to the first pin and the second pin; when the module-side interface is fully inserted into the master-control-side interface, the second pin is connected to the first pin, thereby causing the electrical level of the second pin to become low; when the flight control arbitration master control device detects through the hot-swap detection I / O port that the electrical level of the second pin becomes low, the power control module supplies power to the flight control module corresponding to the master-control-side interface; after the flight control module is powered on, it sends data to the flight control arbitration master control through the CAN bus, and the flight control arbitration master control device updates the flight control status information; when the flight control arbitration master control device detects through the hot-swap detection I / O port that the electrical level of the second pin becomes high, the power control module cuts off the power supply to the corresponding master-control-side interface.

8. The multi-redundant flight control system for an unmanned aerial vehicle according to claim 6, characterized in that: The master control side interface is a PCIE interface with an A / B double-sided pin layout, the first pin is PRSNT1#, defined on the 12th pin of the A side; the second pin is PRSNT2#, defined on the 12th pin of the B side.

Images