Remote boot system and method with dual verification
The remote start system with dual verification utilizes the shadow simulation of the sensing module and the simulation terminal to generate a safety sign, which solves the problem of false start-up of the drone remote start system under abnormal operating conditions and improves safety and battery life.
Patent Information
- Application Number
- CN202611618146.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-10-20
- Publication Date
- 2026-11-17
AI Technical Summary
The remote start system for drones is prone to accidental activation under abnormal operating conditions, leading to safety accidents, and the high standby power consumption shortens the flight time.
The remote start system employs dual verification. The receiving end's sensing module collects a snapshot of the initial state of the real machine and the detection vector of the on-site environment. Combined with the shadow simulation at the simulation end, it generates a first safety sign. The judgment module then generates a second safety sign based on the acoustic wave signal and the back electromotive force signal, and finally generates the start strategy.
It effectively avoids accidental startup and safety accidents under abnormal operating conditions, improves the safety, reliability and environmental adaptability of remote startup, reduces power consumption and extends battery life.
Smart Images

Figure CN123113441A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the fields of unmanned aerial vehicles, wireless remote control and model car technology, and in particular to a remote start system and method with dual verification. Background Technology
[0002] A remote start system for unmanned aerial vehicles (UAVs) typically includes a transmitter and a receiver. The transmitter is equipped with a wireless remote control switch, while the receiver is equipped with an RF receiving module, a power supply module, and a controlled power execution module. The RF receiving module receives the wireless control signals generated by the remote control switch, and the power supply module supplies power to the power execution module according to the received control signals. This enables remote control of the UAV's power system from a distance. This type of remote start is widely used in scenarios such as UAV pre-operation debugging, transportation, and emergency start.
[0003] In existing technologies, the radio frequency receiving module and main control module of the receiver end need to be constantly powered on and listening to respond to the start command issued by the transmitter end. This results in the receiver end being in a high standby power consumption state for a long time, shortening the flight time of the onboard battery. At the same time, after receiving the start signal, the receiver end directly supplies power to the power execution module and immediately starts the power system. The start-up process does not rely on the detection of the drone's own status and the on-site environment. Under abnormal conditions such as insufficient battery voltage, magnetic compass interference, satellite positioning drift, or the presence of obstacles in the near field, the drone will still be started directly, which can easily lead to safety accidents such as false start of the power system, abnormal take-off attitude, or even collisions, posing obvious safety hazards. Summary of the Invention
[0004] In view of the above problems, this application provides a remote boot system and method with dual verification.
[0005] One aspect of this application provides a dual-verification remote start system, comprising: a transmitter that generates a remote start signal in response to a switch control signal from a wireless remote control switch; a receiver that receives the remote start signal and enters a pre-start mode in response to the remote start signal; the receiver includes a sensing module that collects a snapshot of the initial state of the actual machine, a field environment detection vector, and acoustic signals and motor back electromotive force signals during the pre-start process; and a simulation end that is communicatively connected to the receiver, performs shadow simulation based on the snapshot of the initial state of the actual machine and the field environment detection vector, generates a first safety flag, and transmits it to the determination module; wherein the receiver further includes a determination module that generates a second safety flag based on the acoustic signal and the back electromotive force signal, and generates a start strategy based on the first safety flag and the second safety flag.
[0006] Optionally, the sensing module includes a unit sensing module, a first environmental sensing module, and a second environmental sensing module; the unit sensing module collects a snapshot of the initial state of the actual machine; the first environmental sensing module collects the on-site environmental detection vector; and the second environmental sensing module collects the acoustic wave signal and the motor back electromotive force signal during the pre-start process.
[0007] Optionally, the simulation terminal includes a storage module and an assembly module; the storage module stores a twin machine library and a working condition package; the assembly module receives the initial state snapshot of the real machine and the field environment detection vector, retrieves the twin package from the twin machine library, and assembles it with the initial state snapshot of the real machine and the field environment detection vector to generate a shadow session instance and perform shadow simulation.
[0008] Optionally, the assembly module is configured to perform the following steps: receiving the real aircraft model identifier and real aircraft flight control firmware version uploaded by the receiving end, and verifying the real aircraft identity and version of the receiving end; pulling and importing the twin package corresponding to the real aircraft of the receiving end; mapping the real aircraft initial state snapshot uploaded by the receiving end to the twin initial state; loading the operating condition package stored by the storage module; and generating a runnable twin simulation instance.
[0009] Optionally, the simulation terminal further includes a calculation module, which defines the original anomaly metric and performs risk normalization, calculates the comprehensive risk score, and generates the first safety flag based on the comprehensive risk score.
[0010] Optionally, the determination module is configured to calculate the acoustic wave deviation based on the acoustic wave signal, calculate the back electromotive force load deviation based on the back electromotive force signal, and obtain the second safety indicator based on the acoustic wave deviation and the back electromotive force load deviation.
[0011] Optionally, the determination module is configured to: output a start signal when the first safety flag and the second safety flag are both 1; execute an interception start command when the first safety flag and the second safety flag are both 0; and calculate the inconsistency between the risk values of each working condition at the simulation end and the deviation of the receiving end when the first safety flag and the second safety flag are inconsistent, and perform a fusion decision or generate a structured probability report for manual decision-making based on the inconsistency.
[0012] Optionally, the determination module is configured to perform the following steps: establish an association matrix between the risk values of each operating condition at the simulation end and the deviation at the receiving end; calculate the inconsistency degree based on the association matrix; and perform a fusion decision step or generate a structured probability report based on the comparison result of the inconsistency degree and the conflict threshold.
[0013] Optionally, the determination module is configured to perform the following steps: obtain a fusion security score, determine an automatic fusion decision value based on the fusion security score, and output an unlock command or an intercept command.
[0014] One aspect of this application provides a dual-verification remote start method, comprising: a transmitter generating a remote start signal in response to a switch control signal of a wireless remote control switch in a first mode and sending it to a receiver; the receiver's radio frequency receiving module receiving the remote start signal, entering a pre-start mode in response to the remote start signal, acquiring a snapshot of the initial state of the actual device, a field environment detection vector, and acoustic signals and motor back electromotive force signals during the pre-start process in the pre-start mode, and transmitting the snapshot of the initial state of the actual device and the field environment detection vector to a simulation end; the simulation end performing shadow simulation based on the snapshot of the initial state of the actual device and the field environment detection vector, generating a first safety flag and transmitting it to the receiver; the receiver generating a second safety flag based on the acoustic signals and the back electromotive force signal, and generating a start strategy based on the first safety flag and the second safety flag.
[0015] Optionally, the steps for the simulation terminal to perform shadow simulation include: receiving the real machine model identifier and real machine flight control firmware version uploaded by the receiving terminal, and verifying the real machine identity and version of the receiving terminal; pulling and importing the twin package corresponding to the real machine of the receiving terminal; mapping the real machine initial state snapshot uploaded by the receiving terminal to the twin initial state; loading the operating condition package; and generating a runnable twin simulation instance.
[0016] Optionally, the step of generating the first safety flag includes: defining an original anomaly metric and performing risk normalization, calculating a comprehensive risk score, and generating the first safety flag based on the comprehensive risk score.
[0017] Optionally, the step of generating the second safety sign includes: calculating the acoustic wave deviation based on the acoustic wave signal, calculating the back electromotive force load deviation based on the back electromotive force signal, and obtaining the second safety sign based on the acoustic wave deviation and the back electromotive force load deviation.
[0018] Optionally, the step of generating a startup strategy based on the first security flag and the second security flag includes: outputting a startup signal when both the first security flag and the second security flag are 1; executing an intercept startup command when both the first security flag and the second security flag are 0; and calculating the inconsistency between the risk values of each operating condition at the simulation end and the deviation from the receiving end when the first security flag and the second security flag are inconsistent, and performing a fusion decision or generating a structured probability report for manual decision-making based on the inconsistency.
[0019] Optionally, it also includes: establishing a correlation matrix between the risk values of each operating condition at the simulation end and the deviation at the receiving end; calculating the inconsistency degree based on the correlation matrix; and performing a fusion decision based on the comparison result of the inconsistency degree and the conflict threshold, or generating a structured probability report.
[0020] Optionally, it also includes obtaining a fusion security score, determining an automatic fusion decision value based on the fusion security score, and outputting an unlock command or an intercept command.
[0021] The dual-verification remote start system and method provided in this application, by setting up a transmitter, a receiver, and a simulation end, allows the receiver to enter a pre-start mode after responding to a remote start signal. It utilizes a sensing module to collect a snapshot of the initial state of the actual machine, a field environment detection vector, and acoustic signals and motor back EMF signals during the pre-start process. The simulation end then generates a first safety flag based on the snapshot of the initial state of the actual machine and the field environment detection vector through shadow simulation. Simultaneously, the receiver's judgment module generates a second safety flag based on the acoustic signals and back EMF signals. Finally, a start strategy is generated based on the first and second safety flags, achieving dual safety verification of the remote start command. This effectively avoids safety accidents such as false starts, abnormal attitudes, or collisions caused by directly starting the power system under abnormal operating conditions, significantly improving the safety, reliability, and environmental adaptability of remote start.
[0022] By further configuring the perception module to include a crew perception module, a first environmental perception module, and a second environmental perception module, which are used to collect snapshots of the initial state of the real aircraft, on-site environmental detection vectors, and acoustic signals and motor back electromotive force signals during the pre-start process, the classification perception and fine acquisition of the UAV's own state, external environment, and pre-start dynamic process are realized. This provides a multi-dimensional data foundation for subsequent simulation verification and physical detection, and enhances the comprehensiveness and accuracy of safety verification.
[0023] By setting up a storage module and an assembly module for the simulation end, the storage module stores the twin model library and operating condition package, and the assembly module receives the initial state snapshot of the real machine and the on-site environment detection vector, retrieves the corresponding twin package for assembly to generate a shadow session instance and perform shadow simulation, thus realizing virtual simulation verification based on digital twin technology. This enables the simulation process to accurately match the real machine model and real-time operating conditions, improving the authenticity and credibility of the simulation results.
[0024] By configuring the assembly module to perform the following steps: receiving the real aircraft model identifier and flight control firmware version for identity and version verification, pulling the corresponding twin package, mapping the real aircraft initial state snapshot to the twin initial state, loading the operating condition package, and generating a runnable twin simulation instance, the simulation end and the real aircraft are strictly aligned in terms of model identity, firmware version, and initial state. This avoids simulation failures caused by version mismatch or state mismatch from the source, and further improves the reliability and security of shadow simulation.
[0025] By setting up a calculation module on the simulation end to define the original anomaly measurement and perform risk normalization, calculate the comprehensive risk score and generate the first safety indicator accordingly, the quantitative assessment and comprehensive judgment of risks under various working conditions are realized, enabling the simulation end to output quantitative safety indications and enhancing the objectivity and interpretability of safety indicator generation.
[0026] By configuring the judgment module to calculate the acoustic wave deviation based on the acoustic wave signal and the back electromotive force load deviation based on the back electromotive force signal, and obtaining the second safety flag accordingly, real-time analysis and anomaly detection of physical signals during the pre-start process are realized. This enables the receiving end to independently judge the startup safety locally, forming a verification channel that complements the simulation end and improves the redundancy and real-time performance of the dual verification.
[0027] By configuring the judgment module to transmit the start signal when both the first and second safety flags are equal, and to intercept when both are equal, and to calculate the inconsistency between the risk values of each operating condition at the simulation end and the deviation at the receiving end when the two are inconsistent, the module can then make a fusion judgment or generate a structured probability report for manual decision-making based on the inconsistency. This achieves a hierarchical processing strategy for the consistency and inconsistency of the verification results between the simulation end and the receiving end. While ensuring security, it also takes into account the flexibility and traceability of the start decision. In particular, in the case of inconsistency, the fusion judgment or manual intervention effectively avoids the risk of start blocking or false start caused by misjudgment of a single channel.
[0028] By configuring the judgment module to establish a correlation matrix between the risk values of each working condition at the simulation end and the deviation at the receiving end, and calculating the inconsistency degree based on the correlation matrix, the system then decides to execute a fusion decision or generate a structured probability report based on the comparison result of the inconsistency degree and the conflict threshold. This achieves quantitative evaluation and strategy diversion of the degree of conflict between the verification results of the simulation end and the receiving end, and improves the scientificity and rationality of the decision-making process in case of inconsistency.
[0029] By configuring the judgment module to obtain the fusion security score, and determining the automatic fusion judgment value based on the fusion security score to output the unlock or intercept command, a deep fusion judgment of the verification information of the simulation end and the receiving end is realized. In the case of small conflicts, a reliable start or intercept decision can be automatically generated, which further improves the system's autonomous decision-making ability and security under complex working conditions. Attached Figure Description
[0030] The above and other objects, features and advantages of this application will become clearer from the following description of embodiments with reference to the accompanying drawings, in which: Figure 1 A schematic diagram of the structure of a remote start system according to an embodiment of this application is shown; Figure 2 A schematic diagram of the transmitter provided according to an embodiment of this application is shown; Figure 3 A schematic diagram of the structure of a receiver provided according to an embodiment of this application is shown; Figure 4 A schematic diagram of the structure of the simulation terminal provided according to an embodiment of this application is shown. Detailed Implementation
[0031] The present application will now be described in more detail with reference to the accompanying drawings. In the various drawings, the same elements are indicated by similar reference numerals. For clarity, the various parts in the drawings are not drawn to scale. Furthermore, some well-known parts may not be shown.
[0032] This application may be presented in various forms, some of which will be described below.
[0033] Figure 1 A schematic diagram of the structure of a remote start system according to an embodiment of this application is shown, such as... Figure 1 As shown, the remote start system includes a transmitter 100, a receiver 200, and a simulation terminal 300.
[0034] Figure 2 A schematic diagram of the structure of the transmitter 100 according to an embodiment of this application is shown, as follows: Figure 2 As shown, the transmitter 100 includes a main control module 110, a radio frequency transmission module 120 connected to the main control module 110, and a wireless remote control switch 130 connected to the main control module 110. The main control module 110 is configured to generate a remote start signal in response to the switch control signal of the wireless remote control switch 130 in a first mode, and to generate a remote control signal in a second mode. The radio frequency transmission module 120 is connected to the main control module 110 and is used to send the remote start signal or the remote control signal to the receiver 200.
[0035] It is worth noting that in this embodiment, in the first mode, the power module 250 of the receiver 200 only supplies power to a portion of the structure, such as the RF receiving module 210 and the parsing and judgment module 220 described below; or the power module 250 of the receiver 200 supplies power to all modules, including the device function module 240, but the device function module 240 is in a pre-startup mode. In the first mode, the transmitter 100 may also generate a remote control signal, but the device function module 240 of the receiver 200 will not respond to the remote control signal. In the second mode, the power module 250 of the receiver 200 supplies power to all modules, including the device function module 240, and the device function module 240 has entered the startup mode.
[0036] Figure 3 A schematic diagram of the structure of the receiver 200 according to an embodiment of this application is shown, as follows: Figure 3 As shown, the receiver 200 includes an RF receiving module 210, a parsing and judgment module 220, a wireless switch module 230, a device function module 240, a power module 250, and a security module 260. The power module 250 is connected to both the RF receiving module 210 and the parsing and judgment module 220, providing power to them. The power module 250 is also connected to the device function module 240 and the security module 260 via the wireless switch module 230. When the wireless switch module 230 is on, it provides power to the device function module 240 and the security module 260. When the wireless switch module 230 is off, the power module 250 only provides power to the RF receiving module 210 and the parsing and judgment module 220. The on / off state of the wireless switch module 230 is controlled by the wireless remote control switch 130 of the transmitter 100.
[0037] The radio frequency receiving module 210 is connected to the parsing and judgment module 220, which is connected to the wireless switch module 230 and the device function module 240, respectively. The radio frequency receiving module 210 is paired with the radio frequency transmitting module 120 of the transmitter 100 to receive the wireless radio frequency signals transmitted by the transmitter 100. The parsing and judgment module 220 is connected to the radio frequency receiving module 210, receives the wireless radio frequency signal transmitted by the radio frequency receiving module 210, demodulates the received wireless radio frequency signal and restores it to an electrical signal. The parsing and judgment module 220 is configured to send a control command to the wireless switch module 230 when it receives a remote start signal. Upon receiving the remote start signal, the wireless switch module 230 closes under the drive of the remote start signal, allowing the power module 250 to provide power to the device function module 240 and the safety module 260 via the activated wireless switch module 230. Simultaneously, the parsing and judgment module 220 is configured to send a pre-start control command to the device function module 240 when it receives a remote start signal. The device function module 240 receives the pre-start control command and enters the pre-start mode. In the pre-start mode, the blades vibrate slightly at a preset amplitude, referred to as micro-vibration.
[0038] Furthermore, in the second mode, the parsing and judgment module 220 is configured to forward the remote control signal to the device function module 240 when it receives the remote control signal, so as to drive the actuator to move.
[0039] In this embodiment, the device function module 240 is activated in response to the safety sign signal from the safety module 260. The safety module 260 includes a sensing module 2601 and a determination module 2602. The sensing module 2601 includes a unit sensing module, a first environmental sensing module, and a second environmental sensing module. The unit sensing module captures a snapshot of the initial state of the actual machine, the first environmental sensing module captures environmental detection vectors, and the second environmental sensing module captures acoustic signals and motor back electromotive force signals during the pre-start process. After the wireless switch module 230 closes, supplying power to the device function module 240 and the safety module 260 (including the sensing module 2601 and the determination module 2602), the device function module 240 does not start immediately but enters a pre-start mode. The perception module 2601 of the safety module 260 collects a snapshot of the initial state of the UAV and the detection vector of the surrounding environment, and transmits these data to the simulation terminal 300. The simulation terminal 300 performs shadow simulation based on the UAV's initial state snapshot and the environmental detection vector, and generates a first safety flag. The simulation terminal transmits the first safety flag to the judgment module 2602 of the safety module 260. Simultaneously, the perception module 2601 collects acoustic signals and motor back EMF signals during the pre-start process and transmits them to the judgment module 2602. The judgment module 2602 is configured to calculate the deviation based on the received acoustic signals and back EMF signals, and compare the deviation with a preset threshold to obtain a second safety flag. The judgment module 2602 is also configured to generate a start-up strategy based on the first and second safety flags.
[0040] The unit sensing module includes a sensor group in the equipment function module 240, and the unit sensing module collects snapshots of the initial state of the actual unit. This includes raw data items from the actual device, including but not limited to: raw sampled values of the device's battery voltage, the device's spatial positioning information, the number of effective observation satellites and the horizontal precision factor HDOP, raw sampled readings of the magnetometer, raw IMU output and online zero-bias estimation results, raw nearest obstacle distance output by the obstacle perception module, current throttle command value, and raw signal of the device's unlock flag. The first environmental perception module includes, but is not limited to, wind speed perception, temperature perception, humidity perception, and obstacle perception modules, acquiring on-site environmental detection vectors. for:
[0041] In the formula: Indicating wind speed, Characterizing temperature, Characterizing air humidity It represents the intensity of precipitation (the amount of precipitation per unit time). A set of point cloud data or a set of obstacle distance parameters that characterizes near-field obstacles on a real machine.
[0042] The simulation terminal 300 includes a storage module 310 and an assembly module 320. The storage module 310 stores a twin model library. and working condition package Assembly module 320 receives a snapshot of the initial state of the actual device from sensing module 2601. and on-site environmental detection vectors And retrieve the twin packet corresponding to the receiver 200 from the storage module 310. , will twin package The initial state snapshot of the real device uploaded by the receiving end 200 and on-site environmental detection vector Assembly is performed to generate a shadow session instance, and shadow simulation is then conducted. It's worth noting that load case packages are loaded at different stages of the simulation. And observe the simulation results.
[0043] Specifically, assembly module 320 is configured to perform the following steps: S110: Shadow Session Assembly; S120: Shadow pre-startup closed-loop simulation.
[0044] Step S110 includes: S1101: Obtain the real device model identifier IDac and the real device flight controller firmware version vfw uploaded by the receiver 200, and verify the real device identity and version of the receiver 200 based on the real device model identifier IDac and the real device flight controller firmware version vfw. S1102: Retrieve and import the twin packet corresponding to the receiver 200 real device. ; S1103: Snapshot of the initial state of the real device uploaded by the receiving end 200. Mapped to twin initial state; S1104: Load the operating condition package stored in storage module 310. ; S1105: Generate a runnable twin simulation instance.
[0045] In step S1101, the assembly module 320 obtains the real aircraft model identifier (IDac) and the real aircraft flight control firmware version (vfw) uploaded by the receiving end 200. The real aircraft model identifier (IDac) is the real aircraft model code stored in the device function module 240 and reported to the simulation end by the device function module 240 after power-on. The real aircraft model identifier (IDac) is bound one-to-one with the overall hardware configuration and model configuration parameters of the real aircraft on the receiving end 200. The real aircraft model identifier (IDac) is the core index field of the twin model library on the simulation end 300, used to accurately locate the model library entry that uniquely corresponds to the real aircraft on the receiving end 200, achieving resource isolation and precise matching at the model level. The real aircraft flight control firmware version (vfw) is the firmware program version number of the current real-time running flight control system of the real aircraft on the receiving end 200, representing the iterative version of the current real aircraft control logic, state machine mechanism, and hardware driver.
[0046] The twin model library stored in storage module 310 is a database for storing digital twin model resources. It is partitioned and stored using the digital twin model model identifier (Itwin) as the category index. Each database entry stores the following: the digital twin model model identifier (Itwin), the digital twin model firmware version (vtwin), and the twin package. .
[0047] During the twin simulation startup verification phase, assembly module 320 reads the real aircraft model identifier (IDac) and the real aircraft flight control firmware version (vfw) uploaded by receiver 200. It then iterates through the digital twin model model identifier (Itwin) in the twin model library. If no matching entry exists in the digital twin model model identifier (Itwin) of the twin model library, authentication fails. If a matching entry exists, authentication succeeds, and version verification is performed. During version verification, the system searches the twin model library for entries matching the real aircraft model identifier (IDac) and compares the digital twin model firmware version (vtwin) in the library with the real aircraft flight control firmware version (vfw) uploaded by receiver 200. If the real aircraft flight control firmware version (vfw) does not match the digital twin model firmware version (vtwin), a twin simulation version mismatch is determined.
[0048] This application performs dual verification of identity and version. If either identity or version verification fails, the twin simulation assembly process is rejected and an exception is reported; if both identity and version verifications succeed, the twin package within that entry is loaded. This generates a digital twin simulation instance corresponding to the real device. In this embodiment, the identity and version of the receiving end 200 are verified before simulation, thus avoiding simulation failures caused by version mismatches at the source.
[0049] In step S1102, the assembly module 320 retrieves the corresponding twin package from the twin model library stored in the storage module 310. twin bags This is a collection of multi-dimensional aircraft model simulation models, specifically represented as:
[0050] in, It is a flight control simulation model for twin aircraft, used to replicate the control logic and state transition rules of the real aircraft's flight control. It is a simulation model of the twin's body dynamics, used to simulate the dynamic characteristics of the real machine, such as body movement and dynamic response; This is a simulation model of the equivalent circuit of a twin battery, used to reproduce the electrical characteristics of a real battery, such as power supply, voltage drop, and losses. This is a virtual GPS navigation simulation model. This is a virtual magnetometer simulation model; This is a near-field obstacle geometry simulation model for twin aircraft, used to simulate the spatial position, contour, and distance characteristics of obstacles in the real aircraft flight environment; The model calibration parameter set includes the UAV's overall weight, propeller parameters, motor KV value (a core parameter characterizing the motor's speed and torque output trade-off characteristics), battery internal resistance nominal value, and other inherent calibration parameters of the model, providing benchmark parameter support for simulation state mapping and model calculation.
[0051] It is worth noting that the twin battery equivalent circuit simulation model It can also be equipped with virtual voltage sensors and virtual current sensors corresponding to the equivalent circuit to measure the voltage and current of the equivalent circuit; Geometric simulation model of near-field obstacles for twin machines. Virtual distance sensors can also be installed to obtain the distance between the twin and obstacles. (Twin battery equivalent circuit simulation model) The assembled virtual voltage sensor, virtual current sensor, and virtual GPS navigation simulation model Assembled virtual GPS, virtual magnetometer simulation model The assembled virtual magnetometer and near-field obstacle geometry simulation model of the twin machine The assembled virtual distance sensors constitute a virtual sensor group.
[0052] In step S1103, based on the preset state mapping function The receiver 200 will capture a snapshot of the initial state of the real device. Combined with the model calibration parameter set After performing standardized mapping, the initial operating state of the twin simulation system is generated. The initial operating state of the twin simulation system satisfies the following:
[0053] The mapping function It is used to complete the unit conversion, spatial coordinate system alignment, data range clipping and outlier filtering of the original state data of the real machine, so as to achieve accurate replication of the physical state of the real machine to the virtual state of the twin simulation.
[0054] Real device initial state snapshot This is a complete instantaneous state dataset obtained by receiver 200 from the flight control system of the actual aircraft at a specific sampling moment. This is a snapshot of the initial state of the actual aircraft. As the initial input for digital twin simulation initialization, it is input to the state mapping function. Combined with the model calibration parameter set A standardized mapping is performed to obtain a virtual initial state recognizable by the twin simulation system, thus aligning the virtual twin instance with the real machine's state at simulation startup. The mapped initial twin state is... The core components are derived from the data after parsing and processing the initial state snapshot of the real machine, including: the initial open-circuit voltage of the battery Voc(0), the initial GPS positioning information, the number of satellites, the horizontal accuracy factor HDOP, the initial magnetometer reading B(0), the IMU zero bias estimate, the initial nearest obstacle distance dobs(0), the throttle state before unlocking, and the unlock flag; among them, the unlock flag of the twin simulation instance is forcibly configured to be in the unlocked state to ensure the safety and standardization of the initial state of the simulation.
[0055] In step S1104, the assembly module 320 loads the operational condition package according to the actual aircraft flight mission type and the operational site environment strategy. The working condition package For a non-empty subset of the preset working condition type set, satisfying
[0056] in, For the number of working condition types, Injecting a model into satellite navigation drift conditions; Injecting a model for magnetic field interference conditions; Injecting a model into the battery voltage drop condition; Injecting a model into near-field collision scenarios.
[0057] Each type of operating condition injection model configuration is bound to specific fault or environment injection parameters. Inject parameters Including but not limited to sensor drift rate, magnetometer bias amplitude, battery internal resistance increment, obstacle approach trajectory parameters, etc., are used to inject real environmental interference, equipment performance degradation, obstacle disturbance and other working conditions into the twin simulation process to realize full-scene simulation and deduction.
[0058] In step S1105, a shadow session instance is generated.
[0059] Assemble functions using preset instances. Integrating twin packages Initial state of twins Working condition package Injection parameters for corresponding working conditions and on-site environmental detection vector Generate stand-alone airborne shadow simulation session instances. That is, satisfying:
[0060] In step S120, the shadow simulation engine is activated. The flight control replica model reads the simulation output data of the virtual sensor group in real time at a high sampling frequency of 100Hz to 1kHz (sub-kilohertz to kilohertz), iteratively generates flight control quantities, and drives the airframe dynamics simulation model. Complete the simulation of propulsion and body motion.
[0061] The simulation engine's simulated clock supports an accelerated operating mode that is faster than physical wall clocks, and the simulation runtime timing interval completely covers the preset system startup envelope time window. This ensures a complete reproduction of the entire time-series operating process of the real machine from startup to stable operation, achieving full coverage and no omissions in twin simulation.
[0062] The shadow simulation is divided into three stages in chronological order. The total duration of each stage is determined according to the actual device model, ranging from seconds to tens of seconds.
[0063] The first stage, namely the Power-On Self-Test (POST) stage, corresponds to the following time window: During this phase, voltage checks, inertial measurement unit and magnetometer calibrations, and satellite constellation and positioning accuracy checks are performed sequentially in the virtual process. For the main observation items, data on the number of satellites, horizontal accuracy factor (HDOP), magnetic field consistency, and steady-state voltage are collected.
[0064] The second phase, namely the unlock and idle phase (ARM), corresponds to a time window of... During this phase, unlocking logic verification, motor speed ramp-up, and idle speed stabilization control are performed in the virtual process; for the main observation items, heading jitter, voltage drop slope, and unlocking rejection flag are collected.
[0065] In the third phase, namely the takeoff and climb phase (TO), the corresponding time window is: This phase involves performing throttle control, attitude and altitude control, ground clearance determination, and climb control during the virtual process; for key observation items, it collects data on horizontal drift, near-ground distance, and altitude tracking error.
[0066] The operating condition injection spans the entire startup envelope time window mentioned above. For each original anomaly metric ( Its sampling window is bound to the corresponding stage in the above three stages, thereby excluding the error generated in the cruise stage from the start-up risk assessment and avoiding its miscounting in the error statistics of the start-up stage.
[0067] The virtual sensor array monitors the state of the twin. Ideally, the virtual sensors in the array only output true values. During shadow simulation testing, artificially injected operating condition packages are superimposed. Various operating conditions / fault disturbances are added, along with the random noise of the sensor itself, to finally obtain the simulation output of the virtual sensor. This is used to simulate data collected by sensors under faulty or abnormal operating conditions, for subsequent verification of anomaly detection algorithms. The virtual sensor outputs the following after a disturbance is injected:
[0068] in, For the first The virtual sensor outputs the observed values corresponding to the simulated walk; The truth value under ideal conditions is used to represent the first... Step Twin System Status Convert to the true value of an ideal sensor without disturbance or noise; For the first Inject disturbance operators for similar operating conditions; For the first The disturbance parameters corresponding to the type of working condition, For work condition packages, This indicates that all active perturbation operators within the load case are superimposed. For the first Simulates virtual sensor noise under staggered conditions to simulate random measurement noise of real sensors.
[0069] The flight control simulation model of the twin aircraft receives virtual sensor data and, combined with the current flight status (such as self-test, unlock, takeoff / flight mode, etc.), calculates and outputs control commands; the flight control simulation model outputs control quantities:
[0070] in, For twin-aircraft flight control simulation model; This indicates the current flight status of the twin aircraft.
[0071] The dynamic simulation model is based on the current state. and flight control output control Proceeding to the next moment The system state is determined, and the closed-loop iteration is completed. The recursive derivation of the twin's aircraft dynamic state satisfies the following formula:
[0072] in, A dynamic model for twin machines; Set the parameter set for the model; This represents aerodynamic drag.
[0073] In this embodiment, the aerodynamic drag effect needs to be considered in the recursive derivation of the twin's aircraft dynamics state. The aerodynamic drag is determined by the density of moist air. and on-site environmental detection vector wind speed Calculated aerodynamic drag for:
[0074] in, Density of moist air, unit: kg / m³ 3 humid air density Related to temperature and humidity, it can be obtained from on-site environmental detection vectors. Temperature in ,humidity Calculated; is the aerodynamic drag coefficient, which is dimensionless; For reference, the windward area is in meters. 2 ; The velocity of the aircraft relative to the airflow is expressed in m / s (vector). The negative sign indicates that the direction of the aerodynamic force is always opposite to the direction of the relative incoming flow—the airflow blows from the front, and the resistance pushes the aircraft from the front.
[0075] relative airflow speed of the body Represented as:
[0076] in, For the body's ground speed, For on-site environmental detection vectors The wind speed in the middle.
[0077] The simulation terminal 300 also includes a calculation module 330, which is configured to perform the following steps: S210: Define the original anomaly measurement and risk normalization; S220: Calculate the comprehensive risk score and generate the first safety indicator based on the comprehensive risk score. 。
[0078] In step S210, for the working condition package Each startup-specific operating condition Configure triples To characterize the original anomaly measure under this operating condition, where, For the first The disturbance parameters corresponding to the type of working condition, This is the sampling window for this operating condition. The window aggregation functional for this operating condition.
[0079] For the For each working condition, its original anomaly measurement Window aggregation functional For sampling window The original anomaly measurement is obtained by aggregating simulation data within the dataset. satisfy:
[0080] in, For the first A simulated walk of the twin system's state. For the first A virtual sensor, simulating a walk, outputs observations. For the first The flight control output of a simulated twin aircraft is a control quantity for a walk. Indicates sampling window The various simulated walks inside.
[0081] Then the original anomaly measurement is performed using a truncation function. Normalization is performed to obtain the normalized risk value:
[0082] in, Lower bound of measurement under healthy operating conditions; Upper bound of extreme risk; Truncation function, guaranteeing , The larger the value, the higher the risk of the working condition.
[0083] Next, we will measure the original anomalies under each operating condition. The definition and normalization are explained in detail.
[0084] The satellite navigation drift injection model injects position disturbances into the virtual GPS output to simulate virtual GPS positioning deviation, slow drift, and multipath interference in the twin machine. The formula for virtual GPS position disturbance is:
[0085] in, These are the observations output by the virtual GPS. The ideal location of the twin machine obtained by virtual GPS; This is the static offset, which is a constant value and a fixed GPS positioning deviation. It has a slow drift speed and shifts linearly and continuously over time. It is a simulation moment; This is for locating noise.
[0086] Disturbance can be activated at the end of the Power-On Self-Test (POST) phase or during the unlock and idle phase (ARM). Before simulated takeoff, the virtual GPS output position already has a static offset. Slow drift and positioning noise After takeoff, slow drift accumulates over time. It's worth noting that positioning noise... Due to weather conditions, this application introduces the on-site environmental detection vector uploaded by the receiver 200. medium wind speed and precipitation intensity wind speed and precipitation intensity Impact of positioning noise In Chinese, this is specifically expressed as:
[0087] in, The nominal standard deviation for sunny weather. As a factor that worsens the weather, For on-site environmental detection vectors The intensity of precipitation in the middle, For on-site environmental detection vectors The wind speed in the precipitation. The greater the precipitation and the higher the wind speed, the more localized the noise. The larger the value, the more specific the mapping relationship can be obtained through experimental calibration or lookup tables.
[0088] During the rehearsal window from the unlock and idling phase (ARM) to the takeoff and climb phase (TO), the peak drift (in meters) of the virtual GPS horizontal position relative to the sampling origin is the original anomaly measurement of the GPS drift condition. Original anomaly measurement Represented as:
[0089] in, The sampling window is the time from the start of the climb phase to the unlocking point. for The observation values output by the virtual GPS at any given time; The observed values output by the virtual GPS at the start of the sampling window (unlocking time); For horizontal projection, only the east and north horizontal coordinates are retained, while the height direction is discarded; Throughout the sampling window Take the maximum value within the range.
[0090] The above original outlier measures are normalized using the following formula:
[0091] in, This is the normalized risk value under satellite navigation drift conditions, with a value of 0. 1. The higher the value, the more severe the GPS drift fault; Shake for your health; Unacceptable drift during takeoff.
[0092] Magnetic field interference injection model The virtual magnetometer output is injected with a constant bias or a step jump in the magnetic field to simulate magnetic interference caused by the proximity of ferromagnetic objects and geomagnetic anomalies. The disturbance formula is:
[0093] in, Observations acquired by a virtual magnetometer; The geomagnetic field vector obtained by a virtual magnetometer under ideal conditions; Static magnetic bias; This represents the amplitude of the magnetic field step jump. An indicator function that returns 1 if the condition is true and 0 if the condition is false. The simulated moment when the magnetic field jump occurs; This represents random noise from a virtual magnetometer.
[0094] The heading angle is obtained by flight control attitude calculation using a virtual magnetometer. The sampling window captures the unlocked idling phase. Then the original anomaly measurement of the magnetic field interference condition. This represents the peak-to-peak value of the heading angle jitter within the sampling window, expressed in degrees (°). It is the original anomaly measurement for the magnetic field interference condition. Represented as:
[0095] Its normalized risk value is:
[0096] in, This is the normalized risk value under magnetic field interference conditions, with a value of 0. 1; Shake for your health; Drifting is unacceptable for takeoff.
[0097] Battery voltage drop injection model By increasing the battery's equivalent internal resistance or applying a load step, the voltage collapse phenomenon of a low-temperature, aged battery during a high-current unlocking is simulated. The virtual battery voltage model is as follows:
[0098] in, The current open-circuit voltage of the virtual battery, in V; The current output by the virtual current sensor in the k-th simulated staggered step, in A; The virtual battery internal resistance is expressed in Ω. The increment of internal resistance injected, in Ω; The equivalent inductance of the virtual battery circuit; This represents the rate of change of virtual battery current.
[0099] Virtual battery internal resistance Internal resistance is temperature-dependent; it increases at low temperatures. (Battery internal resistance) The mapping relationship between temperature and battery characteristics is calibrated through battery performance experiments and provided in the form of a lookup table. For example, at 25℃, = , At 10℃, ≈2 .
[0100] During the unlocked idling phase, the motor current rises rapidly, and the increased internal resistance of the virtual battery causes a significant voltage drop. The sampling window captures the unlocked and idling phases (ARM). The peak value of the absolute value of the voltage drop slope is the original anomaly measure of the battery voltage drop condition. :
[0101] Its normalized risk value is:
[0102] in, It is the normalized risk value under battery voltage drop conditions. This represents the upper limit of normal voltage fluctuations. This is the severe drop threshold.
[0103] Near-field collision injection model In this context, the obstacle set O originates from the on-site environment detection vector uploaded by the receiver 200. Alternatively, obstacles can be configured within the twin scenario; approach velocity parameters can be applied to obstacles to test the collision risk during takeoff. The spatial position of the twin is denoted as... The closest distance between the twin and all obstacles at the current moment is:
[0104] Where o represents a single obstacle within the obstacle set O; Represents the Euclidean distance between the drone and the obstacle; This represents the closest distance between the twin and all obstacles at the current moment.
[0105] The sampling window is during the takeoff and climb phase. Take the minimum obstacle distance within the sampling window. Perform an inverse quantization mapping to obtain the original anomaly measurement under near-field collision conditions. :
[0106] in, For a safe distance, This is the critical danger distance. This represents the minimum obstacle distance within the sampling window.
[0107] Its physical meaning is: ≥ hour, =0, meaning there is no risk of collision; ≤ hour, =1, meaning the highest collision risk has been reached.
[0108] Original anomaly measurement The normalized risk value under near-field collision conditions has fallen within the [0,1] interval. Take directly:
[0109] In step S220, a comprehensive risk score is calculated, and a first safety indicator is generated based on the comprehensive risk score. Among them, comprehensive risk classification satisfy:
[0110] in, The maximum mixing coefficient is used to avoid the extremely high risk of a single working condition being diluted by the weighted average.
[0111] First safety sign The logic behind the judgment: The first security flag output to the determination module 2602 of the receiving end 200 ; The first security flag output to the determination module 2602 of the receiving end 200 Furthermore, set high-risk thresholds. When the risk value of any working condition Mandate the submission of risk lists and directly designate the first safety indicator. Set to zero.
[0112] The second environmental perception module includes an acoustic wave acquisition unit and a back electromotive force (EMF) sampling unit. The acoustic wave acquisition unit is configured to acquire the acoustic wave signal generated by the blade during micro-vibration when the device function module 240 enters the pre-start mode, and transmit it to the judgment module 2602; the back EMF sampling unit is configured to acquire the back EMF signal of the motor during micro-vibration when the device function module 240 enters the pre-start mode, and transmit it to the judgment module 2602.
[0113] The determination module 2602 is configured to calculate the deviation based on the received acoustic signal and back electromotive force signal, and compare the deviation with a preset threshold to obtain a second safety indicator. The determination module 2602 is configured to perform the following steps to obtain a second security flag. : S310: Calculate acoustic wave deviation and back EMF load deviation ; S320: Based on sound wave deviation and back EMF load deviation Obtain the second security mark ; In step S310, within the micro-jitter window, the acoustic wave signal is acquired, and the short-time energy E is calculated based on the discrete sampling sequence.
[0114] in, The amplitude of the nth acoustic wave sample within the micro-jitter window. This represents the number of sampling points in a single jitter detection.
[0115] Simultaneously, spectral analysis is performed on the acoustic signal to calculate the spectral centroid. Time domain standard deviation Thus, the sound wave feature vector is formed:
[0116] in, It represents the short-time energy of sound waves, reflecting the shift in the centroid of the frequency spectrum. The standard deviation in the time domain reflects the waveform fluctuations.
[0117] The acoustic wave vector is compared with the pre-calibrated no-load reference characteristic vector. Normalized deviations are calculated for each component, and the acoustic wave deviation is obtained by combining them according to predetermined rules (such as normalized Euclidean distance). :
[0118] in, This is the measured characteristic vector of the sound wave. This is the unloaded reference feature vector. The smaller the value, the closer the current acoustic echo environment is to the acoustic environment under no-load safe conditions, and the lower the possibility of foreign objects or obstructions in the startup environment.
[0119] Simultaneously, the back electromotive force of the motor is collected during the micro-vibration process. And calculate its reference value for back electromotive force during no-load micro-vibration. relative deviation
[0120]
[0121] in, This refers to the back electromotive force of the motor during the micro-vibration process. This serves as a reference for the back EMF during no-load micro-vibration. When the motor shaft is subjected to mechanical obstruction or additional load, the back EMF will deviate from the no-load reference value. It increases accordingly; The smaller the value, the closer the load on the motor is to an unloaded state.
[0122] In step S320, the sound wave deviation is... Deviation from the preset acoustic wave characteristics from the decision threshold Compare the back EMF load deviation. Deviation from the preset back electromotive force threshold Compare; take the product of the two indicator functions as the second safety indicator. :
[0123] in, If the acoustic characteristics deviate from the decision threshold, If the back electromotive force deviates from the decision threshold, For indicator functions: 1 for true, 0 for false. When < and < At the same time, the second safety sign =1, in other cases, the second safety sign =0.
[0124] In this embodiment, upon receiving a remote start command, direct high-speed unlocking is prohibited, and parallel dual verification is initiated. Parallel dual verification includes two parallel channels: a 200-bit micro-jitter liveness detection channel at the receiver and a 300-bit shadow pre-start simulation channel at the simulation end. Specifically, the 200-bit micro-jitter liveness detection channel at the receiver performs liveness detection based on the micro-jitter signal collected by the second environmental perception module and outputs a second security flag. The 300 shadow pre-startup simulation channel at the simulation end performs pre-startup simulation in the digital twin environment and outputs the first safety flag. .
[0125] The determination module 2602 obtains the first security flag transmitted by the simulation terminal 300. and the calculated second safety mark Generate automatic unlock flag Automatic unlock icon satisfy:
[0126] when and When both are 1, The value is 1. When the value is 1, the determination module 2602 transmits a start signal to the device function module 240. It is worth noting that in this embodiment, after the device function module 240 enters the start mode, the sensing module 2601 and the determination module 2602 stop working.
[0127] when and When both are 0, the command to intercept startup is executed. In a specific embodiment, when and When both are 0, the interception start instruction is executed, and the simulation end 300 outputs a risk list to the judgment module 2602 of the receiving end 200 (according to...). Sort in descending order.
[0128] when =0, When it is 1, or when =1, When the value is 0, establish a correlation matrix between the risk values of each working condition at the simulation end 300 and the deviation of the receiver end 200, and calculate the inconsistency between the simulation end and the receiver end based on the correlation matrix.
[0129] Specifically, the correlation matrix between the risk values of each operating condition at the simulation end and the deviation at the receiving end. for:
[0130] in, For the simulation end Normalized risk value for each working condition. For the receiving end One deviation index; For working conditions With deviation The mapping function between them (determined by calibration experiments). The correlation matrix. The error between the normalized risk value of each working condition obtained by the simulation terminal 300 and the equivalent risk obtained by mapping the measured deviation of the receiver 200 is characterized. The smaller the value, the smaller the error between the normalized risk value of each working condition of the simulation terminal 300 and the equivalent risk obtained after mapping to the receiver 200, and the higher the matching degree between the normalized risk value of each working condition of the simulation terminal 300 and the measured deviation of the receiver 200. The larger the value, the greater the error between the normalized risk value of each working condition of the simulation terminal 300 and the equivalent risk obtained after mapping the measured deviation of the receiver terminal 200, and the worse the consistency between the normalized risk value of each working condition of the simulation terminal 300 and the measured deviation of the receiver terminal 200.
[0131] During this process, several valid correlation pairs were obtained, as shown in Table 1. Among them, voltage instability caused abnormal motor power supply, resulting in the back electromotive force deviating from the no-load reference value, and the risk value of battery voltage drop under operating conditions. Deviation from back electromotive force Forming effective correlation pairs; near-field obstacles enhance sound wave reflection, leading to increased echo energy and a higher risk value for near-field collisions. Deviation from sound waves Form effective association pairs.
[0132] Table 1 Effective Association Pairs
[0133] It is worth noting that the risk value of satellite navigation drift is... Primarily related to position changes, during the pre-start-up phase, the actual machine will not change position, thus reducing the risk value of magnetic field interference. It mainly affects the accuracy of heading angle calculation, and its anomalies are difficult to detect directly through acoustic waves or back electromotive force signals during the micro-jitter phase. Its deviation from acoustic waves... Or back electromotive force deviation The resulting association pairs are invalid and therefore are not included in the association matrices of the simulation end and the receiver end, but are evaluated independently only within the simulation end channel.
[0134] Next, the inconsistency between the simulation end and the receiving end is calculated based on the correlation matrix. :
[0135] in, Consistency factor Represented as:
[0136] in, For effective correlation logarithms, For the first Normalized mapping function for each association pair.
[0137] Inconsistency The larger the value, the more inconsistent the simulation and receiving ends become at the process value level, leading to more severe conflicts and requiring a more cautious decision-making strategy. When the value is less than the conflict threshold, proceed to the fusion decision step. When the conflict threshold is greater than or equal to the threshold, a structured probability report is generated for human decision-making. In one embodiment, the conflict threshold is, for example, 0.3, but is not limited to this.
[0138] In the process of fusion decision-making, fusion security scores are obtained respectively. According to the integration security division Obtain the fusion decision value .
[0139] Obtaining a fusion security score In the steps, the confidence level of the simulation end is obtained respectively. and receiver confidence Based on the confidence level of the simulation end and receiver confidence Obtaining a fusion security score .
[0140] Specifically, the confidence level of the simulation end for:
[0141] in, For environmental parameter completeness, environmental parameter completeness Represented as:
[0142] in, The completeness of environmental parameters represents the proportion of effective environmental parameters. For effective environmental parameters, This represents the total number of environmental parameters. It's worth noting that the receiver has 200 on-site environmental detection vectors. For on-site environmental detection vectors The various environmental parameters in ( If an environmental parameter can be acquired by the first environmental sensing module, it is determined to be a valid environmental parameter; if it cannot be acquired by the first environmental sensing module, a preset environmental parameter is directly filled in, and it is determined to be an invalid environmental parameter. For example, in one embodiment, the environmental parameter acquired by the first environmental sensing module is wind speed. ,temperature air humidity And precipitation intensity And the point cloud data set or obstacle distance parameter set corresponding to near-field obstacles on the actual device. If this fails to acquire the data, then the on-site environmental detection vector... The precipitation intensity is filled directly using a preset precipitation intensity and a preset set of point cloud data or obstacle distance parameters. In this case, the environmental parameter completeness is 0.6.
[0143] Receiver confidence for:
[0144] in, The signal-to-noise ratio (SNR) quality of a sound wave signal characterizes the actual signal-to-noise ratio of the sound wave. Compared with the reference signal-to-noise ratio The ratio of the actual signal-to-noise ratio of sound waves. Signal-to-noise ratio within the micro-jitter window Mean, signal-to-noise ratio quality of sound wave signal Represented as:
[0145] Environmental interference inhibitory factor Represented as:
[0146] in, This is a reference value for wind speed. For wind speed, This is a reference value for precipitation intensity. For precipitation intensity, wind speed and precipitation intensity The larger The smaller the value, the lower the reliability. In windy or rainy conditions, sound signals are severely affected by environmental noise, naturally reducing the confidence level of the receiver. Conversely, in quiet environments, receiver detection is more reliable.
[0147] Integrated security The system comprehensively considers the security scores and confidence levels of both the receiving and simulation ends; the more trustworthy score receives a higher weight. (This involves) fusing security scores. Represented as:
[0148] in, For the confidence level of the simulation end, For the confidence level of the receiving end, For comprehensive risk assessment at the simulation end, Assess the security of the simulation terminal. Assess the security score at the receiver. Receiver security score Represented as:
[0149] in, For receiver fusion deviation, receiver fusion deviation It is a weighted combination of acoustic wave deviation and back electromotive force deviation; The receiving end fusion decision threshold. This is a weighted combination of acoustic feature deviation from the decision threshold and back electromotive force deviation from the decision threshold. Receiver fusion deviation. Fusion decision threshold with receiver They are represented as follows:
[0150]
[0151] in, For sound wave deviation, The back electromotive force load deviation; If the acoustic characteristics deviate from the decision threshold, If the back electromotive force deviates from the decision threshold, , These are the weighting coefficients; + =1, calibrated according to sensor reliability.
[0152] when Automatic fusion of judgment values If the value is 1, the unlock command will be output. Automatic fusion of judgment values If the value is 0, an intercept command is output. Wherein, To integrate security scoring thresholds. In one embodiment, .
[0153] When inconsistency When the value is greater than or equal to the conflict threshold, the automatic decision-making process is exited, and a structured probability report is generated for human decision-making. The structured probability report is shown in Table 2.
[0154]
[0155] in, To integrate security components; For the confidence level of the simulation end, Confidence level at the receiving end; These are the two operating conditions with the highest risk in the simulation. Receiver fusion deviation; Inconsistency; This is a system suggestion.
[0156] Table 2 Structured Probability Report
[0157] Among them, system recommendations According to the integration security segment It is concluded that when the security is integrated... When the system outputs a suggested operation prompt, indicating a high overall safety score, and if manual operation times out, an automatic release action is executed; when the integrated safety score is high... When this occurs, the system outputs a cautionary prompt, indicating that manual confirmation is required in this scenario. If the manual operation times out, a default blocking action will be executed. When the security is integrated... If the operation is interrupted, the system will output a prompt suggesting that the operation should be abandoned, indicating that the overall security score is low. If the manual operation times out, the default blocking action will be executed.
[0158] As described above, these embodiments of this application do not exhaustively cover all details, nor do they limit the application to merely the specific embodiments described. Clearly, many modifications and variations can be made based on the above description. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of this application, thereby enabling those skilled in the art to effectively utilize this application and its modifications. This application is limited only by the claims and their full scope and equivalents.
Claims
1. A remote boot system with dual verification, characterized in that, include: The transmitting end generates a remote start signal in response to the switch control signal of the wireless remote control switch; The receiving end receives a remote start signal and enters a pre-start mode in response to the remote start signal; the receiving end includes a sensing module, which collects a snapshot of the initial state of the actual machine, a field environment detection vector, and acoustic signals and motor back electromotive force signals during the pre-start process; The simulation end is connected to the receiving end and performs shadow simulation based on the initial state snapshot of the real machine and the on-site environment detection vector to generate a first safety flag and transmit it to the judgment module. The receiving end further includes a determination module, which is used to generate a second safety flag based on the acoustic signal and the back electromotive force signal, and to generate a startup strategy based on the first safety flag and the second safety flag. When the first safety flag and the second safety flag are both 1, a start signal is output; when the first safety flag and the second safety flag are both 0, an interception start command is executed; when the first safety flag and the second safety flag are inconsistent, the inconsistency between the risk values of each working condition at the simulation end and the deviation of the receiving end is calculated, and a fusion decision is made or a structured probability report is generated for manual decision-making based on the inconsistency.
2. The remote boot system with dual verification according to claim 1, characterized in that, The sensing module includes a unit sensing module, a first environmental sensing module, and a second environmental sensing module; the unit sensing module collects a snapshot of the initial state of the actual machine; the first environmental sensing module collects the on-site environmental detection vector; and the second environmental sensing module collects the acoustic wave signal and the motor back electromotive force signal during the pre-start process.
3. The remote boot system with dual verification according to claim 1, characterized in that, The simulation terminal includes a storage module and an assembly module; the storage module stores a twin machine library and a working condition package; the assembly module receives the initial state snapshot of the real machine and the field environment detection vector, retrieves the twin package from the twin machine library, assembles it with the initial state snapshot of the real machine and the field environment detection vector, generates a shadow session instance and performs shadow simulation, and loads the working condition package during the simulation process.
4. The remote boot system with dual verification according to claim 3, characterized in that, The assembly module is configured to perform the following steps: Obtain the real device model identifier and real device flight control firmware version uploaded by the receiving end, and verify the real device identity and version of the receiving end; Retrieve and import the twin packet corresponding to the real device of the receiving end; Map the snapshot of the real device's initial state uploaded by the receiving end to a twin initial state; Load the operating condition package stored in the storage module; Generate a runnable twin simulation instance.
5. The remote boot system with dual verification according to claim 3, characterized in that, The simulation terminal also includes a calculation module, which defines the original anomaly metric and performs risk normalization, calculates the comprehensive risk score, and generates the first safety flag based on the comprehensive risk score.
6. The remote boot system with dual verification according to claim 1, characterized in that, The determination module is configured to calculate the acoustic wave deviation based on the acoustic wave signal, calculate the back electromotive force load deviation based on the back electromotive force signal, and obtain the second safety indicator based on the acoustic wave deviation and the back electromotive force load deviation.
7. The remote boot system with dual verification according to claim 1, characterized in that, The determination module is configured to perform the following steps: Establish a correlation matrix between the risk values of each operating condition at the simulation end and the deviation at the receiving end; The degree of inconsistency is calculated based on the aforementioned correlation matrix; as well as The steps of performing a fusion decision or generating a structured probability report are performed based on the comparison result between the inconsistency degree and the conflict threshold.
8. The remote boot system with dual verification according to claim 7, characterized in that, The determination module is configured to perform the following steps: Obtain the fusion security score, determine the automatic fusion decision value based on the fusion security score, and output the unlock command or the intercept command.
9. A remote startup method with dual verification, characterized in that, include: In the first mode, the transmitter responds to the switch control signal of the wireless remote control switch, generates a remote start signal, and sends it to the receiver. The radio frequency receiving module of the receiving end receives the remote start signal, and enters the pre-start mode in response to the remote start signal. In the pre-start mode, it collects the initial state snapshot of the real machine, the on-site environment detection vector, and the acoustic wave signal and motor back electromotive force signal during the pre-start process, and transmits the initial state snapshot of the real machine and the on-site environment detection vector to the simulation end. The simulation end performs shadow simulation based on the initial state snapshot of the real machine and the detection vector of the on-site environment, generates the first safety flag, and transmits it to the receiving end; The receiving end generates a second safety flag based on the acoustic signal and the back electromotive force signal, and generates a startup strategy based on the first safety flag and the second safety flag; The steps for generating a startup strategy based on the first security flag and the second security flag include: When both the first safety flag and the second safety flag are 1, a start signal is output. When both the first security flag and the second security flag are 0, the command to initiate interception is executed; When the first safety flag is inconsistent with the second safety flag, the degree of inconsistency between the risk values of each operating condition at the simulation end and the deviation of the receiving end is calculated, and a fusion decision is made or a structured probability report is generated for manual decision-making based on the degree of inconsistency.
10. The remote startup method with dual verification according to claim 9, characterized in that, The steps for performing shadow simulation at the simulation terminal include: Obtain the real device model identifier and real device flight control firmware version uploaded by the receiving end, and verify the real device identity and version of the receiving end; Retrieve and import the twin packet corresponding to the real device of the receiving end; Map the snapshot of the real device's initial state uploaded by the receiving end to a twin initial state; Load the work status package; Generate a runnable twin simulation instance.
11. The remote startup method with dual verification according to claim 9, characterized in that, The steps for generating the first safety flag include: defining an original anomaly metric and performing risk normalization, calculating a comprehensive risk score, and generating the first safety flag based on the comprehensive risk score.
12. The remote startup method with dual verification according to claim 9, characterized in that, The step of generating the second safety mark includes: calculating the acoustic wave deviation based on the acoustic wave signal, calculating the back electromotive force load deviation based on the back electromotive force signal, and obtaining the second safety mark based on the acoustic wave deviation and the back electromotive force load deviation.
13. The remote startup method with dual verification according to claim 9, characterized in that, Also includes: Establish a correlation matrix between the risk values of each operating condition at the simulation end and the deviation at the receiving end; The degree of inconsistency is calculated based on the aforementioned correlation matrix; as well as The steps of performing a fusion decision or generating a structured probability report are performed based on the comparison result between the inconsistency degree and the conflict threshold.
14. The remote startup method with dual verification according to claim 13, characterized in that, It also includes obtaining a fusion security score, determining an automatic fusion decision value based on the fusion security score, and outputting an unlock command or an intercept command.
