A method, equipment, and medium for training flight skills based on androids

Abstract

This invention discloses a method, device, and medium for flight skill training based on embodied robots, relating to the field of flight training technology. The method includes: performing manipulation correction and convergence teaching analysis on embodied flight training time-series data; determining the current flight training stage through stage anchor point convergence to generate a training stage determination map; generating state change characterization indicators based on the training stage determination map and embodied flight training time-series data through continuous quantitative evaluation; identifying unstable intervals of control strategies during flight training based on the state change characterization indicators; and selecting and executing embodied manipulation based on the uncertain intervals of strategy switching using a state change continuous control strategy selection algorithm, generating a flight skill training evaluation report. This invention improves the stability, relevance, and reliability of flight skill training and evaluation by temporally correlating flight state data and embodied manipulation data.

Description

Technical Field

[0001] This invention relates to the field of flight training technology, and in particular to a flight skills training method, equipment and medium based on an embodied robot. Background Technology

[0002] In recent years, embodied robots, with their physical execution capabilities, motion perception capabilities, and direct interaction with flight control interfaces, have enabled training processes to not only record changes in flight status but also simultaneously collect control behaviors and feedback results, thus providing support for detailed analysis of the training process. Especially in flight operation skills training, establishing training analysis mechanisms around continuous processes such as attitude control, power adjustment, control correction, and status feedback has become an important direction for improving the standardization and reproducibility of training. Flight training technology is geared towards flight skills training and instructional analysis scenarios, excluding the manufacturing of aviation-related equipment. The technological focus is on data organization, stage identification, strategy determination, and training evaluation during the training process, without involving aircraft, aero-engines, airborne equipment, or related manufacturing processes.

[0003] However, existing technologies still have two shortcomings in flight skill training involving embodied robots. Firstly, current training techniques largely focus on judging flight results or single-shot control performance, lacking detailed analysis of control correction convergence and the formation of patterns in training stages, making it difficult to support phased and targeted training. Secondly, existing technologies are insufficient in identifying critical periods of strategy switching during flight training. When flight status fluctuates continuously or the strategy is not yet stable, they still rely heavily on single-point thresholds or static rules for judgment, leading to inaccurate identification of key training segments and low precision in training feedback and evaluation. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a flight skill training method based on an embodied robot to solve the problems of difficulty in accurately characterizing the training phase and difficulty in effectively identifying key sections of strategy switching.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: In a first aspect, the present invention provides a flight skill training method based on an embodied robot, comprising: collecting flight state data and embodied control data, performing time alignment processing to generate embodied flight training time series data; performing control correction convergence teaching analysis on the embodied flight training time series data, and determining the current flight training stage through stage anchor point convergence to generate a training stage determination map; generating a state change characterization index through continuous quantitative evaluation based on the training stage determination map and the embodied flight training time series data; identifying the unstable interval of the control strategy during flight training based on the state change characterization index, and obtaining the uncertain interval of strategy switching; selecting a strategy and executing embodied control through a state change continuous control strategy selection algorithm based on the uncertain interval of strategy switching, and generating a flight skill training evaluation report.

[0007] As a preferred embodiment of the flight skill training method based on embodied robots described in this invention, the embodied flight training time sequence data is generated by aligning and registering the flight state data and embodied control data with timestamps and then associating and organizing them according to the chronological order.

[0008] As a preferred embodiment of the flight skill training method based on android described in this invention, the specific steps of manipulating and correcting convergent teaching analysis of the android flight training time series data are as follows: The time-series data of the embodied flight training is continuously segmented according to the time sequence to obtain training analysis segments; The flight state sequence and embodied control sequence are extracted from the training analysis segment, and the control correction effect of the corresponding training analysis segment is determined based on the changes in the flight state sequence before and after the embodied control sequence is applied. The manipulation and correction effects on the training analysis segments are continuously compared in chronological order to determine the convergence trend of the manipulation.

[0009] As a preferred embodiment of the flight skill training method based on embodied robots described in this invention, the training stage determination map is generated by extracting stage anchor points based on the control convergence trend, performing convergence determination on the training analysis segments around the stage anchor points, determining the flight training stage to which the training analysis segments belong, and performing associated registration.

[0010] As a preferred embodiment of the flight skill training method based on embodied robots described in this invention, the specific steps for generating state change characterization indicators through continuous quantitative evaluation based on the training phase judgment map and embodied flight training time series data are as follows: Based on the training phase determination map, the time series data of the embodied flight training is segmented and extracted to obtain the phase state analysis sequence. Perform differential calculations on the flight states at adjacent time points in the phase state analysis sequence to obtain the flight state change sequence; Continuous calculations are performed on the flight state change sequence to generate state change characterization indicators.

[0011] As a preferred embodiment of the flight skill training method based on android described in this invention, the specific steps for identifying the unstable interval of the control strategy during flight training based on state change characterization indicators are as follows: The state change characterization indicators are continuously scanned in chronological order to obtain a set of abnormal fluctuation segments. The consistency analysis of the direction of change and the stability analysis of the amplitude of change are performed on the abnormal fluctuation fragment set to determine the state swing characteristics, and the unstable interval of the control strategy is identified based on the state swing characteristics.

[0012] As a preferred embodiment of the flight skill training method based on embodied robots described in this invention, the uncertain interval of strategy switching is obtained by sorting out the position association and registering the duration of the unstable interval of the control strategy.

[0013] As a preferred embodiment of the flight skill training method based on embodied robots described in this invention, the specific steps of selecting a strategy and executing embodied control based on a continuous state change control strategy selection algorithm according to the uncertain interval of strategy switching, and generating a flight skill training evaluation report are as follows: Extract flight state sequences and embodied control sequences corresponding to the uncertainty intervals of strategy switching from embodied flight training time series data; Perform continuity calculations on the flight state sequence to obtain the degree of continuity improvement; Compare the flight state sequences before and after the application of the embodied control sequence to obtain state correction information; The target control strategy is determined based on the degree of continuous improvement and state correction information, and the embodied robot is driven to execute the target control strategy to obtain flight state feedback. The target control strategy and flight status feedback are correlated and organized to generate a flight skills training evaluation report.

[0014] In a second aspect, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, it implements any step of the flight skill training method based on an embodied robot as described in the first aspect of the present invention.

[0015] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the flight skill training method based on an embodied robot as described in the first aspect of the present invention.

[0016] The beneficial effects of this invention are as follows: by temporally correlating the flight status data and embodied control data during the flight training process of the embodied robot, and combining the training stage determination, the quantification of the continuity of state changes and the identification of the unstable interval of the strategy, the invention achieves accurate positioning of key sections and targeted control strategy selection without excluding the manufacturing of aviation-related equipment, thereby improving the stability, relevance and reliability of flight skill training. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of a flight skills training method based on androids.

[0019] Figure 2 A flowchart for generating the decision graph during the training phase.

[0020] Figure 3 A flowchart for generating indicators representing state changes.

[0021] Figure 4 A flowchart for generating a flight skills training assessment report. Detailed Implementation

[0022] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0023] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0024] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0025] Reference Figures 1-4 As one embodiment of the present invention, this embodiment provides a flight skill training method based on an embodied robot, comprising the following steps: S1: Collect flight status data and embodied control data, perform time alignment processing, and generate embodied flight training time sequence data.

[0026] S1.1: Flight status data includes attitude angle, angular velocity, airspeed, altitude, vertical velocity, power output status, and alarm status.

[0027] Specifically, the attitude measurement results are collected by reading the attitude measurement results from the flight training equipment. The attitude measurement results include pitch angle, roll angle and yaw angle, and the pitch angle, roll angle and yaw angle at the corresponding time are recorded as attitude angles.

[0028] The data is collected by reading the angular velocity measurement results from the flight training equipment. The angular velocity measurement results include the rotational speeds around the pitch axis, roll axis, and yaw axis, and the rotational speeds of each axis at the corresponding time are recorded as angular velocities.

[0029] The airspeed data is collected by reading the airspeed measurement results from the flight training equipment. The airspeed measurement result is the speed relative to the surrounding air, and the speed value at the corresponding moment is recorded as the airspeed.

[0030] The altitude is collected by reading the altitude measurement results from the flight training equipment. The altitude measurement result is the vertical position relative to the reference plane, and the altitude value at the corresponding moment is recorded as the altitude.

[0031] The data is collected by reading the vertical motion measurement results from the flight training equipment. The vertical motion measurement results are the upward or downward speed in the vertical direction, and the vertical speed value at the corresponding moment is recorded as the vertical speed.

[0032] The operating parameters of the power unit are collected by reading the operating parameters of the flight training equipment. These parameters include thrust output status, power output status, or speed output status. The operating parameters of the power unit at the corresponding moment are then recorded as the power output status.

[0033] The alarm information is collected by reading the output results of the flight training equipment. The output results include whether the alarm is triggered, the alarm type, and the alarm duration. The alarm information at the corresponding time is then registered as the alarm status.

[0034] S1.2: The specific control data includes joystick displacement, throttle displacement, control surface control quantity, control force feedback, and control action occurrence time information.

[0035] Specifically, the position change results of the robot end effector acting on the joystick are obtained, and the deflection position of the joystick at each acquisition moment is recorded as the joystick displacement.

[0036] The positional changes of the throttle device during the push-pull adjustment process by the embodied robot are obtained, and the movement position of the throttle device at each acquisition time is recorded as the throttle displacement.

[0037] The control outputs corresponding to the current maneuver of the embodied robot are obtained by reading the control surface outputs from the flight training equipment, and the aileron control, elevator control, and rudder control at the corresponding moment are registered as control surface control quantities.

[0038] The results of force changes experienced by the embodied robot during joystick and throttle operations are collected, and the force feedback values ​​at each collection moment are recorded as control force feedback.

[0039] The start time, duration, and end time of each control action are recorded when the embodied robot performs various control actions. The time corresponding to each control action is then recorded as the control action occurrence time information.

[0040] S1.3: Time-stamp alignment and sequence registration of flight status data and embodied control data, and correlation and organization according to time sequence to generate embodied flight training time sequence data.

[0041] Specifically, the data collection times corresponding to the flight status records in the flight status data and the data collection times corresponding to the embodied control records in the embodied control data are read separately, and flight status time series and embodied control time series are formed from morning to night according to the data collection times; based on the data collection times in the flight status time series, embodied control records with the same data collection times are searched in the embodied control time series, and a correspondence is established between the embodied control records with the same data collection times and the corresponding flight status records.

[0042] When there are no embodied control records with the same acquisition time in the embodied control time series, the time interval values ​​corresponding to the adjacent embodied control records before the current acquisition time and the adjacent embodied control records after the current acquisition time are determined respectively, and the embodied control record with the first ranking according to the sorting result of the time interval values ​​is selected to establish a correspondence; when the two time interval values ​​are the same, the embodied control record with the earlier acquisition time is selected to establish a correspondence; when the time difference between the acquisition time corresponding to the flight status record and the acquisition time corresponding to the embodied control record is not the same, no correspondence is established and the corresponding record is removed, thereby completing the timestamp alignment.

[0043] After completing the timestamp alignment, the flight status records and embodied control records that have established corresponding relationships are uniformly sorted according to the collection time to complete the sequential registration; the sequentially registered flight status records, embodied control records and corresponding collection times are written into the same associated record and arranged continuously in chronological order to generate embodied flight training time series data.

[0044] S2: Perform manipulation, correction, convergence, and teaching analysis on the time-series data of embodied flight training, and determine the current flight training stage through stage anchor point convergence, generating a training stage determination map.

[0045] S2.1: Perform continuous segmentation processing on the time-series data of the embodied flight training to obtain training analysis segments.

[0046] Specifically, first read each associated record in the time sequence data of the embodied flight training, and then arrange them continuously from morning to night according to the collection time corresponding to each associated record.

[0047] The acquisition times corresponding to adjacent associated records are compared sequentially, and it is determined whether the adjacent associated records maintain a continuous temporal relationship based on whether there is a time interruption between the acquisition times of the adjacent associated records. When the adjacent associated records maintain a continuous temporal relationship, the adjacent associated records are grouped into the same continuous record segment. When the adjacent associated records do not maintain a continuous temporal relationship, the current continuous record segment is determined as a training analysis segment, and the segments are arranged continuously in chronological order to obtain the training analysis segment.

[0048] S2.2: Extract flight state sequences and embodied control sequences from training analysis segments, and determine the control correction effect of the corresponding training analysis segments based on the changes in flight state sequences before and after the embodied control sequences take effect.

[0049] Specifically, the associated records in the training analysis segment are arranged from morning to night according to the collection time, and the flight status data in the associated records are continuously organized into a flight status sequence, and the embodied control data in the associated records are continuously organized into an embodied control sequence.

[0050] Read the embodied control records one by one according to the acquisition time in the embodied control sequence, and read the flight state records before and after the current embodied control record in the flight state sequence. Compare the attitude angle, angular velocity, airspeed, altitude, vertical velocity, power output status and alarm status in the flight state records before and after the action to obtain the flight state change results of the corresponding embodied control record.

[0051] The flight state changes were continuously recorded according to the order of the acquisition time in the embodied control sequence, and the continuous registration results were determined as the control correction effect corresponding to the training analysis segment.

[0052] S2.3: Perform continuous comparisons of the manipulation correction effects on the training analysis segments in chronological order to determine the manipulation convergence trend.

[0053] Specifically, the control correction effect corresponding to the previous training analysis segment is compared with the control correction effect corresponding to the next training analysis segment in sequence. During the comparison, based on the same flight state items in the flight state change results, it is determined whether the flight state change results corresponding to the next training analysis segment converge relative to the flight state change results corresponding to the previous training analysis segment in terms of the number of repetitions, the dispersion of changes, and the stability of changes.

[0054] Comparison results that converge are recorded as convergent change results, and comparison results that do not converge are recorded as non-convergent change results. The results of convergent change results and non-convergent change results are continuously sorted according to their chronological order, and the continuous sorting results are determined as the control convergence trend.

[0055] S2.4: Extract stage anchor points based on the control convergence trend, and determine the convergence of training analysis segments around the stage anchor points to identify the flight training stage to which the training analysis segments belong, and register the association to generate a training stage determination map.

[0056] Specifically, according to the chronological order of the training analysis segments in the embodied flight training time series data, the control convergence trend corresponding to each training analysis segment is read sequentially, and the control convergence trend corresponding to each training analysis segment is judged in sequence.

[0057] When the trend of convergence is manifested by continuous non-convergent changes, the corresponding training analysis segment is determined to be in the initial adjustment stage; when the trend of convergence is manifested by alternating convergent and non-convergent changes, the corresponding training analysis segment is determined to be in the convergence formation stage; when the trend of convergence is manifested by continuous convergent changes, the corresponding training analysis segment is determined to be in the stable mastery stage.

[0058] According to the chronological order of the training analysis segments in the embodied flight training time series data, the start time, end time, flight training stage, and control convergence trend corresponding to the training analysis segments are continuously registered, and the continuous registration results are determined as the training stage judgment map.

[0059] S3: Based on the training phase judgment map and the time series data of the embodied flight training, a state change characterization index is generated through continuous quantitative evaluation.

[0060] S3.1: Based on the training phase judgment map, the time series data of the embodied flight training is segmented and extracted to obtain the phase state analysis sequence.

[0061] Specifically, the start and end times of each training analysis segment and the flight training stage are read sequentially according to the time sequence in the training stage determination map. Based on the start and end times of each training analysis segment, the associated records within the corresponding time range are searched from the embodied flight training time series data. The associated records within the same flight training stage are then extracted continuously from the beginning to the end of the time.

[0062] The flight status data extracted from the associated records are continuously organized according to the order of collection time, and the results of the continuous organization are determined as the stage status analysis sequence.

[0063] S3.2: Perform differential calculation on the flight state at adjacent time points in the phase state analysis sequence to obtain the flight state change sequence.

[0064] Specifically, each flight status record is read sequentially from morning to night according to the collection time in the phased state analysis sequence, and the previous flight status record is paired with the next flight status record in chronological order.

[0065] Differential calculations are performed on the attitude angle, angular velocity, airspeed, altitude, vertical velocity, power output status, and alarm status in the previous and subsequent flight status records, respectively. The differential calculations calculate the difference between the subsequent flight status record and the previous flight status record for the attitude angle, angular velocity, airspeed, altitude, vertical velocity, power output status, and alarm status.

[0066] The difference between the flight status records corresponding to each group of adjacent flight status records is recorded as a flight status change result; all flight status change results are continuously arranged from morning to night according to the collection time, and the continuous arrangement result is determined as the flight status change sequence.

[0067] S3.3: Perform continuous calculations on the flight state change sequence to generate state change characterization indicators.

[0068] Specifically, the flight status change results are read sequentially from morning to night according to the collection time in the flight status change sequence, and the previous flight status change result is paired with the next flight status change result according to the time sequence.

[0069] For each group of adjacent flight state change results, continuous calculations are performed on the attitude angle difference, angular velocity difference, airspeed difference, altitude difference, vertical velocity difference, power output state difference, and alarm state difference to obtain state change characterization indicators. The state change characterization indicators corresponding to each flight state item are then continuously arranged from morning to night according to the collection time.

[0070] The expression for calculating the index characterizing state change is: ; in, It serves as a characterization index for state changes. For the number of flight status items, This is the sequence number of the flight status item. This refers to the time sequence number of the flight state change result within the flight state change sequence. For the first The first result of the flight status change The difference corresponding to each flight status item To indicate the first The first result of the flight status change The difference corresponding to each flight status item It is a symbolic function.

[0071] It should be noted that the sign function is used to determine the consistency of the direction of change between flight state change results. It takes the value 1 when adjacent flight state change results are in the same direction, -1 when they are in opposite directions, and 0 when there is no change.

[0072] State change characterization indicators are used to quantitatively characterize the degree, direction, and continuity of changes in flight state and embodied control state during flight training. They can transform originally scattered state fluctuations into comparable and identifiable continuous data, which helps to more accurately identify unstable control strategy intervals and uncertain strategy switching intervals, thereby improving the pertinence and reliability of subsequent strategy selection and training evaluation.

[0073] S4: Based on state change characterization indicators, identify the unstable interval of control strategy during flight training and obtain the uncertain interval of strategy switching.

[0074] S4.1: Perform a continuous scan of the state change characterization indicators in chronological order to obtain a set of abnormal fluctuation segments.

[0075] Specifically, each state change indicator is read sequentially from morning to night according to the collection time corresponding to the state change indicator, and the previous state change indicator is compared with the next state change indicator in chronological order.

[0076] During the adjacent comparison process, it is determined whether the change direction of the previous state change indicator is consistent with that of the change direction of the next state change indicator, and whether the change amplitude of the change direction of the next state change indicator is fluctuating relative to that of the previous state change indicator. When the change direction is inconsistent, or the change amplitude shows a sudden increase, a sudden decrease, or fluctuating change, the corresponding time position is registered as an abnormal fluctuation result. When the change direction is consistent, and the change amplitude does not show a sudden increase, a sudden decrease, or fluctuating change, the corresponding time position is registered as a non-abnormal fluctuation result.

[0077] The results of consecutive abnormal fluctuations are merged into segments, and the merged segments are determined as a set of abnormal fluctuation segments.

[0078] S4.2: Perform consistency analysis of change direction and stability analysis of change amplitude on the abnormal fluctuation fragment set to determine the state swing characteristics, and identify the unstable interval of the control strategy based on the state swing characteristics.

[0079] Specifically, each state change indicator is compared according to its adjacent arrangement to determine whether adjacent state change indicators maintain the same direction of change and whether adjacent state change indicators remain stable in terms of change magnitude.

[0080] When the state change characterization indicators in the same abnormal fluctuation segment repeatedly show inconsistencies in the direction of change, or repeatedly show sudden increases, sudden decreases, or reciprocal changes in the amplitude of change, the corresponding abnormal fluctuation segment is identified as an abnormal fluctuation segment with state swing characteristics. After completing the consistency analysis of the direction of change and the stability analysis of the amplitude of change of all abnormal fluctuation segments, the start and end acquisition times corresponding to the abnormal fluctuation segments with state swing characteristics are registered, and the registration results are determined as the unstable interval of the control strategy.

[0081] It should be noted that state oscillation characteristics refer to the repeated deviations and returns of flight state parameters and embodied control parameters around a certain target state within a short period of time during flight training. This is usually manifested as continuous back-and-forth fluctuations in direction, amplitude, or adjustment rhythm. State oscillation characteristics can reflect that the control strategy is not yet stable and that control corrections are still in a stage of repeated trial and error. Therefore, it can be used to identify signs of instability such as overcorrection, delayed correction, or strategy switching during the training process.

[0082] S4.3: Perform position correlation and duration registration for unstable control strategies to form uncertain intervals for strategy switching.

[0083] Specifically, the unstable intervals of the previous control strategy are compared with the unstable intervals of the next control strategy in chronological order. During the comparison, it is determined whether there is an interruption time between the end of the acquisition of the unstable interval of the previous control strategy and the start of the acquisition of the unstable interval of the next control strategy.

[0084] When there is no interruption time between the end of the data collection and the start of the data collection, the unstable intervals of the previous control strategy and the unstable intervals of the next control strategy are correlated and organized, and the continuous time range after the correlation and organization is determined as the same uncertain interval for switching strategies; when there is an interruption time between the end of the data collection and the start of the data collection, the unstable intervals of the previous control strategy and the unstable intervals of the next control strategy are kept to correspond to different uncertain intervals for switching strategies.

[0085] Read the start and end times of each strategy switching uncertainty interval, and register the duration based on the start and end times. Then, determine the location association and duration registration results as the strategy switching uncertainty interval.

[0086] S5: Based on the uncertain interval of strategy switching, the strategy selection algorithm is used to select a strategy and execute the embodied control, generating a flight skill training evaluation report.

[0087] S5.1: Extract the flight state sequence and embodied control sequence corresponding to the strategy switching uncertainty interval from the embodied flight training time series data.

[0088] Specifically, each uncertain interval of strategy switching is read sequentially from morning to night according to the start time corresponding to the uncertain interval of strategy switching, and the corresponding end time of acquisition is read for each uncertain interval of strategy switching.

[0089] Based on the start and end times of data collection, search for related records within the corresponding time range in the embodied flight training time series data; after completing the search for related records, continuously extract related records within the time range corresponding to the same strategy switching uncertainty interval according to the collection time from early to late.

[0090] The flight status data extracted from the associated records are continuously organized according to the order of collection time, and the continuous organization result is determined as the flight status sequence corresponding to the uncertainty interval of strategy switching. At the same time, the embodied control data extracted from the associated records are continuously organized according to the order of collection time, and the continuous organization result is determined as the embodied control sequence corresponding to the uncertainty interval of strategy switching.

[0091] S5.2: Perform continuity calculations on the flight state sequence to obtain the degree of continuity improvement.

[0092] Specifically, the flight status records are read sequentially from morning to night according to the acquisition time in the flight status sequence, and the previous flight status record is paired with the next flight status record in chronological order. The attitude angle, angular velocity, airspeed, altitude, vertical velocity, power output status and alarm status in the previous and next flight status records are compared to obtain the change direction, change magnitude and change reversal of each flight status item.

[0093] The number of times the change direction is consistent, the difference in change amplitude, the number of times the change reverses, and the length of continuous recording are counted for each flight state item. Then, the degree of continuity improvement corresponding to the flight state sequence is calculated based on the number of times the change direction is consistent, the difference in change amplitude, the number of times the change reverses, and the length of continuous recording.

[0094] The expression for calculating the degree of continuous improvement is: ; in, To ensure continuous improvement The number of times the direction of change is consistent. For the number of reversals, The number of abnormal fluctuations. The length of the continuous record.

[0095] S5.3: Compare the flight state sequences before and after the implementation of the embodied control sequence to obtain state correction information.

[0096] Specifically, the attitude angle, angular velocity, airspeed, altitude, vertical velocity, power output status, and alarm status in the previous and subsequent flight status records are compared item by item to determine the attitude angle correction result, angular velocity correction result, airspeed correction result, altitude correction result, vertical velocity correction result, power output status correction result, and alarm status correction result respectively.

[0097] After determining the correction results corresponding to each embodied control record, the attitude angle correction results, angular velocity correction results, airspeed correction results, altitude correction results, vertical velocity correction results, power output state correction results, and alarm state correction results are continuously registered according to the order of collection time in the embodied control sequence. The results of attitude angle, angular velocity, airspeed, altitude, vertical velocity, power output state, and alarm state that move toward the target training state are determined as improved results, and the results of deviating from the target training state are determined as deteriorated results. The state correction information is formed by summarizing the number of improved results and deteriorated results.

[0098] S5.4: Determine the target control strategy based on the degree of continuous improvement and state correction information, and drive the embodied robot to execute the target control strategy to obtain flight state feedback.

[0099] Specifically, each embodied control record is read sequentially from morning to night according to the collection time in the embodied control sequence. Each embodied control record is then registered with the corresponding continuous improvement degree and state correction information to form the continuous improvement degree and state correction information corresponding to the embodied control record.

[0100] According to the chronological order in the embodied control sequence, the degree of continuous improvement corresponding to each embodied control record is compared sequentially, and the state correction information corresponding to each embodied control record is compared sequentially. The embodied control record in which the degree of continuous improvement increases continuously relative to the previous embodied control record and the number of improvement results in the state correction information increases continuously relative to the previous embodied control record is determined as the target control strategy.

[0101] The robot is driven to perform corresponding embodied control based on the joystick displacement, throttle displacement, control surface control quantity, control force feedback, and control action occurrence time information corresponding to the target control strategy. According to the acquisition time corresponding to the target control strategy, the flight status record located after the time position corresponding to the target control strategy is read from the embodied flight training time sequence data. The acquired attitude angle, angular velocity, airspeed, altitude, vertical velocity, power output status, and alarm status are continuously organized from morning to night according to the acquisition time to obtain flight status feedback.

[0102] S5.5: Correlate and organize target control strategies and flight status feedback to generate a flight skills training evaluation report.

[0103] Specifically, each target control strategy is read sequentially from morning to night according to the acquisition time corresponding to the target control strategy. Then, based on the acquisition time corresponding to each target control strategy, the flight status record after the time position corresponding to the target control strategy is searched in the flight status feedback.

[0104] The stick displacement, throttle displacement, control surface control quantity, control force feedback, and control action occurrence time information in the target control strategy are matched one-to-one with the attitude angle change results, angular velocity change results, airspeed change results, altitude change results, vertical velocity change results, power output status change results, and alarm status change results in the flight status record.

[0105] The registration results are arranged continuously from morning to night according to the collection time, and the continuous arrangement results are determined as the flight skills training assessment report.

[0106] This embodiment also provides a computer device applicable to the flight skill training method based on embodied robots, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the flight skill training method based on embodied robots as proposed in the above embodiment.

[0107] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0108] This embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements the flight skill training method based on an embodied robot as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0109] In summary, this invention improves the stability, relevance, and reliability of flight skill training by: sequentially correlating flight status data and embodied control data during the flight training process of the embodied robot, combining training phase determination, quantification of the continuity of state changes, and identification of unstable policy intervals, thereby achieving precise positioning of key segments and targeted control strategy selection without excluding the manufacturing of aviation-related equipment.

[0110] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for training flight skills based on an embodied robot, characterized in that, include: Collect flight status data and embodied control data, and perform time alignment processing to generate embodied flight training time sequence data; Manipulate and correct convergent teaching analysis on the time sequence data of embodied flight training, and determine the current flight training stage by stage anchor point convergence, and generate a training stage determination map. Based on the training phase judgment map and the time series data of the embodied flight training, a state change characterization index is generated through continuous quantitative evaluation. Based on state change characterization indicators, identify the unstable interval of control strategy during flight training and obtain the uncertain interval of strategy switching. Based on the uncertainty interval of strategy switching, the strategy selection algorithm is used to select a strategy and execute embodied control, generating a flight skill training evaluation report.

2. The flight skill training method based on a android as described in claim 1, characterized in that, The embodied flight training time sequence data is generated by aligning and registering the flight status data and embodied control data with timestamps, and then associating and organizing them in chronological order.

3. The flight skill training method based on a android as described in claim 2, characterized in that, The specific steps for manipulating and correcting convergent teaching analysis of the embodied flight training time series data are as follows: The time-series data of the embodied flight training is continuously segmented according to the time sequence to obtain training analysis segments; The flight state sequence and embodied control sequence are extracted from the training analysis segment, and the control correction effect of the corresponding training analysis segment is determined based on the changes in the flight state sequence before and after the embodied control sequence is applied. The manipulation and correction effects on the training analysis segments are continuously compared in chronological order to determine the convergence trend of the manipulation.

4. The flight skill training method based on a android as described in claim 1 or 3, characterized in that, The training phase determination map is generated by extracting phase anchor points based on the control convergence trend, and performing convergence determination on the training analysis segments around the phase anchor points to determine the flight training phase to which the training analysis segments belong, and then performing associated registration.

5. The flight skill training method based on a android as described in claim 1, characterized in that, The process involves generating state change characterization indicators based on the training phase judgment map and the time series data of embodied flight training through continuous quantitative evaluation. The specific steps are as follows: Based on the training phase determination map, the time series data of the embodied flight training is segmented and extracted to obtain the phase state analysis sequence. Perform differential calculations on the flight states at adjacent time points in the phase state analysis sequence to obtain the flight state change sequence; Continuous calculations are performed on the flight state change sequence to generate state change characterization indicators.

6. The flight skill training method based on a android as described in claim 5, characterized in that, The specific steps for identifying unstable control strategies during flight training based on state change characterization indicators are as follows: The state change characterization indicators are continuously scanned in chronological order to obtain a set of abnormal fluctuation segments. The consistency analysis of the direction of change and the stability analysis of the amplitude of change are performed on the abnormal fluctuation fragment set to determine the state swing characteristics, and the unstable interval of the control strategy is identified based on the state swing characteristics.

7. The flight skill training method based on a android as described in claim 6, characterized in that, The uncertain interval for strategy switching is obtained by associating and recording the location and duration of the unstable interval of the control strategy.

8. The flight skill training method based on a android as described in claim 1, characterized in that, The specific steps for switching uncertain intervals based on strategy, selecting a strategy and executing embodied control through a state change continuity control strategy selection algorithm, and generating a flight skill training evaluation report are as follows: Extract flight state sequences and embodied control sequences corresponding to the uncertainty intervals of strategy switching from embodied flight training time series data; Perform continuity calculations on the flight state sequence to obtain the degree of continuity improvement; Compare the flight state sequences before and after the application of the embodied control sequence to obtain state correction information; The target control strategy is determined based on the degree of continuous improvement and state correction information, and the embodied robot is driven to execute the target control strategy to obtain flight state feedback. The target control strategy and flight status feedback are correlated and organized to generate a flight skills training evaluation report.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the flight skill training method based on a holographic robot as described in any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the flight skill training method based on a android as described in any one of claims 1 to 8.

Images