Intelligent robot target control system with counterattack function

Abstract

The application discloses a kind of intelligent robot target control systems with reverse attack function, belong to live ammunition tactical training equipment control technical field;The system aims at solving the technical problems in the prior art, such as the precision of being hit and the operation load cannot be considered, motor control cannot simultaneously satisfy the mobile response and firing steady state, firing timing and chassis steady state misplacement, full-link no closed-loop collaboration.Its technical scheme includes being hit sensing module, mobile control module and firing control module, being hit sensing module determines the zero point of being hit time by slope characteristic and completes timing alignment, mobile control module dynamically switches motor current limiting interval and generates chassis steady state window, and firing control module determines firing opportunity by double window matching and completes timing compensation.The application realizes full-link closed-loop collaborative management and control, greatly improves the precision of being hit positioning, chassis stability and dynamic scene counterattack hit rate.

Description

Technical Field

[0001] This invention relates to the field of control technology for live-fire tactical training equipment, and in particular to an intelligent robot target control system with a reverse attack function. Background Technology

[0002] As live-fire tactical combat training develops towards realism and precision, intelligent robotic targets with reverse attack capabilities have become core training equipment in military tactical training bases and civilian professional shooting ranges. They can simulate combat targets through autonomous movement and perform evasive maneuvers and reverse attacks after being hit, achieving two-way dynamic combat training and improving training effectiveness.

[0003] The existing technology has the following core technical problems in practical applications: First, the impact signal acquisition adopts a global fixed frequency scheme. Fixed low-frequency acquisition cannot accurately capture the nanosecond-level transient impact signal, resulting in insufficient impact positioning accuracy. On the other hand, global fixed high-frequency acquisition will generate a large amount of invalid redundant data, increasing the computing load of the controller. Second, the target moving motor adopts a fixed current parameter control mode, which cannot simultaneously take into account the high torque and fast response required for evasive maneuvers and the high stability of the chassis required for counterattack firing, resulting in poor stability of the counterattack trajectory. Third, the counterattack firing adopts an open-loop fixed time delay calibration mode, which does not adjust the firing sequence according to the real-time steady-state state of the target. This easily leads to the problem of misalignment between the firing sequence and the chassis steady-state window, resulting in a low hit rate of the counterattack. Furthermore, it cannot achieve closed-loop coordinated management of the entire process of impact perception, maneuver control, and firing execution, and cannot meet the usage requirements of high-dynamic live-fire combat training. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and propose an intelligent robot target control system with a reverse attack function.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: an intelligent robot target control system with reverse attack function includes: The impact perception module is used to acquire the impact transient electrical signal collected by the sensor, determine the impact time zero point by analyzing the slope change characteristics of the impact transient electrical signal, extract the impact features based on the time offset alignment of the multi-channel signal, generate the target firing time window, and generate an impact flag bit to send to the motion control module. The motor control module is used to dynamically switch the drive current limiting range of the chassis motor according to the impact flag, and to determine the physical stability of the system based on the chassis multi-dimensional state convergence data, and generate a chassis steady-state window. The firing control module is used to calculate the overlap between the target firing time window and the chassis steady-state window to determine the actual firing time point, and to perform timing advance compensation on the actual firing time point in combination with the real-time state variables of the physical action actuator, and output firing control commands.

[0006] As a further aspect of the present invention, the impact sensing module includes a dual-frequency sampling unit and a timing alignment unit; The dual-frequency sampling unit is used to poll data at a first sampling frequency in standby mode, and to switch to a second sampling frequency for acquisition when the slope change characteristics of the impact transient electrical signal meet the trigger condition. The second sampling frequency is greater than the first sampling frequency. The timing alignment unit is used to calculate the timing delay compensation of the remaining impacted sensor channels based on the first sensor channel that meets the triggering condition, to pull forward and align the sampling time axis of all channels, and to extract the target firing time window centered on the average peak time after alignment.

[0007] As a further aspect of the present invention, the maneuver control module incorporates a multi-stage current limiting strategy, wherein the current limiting range includes a steady-state operating range, an evasive maneuver range, and a firing steady-state range. After receiving the impact flag from the impact sensing module, the maneuver control module raises the limiting threshold of the chassis motor to the evasive maneuver range; after receiving the firing pre-trigger signal, it lowers the limiting threshold to the firing steady-state range, and enters a stable control state by eliminating the synchronous current difference between each drive wheel group.

[0008] As a further embodiment of the present invention, the physical action actuator is an electromagnetic pneumatic valve, and the real-time state variable includes the real-time pressure of the air path; The firing control module includes a feedforward compensation unit; The feedforward compensation unit is used to calculate the dynamic physical delay increment of the electromagnetic pneumatic valve based on the deviation between the real-time pressure of the air circuit and the standard calibration pressure, and to add the dynamic physical delay increment to the basic mechanical delay of the system as a timing advance to pre-correct the output timing of the actual firing time.

[0009] As a further aspect of the present invention, the determination logic for the triggering condition is as follows: The transient slope of the sampling points is calculated using the sliding window difference algorithm, and the formula is as follows: , in, This is the current voltage value. The sampling period is This is the voltage value two sampling periods prior to the current moment; When the absolute value of the transient slope for a consecutive preset number of sampling periods is greater than a preset slope threshold, the trigger condition is determined to be met.

[0010] As a further aspect of the present invention, within the evasive maneuver zone, the maneuver control module calculates the current adjustment coefficient in real time based on the real-time steering angle and real-time acceleration using the following nonlinear weighted model: , in, This is the current adjustment coefficient. The maximum physical steering angle, For maximum physical acceleration, and To pre-set weight calibration values, For real-time acceleration, This is the real-time steering angle.

[0011] As a further embodiment of the present invention, the firing control module further includes a feedback calibration unit; The feedback calibration unit is used to collect historical actual firing times, calculate the difference between the actual firing time and the time when the system issues the firing control command as the timing error, and use a proportional, integral, and derivative closed-loop algorithm to process the timing error, and add the processed compensation amount to the timing advance amount calculated in the next iteration.

[0012] As a further aspect of the present invention, when the impact sensing module receives multiple impact signal sources simultaneously, the response target is determined by calculating the timing priority: , in, For timing priority, The amplitude of the impact wave crest. This is the reference value for the maximum amplitude of the impact wave peak as calibrated by the system. It is the reciprocal of the distance from the point of impact to the pre-set core area. Pre-determine the hazard level for armor plates. These are the weighting coefficients; The impact sensing module extracts The firing time window of the target is derived from the highest value of the hit signal source.

[0013] As a further aspect of the present invention, the triggering condition for the motion control module to determine the physical stability of the system and generate the chassis steady-state window is that the following characteristics are simultaneously met and the duration is greater than a preset determination period: The variance of phase current fluctuation in each drive motor is less than the first threshold. The pitch and yaw angular velocities output by the chassis inertial measurement unit are both less than the second threshold. The spatial pointing deviation angle of the target aiming line is less than the third threshold.

[0014] As a further aspect of the present invention, the system adopts a heterogeneous processing architecture; The impact sensing module runs in a field-programmable gate array processor and is used to perform hardware-level parallel calculation and time-axis alignment of impact signals. The motion control module and the firing control module operate within a microcontroller; the microcontroller and the actuator motor communicate via a controller local area network bus with deterministic delay.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, by constructing a closed-loop collaborative management and control architecture that integrates impact perception, maneuver control, and firing execution, the limitations of independent open-loop operation of each link in the existing technology are broken. By leveraging the signal linkage and data interaction of the three modules, the entire system achieves time-series collaborative control, successfully adapting to the usage requirements of high-dynamic live-fire combat training, and fundamentally solving the problems of time sequence misalignment and performance conflict in each link in the existing technology.

[0016] By using the slope trigger timing zero-point determination and multi-channel timing offset alignment technology of the impact sensing module, the contradiction between insufficient accuracy and large amount of redundant data caused by the fixed frequency acquisition of the existing technology is accurately solved, which greatly improves the accuracy of impact positioning while reducing the computing load of the controller. By dynamically switching the multi-stage current limiting range of the maneuver control module, the conflicting requirements of rapid response in maneuvering and high stability in firing steady state are balanced, effectively solving the technical pain point that fixed motor parameters cannot achieve both. Through the dual-window overlapping matching and timing advance compensation of the firing control module, the precise matching of firing timing and chassis steady state is achieved, which completely improves the problems of firing timing misalignment and low hit rate caused by the open-loop fixed timing of the existing technology, and significantly improves the hit rate of reverse attack in dynamic scenarios.

[0017] By employing high- and low-frequency graded acquisition by the dual-frequency sampling unit and benchmark calibration by the timing alignment unit, the accuracy of impact signal acquisition and window precision are further improved. Multi-stage current limiting segment switching and elimination of synchronous current difference in the drive wheel group ensure precise matching of motor operating status and ballistic consistency. The dynamic delay calculation of air path pressure by the feedforward compensation unit and the PID closed-loop algorithm of the feedback calibration unit respectively eliminate the influence of air path pressure fluctuations and timing drift, improving the long-term operational stability of the system. The sliding window differential algorithm, nonlinear weighted model, timing priority calculation of multiple impact signal sources, multi-dimensional physical stability determination, heterogeneous processing architecture, and deterministic delay bus communication improve system performance in terms of anti-interference capability, motor load adaptation, complex scene response, steady-state identification accuracy, and signal processing speed, ensuring that the system can operate stably and efficiently in various training scenarios. Attached Figure Description

[0018] Figure 1 This is a system flowchart of the present invention; Figure 2 This is the system timing logic diagram of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0020] like Figure 1-2 As shown, the intelligent robot target control system with reverse attack function described in this embodiment is applied to the live-fire dynamic confrontation training ground of military tactical training base and civilian professional shooting range. The corresponding physical entity is an intelligent robot target with reverse attack function. The target is equipped with 8 piezoelectric ceramic stress film sensors mounted on the armor plate, STM32H743IIT6 microcontroller, XC7A35T field programmable gate array, four independent hub motors, pneumatic solenoid valves, counterattack mechanical launch mechanism, six-axis IMU chassis attitude sensor, and path planning control unit. All functions of this system are implemented based on the above hardware architecture, without the need for additional hardware components.

[0021] Considering the problems of inaccurate acquisition of transient impact signals, insufficient positioning accuracy, and large amount of redundant data caused by fixed-frequency global acquisition in existing technologies, the transient impact signal generated by live ammunition hitting the armor plate of the target is a steep rising edge signal in the nanosecond to millisecond range. Its signal slope is much greater than the interference signal caused by vibration in the normal environment and target movement. At the same time, the distance between the piezoelectric ceramic sensor and the impact point is different at different installation positions, and there is an inherent time difference in the propagation of the impact signal. Fixed timing acquisition will cause the peak values ​​of multiple signals to be misaligned, which directly affects the accuracy of impact positioning. Therefore, an impact sensing module is constructed. This module determines the zero point of impact time through slope change characteristics and achieves accurate acquisition of impact signals by combining multi-channel timing offset alignment. The input to the impact sensing module is the raw voltage signal output in real time from the 8 piezoelectric ceramic sensors on the target armor plate. The output is the impact event trigger flag, impact signal characteristic parameters, and target firing time window. At the same time, the impact flag is sent to the motion control module via a 50MHz SPI high-speed bus. The impact sensing module includes a dual-frequency sampling unit and a timing alignment unit; Considering the contradiction between low standby load and high-precision impact acquisition in existing technologies, if high-frequency acquisition is used throughout, a massive amount of invalid data will be generated, occupying more than 90% of the controller's computing resources. If low-frequency acquisition is used throughout, it will be impossible to capture the nanosecond-level transient peak signal of impact. Therefore, a dual-frequency sampling unit is constructed to solve the problems described in the existing technology by adopting a tiered dual-frequency sampling mode. In the system standby state, the raw voltage signals of the 8 piezoelectric ceramic sensors are continuously polled at the first sampling frequency. In this embodiment, the first sampling frequency is 10kHz, which meets the requirements of conventional environmental vibration monitoring and conforms to the Nyquist sampling theorem. Each sensor generates only 100,000 sets of data per second, and the controller's computing utilization rate is less than 10%, which greatly reduces the system load in the standby state.

[0022] During the polling acquisition process, the dual-frequency sampling unit analyzes the slope change characteristics of the transient electrical signal under impact in real time to determine whether the triggering condition is met. The sliding window differential algorithm is used here to calculate the transient slope of the sampling point. The design motivation is that the differential calculation of a single sampling point is easily affected by signal noise, which can lead to false triggering. The sliding window differential algorithm with a dual sampling period can effectively filter out high-frequency noise and improve the anti-interference ability of the triggering judgment. The formula for calculating the transient slope is: , in, This is the current voltage value. The sampling period is This is the voltage value of the second sampling period prior to the current moment. In this embodiment, it is in standby mode. The value is 100μs, corresponding to a first sampling frequency of 10kHz. When the absolute value of the transient slope of a consecutive preset number of sampling periods is greater than the preset slope threshold, the triggering condition is determined to be met. In this embodiment, the preset number of consecutive sampling periods is 3. The design motivation is to avoid false triggering caused by signal spike interference in a single sampling period. If the threshold requirement is met for 3 consecutive sampling periods, the interference of random noise can be eliminated 100%. The preset slope threshold is 12.5V / ms. The basis for determining this threshold is: when a 9mm pistol bullet or a 5.56mm rifle bullet hits the target, the minimum measured slope of the front edge of the impact signal is 8.2V / ms, while the maximum signal slope caused by normal target movement and environmental vibration is 4.1V / ms. This threshold is 3 times the maximum interference slope, which can completely avoid false triggering by non-impact signals, while ensuring 100% identification of effective impact signals. When the triggering condition is met, the dual-frequency sampling unit immediately switches the sampling frequency of the corresponding channel from the first sampling frequency to the second sampling frequency. The second sampling frequency is greater than the first sampling frequency. In this embodiment, the second sampling frequency is 2MHz. The basis for determining this frequency is that the main spectrum of the live impact signal is concentrated within 500kHz. According to the Nyquist sampling theorem, a sampling frequency of 2MHz can completely capture the transient peak signal at the nanosecond level, ensuring the acquisition accuracy of the impact signal. Considering the peak timing misalignment caused by the signal propagation delay of different channel sensors in the existing technology, and in order to improve the impact positioning accuracy, the propagation speed of the impact signal in the armor plate is about 5000m / s, the distance between the two farthest sensors on the target is 0.8m, and the maximum propagation delay is 160μs. Fixed timing acquisition will cause the peak values ​​of multiple signals to be not on the same time axis, directly causing the impact positioning error to exceed 20mm. Therefore, a timing alignment unit is constructed to perform multi-channel timing offset alignment to solve the above problems. Specifically, the first sensor channel that meets the triggering conditions is taken as the time reference. This channel is the one closest to the impact point and receives the impact signal first. The triggering time of this channel is determined as the zero point of the impact time. For all other sensor channels that trigger high-frequency acquisition, calculate the time difference between their signal trigger time and the zero point of the impact time. This difference is the timing delay compensation amount for the corresponding channel. For example, if the trigger time of a certain channel is 80μs later than the zero point of the impact time, then the timing delay compensation amount for that channel is 80μs. Based on the calculated timing delay compensation, the sampling time axis of all channels is pulled forward and aligned. Specifically, the sampling time axis of the corresponding channel is shifted forward by the corresponding compensation amount, so that the peak acquisition window of the impact signal of all triggered channels is on the same timing reference axis, thus eliminating the timing misalignment caused by the signal propagation delay.

[0023] After timing alignment is completed, the peak occurrence time of the impact signal of all channels is extracted, and the average of all peak occurrence times is calculated to obtain the peak reference time. The target firing time window is extracted with the peak reference time as the center. In this embodiment, the target firing time window is the interval of ±200μs before and after the peak reference time. The basis for determining this window is that the ±200μs interval can completely cover the effective rising edge, peak value and falling edge of the impact signal. At the same time, invalid and redundant data outside the window can be eliminated, which can eliminate more than 92% of invalid data and greatly reduce the computing load of the controller. After capturing the target firing time window, the timing alignment unit extracts the characteristic parameters of the impact signal within the window, including the peak amplitude of the impact signal, the spatial coordinates of the impact point, and the effective impact level. Simultaneously, it generates an impact flag and sends it to the motion control module. The formula for calculating the spatial coordinates of the impact point is: , in, The spatial coordinates of the point of impact. Let be the peak amplitude of the signal from the i-th sensor. Using the installation coordinates of the i-th sensor, and through weighted calculation of the peak amplitude after time alignment, the positioning accuracy of the impact point can be improved to within ±2mm, which is an order of magnitude higher than the existing technology. When the hit detection module receives multiple hit signal sources simultaneously, such as when the target is hit by multiple munitions in a short period of time, generating multiple effective trigger channel groups, this module determines the response target by calculating the timing priority P. The motivation for this design is to solve the problem that the system cannot determine the priority hit event in the scenario of multiple consecutive hits, which leads to confusion in the maneuver and firing control logic. The formula for calculating timing priority is: , in, This represents the timing priority; the higher the value, the higher the response priority. The amplitude of the impact wave crest. In this embodiment, the maximum amplitude of the impact wave peak is used as a reference value for system calibration. It is 5V, corresponding to the full-scale output of the piezoelectric sensor; The distance from the point of impact to the pre-set core area is the reciprocal of the distance from the point of impact to the pre-defined core area. In this embodiment, the pre-defined core area is the chest region of the target. The closer the point of impact is to the core area, the better. The larger the value; The armor plates are pre-set with hazard levels. In this embodiment, the armor plates are divided into 3 hazard levels: the core area is hazard level 3, the sub-core area is hazard level 2, and the edge area is hazard level 1. The weighting coefficients are 0.2, 0.5, and 0.3 respectively in this embodiment. The weighting is based on the following: whether the point of impact is in the core area directly determines the priority of the counterattack, so it is given the highest weight, followed by the level of impact danger, and the impact amplitude has the lowest weight. After calculating the timing priority of all hit signal sources, the hit perception module extracts the hit signal source with the highest value to deduce the target firing time window, and at the same time generates the corresponding hit flag and sends it to the maneuver control module to ensure the orderly response of the system in the scenario of multiple hit events. Considering that the existing technology with a fixed motor current range cannot simultaneously address the core contradiction of rapid response in evasive maneuvers and chassis steady-state response in counter-attack firing, evasive maneuvers require the motor to output high torque and respond quickly, which requires the upper and lower limits of the motor current range to be sufficiently large. On the other hand, counter-attack firing requires the chassis to have extremely high stability, which requires the motor current fluctuation threshold to be sufficiently small. The two requirements place completely opposite demands on the motor current parameters, which the existing fixed range cannot meet at the same time. Therefore, a maneuver control module is constructed. This module adopts a multi-stage current limit range dynamic switching method to achieve precise matching of the two requirements. The inputs to the motion control module are the impact flag sent by the impact perception module, the evasive maneuver path planning parameters output by the path planning control unit, and the firing pre-trigger signal sent by the firing control module. The outputs are the stator phase current adjustment commands of the four wheel hub motors and the chassis steady-state window. At the same time, the chassis steady-state window parameters are sent to the firing control module. The motor control module has a built-in multi-stage current limiting strategy. The current limiting range includes the steady-state operation range, the evasive maneuver range, and the firing steady-state range. The parameters of the three ranges are all determined based on the actual measurement of the four-wheel independent hub motor with a rated power of 400W and a rated current of 8A used in this embodiment. The steady-state operating range has parameters of 0~5A and a current fluctuation threshold of ≤±0.8A. This range is suitable for the system's standby and routine patrol scenarios, balancing chassis stability and endurance. The evasive maneuver range has parameters of 0~12A and a current fluctuation threshold of ≤±2A. The upper limit of the current in this range is 1.5 times the rated current of the motor, which can provide twice the rated torque to meet the needs of rapid acceleration and sharp turns for quick evasive maneuvers. Short-term operation will not cause the motor to overheat. The firing steady-state range has parameters of 2~4A and a current fluctuation threshold of ≤±0.2A. This range is designed with a narrow range and low fluctuation, strictly limiting the current fluctuation of the motor, ensuring the output synchronization of the four motors, and making the chassis attitude angle fluctuation ≤0.05°, which fully meets the requirements of firing trajectory stability. Specifically, upon receiving the impact flag sent by the impact sensing module, a current range switching command is immediately generated to raise the current limit threshold of the four wheel hub motors of the chassis from the steady-state operating range to the evasive maneuver range. The design motivation is that evasive maneuvers need to be executed immediately after an impact. The high current range can ensure the motor's rapid response and shorten the maneuver response time. In this embodiment, after switching to the evasive maneuver range, the response time for the motor to reach maximum acceleration can be shortened to 85ms, which is 2.76 times better than the 320ms of the prior art. Within the avoidance maneuver range, the maneuver control module calculates the current adjustment coefficient in real time based on the real-time steering angle and real-time acceleration using a nonlinear weighted model. The motivation for this design is that different steering angles and accelerations require different motor torques. A fixed current range would cause the motor to remain in the high current range even in scenarios with small steering angles and low acceleration, resulting in unnecessary energy waste and motor overheating. Therefore, a nonlinear weighted model is used to achieve dynamic adaptation of the current range. The formula for calculating the current adjustment coefficient is: , in, This is the current adjustment coefficient. The final real-time current range upper and lower limits are the reference upper and lower limits of the avoidance maneuver range multiplied by this adjustment coefficient. This is the maximum physical steering angle, which is 30° in this embodiment. The maximum physical acceleration is 4 m / s² in this embodiment. 2 , and In order to pre-set the weight calibration values, in this embodiment, the technicians measured and set the weight calibration values. Set to 0.4, The value is set to 0.6 to ensure a reasonable weighted response of the system to acceleration and steering angle. For real-time acceleration, This is the real-time steering angle; The design principle of this nonlinear weighted model is that the larger the steering angle, the closer the value of the exponent term is to 1, the larger the current adjustment coefficient, and the higher the upper limit of the current range; the larger the acceleration, the larger the value of the square term, and the larger the current adjustment coefficient. This realizes real-time dynamic adaptation between maneuvering action and current range, ensuring maneuvering response speed while avoiding the motor from being in a high-load operation state for a long time. Actual measurements show that the temperature rise of the motor after 2 hours of continuous operation can be reduced from 68K to 32K, which greatly extends the service life of the motor. Upon receiving the firing pre-trigger signal from the firing control module, and when the advance time window of the firing pre-trigger signal arrives, the motion control module generates a secondary interval switching command to lower the current limiting threshold of the chassis motor from the avoidance maneuver interval to the firing steady-state interval. In this embodiment, the advance time window of the firing pre-trigger is 150ms. The basis for determining this value is that the maximum response time for the hub motor to switch from the high-current maneuver interval to the firing steady-state interval and reach the target steady state is 120ms. The 150ms advance ensures that the chassis is fully in the steady-state window before firing.

[0024] After switching to the firing steady-state range, a stable control state is entered by eliminating the synchronous current difference between each drive wheel group. Specifically, the stator phase current values ​​of the four hub motors are read in real time, and the absolute value of the difference between the real-time phase current values ​​of any two motors is calculated to obtain the phase current synchronization difference of the four motors. The steady-state difference threshold is preset to 0.3A. The basis for determining this threshold is that when the current difference between two motors exceeds 0.3A, the chassis will exhibit a perceptible attitude deviation, affecting ballistic stability. When the phase current synchronization difference of a certain group is greater than the steady-state difference threshold, the phase current amplitude of the corresponding motor is finely adjusted and compensated to ensure that the phase current synchronization difference of the four motors is always less than the steady-state difference threshold, thereby ensuring the synchronous operation of the four motors and improving the stability of the chassis. Within the firing steady-state range, the physical stability of the system is determined based on the multi-dimensional state convergence data of the chassis, and a chassis steady-state window is generated. The motivation for this design is that the motor current parameters alone cannot fully characterize the stable state of the chassis. It is necessary to combine multi-dimensional state parameters to accurately identify the chassis steady-state window and ensure the accuracy of the firing timing. The physical stability determination is triggered when the following three characteristics are met simultaneously, and the duration of all characteristics is greater than the preset determination period: The first characteristic is that the variance of the phase current fluctuation of each drive motor is less than a first threshold, which is 0.01A in this embodiment. 2 The judgment period is 50ms, and this feature is used to characterize the smoothness of motor operation. The second feature is that the pitch and yaw angular velocities output by the chassis inertial measurement unit are both less than the second threshold. In this embodiment, the second threshold is 0.02° / s. This feature is used to characterize the attitude stability of the chassis. The third feature is that the spatial pointing deviation angle of the target aiming line is less than the third threshold. In this embodiment, the third threshold is 0.03°. This feature is used to characterize the stability of the aiming line and directly determines the hit accuracy of the counterattack trajectory.

[0025] When the above three features are satisfied simultaneously and the duration exceeds the 50ms judgment period, the judgment system enters a physically stable state and generates a chassis steady-state window. The start time of the chassis steady-state window is the initial time when all features are satisfied, and the duration is the time when the features are maintained. The motion control module sends the start time and duration parameters of the chassis steady-state window to the firing control module in real time. When multiple intelligent robot targets are networked for training, the motion control module can receive the unified synchronization clock from the training management platform through the global timing synchronization unit, so as to realize the global synchronization of the motion actions and firing timing of multiple target robots, avoid motion interference and ballistic crossing between multiple target robots, and adapt to the usage requirements of squad-level combat training scenarios. This mechanism can be realized through the time-triggered CANFD bus, and the timing synchronization error between multiple target robots can be controlled within 1ms. The purpose of building the firing control module is to solve the problems caused by the misalignment of firing timing with the chassis steady-state window, low hit rate of reverse attack, and inability to eliminate timing drift due to the open-loop fixed delay calibration of existing technologies. Existing technologies only consider the inherent delay of pneumatic actuators, without taking into account the response requirements of impact events and the real-time steady-state state of the chassis. They cannot find the optimal firing timing in dynamic scenarios, and open-loop calibration cannot compensate for timing drift caused by air pressure fluctuations and device aging. Therefore, this module uses dual-window overlapping matching to determine the firing timing, and combines feedforward compensation and closed-loop calibration to achieve precise timing control. The inputs to the firing control module are the target firing time window sent by the impact sensing module, the chassis steady-state window sent by the mobility control module, the real-time air pressure output by the air pressure sensor, and the firing position detection signal. The output is the firing control PWM command sent to the physical action actuator. In this embodiment, the physical action actuator is an electromagnetic pneumatic valve, which is used to drive the counterattack mechanical launching mechanism to complete the firing action. The firing control module first reads the target firing time window and the chassis steady-state window, calculates the overlapping interval of the two windows, and takes the center point of the overlapping interval as the actual firing time point. The design motivation is that the target firing time window is the counterattack response time range determined based on the impact event, and the chassis steady-state window is the firing time range determined based on the chassis stable state. Only the overlapping interval of the two windows can simultaneously meet the requirements of counterattack response speed and firing trajectory stability. The center point of the overlapping interval is the time point when the chassis is most stable, which is the optimal firing time. Specifically, let the target firing time window be [T1s, T1e], and the chassis steady-state window be [T2s, T2e]. Then the overlapping interval of the two windows is [Ts, Te], where Ts is the larger value of T1s and T2s, and Te is the smaller value of T1e and T2e. When Ts < Te, there is a valid overlapping interval, and the actual firing time point is (Ts + Te) / 2. When Ts ≥ Te, there is no valid overlapping interval, and it is postponed to the center point of the next chassis steady-state window as the actual firing time point. After determining the actual firing time point, a timing advance compensation is performed on the actual firing time point in combination with the real-time state variables of the physical action execution mechanism. The firing control module includes a feedforward compensation unit. The motivation for the design is that there is an inherent mechanical delay from when the electromagnetic pneumatic valve receives the control signal until the valve is fully opened and the firing mechanism completes its action, and this delay changes with the change of the gas path pressure. If no advance compensation is performed, the actual firing time will deviate from the optimal firing time point, resulting in a decrease in the hit rate. In this embodiment, the real-time state variables of the physical action execution mechanism include the real-time gas path pressure. The specific implementation method of the feedforward compensation unit is as follows: The system basic mechanical delay of the electromagnetic pneumatic valve under the standard calibration pressure is pre-calibrated. In this embodiment, the standard calibration pressure is 0.7 MPa, which is the rated working pressure of the pneumatic launching mechanism, and the corresponding system basic mechanical delay is 28 ms. This value is the average delay of 1000 repeated measurements, and the standard deviation ≤ ±2 ms. The feedforward compensation unit reads the real-time gas path pressure output by the gas path pressure sensor in real time, calculates the deviation amount between the real-time gas path pressure and the standard calibration pressure, and calculates the dynamic physical delay increment of the electromagnetic pneumatic valve based on this deviation amount. Through actual measurement, for every 0.1 MPa deviation of the gas path pressure from the standard calibration pressure, the action delay of the electromagnetic pneumatic valve changes by 3 ms. When the pressure is lower than the standard value, the delay increases; when the pressure is higher than the standard value, the delay decreases. Therefore, the calculation formula for the dynamic physical delay increment is: , where, is the dynamic physical delay increment, is the real-time gas path pressure, is the standard calibration pressure; The calculated dynamic physical delay increment is superimposed on the system basic mechanical delay to obtain the total timing advance amount. The total timing advance amount = system basic mechanical delay + dynamic physical delay increment. Based on the total timing advance amount, the output timing of the actual firing time point is pre-corrected, that is, the output time of the PWM signal = actual firing time point - total timing advance amount, ensuring that the occurrence time of the actual firing action coincides exactly with the optimal firing time point and eliminating the timing deviation caused by the gas path pressure fluctuation. [[ID=二十]] The firing control module also includes a feedback calibration unit. Considering that in the existing technology, feedforward compensation can only eliminate the instantaneous timing deviation caused by gas pressure fluctuations, but cannot eliminate the long-term timing drift caused by device aging, wear, and gas leakage, it is necessary to build a feedback calibration unit to achieve long-term accurate timing control through feedback closed-loop calibration. Specifically, the actual firing time is collected by the firing position detection switch. The difference between the actual firing time and the time when the system issues the firing control command is calculated as the actual total delay. The difference between the actual total delay and the total timing advance calculated in this case is used as the timing error. For example, if the total timing advance calculated in this case is 28ms and the actual total delay is 32ms, then the timing error is +4ms, which means that the actual firing time is 4ms later than expected. A proportional-integral-derivative (PID) closed-loop algorithm is used to process timing errors. In this embodiment, the proportional coefficient of the PID algorithm is 0.3, the integral coefficient is 0.1, and the derivative coefficient is 0.05. The parameter settings are based on avoiding compensation overshoot and ensuring the stability of timing calibration. The compensation amount processed by the PID algorithm is added to the timing advance amount calculated in the next iteration to achieve closed-loop compensation of firing timing. In this embodiment, through closed-loop calibration, the cumulative timing drift after 1000 firings can be reduced from ±28ms to ±2ms, completely eliminating long-term timing drift and improving the long-term stability of the system. After the firing action is completed, the firing control module generates a firing completion flag and sends it to the motor control module. After receiving the firing completion flag, the motor control module resets the motor current range from the firing steady-state range to the steady-state operating range, thus completing the closed-loop control of the entire system. This system adopts a heterogeneous processing architecture. In order to solve the problem that the acquisition of impact signals, slope calculation and timing alignment in the existing technology require nanosecond-level parallel processing capabilities, which conventional microcontrollers cannot meet, while motor control and firing timing logic operations require rich peripheral interfaces and stable serial processing capabilities, a heterogeneous processing architecture is adopted to take into account the needs of high-speed parallel processing and stable control. Specifically, the impact sensing module runs in a field-programmable gate array processor (FPGA) to perform hardware-level parallel calculation of impact signals and time axis alignment. In this embodiment, an XC7A35T FPGA is used, whose parallel processing capability can realize the synchronous acquisition and slope calculation of 8-channel sensor signals with a processing delay of ≤1μs, ensuring the accuracy of the zero point determination of the impact time. The motion control module and the firing control module run in a microcontroller. In this embodiment, an STM32H743IIT6 MCU is used, whose rich CAN and serial port peripherals can meet the needs of motor drive and sensor data acquisition, and its stable computing power can ensure the reliable execution of the control logic.

[0026] The microcontroller and the actuator motor communicate with each other via a controller area network bus with deterministic delay. In this embodiment, a CANFD bus is used with a communication rate of 8Mbps and a communication period of 1ms. The time-triggered communication mechanism can guarantee the determinism of the communication delay, with a maximum communication jitter of ≤100μs, eliminating the timing misalignment caused by the communication delay and ensuring the timing synchronization accuracy of the entire system.

[0027] Based on the above core control logic, as an optional preferred embodiment of the present invention, the system also has a built-in abnormal operating condition fault tolerance module, which is used to classify and handle abnormal operating conditions throughout the entire link, so as to ensure the safe and stable operation of the equipment in the live-fire training scenario. The abnormal operating condition fault tolerance module divides abnormal operating conditions into first-level faults and second-level faults, and executes corresponding handling strategies for different levels of faults.

[0028] Level 2 faults are non-safety-related functional abnormalities, including single-channel piezoelectric sensor failure, timing calibration failure, motor synchronization failure, and abnormal air pressure. For these faults, the system performs adaptive parameter adjustment and degraded operation without interrupting normal training. For example, if a single-channel piezoelectric sensor fails to output a valid signal after three consecutive impact events, it is determined that the sensor has failed. The system immediately blocks the signal of that channel and completes impact location and timing calibration based on the remaining seven normal sensors. At the same time, it sends an alarm signal to the training control platform to maintain normal system operation. Level 1 faults are safety-related serious faults, including motor overcurrent and overheating, chassis attitude deviation, firing mechanism jamming, and timing deviation exceeding the standard. For these faults, the system immediately performs emergency shutdown and functional interlock, cuts off the power supply to the motor and the air circuit of the firing mechanism, locks all movement and firing actions, and sends an emergency shutdown signal to the training control platform to ensure absolute safety in live-fire training scenarios.

[0029] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A smart robot target control system with reverse attack function, characterized in that, The system includes: The impact perception module is used to acquire the impact transient electrical signal collected by the sensor, determine the impact time zero point by analyzing the slope change characteristics of the impact transient electrical signal, extract the impact features based on the time offset alignment of the multi-channel signal, generate the target firing time window, and generate an impact flag bit to send to the motion control module. The motor control module is used to dynamically switch the drive current limiting range of the chassis motor according to the impact flag, and to determine the physical stability of the system based on the chassis multi-dimensional state convergence data, and generate a chassis steady-state window. The firing control module is used to calculate the overlap between the target firing time window and the chassis steady-state window to determine the actual firing time point, and to perform timing advance compensation on the actual firing time point in combination with the real-time state variables of the physical action actuator, and output firing control commands.

2. The intelligent robot target control system with reverse attack function according to claim 1, characterized in that, The impact sensing module includes a dual-frequency sampling unit and a timing alignment unit; The dual-frequency sampling unit is used to poll data at a first sampling frequency in standby mode, and to switch to a second sampling frequency for acquisition when the slope change characteristics of the impact transient electrical signal meet the trigger condition. The second sampling frequency is greater than the first sampling frequency. The timing alignment unit is used to calculate the timing delay compensation of the remaining impacted sensor channels based on the first sensor channel that meets the triggering condition, to pull forward and align the sampling time axis of all channels, and to extract the target firing time window centered on the average peak time after alignment.

3. The intelligent robot target control system with reverse attack function according to claim 1, characterized in that, The maneuver control module incorporates a multi-stage current limiting strategy, and the current limiting range includes a steady-state operating range, an evasive maneuver range, and a firing steady-state range. After receiving the impact flag from the impact sensing module, the maneuver control module raises the limiting threshold of the chassis motor to the evasive maneuver range; after receiving the firing pre-trigger signal, it lowers the limiting threshold to the firing steady-state range, and enters a stable control state by eliminating the synchronous current difference between each drive wheel group.

4. The intelligent robot target control system with reverse attack function according to claim 1, characterized in that, The physical action actuator is an electromagnetic pneumatic valve, and the real-time state variable includes the real-time pressure of the air circuit. The firing control module includes a feedforward compensation unit; The feedforward compensation unit is used to calculate the dynamic physical delay increment of the electromagnetic pneumatic valve based on the deviation between the real-time pressure of the air circuit and the standard calibration pressure, and to add the dynamic physical delay increment to the basic mechanical delay of the system as a timing advance to pre-correct the output timing of the actual firing time.

5. The intelligent robot target control system with reverse attack function according to claim 2, characterized in that, The logic for determining the triggering condition is as follows: The transient slope of the sampling points is calculated using the sliding window difference algorithm, and the formula is as follows: ， in, This is the current voltage value. The sampling period is This is the voltage value two sampling periods prior to the current moment; When the absolute value of the transient slope for a consecutive preset number of sampling periods is greater than a preset slope threshold, the trigger condition is determined to be met.

6. The intelligent robot target control system with reverse attack function according to claim 3, characterized in that, Within the evasive maneuver zone, the maneuver control module calculates the current adjustment coefficient in real time based on the real-time steering angle and real-time acceleration using the following nonlinear weighted model: ， in, This is the current adjustment coefficient. The maximum physical steering angle, For maximum physical acceleration, and To pre-set weight calibration values, For real-time acceleration, This is the real-time steering angle.

7. The intelligent robot target control system with reverse attack function according to claim 4, characterized in that, The firing control module also includes a feedback calibration unit; The feedback calibration unit is used to collect historical actual firing times, calculate the difference between the actual firing time and the time when the system issues the firing control command as the timing error, and use a proportional, integral, and derivative closed-loop algorithm to process the timing error, and add the processed compensation amount to the timing advance amount calculated in the next iteration.

8. The intelligent robot target control system with reverse attack function according to claim 1, characterized in that, When the impact sensing module receives multiple impact signal sources simultaneously, it determines the response target by calculating the timing priority: ， in, For timing priority, The amplitude of the impact wave crest. This is the reference value for the maximum amplitude of the impact wave peak as calibrated by the system. It is the reciprocal of the distance from the point of impact to the pre-set core area. Pre-determine the hazard level for armor plates. These are the weighting coefficients; The impact sensing module extracts The firing time window of the target is derived from the highest value of the hit signal source.

9. The intelligent robot target control system with reverse attack function according to claim 1, characterized in that, The trigger condition for the motion control module to determine the physical stability of the system and generate the chassis steady-state window is that the following characteristics are met simultaneously and the duration is greater than the preset determination period: The variance of phase current fluctuation in each drive motor is less than the first threshold. The pitch and yaw angular velocities output by the chassis inertial measurement unit are both less than the second threshold. The spatial pointing deviation angle of the target aiming line is less than the third threshold.

10. The intelligent robot target control system with reverse attack function according to any one of claims 1-9, characterized in that, The system adopts a heterogeneous processing architecture; The impact sensing module runs in a field-programmable gate array processor and is used to perform hardware-level parallel calculation and time-axis alignment of impact signals. The motion control module and the firing control module operate within a microcontroller; the microcontroller and the actuator motor communicate via a controller local area network bus with deterministic delay.

Images