Intelligent target range cooperative control system based on multi-modal perception

Abstract

The present application relates to the technical field of automatic control and sensing data processing, in particular to a wisdom target range cooperative control system based on multi-modal perception, the system comprises a cooperative control module, which generates cluster cooperative instructions based on tactical logic and issues them to target nodes; a multi-modal perception execution module, which drives target actions in response to the instructions, and generates accurate impact point coordinates by multi-modal fusion calculation of acoustic shock time difference signals, taking the on-off signal of the target surface medium as the trigger reference; and a state feedback module, which receives hit data and renders local or global battle damage views of the target site; the present application solves the problems of poor anti-interference and inability of dynamic cooperation of single target reporting by multi-modal feature fusion calculation and wireless ad hoc network node scheduling, and has the beneficial effects of high target reporting accuracy, strong anti-interference and support for large-scale high-fidelity tactical exercises.

Description

Technical Field

[0001] This invention relates to the field of automatic control and sensor data processing technology, and in particular to a smart target range collaborative control system based on multimodal perception. Background Technology

[0002] In modern light weapons live-fire exercises and tactical combat drills, it is mainly used for automated collaborative control of multi-target clusters and precise acquisition and feedback of shooting results. By integrating control, execution, and display feedback, it provides objective data support and highly realistic tactical combat environment simulation for shooting training.

[0003] Currently, existing technologies employ automatic target reporting and range control systems, which consist of independently operating individual target devices and corresponding host computer software. In the target reporting phase, traditional equipment typically uses a single acoustic microphone array to capture the sound waves generated during supersonic projectile flight, or uses only a conductive rubber target surface to detect short-circuit signals during physical penetration, thereby calculating the impact point. In the range scheduling phase, existing technologies typically use rigid cables or point-to-point one-way communication modules, with the control console sending fixed-time start / stop commands to specific target launchers or monorail moving target launchers. Trainees then fire at stationary targets or targets moving along a fixed straight-line trajectory from a fixed firing ground. Furthermore, after a consumable target plate is worn down, a new target plate needs to be directly fastened to the main frame containing the acoustic detection elements using bolts or metal clips.

[0004] However, the aforementioned existing technologies have the following drawbacks: First, the single acoustic detection method is highly susceptible to wind noise, echoes, and sound crosstalk from concurrent firing at adjacent targets in the open environment of the firing range, leading to false alarms or missed alarms in the bullet impact calculation. The purely conductive detection method is prone to impedance changes after repeated damage to the target surface, resulting in detection failure. Due to the lack of a multi-dimensional physical quantity data verification mechanism, the overall target reporting accuracy and data robustness of the system are extremely poor. Second, the existing target machine's structural design results in rigid transmission between the consumable target plate and the main frame of the acoustic detection element. The severe physical vibration of the target plate when hit by a bullet is directly transmitted to the underlying detection element. Furthermore, during frequent disassembly and replacement of the consumable target plate, changes in installation stress and tolerances can easily cause mechanical displacement of the acoustic detection reference origin, necessitating time-consuming recalibration of the reference coordinates after each target plate replacement, severely reducing the efficiency of range maintenance and operation. Finally, traditional range equipment is isolated from each other and has a single control link, which makes it impossible to achieve large-scale dynamic cascading of multiple target nodes and complex tactical timing coordination. This results in highly homogenized training scenarios and makes it impossible to build a realistic adversarial environment with high maneuverability and two-way state awareness.

[0005] Therefore, this application aims to solve the problems of poor anti-interference capability of the single target reporting method in the prior art and the easy deviation of the reference caused by the force of board replacement, resulting in low maintenance efficiency. At the same time, it also solves the problem of isolated target node information, which makes it impossible to achieve large-scale cluster tactical collaborative control. Summary of the Invention

[0006] The purpose of this invention is to address the shortcomings of existing technologies by proposing a smart target range collaborative control system based on multimodal perception.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: a smart target range collaborative control system based on multimodal perception includes: The collaborative control module is used to generate target cluster collaborative instructions containing time series and action types through pre-set tactical exercise logic, and to distribute the target cluster collaborative instructions to each distributed target node through a wireless ad hoc network. A multimodal perception and execution module is deployed on the target node to respond to the target cluster's collaborative instructions, drive the corresponding target node to perform corresponding maneuvering displacement or attitude switching actions; and simultaneously collect the acoustic shock wave time difference signal and the electrical on / off signal of the target surface medium generated when the projectile is fired, and use the electrical on / off signal as a trigger reference to perform multimodal fusion calculation on the acoustic shock wave time difference signal to generate the projectile impact point coordinates and hit event data; The status feedback module is communicatively connected to the collaborative control module and the multimodal perception execution module. It is used to receive the hit event data and target node running status, and to perform target-level local view rendering of the bullet impact point or aggregate view rendering of the global battle damage situation of the target range according to different terminal service permissions.

[0008] As a further aspect of the present invention, the multimodal perception execution module includes: The signal synchronization window unit is used to take the sudden pulse of the acquired target surface dielectric electrical on / off signal as a time reference trigger signal, and retrieve the acoustic wave timestamp data acquired by the sensor array within a preset time window before and after the time reference trigger signal. The fusion computing unit is used to calculate the time difference of arrival coordinates to generate the impact point coordinates when there is sound wave timestamp data that conforms to the characteristic frequency within the preset time window, and to remove environmental noise interference signals outside the preset time window.

[0009] As a further aspect of the present invention, the multimodal sensing and execution module also includes a calibration-free mechanical decoupling structure: The calibration-free mechanical decoupling structure includes a rigid inner frame and a floating outer frame suspended on the rigid inner frame by a damping element. The sensor array is fixedly installed within the rigid inner frame, forming an absolutely static detection reference coordinate system; The outer side of the floating outer frame is connected to an elastic clamping member. The consumable target plate is detachably plugged and pulled to the front of the floating outer frame through the elastic clamping member, so that the physical vibration of the consumable target plate when it is hit by a ball is mechanically filtered through the floating outer frame and is not transmitted to the rigid inner frame; and when the consumable target plate is replaced, the origin of the detection reference coordinate system remains unchanged.

[0010] As a further aspect of the present invention, the underlying signal processing link of the multimodal sensing execution module sequentially includes: The acoustic and electrical physical sensors on the target node; A single-power rail-to-rail dual operational amplifier connected to the output of the physical sensor device; A high-speed voltage comparator connected to the output of the operational amplifier; A main control microprocessor connected to the output of the high-speed voltage comparator has an embedded high-speed capture timer interface for capturing the edge signal of the output of the high-speed voltage comparator and generating a precise hardware timestamp to perform the multimodal fusion calculation. The communication connection is to the self-organizing network radio frequency transceiver front end of the main control microprocessor, used to transmit the calculated data back to the collaborative control module.

[0011] As a further aspect of the present invention, some of the target nodes are configured as all-terrain tracked robot targets with independent power, and their multimodal perception and execution modules further include: A high-precision positioning unit is used to obtain the real-time geographical coordinates of the tracked robot target; The differential drive execution unit, which uses a brushless motor, is used to parse the target waypoint and speed commands issued by the cooperative control module, and perform on-the-spot turning, forward movement or maneuvering evasion actions in combination with the real-time geographical coordinates.

[0012] As a further aspect of the present invention, the system also includes a human-machine bidirectional adversarial perception module, the human-machine bidirectional adversarial perception module comprising: The simulated fire launch unit, integrated on the target node, is used to launch laser beam pulses with coded identifiers into a pre-set area of ​​the target range under the triggering of the cooperative control module. The individual soldier wearable sensing unit is configured as a wearable carrier with an embedded distributed sensor network. It is used to receive the laser beam pulses and parse the coded identifier, determine the hit location, and report the battle damage signal to the cooperative control module through a wireless communication link.

[0013] As a further aspect of the present invention, the system also includes a wireless video calibration module that assists the status feedback module in establishing a shooting reference, comprising: The image acquisition and transmission end is coaxially fixed inside the firing tube of the light weapon via a multi-caliber adaptable quick-connect shaft, and is used to acquire real-time video streams of the firing target surface; A dynamic reticle calibration terminal is used to display the electronic crosshair on the real-time video stream; in response to the user's calibration translation command, the center origin of the electronic crosshair is dynamically translated to the geometric average center of the impact points of multiple consecutive live bullets, so as to construct a shooting aiming reference point based on software compensation.

[0014] As a further embodiment of the present invention, the multi-diameter adaptable quick-connect shaft includes a first outer diameter reference section and a second outer diameter reference section, which are connected to the housing base of the image acquisition and transmission end by elastic snap-fit, so as to ensure that the quick-connect shafts with different outer diameters coincide with the optical central axis of the image acquisition and transmission end when they are replaced.

[0015] As a further aspect of the present invention, the status feedback module is specifically configured as follows: On the individual shooter terminal, the local view rendering includes applying different colors or symbols to the historical bullet impact points and the latest bullet impact point of the currently assigned target position; On the global command terminal, the aggregated view rendering includes scattered domain data of all target nodes clustered based on time windows, and overlays and outputs a real-time live video stream of the target range.

[0016] As a further aspect of the present invention, the following system operating parameters are provided: Under a pre-set standard test environment, the target reporting response time threshold of the multimodal fusion calculation is less than or equal to 0.02 seconds, and the accuracy error limit of the calculated bullet impact point coordinates is less than or equal to 5 millimeters. When the collaborative control module issues commands to multiple points through the wireless self-organizing network, the maximum number of target nodes that can be cascaded in a single network is no less than 50.

[0017] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, under the instruction drive of the collaborative control module, the multi-modal perception execution module deployed on the target node synchronously collects multi-source signals, and strictly uses the electrical on / off signal of the target surface medium as the trigger reference to perform multi-modal fusion calculation on the acoustic shock wave time difference signal to generate the impact point coordinates. This enables the single acoustic microphone array and the conductive rubber target surface to cross-verify and collaboratively filter invalid environmental noise in the time dimension. Furthermore, a wireless scheduling link supporting distributed node cascading is built at the bottom layer, thereby effectively solving the core defects of the traditional single target reporting method, which is easily affected by crosstalk interference in open environments, leading to false alarms and missed alarms, and the traditional target range equipment information isolation, which cannot achieve large-scale tactical timing coordination. This invention achieves the beneficial effects of extremely high target reporting accuracy, extremely strong data robustness, and the ability to support highly realistic multi-target cluster tactical exercises.

[0018] By setting up a calibration-free mechanical decoupling structure that includes a rigid inner frame and a floating outer frame with damping components suspended on the inner frame, the severe physical vibration generated when the consumable target plate is hit by a bullet can be mechanically filtered through the floating frame. This completely isolates the transmitted stress from the rigid inner frame of the underlying acoustic wave detection element, ensuring that the origin of the detection reference coordinate system remains absolutely stationary when the target plate is frequently plugged and unplugged. This eliminates the time-consuming recalibration caused by mechanical stress offset and greatly improves the daily operation and maintenance efficiency of the range.

[0019] By setting up a human-machine two-way adversarial perception module that includes simulated fire launch units and individual soldier wearable perception units, the target node can actively emit laser counterattack pulses with coded identifiers into the training area under the control of the system. The system can also accurately capture the hit points and report the battle damage status through the individual soldier distributed sensor network, thus transforming the target range from a traditional one-way shooting environment to a real adversarial environment with high mobility evasion and two-way status perception. Attached Figure Description

[0020] Figure 1 This is a system flowchart of the present invention; Figure 2 This is a system workflow diagram of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] In the description of this application, the collaborative control module refers to the core control center of the smart range, which is a hardware device integrating control programs and communication modules, and can realize the functions of tactical exercise logic editing, command generation and issuance, and data reception and aggregation. The multimodal perception and execution module refers to an integrated module deployed on each target node that combines action execution with multi-source signal acquisition and calculation functions; The status feedback module refers to a hardware and software combination module that enables the visualization of range data and multi-terminal permission-based display; Wireless ad hoc networks refer to decentralized distributed communication networks built on Zigbee and LoRa protocols, which support multi-node cascading and bidirectional data transmission. The calculation of time difference of arrival coordinates refers to an algorithm based on the TDOA algorithm, which calculates the spatial coordinates of the impact point by using the time difference of signals collected by a sensor array.

[0023] Please see Figures 1-2 A smart target range collaborative control system based on multimodal perception includes: The collaborative control module is used to generate target cluster collaborative instructions containing time series and action types through pre-set tactical exercise logic, and to distribute the target cluster collaborative instructions to each distributed target node through a wireless ad hoc network. The collaborative control module is mounted on the industrial control computer of the smart target range management platform. Operators can pre-set tactical exercise logic through the human-machine interface of the management platform. This logic includes the action time sequence, action type, and linkage rule information of the target cluster. The collaborative control module generates standardized target cluster collaborative instructions based on this logic. The instructions carry the target node number, action instructions, and execution time information. Then, the instructions are sent to each distributed target node point-to-point or point-to-multipoint through a wireless self-organizing network. A single network supports the simultaneous transmission of instructions to up to 50 or more target nodes. The multimodal perception and execution module is deployed on the target node to respond to the target cluster's collaborative commands, drive the corresponding target node to perform corresponding maneuvering displacement or attitude switching actions; and simultaneously collect the acoustic shock wave time difference signal and the electrical on / off signal of the target surface medium generated when the projectile is fired. Using the electrical on / off signal as the trigger reference, the acoustic shock wave time difference signal is fused and calculated in a multimodal manner to generate the projectile's impact point coordinates and hit event data. The multimodal perception and execution module has a modular integrated structure and is directly deployed inside each target node. This module first receives the collaborative instructions issued by the collaborative control module, and then drives the corresponding target node to perform corresponding maneuvering displacement or attitude switching actions through the internal execution drive unit, such as target tilting, lateral movement, in-situ turning, and concealment switching. At the same time, this module also has a built-in signal acquisition unit, which can synchronously acquire the acoustic shock wave time difference signal and the electrical on / off signal of the target surface medium generated when the projectile is fired. The electrical on / off signal serves as the trigger reference for effective hit. The multimodal perception and execution module uses this trigger reference as the time origin and performs multimodal fusion calculation on the acoustic shock wave time difference signal to finally generate accurate projectile impact point coordinates and hit event data including hit time, hit location, and projectile type. The status feedback module is connected to the collaborative control module and the multimodal perception and execution module. It is used to receive the hit event data and target node operation status, and to perform target-level local view rendering of the bullet impact point or aggregate view rendering of the global battle damage situation of the target range according to different terminal service permissions.

[0024] The status feedback module is communicatively connected to both the collaborative control module and the multimodal perception and execution module. On one hand, it receives hit event data from each target node from the multimodal perception and execution module, and simultaneously obtains the real-time operating status of each target node from the collaborative control module. This status includes multiple status items such as battery level, online status, and action execution progress. On the other hand, based on pre-set terminal service permissions, this module performs differentiated visualization rendering of the above data, providing target-level local view rendering of the bullet impact point and aggregated view rendering of the overall battle damage situation of the firing range for both individual shooters and global command service terminals, thereby achieving permission-based display of data.

[0025] Please see Figures 1-2 The multimodal perception execution module includes: The signal synchronization window unit is used to take the sudden pulse of the acquired target surface dielectric electrical on / off signal as the time reference trigger signal, and retrieve the acoustic wave timestamp data acquired by the sensor array within a preset time window before and after the time reference trigger signal. The signal synchronization window unit is a combination of hardware circuitry and software program integrating signal triggering and data retrieval logic. Its core function is to capture the sudden pulse of the electrical on / off signal of the target surface medium and use the sudden pulse as a time reference trigger signal. The sudden pulse is generated by the projectile penetrating the consumable target plate and causing a change in the on / off state of the conductive circuit of the target surface, which is a direct signal of effective hit. At the same time, the signal synchronization window unit has a time window parameter preset. After receiving the time reference trigger signal, the unit only retrieves the acoustic wave timestamp data collected by the sensor array within the preset time window before and after the trigger signal. The acoustic wave data outside the time window will be temporarily blocked. The range of the preset time window can be flexibly adjusted according to the projectile velocity and target type. It is generally set to ±5ms. This data setting can be flexibly adjusted according to the actual operation.

[0026] The fusion computing unit is used to calculate the time difference of arrival coordinates to generate the impact point coordinates when there is sound wave timestamp data that meets the characteristic frequency within a pre-set time window, and to remove environmental noise interference signals outside the pre-set time window.

[0027] The fusion computing unit is a main control chip equipped with the TDOA calculation algorithm and is electrically connected to the signal synchronization window unit. It receives the acoustic wave timestamp data retrieved by the signal synchronization window unit. The unit first performs characteristic frequency matching on the acoustic wave timestamp data to determine whether the acoustic wave data within the preset time window matches the characteristic frequency of the acoustic shock wave generated by the supersonic flight of the projectile. If it matches, it is determined to be a valid signal. Then, it performs the arrival time difference coordinate calculation and calculates the two-dimensional or three-dimensional coordinates of the impact point by combining the position parameters of the sensor array. If it does not match or there is no acoustic wave timestamp data within the preset time window, it is determined to be an environmental noise interference signal and is rejected. No coordinate calculation is performed to avoid false alarms.

[0028] Please see Figures 1-2 The multimodal perception and execution module also includes a calibration-free mechanical decoupling structure: The calibration-free mechanical decoupling structure comprises a rigid inner frame and a floating outer frame suspended from the rigid inner frame by damping elements. The sensor array is fixedly installed within a rigid inner frame, forming an absolutely static detection reference coordinate system; The outer side of the floating outer frame is connected to an elastic clamping component. The consumable target plate is detachably plugged and pulled to the front of the floating outer frame through the elastic clamping component, so that the physical vibration of the consumable target plate when hit by a bullet is mechanically filtered through the floating outer frame and is not transmitted to the rigid inner frame; and when the consumable target plate is replaced, the origin of the detection reference coordinate system remains unchanged. The calibration-free mechanical decoupling structure is a purely mechanical structure used to achieve mechanical decoupling between the sensor array and the consumable target plate, avoiding the impact of target plate vibration on detection accuracy, and at the same time enabling target plate replacement without calibration. The calibration-free mechanical decoupling structure includes a rigid inner frame, damping components, and a floating outer frame. The damping components are rubber shock absorbers or spring shock absorbers, and there are four of them, which are connected to the four inner corners of the rigid inner frame. The floating outer frame is suspended outside the rigid inner frame through the damping components, and a pre-set gap is maintained between it and the rigid inner frame, with no direct rigid contact. The sensor array is bolted to the inside of the rigid inner frame, forming an integrated structure with the rigid inner frame and constituting an absolutely static detection reference coordinate system. The origin of this coordinate system coincides with the geometric center of the sensor array. An elastic clamping component is fixed to the outer side of the floating outer frame by welding. The elastic clamping component adopts a spring clamp structure and has elastic clamping force. The consumable target plate is detachably plugged and pulled to the front of the floating outer frame through the elastic clamping component, maintaining a preset detection distance from the sensor array. When the consumable target plate is subjected to physical vibration by a bullet, the vibration is first transmitted to the floating outer frame. After mechanical filtering and shock absorption by the damping component, the vibration cannot be transmitted to the rigid inner frame, and the sensor array remains stationary, ensuring detection accuracy. When replacing the consumable target plate, it is only necessary to insert or remove the target plate through the elastic clamping component. The positions of the rigid inner frame and the sensor array do not change, and the origin of the detection reference coordinate system remains unchanged, so there is no need to recalibrate.

[0029] Please see Figures 1-2 The underlying signal processing chain of the multimodal perception execution module includes, in sequence: Acoustic and electrical physical sensors on target nodes; A single-supply rail-to-rail dual operational amplifier connected to the output of a physical sensor device; A high-speed voltage comparator connected to the output of an operational amplifier; The main control microprocessor is connected to the output of the high-speed voltage comparator. The main control microprocessor has an embedded high-speed capture timer interface, which is used to capture the edge signal of the output of the high-speed voltage comparator and generate a precise hardware timestamp to perform multi-modal fusion calculation. The communication is connected to the self-organizing network radio frequency transceiver front end of the main control microprocessor, which is used to transmit the calculated data back to the collaborative control module; The underlying signal processing link of the multimodal sensing execution module is a serial hardware circuit structure, which includes acoustic and electrical physical sensors, a single-power rail-to-rail dual-channel operational amplifier, a high-speed voltage comparator, a main control microprocessor, and a self-organizing network RF transceiver front end to realize signal acquisition, amplification, shaping, calculation and transmission. Acoustic and electrical physical sensors are the basic components for signal acquisition. The acoustic physical sensor uses an 18mm ultrasonic shock wave sensor array to acquire the acoustic shock wave signal generated by the projectile. The electrical physical sensor uses a conductive circuit and voltage detection sensor laid on the consumable target plate to acquire the electrical on / off signal of the target surface medium. The output terminals of both types of sensors are electrically connected to the input terminals of a single-power rail-to-rail dual-channel operational amplifier. The single-power rail-to-rail dual operational amplifier adopts a single power supply mode and is compatible with the DC power supply system of the target node. It amplifies the weak analog signals collected by the sensor device, and the amplification factor can be adjusted according to actual needs. The amplified analog signal is output to a high-speed voltage comparator. The high-speed voltage comparator compares the amplified analog signal with a preset voltage threshold, shapes the analog signal into a digital pulse signal, and captures the edge characteristics of the signal. The shaped digital pulse signal is then output to the main control microprocessor. The main control microprocessor uses a 32-bit ARM processor, specifically the STM32F103. This chip has an embedded high-speed capture timer interface, which can achieve microsecond-level edge signal capture and generate a precise hardware timestamp. The hardware timestamp provides a time reference for the calculation of acoustic shock wave time difference signal. The main control microprocessor performs multimodal fusion calculation based on the hardware timestamp to generate the impact point coordinates and hit event data. The self-organizing network RF transceiver front end uses a Zigbee wireless communication module, which is electrically connected to the serial communication pin of the main control microprocessor. It modulates the digital data calculated by the main control microprocessor and transmits it back to the collaborative control module through the wireless self-organizing network to realize wireless data transmission.

[0030] Please see Figures 1-2 Some target nodes are configured as all-terrain tracked robot targets with independent power, and their multimodal perception and execution modules also include: A high-precision positioning unit is used to obtain the real-time geographical coordinates of the tracked robot target; The differential drive actuator, which uses a brushless motor, is used to parse the target waypoint and speed commands issued by the cooperative control module, and combine them with real-time geographic location coordinates to perform on-the-spot turning, forward movement or maneuvering evasion actions. Some target nodes are configured as all-terrain tracked robot targets with independent power. The multimodal perception and execution module of this type of target node also adds a high-precision positioning unit and a differential drive execution unit to the basic structure to achieve all-terrain mobility. The high-precision positioning unit uses an RTK satellite positioning module, which can achieve centimeter-level positioning accuracy. It is electrically connected to the main control microprocessor to acquire the three-dimensional geographical coordinates of the tracked robot target in real time and transmit the coordinate data to the main control microprocessor in real time. The differential drive actuator adopts a combination structure of brushless DC motor and differential. There are two brushless DC motors, which drive the left and right tracks of the tracked robot target respectively. The unit is electrically connected to the main control microprocessor and receives the target waypoint and speed commands issued by the collaborative control module after being parsed by the main control microprocessor. Combined with the real-time geographical coordinates obtained by the high-precision positioning unit, the unit controls the speed and direction of the two brushless DC motors to realize the tracked robot target's turning in place, straight-line forward movement, variable speed movement and maneuver avoidance. It is suitable for multiple complex terrain environments such as grassland, mountain road and construction site. The climbing angle can reach 30° and above, and the obstacle crossing height is ≥150mm. Please see Figures 1-2 The system also includes a human-machine two-way adversarial perception module, which includes: The simulated fire launch unit, integrated on the target node, is used to launch laser beam pulses with coded identifiers into a pre-set area of ​​the target range under the triggering of the cooperative control module; The individual soldier wearable sensing unit is configured as a wearable carrier with an embedded distributed sensor network. It is used to receive laser beam pulses and parse the encoded identifiers, determine the hit location, and report the battle damage signal to the collaborative control module through a wireless communication link. In this embodiment, the system is also equipped with a human-machine two-way confrontation perception module. This module is wirelessly connected to the collaborative control module to realize the human-machine two-way confrontation exercise function of the target range, including a simulated fire launch unit and a soldier wearable perception unit. Specifically, the simulated fire launch unit is integrated into the target node's body and electrically connected to the target node's main control microprocessor. Its core is a laser emission device, and the laser power can be adjusted within the range of 5mW-50mW. Under the trigger of the collaborative control module, the simulated fire launch unit emits laser beam pulses with coded identifiers into a pre-set area of ​​the target range according to the pre-set tactical exercise logic. The coded identifiers contain the target node number, launch time, and simulated ammunition type information. The emission frequency of the laser beam pulses can be adjusted within the range of 1-10 shots / second. The individual soldier wearable sensing unit is a wearable carrier with an embedded distributed sensor network, including a recognition vest and a recognition helmet. The distributed sensor network consists of multiple laser receiving sensors, which are respectively deployed on the head, chest, abdomen, and limbs of the wearable carrier, enabling accurate identification of the hit site. The unit receives laser beam pulses emitted by the simulated firepower unit and analyzes the encoded identifiers in the laser beam pulses through an internal decoding module to determine whether it has been hit and the specific hit site. Then, it reports the battle damage signal, which includes the hit site, hit time, and shooter number, to the collaborative control module through a wireless communication link. The collaborative control module then summarizes the information and transmits it to the status feedback module for visualization. Please see Figures 1-2The system also includes a wireless video calibration module for establishing shooting benchmarks using an auxiliary status feedback module, including: The image acquisition and transmission end is coaxially fixed inside the firing tube of the light weapon via a multi-caliber adaptable quick-connect shaft, and is used to acquire real-time video streams of the firing target surface; The dynamic reticle calibration terminal is used to display the electronic crosshair on the real-time video stream; in response to the user's calibration translation command, the center origin of the electronic crosshair is dynamically translated to the geometric average center of the impact points of multiple consecutive live bullets, so as to construct a shooting aiming reference point based on software compensation. The wireless video calibration module communicates with the status feedback module to assist the status feedback module in establishing an accurate shooting reference, including an image acquisition and transmission end and a dynamic reticle calibration terminal.

[0031] The image acquisition and transmission end is a hardware device integrating a high-definition camera, a wireless image transmission module, and a quick-connect structure. It is coaxially fixed inside the firing tube of the light weapon through a multi-caliber adaptable quick-connect shaft, ensuring that the optical center axis of the image acquisition and transmission end coincides with the barrel axis of the light weapon. The device acquires real-time video streams of the firing target surface in real time and wirelessly transmits the video streams to the dynamic reticle calibration terminal through the wireless image transmission module. The working frequency of the wireless image transmission is 2.4G, and the transmission distance is ≥15m. The dynamic reticle calibration terminal is a high-brightness display terminal equipped with calibration software, with a brightness of ≥1200cd / ㎡, which can be clearly displayed in strong outdoor light environments. It receives real-time video streams transmitted from the image acquisition and transmission end, and overlays electronic crosshairs on the video stream. The electronic crosshairs serve as the reference for shooting aiming. Operators can issue calibration translation commands through the terminal's operation interface. According to the commands, the dynamic reticle calibration terminal dynamically translates the center origin of the electronic crosshairs to the geometric average center of the impact points of multiple consecutive live bullets, constructing a shooting aiming reference point based on software compensation, and realizing the accurate establishment of zero-bullet gun calibration and shooting reference. Please see Figures 1-2 The multi-diameter adaptable quick-connect shaft includes a first outer diameter reference section and a second outer diameter reference section, which are connected to the housing base of the image acquisition and transmission end by elastic buckle to ensure that quick-connect shafts with different outer diameters are aligned with the optical central axis of the image acquisition and transmission end when being replaced. The multi-caliber adaptable quick-connect shaft is a machined metal part used to achieve coaxial fixation between the image acquisition and transmission end and the launch tubes of small arms of different calibers. It includes a first outer diameter reference section and a second outer diameter reference section. The two reference sections are an integrated structure and are coaxially set. The first outer diameter reference section is adapted to small-caliber small arms launch tubes, such as 5.8mm and 9mm, and the second outer diameter reference section is adapted to large-caliber small arms launch tubes, such as 7.62mm and 12.7mm. The outer diameters of both reference sections are precision machined to ensure a tight fit with the inner wall of the launch tube of the corresponding caliber. The quick-connect shaft has an elastic snap-fit ​​structure at its tail end, and the housing base of the image acquisition and transmission end has a corresponding snap-fit ​​groove. The quick-connect shaft is detachably connected to the housing base of the image acquisition and transmission end through the cooperation of the elastic snap-fit ​​and the groove. The clamping force of the elastic snap-fit ​​ensures that there is no relative displacement between the quick-connect shaft and the housing base, ensuring that when quick-connect shafts with different outer diameters are replaced, their central axis always coincides with the optical central axis of the image acquisition and transmission end, without the need to recalibrate the coaxiality. Please see Figures 1-2 The specific configuration of the status feedback module is as follows: On the individual shooter's terminal, the local view rendering includes applying different colors or symbols to the historical bullet impact points and the latest bullet impact point of the currently assigned target position. On the global command terminal, the aggregated view rendering includes scattered domain data of all target nodes clustered based on time windows, and overlays and outputs real-time live video stream of the target range. The visualization rendering function of the status feedback module adopts differentiated configuration, and different view rendering effects are achieved according to the different business permissions of the individual shooter business terminal and the global command business terminal. On the individual shooter's terminal, which is a dedicated tablet display device with a protection level of ≥IP65, the local view rendering of the status feedback module mainly displays the target position data currently assigned to the shooter, including real-time bullet impact coordinates, historical shooting data, and hit ring count. Different colors or symbols are applied to historical bullet impact points and the latest bullet impact point. For example, historical bullet impact points are marked as blue dots, and the latest bullet impact point is marked as a red pentagram, which makes it easier for the shooter to intuitively observe shooting deviations and adjust shooting posture in a timely manner. On the global command terminal, which serves as a dedicated control platform and large-screen display device for commanders, the aggregated view rendering of the status feedback module mainly summarizes and displays the global data of the entire firing range. First, it performs cluster analysis on the bullet impact dispersion data of all target nodes based on a time window, which can be set to 1 minute, 5 minutes, or 10 minutes. After the cluster analysis, it generates statistical data on the shooting accuracy and dispersion range of each target node. At the same time, it overlays the statistical data with the real-time live video stream of the firing range and marks the location, damage status, and shooter performance information of each target node on the live video stream, so that commanders can grasp the overall damage situation of the firing range in real time. Please see Figures 1-2 It has the following system operating parameters: Under a pre-set standard test environment, the target reporting response time threshold of the multimodal fusion calculation is less than or equal to 0.02 seconds, and the accuracy error limit of the calculated bullet impact point coordinates is less than or equal to 5 millimeters. When the collaborative control module issues commands to multiple points through a wireless self-organizing network, the maximum number of target nodes that can be cascaded in a single network is no less than 50. In this embodiment, the system operates under a pre-set standard test environment, which is: ambient temperature 25℃±5℃, ambient humidity 40%-60%, no significant ambient noise, no strong electromagnetic interference, projectile velocity ≥250 m / s, and test distance 100 meters. The system's operating parameters meet the following requirements: The target reporting response time threshold of multimodal fusion computing is less than or equal to 0.02 seconds, that is, the total time from the generation of the signal when the projectile hits the target plate to the calculation and generation of the impact point coordinates by the multimodal perception execution module and the transmission back to the collaborative control module does not exceed 0.02 seconds; the accuracy error limit of the calculated impact point coordinates is less than or equal to 5 millimeters, ensuring that the target reporting accuracy meets the requirements of high-precision shooting training. When the collaborative control module issues commands to multiple points through the wireless self-organizing network, it supports a maximum of 50 target nodes in a single network and the difference in command reception latency between each target node does not exceed 100ms, ensuring the consistency of collaborative linkage among multiple target nodes and meeting the needs of large-scale tactical confrontation exercises.

[0032] To further improve the stability, accuracy, and intelligence of this system, this application adds the following innovative optimizations based on the above embodiments. These optimizations are optional configurations of the system and can be implemented according to actual training needs: The anomaly handling mechanism of the multimodal perception execution module adds signal anomaly judgment logic to the main control microprocessor. When multiple consecutive sudden pulses of electrical on / off signals are detected, but no acoustic shock wave signal matching the characteristic frequency is found within the preset time window, it is determined that the consumable target plate is damaged or the electrical sensor device is faulty. The main control microprocessor immediately generates a fault alarm signal and transmits it back to the collaborative control module through a wireless self-organizing network. The status feedback module provides a visual alarm and automatically blocks the target node's reporting data to avoid false alarms affecting the training results.

[0033] The collaborative control module has a command retransmission mechanism. After issuing a collaborative command to the target cluster, if the collaborative control module does not receive a command reception confirmation signal from some target nodes within a preset time, it is determined that the command transmission has failed. The collaborative control module will automatically retransmit the command to those target nodes. The number of retransmissions can be preset to 3. If the retransmission still fails after 3 retransmissions, it is determined that the target node is offline and an offline alarm signal is generated to ensure the reliability of command issuance.

[0034] The bullet impact dispersion analysis function of the status feedback module adds an intelligent analysis function to the aggregated view rendering of the global command terminal. It analyzes the bullet impact dispersion data of each shooter and each target node through big data algorithms, and generates shooting accuracy trend charts and deviation analysis reports to provide reference data for commanders to formulate subsequent improvement training plans.

[0035] The aforementioned innovative optimizations are all based on the core technical solutions of this application and do not depart from the protection scope of this application. They can also be arbitrarily combined with the above embodiments to achieve functional upgrades of the system. 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 use the above-disclosed technical content to make changes or modifications to create equivalent embodiments for application in 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 technical solution of the present invention still fall within the protection scope of the present invention.

Claims

1. A smart target range collaborative control system based on multimodal perception, characterized in that: The system includes: The collaborative control module is used to generate target cluster collaborative instructions containing time series and action types through pre-set tactical exercise logic, and to distribute the target cluster collaborative instructions to each distributed target node through a wireless ad hoc network. A multimodal perception and execution module is deployed on the target node to respond to the target cluster's collaborative instructions, drive the corresponding target node to perform corresponding maneuvering displacement or attitude switching actions; and simultaneously collect the acoustic shock wave time difference signal and the electrical on / off signal of the target surface medium generated when the projectile is fired, and use the electrical on / off signal as a trigger reference to perform multimodal fusion calculation on the acoustic shock wave time difference signal to generate the projectile impact point coordinates and hit event data; The status feedback module is communicatively connected to the collaborative control module and the multimodal perception execution module. It is used to receive the hit event data and target node running status, and to perform target-level local view rendering of the bullet impact point or aggregate view rendering of the global battle damage situation of the target range according to different terminal service permissions.

2. The intelligent target range collaborative control system based on multimodal perception according to claim 1, characterized in that: The multimodal perception execution module includes: The signal synchronization window unit is used to take the sudden pulse of the acquired target surface dielectric electrical on / off signal as a time reference trigger signal, and retrieve the acoustic wave timestamp data acquired by the sensor array within a preset time window before and after the time reference trigger signal. The fusion computing unit is used to calculate the time difference of arrival coordinates to generate the impact point coordinates when there is sound wave timestamp data that conforms to the characteristic frequency within the preset time window, and to remove environmental noise interference signals outside the preset time window.

3. The intelligent target range collaborative control system based on multimodal perception according to claim 2, characterized in that: The multimodal sensing and execution module also includes a calibration-free mechanical decoupling structure: The calibration-free mechanical decoupling structure includes a rigid inner frame and a floating outer frame suspended on the rigid inner frame by a damping element. The sensor array is fixedly installed within the rigid inner frame, forming an absolutely static detection reference coordinate system; The outer side of the floating outer frame is connected to an elastic clamping member. The consumable target plate is detachably plugged and pulled to the front of the floating outer frame through the elastic clamping member, so that the physical vibration of the consumable target plate when it is hit by a ball is mechanically filtered through the floating outer frame and is not transmitted to the rigid inner frame; and when the consumable target plate is replaced, the origin of the detection reference coordinate system remains unchanged.

4. The intelligent target range collaborative control system based on multimodal perception according to claim 1, characterized in that: The underlying signal processing chain of the multimodal sensing execution module includes, in sequence: The acoustic and electrical physical sensors on the target node; A single-power rail-to-rail dual operational amplifier connected to the output of the physical sensor device; A high-speed voltage comparator connected to the output of the operational amplifier; A main control microprocessor connected to the output of the high-speed voltage comparator has an embedded high-speed capture timer interface for capturing the edge signal of the output of the high-speed voltage comparator and generating a precise hardware timestamp to perform the multimodal fusion calculation. The communication connection is to the self-organizing network radio frequency transceiver front end of the main control microprocessor, used to transmit the calculated data back to the collaborative control module.

5. The intelligent target range collaborative control system based on multimodal perception according to claim 1, characterized in that: Some of the target nodes are configured as independently powered all-terrain tracked robot targets, and their multimodal perception and execution modules further include: A high-precision positioning unit is used to obtain the real-time geographical coordinates of the tracked robot target; The differential drive execution unit, which uses a brushless motor, is used to parse the target waypoint and speed commands issued by the cooperative control module, and perform on-the-spot turning, forward movement or maneuvering evasion actions in combination with the real-time geographical coordinates.

6. The intelligent target range collaborative control system based on multimodal perception according to claim 1, characterized in that: The system also includes a human-machine two-way adversarial perception module, which includes: The simulated fire launch unit, integrated on the target node, is used to launch laser beam pulses with coded identifiers into a pre-set area of ​​the target range under the triggering of the cooperative control module. The individual soldier wearable sensing unit is configured as a wearable carrier with an embedded distributed sensor network. It is used to receive the laser beam pulses and parse the coded identifier, determine the hit location, and report the battle damage signal to the cooperative control module through a wireless communication link.

7. The intelligent target range collaborative control system based on multimodal perception according to claim 1, characterized in that: The system also includes a wireless video calibration module that assists the status feedback module in establishing a shooting reference, including: The image acquisition and transmission end is coaxially fixed inside the firing tube of the light weapon via a multi-caliber adaptable quick-connect shaft, and is used to acquire real-time video streams of the firing target surface; A dynamic reticle calibration terminal is used to display the electronic crosshair on the real-time video stream; in response to the user's calibration translation command, the center origin of the electronic crosshair is dynamically translated to the geometric average center of the impact points of multiple consecutive live bullets, so as to construct a shooting aiming reference point based on software compensation.

8. The intelligent target range collaborative control system based on multimodal perception according to claim 7, characterized in that: The multi-diameter adaptable quick-connect shaft includes a first outer diameter reference section and a second outer diameter reference section, which are connected to the housing base of the image acquisition and transmission end by elastic snap-fit, so as to ensure that the quick-connect shafts with different outer diameters are aligned with the optical central axis of the image acquisition and transmission end when they are replaced.

9. The intelligent target range collaborative control system based on multimodal perception according to claim 1, characterized in that: The specific configuration of the status feedback module is as follows: On the individual shooter terminal, the local view rendering includes applying different colors or symbols to the historical bullet impact points and the latest bullet impact point of the currently assigned target position; On the global command terminal, the aggregated view rendering includes scattered domain data of all target nodes clustered based on time windows, and overlays and outputs a real-time live video stream of the target range.

10. The intelligent target range collaborative control system based on multimodal perception according to claim 2, characterized in that: The system has the following operating parameters: Under a pre-set standard test environment, the target reporting response time threshold of the multimodal fusion calculation is less than or equal to 0.02 seconds, and the accuracy error limit of the calculated bullet impact point coordinates is less than or equal to 5 millimeters. When the collaborative control module issues commands to multiple points through the wireless self-organizing network, the maximum number of target nodes that can be cascaded in a single network is no less than 50.

Images