A transient speed driving high-mobility soft pursuit robot and a control method thereof

Abstract

The application discloses a kind of transient speed drive high mobility soft pursuit robot and its control method, including bionic mechanical main body and intelligent control system, bionic mechanical main body includes flexible bionic spine, flexible air pipe, gliding membrane drive mechanism, foldable bionic gliding membrane, image acquisition module and several electric wheels.The flexible skeleton and embedded flexible air pipe are deeply fused in the application, the perfect synergy of structural compliance and high power density drive is realized;The structure not only can support low-energy consumption smooth cruise, but also can accurately convert transient knock energy into directional driving force for driving spine deformation in the ignition moment, give the robot millisecond level instantaneous ejection force, and form gliding motion after gliding membrane is unfolded, to expand the movement displacement of robot in horizontal direction, so as to improve the mobility and dynamic response performance of soft system in complex environment.

Description

Technical Field

[0001] This invention relates to the field of robotics, specifically to a high-mobility soft pursuit robot driven by instantaneous speed change and its control method. Background Technology

[0002] With the continuous development of robotics technology in applications such as field inspection, target tracking, disaster relief, and maneuvering in complex terrain, higher demands are being placed on the mobility of robots in unstructured environments. Although traditional wheeled or tracked robots have mature structures and stable control, their movement is limited when facing ditches, obstacles, or high-speed moving targets, making it difficult to balance traversing ability and energy efficiency.

[0003] To improve the robot's ability to navigate complex environments, some studies have introduced jumping or launching mechanisms. However, existing jumping robots mostly rely on rigid links or high-power motors for propulsion, which suffers from problems such as high structural rigidity, high impact loads, and limited horizontal displacement per unit energy consumption, making it difficult to meet the combined requirements of long-distance maneuverability and structural compliance.

[0004] Highly maneuverable gliding animals in nature, characterized by rapid, explosive speeds, can deploy a gliding membrane connecting their forelimbs and hindlimbs after takeoff. Through gliding motion, they can achieve a large range of horizontal displacement without continuous active propulsion, thus completing leaps and maneuvers with low energy consumption. This "jump-gliding" composite movement provides important insights for improving the efficiency of robot movement.

[0005] Some studies have attempted to introduce membrane structures or gliding mechanisms into robots, but the following shortcomings still exist: (1) Insufficient initial kinetic energy: Most systems lack a high power density instantaneous drive mode, making it difficult to obtain stable and controllable initial kinetic energy during the take-off phase; (2) The problem of response lag is prominent: software-driven methods usually have physical charging time. When the target enters a high-speed motion state, the robot has difficulty in completing the jump and attitude adjustment in time. (3) Simple motion mode switching strategy: Existing systems mostly use fixed thresholds to determine take-off or gliding. When the target state changes frequently, it is easy to trigger unnecessary mode switching, which affects the stability and energy efficiency of the system.

[0006] Therefore, there is an urgent need for a transient speed-driven high-mobility soft robot that can achieve rapid take-off, stable gliding, and intelligent mode switching while ensuring structural compliance, so as to improve the robot's mobility and pursuit efficiency in complex environments. Summary of the Invention

[0007] The purpose of this invention is to provide a high-mobility soft pursuit robot driven by instantaneous speed and its control method, so as to solve the problems mentioned in the background art.

[0008] To achieve the above objectives, the present invention provides the following technical solution: A high-mobility soft pursuit robot with instantaneous speed drive includes a bionic mechanical body and an intelligent control system mounted on the bionic mechanical body. The bionic mechanical body includes a flexible bionic spine, a flexible trachea arranged along the flexible bionic spine, two sets of gliding membrane devices respectively arranged on the left and right sides of the flexible bionic spine, an image acquisition module located on the bionic mechanical body, and several electric wheels located at the bottom of the bionic mechanical body. The gliding membrane device includes two sets of gliding membrane drive mechanisms respectively arranged at the front and rear ends of the flexible bionic spine, and a foldable bionic gliding membrane connected between the two sets of gliding membrane drive mechanisms. A chemical energy release reaction device is arranged on the flexible trachea. The intelligent control system controls the operation of the chemical energy release reaction device according to the data information acquired by the image acquisition module. The chemical energy release reaction device inflates the flexible trachea through a chemical energy release reaction. When the flexible trachea expands, it drives the flexible bionic spine to switch from a retracted state to an extended state. At the same time, the expansion drives the foldable bionic gliding membrane to open. The gliding membrane drive mechanism controls the retraction and closure of the foldable bionic gliding membrane.

[0009] Furthermore, the side of the foldable bionic gliding membrane away from the gliding membrane drive mechanism is detachably connected to the flexible bionic spine. When the flexible trachea expands, it squeezes to separate the side of the foldable bionic gliding membrane away from the gliding membrane drive mechanism from the flexible bionic spine.

[0010] Furthermore, the side of the foldable bionic gliding membrane away from the gliding membrane drive mechanism is magnetically connected to the flexible bionic spine.

[0011] Furthermore, the gliding membrane drive mechanism includes a drive motor and an unfolding bracket driven to rotate by the drive motor, the unfolding bracket being connected to the end of the foldable biomimetic gliding membrane.

[0012] Furthermore, the flexible bionic spine is an arc-shaped strip structure, and the foldable bionic gliding membrane and the flexible bionic spine form a D-shaped structure.

[0013] Furthermore, the image acquisition module includes a camera; the intelligent control system includes an advance prediction module, a multimodal game decision module, and an execution control module; the advance prediction module acquires image sequences through the camera; the intelligent control system executes target recognition and coordinate transformation algorithms based on the image sequences acquired by the camera to calculate the motion state of the target to be pursued and to calculate the predicted speed threshold index; the multimodal game decision module is used to establish a cost function including a switching penalty term for path planning; the execution control module is used to control the movement of the software pursuit robot according to the decision results.

[0014] The present invention also provides a control method for a transient speed-driven high-mobility soft pursuit robot as described above, comprising: Step 1: Periodically acquire the three-dimensional coordinate position of the target object at the current sampling time through the image acquisition module; Step 2: Utilize the advanced prediction module to execute the pre-trigger judgment mechanism to calculate the future motion state of the target and determine whether to initiate the pre-charging procedure; Step 3: Using MPC mode switching logic with cost penalty, the path comprehensive evaluation value of the soft pursuit robot under different motion modes is calculated in parallel. Step 4: Based on the comparison results of the comprehensive evaluation values, output the optimal control command; when it is determined that maintaining the current state is better, maintain wheeled cruise; when it is determined that the jump-glide maneuver has a better comprehensive evaluation value than wheeled cruise, thus enabling effective interception of the escaping target, execute the explosive jump and deploy the gliding membrane command.

[0015] Furthermore, the advance prediction triggering logic in step 2 includes the following steps: S201, using the extended Kalman filter algorithm, calculates the target's current velocity V based on the historical coordinate sequence. t and acceleration a t ; S202, Introducing the prediction time constant τ, calculate the target's predicted velocity threshold index V. pred The calculation formula is as follows: V pred =V t + a t *τ Simultaneously, the computer dynamic assessment factor M target The calculation formula is as follows: , in, P is the rate of change of acceleration; t is the current sampling time; T is the observation window time; target P represents the three-dimensional coordinate position of the target object. robot This represents the robot's three-dimensional coordinate position. The distance between the two is the Euclidean distance, and σ is the preset weighting coefficient; S203, λ up If V is a preset escape velocity threshold, then... pred ≥λ up And the target mobility assessment factor M target If the trigger threshold is reached, the intelligent control system determines that the target is about to escape in a high-dynamic manner, even if V... t Even before the physical threshold is reached, the intelligent control system still initiates the pre-charging procedure of the flexible air tube ahead of schedule. The pre-charging procedure refers to the process where the intelligent control system determines the target's mobility assessment factor M. target When the trigger threshold is reached, a preset proportion of reaction medium is injected into the flexible trachea in advance, and an initial pressure environment suitable for the occurrence of detonation is established.

[0016] Furthermore, step 3 includes the following steps: S301, within the current decision-making cycle, the intelligent control system calculates two candidate paths in parallel: Path A: Maintaining the wheel rolling mode, the intelligent control system controls the electric vehicle wheels to cruise on the ground, and calculates the cumulative energy consumption E over the next N cycles. wheel ; Path B: Switch to explosive jump mode. The intelligent control system drives the bionic spine to deform and jump by the expansion force generated by the transient detonation, and calculates the cumulative execution energy consumption E over the next N cycles. jump ; S302, calculate the comprehensive evaluation value J for the two candidate paths respectively: , Where K is the time step index, representing each prediction time from step 1 to step N; P target,k and P robot,k These represent the predicted target position and robot position in the time domain, respectively; Q is the tracking error weight matrix; E mode The energy consumption cost of execution under different modes; When path A is selected, E mode =E wheel , representing the electrical energy loss required for the electric vehicle wheels to roll; When path B is selected, E mode =E jump This represents the chemical energy cost required for a flexible trachea to undergo transient detonation. I is a state transition indicator function, which takes a value of 1 when a motion mode transition occurs, and 0 otherwise. switch To switch the penalty constant.

[0017] Compared with the prior art, the beneficial effects of the present invention are: (1) Anticipatory response capability: By introducing the energy storage time constant τ and acceleration prediction algorithm, the physical lag of the soft actuator is effectively compensated, and the robot's response time to highly dynamic targets is improved.

[0018] (2) Intelligent game decision-making: The design of the penalty cost function based on MPC makes the robot no longer mechanically switch modes according to distance, but will weigh "motion gain" and "execution cost", avoiding frequent switching caused by simple threshold judgment, and improving the stability and energy efficiency of the system.

[0019] (3) Advantages of the high-mobility rigid-flexible coupling structure under transient speed drive: The present invention deeply integrates the flexible skeleton with the embedded flexible trachea, realizing the perfect synergy between structural compliance and high power density drive; This structure can not only support stable cruise with low energy consumption, but also accurately convert transient detonation energy into directional driving force to drive spinal deformation at the moment of ignition, giving the robot a millisecond-level instantaneous ejection force, and forming gliding motion after the gliding membrane unfolds, so as to expand the robot's horizontal displacement, thereby improving the maneuverability and dynamic response performance of the soft system in complex environments. Attached Figure Description

[0020] Figure 1 A schematic diagram of the normal state structure of a transient speed-driven, highly maneuverable soft-body pursuit robot; Figure 2 This is a schematic diagram of the drive state structure of the transient speed-driven high-mobility soft pursuit robot; Figure 3 Flowchart for switching between robot multimodal motion (cruising / jumping-gliding); Figure 4 The flowchart of the algorithm for predictive triggering logic and MPC mode switching logic.

[0021] In the diagram: 1 is a flexible bionic spine, 100 is a partition, 2 is a flexible trachea, 3 is a foldable bionic gliding membrane, 4 is a chassis, 5 is an electric vehicle wheel, 6 is a camera, 7 is a lithium battery, 8 is a chip, and 9 is a motor. Detailed Implementation

[0022] 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.

[0023] Please see Figure 1A high-mobility soft pursuit robot with instantaneous speed drive includes a bionic mechanical body and an intelligent control system mounted on the bionic mechanical body. The bionic mechanical body includes a flexible bionic spine 1, a flexible air tube 2 arranged along the flexible bionic spine 1, two sets of gliding membrane devices respectively arranged on the left and right sides of the flexible bionic spine 1, an image acquisition module located on the bionic mechanical body, and several electric wheels 5 located at the bottom of the bionic mechanical body. The gliding membrane device includes two sets of gliding membrane drive mechanisms respectively arranged at the front and rear ends of the flexible bionic spine 1, and a foldable bionic gliding membrane 3 connected between the two sets of gliding membrane drive mechanisms. A chemical energy release reaction device is set on the flexible air tube 2. The intelligent control system controls the operation of the chemical energy release reaction device according to the data information collected by the image acquisition module. The chemical energy release reaction device inflates the flexible air tube 2 through chemical energy release reaction. When the flexible air tube 2 expands, it drives the flexible bionic spine 1 to switch from a retracted state to an extended state. At the same time, the expansion drives the foldable bionic gliding membrane 3 to open. The gliding membrane drive mechanism controls the retraction and closure of the foldable bionic gliding membrane 3. There are two sets of foldable bionic gliding membranes 3, which are arranged on both sides of the flexible bionic spine 1. When the foldable bionic gliding membranes 3 are opened, they form a force-bearing plane, similar to the function of wings.

[0024] The intelligent control system controls the operation of the chemical energy release reaction device based on the data information collected by the image acquisition module. The chemical energy release reaction device inflates the flexible air tube 2 through chemical energy release reaction. When the flexible air tube 2 expands, it drives the flexible bionic spine 1 to switch from a retracted state to an extended state. The structure and principle of the chemical energy release reaction device are existing technologies and will not be described in detail.

[0025] Continue reading Figure 1 The gliding membrane drive mechanism includes a drive motor 9 and an unfolding support 11 driven by the drive motor 9 to rotate. The unfolding support 11 is connected to the end of the foldable bionic gliding membrane 3. The drive motor 9 drives the foldable bionic gliding membrane 3 to flip through the unfolding support 11. When the foldable bionic gliding membrane 3 is opened, the gliding membrane drive mechanism can drive the foldable bionic gliding membrane 3 to flip to a posture that fits with the flexible bionic spine 1, thereby realizing the folding and closing of the foldable bionic gliding membrane 3.

[0026] Continue reading Figure 1 The foldable bionic gliding membrane 3 is detachably connected to the flexible bionic spine 1 on the side away from the gliding membrane drive mechanism. When the flexible trachea expands, it separates the foldable bionic gliding membrane from the flexible bionic spine 1 by squeezing the side away from the gliding membrane drive mechanism.

[0027] The detachable connection between the foldable bionic gliding membrane 3 and the flexible bionic spine 1 is preferably achieved through magnetic attraction. Specifically, an iron block is installed on the side of the foldable bionic gliding membrane 3 away from the gliding membrane drive mechanism, and a magnet is installed at the corresponding position on the flexible bionic spine 1. When the foldable bionic gliding membrane 3 is folded, the iron block and the magnet are magnetically attracted and fixed. When the flexible air tube 2 is inflated, the iron block and the magnet are separated by compression, causing the foldable bionic gliding membrane 3 to rotate downwards. The maximum downward rotation angle of the foldable bionic gliding membrane 3 is limited by a known method to ensure it just opens. A one-way bearing is installed on the output shaft of the motor 9, ensuring that the motor 9 can only drive the foldable bionic gliding membrane 3 to fold, without affecting its opening.

[0028] Continue reading Figure 1 The flexible bionic spine 1 is an arc-shaped strip structure composed of several elastic bone segments connected in series. The foldable bionic gliding membrane 3 and the flexible bionic spine 1 form a D-shaped structure, and the flexible trachea 2 is located on the concave side of the arc structure of the flexible bionic spine 1. Specifically, several partitions 100 are spaced apart along the length of the concave side of the flexible bionic spine 1, and mounting holes are provided on the partitions 100. The flexible trachea 2 is inserted and fitted into the mounting holes on all the partitions.

[0029] Continue reading Figure 1 The image acquisition module includes a camera 6 and a chip 8. The intelligent control system includes a predictive module, a multimodal game decision-making module, and an execution control module, all of which are computer programs stored in the chip 8. The predictive module acquires image sequences through the camera 6 and executes target recognition and coordinate transformation algorithms in the chip 8 to calculate the motion state of the target being pursued and to calculate the predicted speed threshold. The multimodal game decision-making module is used to establish a cost function including a switching penalty term for path planning. The execution control module is used to control the movement of the software pursuit robot based on the decision results.

[0030] like Figure 1 , Figure 2 and Figure 3 As shown, when performing the explosive jump-gliding maneuver, the flexible trachea 2 generates high-pressure expansion force through the transient explosion of the reaction medium. On the one hand, the flexible bionic spine 1 constrains and guides the pressure, accurately converting the omnidirectional expansion energy into a directional driving force that drives the deformation of the spine. On the other hand, the expansion of the flexible trachea 2 drives the foldable bionic gliding membrane 3 to switch from a closed state to an open state. This transient driving mechanism enables the flexible bionic spine 1 to generate extremely high instantaneous impulse within milliseconds, thereby simulating the rapid extension of the trunk and opening of the gliding membrane during the take-off of a highly mobile organism with transient speed burst characteristics, realizing ejection jump and gliding motion under high load.

[0031] The present invention provides a control method for a transient speed-driven high-mobility soft pursuit robot as described above, the logic flow of which is as follows: Figure 4 As shown, the specific steps are as follows: Step 1: Periodically acquire the three-dimensional coordinate position P of the target object to be tracked at the current sampling time t using the camera in the image acquisition module. target ,include: S101, the robot continuously acquires image sequences of the target to be pursued through the camera in the image acquisition module; S102, the intelligent control system performs target recognition and coordinate transformation on the image sequence. By mapping the image pixel coordinates to the robot's current three-dimensional coordinate system, it calculates and obtains the three-dimensional coordinate position P of the target to be pursued at the current sampling time t in real time. target ; Step 2: Utilize the advanced prediction module to execute a pre-triggering determination mechanism to calculate the target's future motion state and determine whether to initiate the pre-charging procedure, including: S201, the extended Kalman filter algorithm is used to solve the target's motion state, and the target's velocity V at the current moment is obtained. t acceleration a t and the rate of change of acceleration rate of change of acceleration The instantaneous explosive force used to characterize the change in motion state of a target is existing technology and will not be elaborated upon. Simultaneously, the system allocates an observation time window of length T in memory, and calculates the... The sequence is pushed in order according to timestamps to complete the encapsulation of the target from instantaneous motion parameters to historical maneuver characteristics. The observation time window T is used to limit the time domain range of the integration operation.

[0032] S202 introduces a prediction time constant τ and calculates the target's predicted velocity index V based on the current motion state. pred The calculation formula is as follows: V pred =V t + a t *τ Here, τ represents the time scale required for the blasting drive mechanism to go from startup to having effective output capability.

[0033] Secondly, in order to quantify the complexity of the target's escape behavior, the maneuverability assessment factor M is calculated using the following model. target :

[0034] Among them, P robot This represents the robot's three-dimensional coordinate position. σ represents the Euclidean distance between the target and the robot; σ is a preset weighting coefficient.

[0035] S203; When V is satisfied pred ≥λ up When, where λ up Based on a preset escape velocity threshold, the system determines that the target is about to accelerate and escape. Even if the target's current velocity V is... t Even before the escape threshold is reached, the control system still initiates the pre-charging program of the explosive jump-glide mode in advance to avoid missing the pursuit opportunity due to the lag in the response of the soft structure.

[0036] Real-time monitoring and prediction speed V pred With mobility assessment factor M target When any indicator exceeds the preset dynamic escape threshold, the target is determined to have entered the high-value interception window. To compensate for the inherent physical response lag of the software-driven mechanism, the system immediately triggers a pre-charge program. The pre-charge program refers to the process where the intelligent control system determines the target's maneuverability assessment factor M. target When the trigger threshold is reached, a preset proportion of reaction medium is injected into the flexible trachea in advance, and an initial pressure environment suitable for the occurrence of detonation is established, so that the flexible trachea enters a high-pressure ready state in advance, ensuring that the robot can achieve a millisecond-level transient response when the decision-making brain outputs instructions.

[0037] Step 3: Employing MPC mode switching logic with penalty, the system calculates the comprehensive path evaluation value of the soft pursuit robot under different motion modes in parallel, including: S301, within the current decision-making cycle, the intelligent control system calculates two candidate paths in parallel: Path A: Maintaining the wheel rolling mode, the intelligent control system controls the electric vehicle wheels to cruise on the ground, and calculates the cumulative energy consumption E over the next N cycles. wheel .

[0038] Path B: Switch to explosive jump mode. The intelligent control system drives the bionic spine to deform and jump by the expansion force generated by the transient explosion, and calculates the cumulative execution energy consumption E over the next N cycles. jump .

[0039] S302, Calculate the comprehensive path evaluation value J. The comprehensive path evaluation value J is used to quantify the execution cost under different motion modes: The intelligent control system adopts MPC mode switching logic with cost penalties, enabling the robot to make decisions between wheeled cruising mode and explosive jump-glide mode, ensuring that the robot always adopts a motion mode that is more conducive to approaching the target or blocking its escape path during the pursuit. This logic calculates the comprehensive path evaluation value J in parallel for the two paths of "maintaining wheeled rolling" and "switching to explosive jump-glide" in each decision cycle. The system constructs the following discrete time-domain cost function J within the prediction time domain N:

[0040] Where N is the prediction time domain, representing the total number of cycles in the forward simulation of the system; K is the time step index, representing each prediction time from step 1 to step N; P target,k and P robotz These represent the predicted target position and robot position in the time domain, respectively; Q is the tracking error weight matrix; E mode This represents the energy consumption cost of execution under different modes. When the simulation path is a wheeled cruise, E... mode =E wheel E represents the electrical energy loss required for the electric vehicle's wheels to roll. When the simulation path is a blast jump, E mode =E jump , representing the chemical energy cost required for transient detonation of the flexible trachea. I is a state switching indicator function; I=1 when the motion mode switches from wheeled cruise to explosive jump, otherwise I=0. C switch This is a switching penalty term used to suppress ineffective oscillations in the system under critical conditions.

[0041] Step 4: Based on the comparison results of the comprehensive evaluation values, output the optimal control command; The system calculates the wheeled cruising cost J in real time using the functions described above. wheel With explosive jump value J jump The intelligent control system follows the decision-making criterion that the interception benefits significantly outweigh the switching costs: when it is determined that maintaining the current state is better, it maintains wheeled cruising; when it is determined that the jump-glide maneuver has a better multi-target path comprehensive evaluation value J than wheeled cruising, thus enabling effective interception of the escaping target, it executes the explosive jump and deploys the gliding membrane command.

[0042] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A high-mobility soft-body pursuit robot driven by instantaneous speed change, characterized in that, The system includes a bionic mechanical body and an intelligent control system mounted on the bionic mechanical body. The bionic mechanical body includes a flexible bionic spine, a flexible trachea arranged along the flexible bionic spine, two sets of gliding membrane devices respectively arranged on the left and right sides of the flexible bionic spine, an image acquisition module located on the bionic mechanical body, and several electric vehicle wheels located at the bottom of the bionic mechanical body. The gliding membrane device includes two sets of gliding membrane drive mechanisms respectively arranged at the front and rear ends of the flexible bionic spine, and a foldable bionic gliding membrane connected between the two sets of gliding membrane drive mechanisms. A chemical energy release reaction device is set on the flexible trachea. The intelligent control system controls the operation of the chemical energy release reaction device according to the data information acquired by the image acquisition module. The chemical energy release reaction device inflates the flexible trachea through a chemical energy release reaction. When the flexible trachea expands, it drives the flexible bionic spine to switch from a retracted state to an extended state. At the same time, the expansion drives the foldable bionic gliding membrane to open. The gliding membrane drive mechanism controls the retraction and closure of the foldable bionic gliding membrane. The flexible bionic spine is an arc-shaped structure composed of several elastic bone segments connected in series. The foldable bionic gliding membrane and the flexible bionic spine form a D-shaped structure, and the flexible trachea is located on the concave side of the arc structure of the flexible bionic spine.

2. The transient speed-driven high-mobility soft pursuit robot according to claim 1, characterized in that, The side of the foldable bionic gliding membrane away from the gliding membrane drive mechanism is detachably connected to the flexible bionic spine. When the flexible trachea expands, it squeezes the side of the foldable bionic gliding membrane away from the gliding membrane drive mechanism to separate from the flexible bionic spine.

3. The transient speed-driven high-mobility soft pursuit robot according to claim 2, characterized in that, The side of the foldable bionic gliding membrane away from the gliding membrane drive mechanism is magnetically connected to the flexible bionic spine.

4. The transient speed-driven high-mobility soft pursuit robot according to claim 1, characterized in that, The gliding membrane drive mechanism includes a drive motor and an unfolding support driven by the drive motor to rotate. The unfolding support is connected to the end of the foldable biomimetic gliding membrane.

5. The transient speed-driven high-mobility soft pursuit robot according to claim 1, characterized in that, The image acquisition module includes a camera; the intelligent control system includes a predictive module, a multimodal game decision-making module, and an execution control module; the predictive module acquires image sequences through the camera; the intelligent control system executes target recognition and coordinate transformation algorithms based on the image sequences acquired by the camera to calculate the motion state of the target to be pursued and to calculate the predicted speed threshold index; the multimodal game decision-making module is used to establish a cost function including a switching penalty term for path planning; the execution control module is used to control the movement of the software pursuit robot according to the decision results.

6. A control method for a transient speed-driven high-mobility soft pursuit robot as described in any one of claims 1-5, characterized in that, include: Step 1: Periodically acquire the three-dimensional coordinate position of the target object at the current sampling time through the image acquisition module; Step 2: Utilize the advanced prediction module to execute the pre-trigger judgment mechanism to calculate the future motion state of the target and determine whether to initiate the pre-charging program; Step 3: Using MPC mode switching logic with cost penalty, the path comprehensive evaluation value of the soft pursuit robot under different motion modes is calculated in parallel. Step 4: Based on the comparison results of the comprehensive evaluation values, output the optimal control command; when it is determined that maintaining the current state is better, maintain wheeled cruise; when it is determined that the jump-glide maneuver has a better comprehensive evaluation value than wheeled cruise, thus enabling effective interception of the escaping target, execute the explosive jump and deploy the gliding membrane command.

7. The control method according to claim 6, characterized in that, The advance prediction triggering logic in step 2 includes the following steps: S201, using the extended Kalman filter algorithm, calculates the target's current velocity V based on the historical coordinate sequence. t and acceleration a t ; S202, Introducing the prediction time constant τ, calculate the target's predicted velocity threshold index V. pred The calculation formula is as follows: V pred =V t + a t *τ Simultaneously, the computer dynamic assessment factor M target The calculation formula is as follows: ， in, P is the rate of change of acceleration; t is the current sampling time; T is the observation window time; target P represents the three-dimensional coordinate position of the target object. robot This represents the robot's three-dimensional coordinate position. The distance between the two is the Euclidean distance, and σ is the preset weighting coefficient; S203, λ up If V is a preset escape velocity threshold, then... pred ≥λ up And the target mobility assessment factor M target If the trigger threshold is reached, the intelligent control system determines that the target is about to escape in a high-dynamic manner, even if V... t Even before the physical threshold is reached, the intelligent control system still initiates the pre-charging procedure of the flexible air tube ahead of schedule. The pre-charging procedure refers to the process where the intelligent control system determines the target's mobility assessment factor M. target When the trigger threshold is reached, a preset proportion of reaction medium is injected into the flexible trachea in advance, and an initial pressure environment suitable for the occurrence of detonation is established.

8. The control method according to claim 6, characterized in that, Step 3 includes the following steps: S301, within the current decision-making cycle, the intelligent control system calculates two candidate paths in parallel: Path A: Maintaining the wheel rolling mode, the intelligent control system controls the electric vehicle wheels to cruise on the ground, and calculates the cumulative energy consumption E over the next N cycles. wheel ; Path B: Switch to explosive jump mode. The intelligent control system drives the bionic spine to deform and jump by the expansion force generated by the transient detonation, and calculates the cumulative execution energy consumption E over the next N cycles. jump ; S302, calculate the comprehensive evaluation value J for the two candidate paths respectively: ， Where K is the time step index, representing each prediction time from step 1 to step N; P target,k and P robot,k These represent the predicted target position and robot position in the time domain, respectively; Q is the tracking error weight matrix; E mode The energy consumption cost of execution under different modes; When path A is selected, E mode =E wheel , representing the electrical energy loss required for the electric vehicle wheels to roll; When path B is selected, E mode =E jump This represents the chemical energy cost required for a flexible trachea to undergo transient detonation. I is a state transition indicator function, which takes a value of 1 when a motion mode transition occurs, and 0 otherwise. switch To switch the penalty constant.

Images