Double-software mechanical arm cooperation method for underwater grabbing task

Abstract

The invention discloses an underwater grabbing task-oriented double-software mechanical arm cooperation method, which is characterized by comprising the following steps of S1, establishing an automatic machine model of an underwater cooperative grabbing task of a double-software mechanical arm system; s2, on the basis of the automaton model established in the step S1, an upper-layer task planner is established in combination with a supervision and control theory; and S3, when the double-software mechanical arm system works, the sensed underwater sensing signals are abstracted into discrete events, the discrete events serve as input of an upper-layer task planner to trigger the upper-layer task planner to generate action instructions, and the generated action instructions are issued to independent motion controllers of the two mechanical arms respectively. According to the method, a perception-action sequence is described in a formalized mode based on an automaton model, an upper-layer task planning problem is converted into a supervision control problem to obtain a feasible task planner, and the upper-layer planning requirement of an underwater grabbing task is met.

Description

Technical Field

[0001] The present invention relates to the field of robotic arm control technology, and in particular to a dual-soft robotic arm collaboration method for underwater grasping tasks. Background Art

[0002] Considering the safe and flexible manipulation capabilities demonstrated by human dual-arm collaboration, researchers have extensively studied and applied dual-arm motion skills to bimanual rigid manipulator (BRM) systems. In recent years, BRM systems have demonstrated excellent manipulation capabilities in numerous industrial applications, such as automotive parts assembly. Although BRM systems can perform dexterous collaboration in free environments, rigid manipulators can collide with other rigid manipulators or with the environment in confined environments, causing irreversible damage to the mechanical structure and even the environment. Therefore, to reduce the environmental damage caused by rigid manipulators, researchers have proposed using soft manipulators to perform sustainable and autonomous exploration tasks in fragile interactive environments. The flexibility of bimanual soft manipulators (BSM) systems in confined spaces is more similar to that of human arms, enabling them to effectively avoid collisions. In other words, BSM systems can exhibit human-like safety in confined spaces. The BSM system working underwater is called the Underwater Bimanual Soft Manipulators (UBSM) system.

[0003] The underwater interactive working environment presents two major challenges for collaborative grasping in UBSM systems. The first is the need to coordinate multiple controlled objects during the underwater grasping process. The second is that the target objects in the interactive underwater environment may drift during the grasping process, meaning their motion state may change. Taking all these factors into consideration, designing a coordination framework and upper-level task planner for underwater collaborative grasping tasks for UBSM systems is an urgent challenge. Summary of the Invention

[0004] The purpose of the present invention is to provide a dual soft robotic arm collaboration method for underwater grasping tasks, which is used to solve the technical problems existing in the background technology.

[0005] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] A dual-soft robotic arm collaboration method for underwater grasping tasks includes the following steps:

[0007] S1. Establish an automaton model for underwater collaborative grasping tasks of a dual soft robotic arm system;

[0008] S2. Based on the automaton model established in step S1, an upper-level task planner is established in combination with supervisory control theory;

[0009] S3. When the dual soft robotic arm system is working, the sensed underwater sensor signals are abstracted into discrete events, and the discrete events are used as the input of the upper-level task planner to trigger the upper-level task planner to generate action instructions, and the generated action instructions are sent to the independent motion controllers of the two robotic arms respectively.

[0010] Furthermore, in step S1, when establishing the automaton model, the control strategy of the dual soft robotic arm system for underwater collaborative grasping is determined, and the controllable action events and uncontrollable observation events of the dual soft robotic arm system for underwater collaborative grasping are determined.

[0011] Furthermore, in step S1, the dual-soft robotic arm system includes a master soft robotic arm and a slave soft robotic arm, and the control strategy for the dual-soft robotic arm system to perform underwater collaborative grasping includes two collaborative grasping modes, namely, an asynchronous master-slave collaborative grasping mode and a synchronous master-slave collaborative grasping mode;

[0012] In the asynchronous master-slave collaborative grasping mode, for target objects whose size is smaller than a preset threshold, the master soft manipulator is responsible for grasping the target object, and the slave soft manipulator is responsible for blocking the target object to prevent it from slipping;

[0013] In the synchronous master-slave collaborative grasping mode, two soft manipulators jointly grasp the target object for a size not less than a preset threshold.

[0014] Furthermore, in step S1, the set of action events is represented as , its mathematical expression is as follows:

[0015] ;

[0016] In the above formula, the action event Explore represents the dual soft robotic arm system exploring the target object in the underwater environment through the endoscope, the action event Carry_Out_Confirmation represents the dual soft robotic arm system using the information from the endoscope to confirm whether the target object has been detected, the action event Measure represents the dual soft robotic arm system measuring the size of the detected target object through the endoscope, the action event Grasp_Support represents the dual soft robotic arm system in asynchronous master-slave collaborative grasping mode, the action event Syn_Grasp represents the dual soft robotic arm system in synchronous master-slave collaborative grasping mode, the action event Move_To_Box represents the dual soft robotic arm system moving the grasped target object to the recycling box, and the action event Release represents the dual soft robotic arm system releasing the grasped target object.

[0017] Furthermore, in step S1, the set of observed events is represented as , its mathematical expression is as follows:

[0018] ;

[0019] In the above formula, the observation events Object_Detect and No_Object_Detect represent the dual soft robotic arm system detecting and failing to detect the suspected target object through the endoscope, respectively. <T co and>=T co They represent the dual soft robotic arm system confirming that the size of the detected target object is less than or not less than the size threshold T. co The observation events Target_Not_Confirmed and Target_Confirmed respectively represent that the dual soft robotic arm system cannot confirm and confirm that the suspected object is the target object. The observation events Slave_Arrive and Slave_Not_Arrive respectively represent that the slave robotic arm reaches the target position and has not yet reached the target position when the dual soft robotic arm system is in the master-slave asynchronous collaborative grasping mode. The observation events Grasp_Complete and Grasp_Not_Complete respectively represent that the dual soft robotic arm system is in the master-slave synchronous grasping mode and has successfully grasped and has not yet successfully grasped. The observation events Reach_Box and Not_Reach_Box respectively represent that the dual soft robotic arm system collaborates to carry the target object to the recycling box and has not yet reached the recycling box.

[0020] Furthermore, in step S1, the event set of the automaton model is represented as Σ = Σ c Σ uc .

[0021] Furthermore, in step S2, an upper-level task planner is established to implement the desired perception-action strategy;

[0022] Specifically, in each underwater grasping mission, the dual-soft robotic arm system repeatedly experiences the following events after startup:

[0023] First, enter the action event Explore, explore the target object through the endoscope in the underwater environment, and then go through the observation events No_Object_Detect and Object_Detect, from never detecting to detecting a suspected target object;

[0024] Then enter the action event Carry_Out_Confirmation to confirm whether the suspected target object is the target object, until the observation event Target_Confirmed occurs, indicating that the object detected by the endoscope is the target object;

[0025] After the observation event Target_Confirmed occurs, the upper-level task planner issues a control command to enter the action event Measure to measure the size of the target object;

[0026] If the measured size of the target object is smaller than the size threshold T co , using observed events <T co Indicates that the upper-level task planner issues a control command to enter the action event Grasp_Support and enter the asynchronous master-slave collaborative grasping mode. In the asynchronous master-slave collaborative grasping mode, the upper-level task planner requires the dual-soft robotic arm system to continuously trigger the observation event Slave_Not_Arrive, and the action event remains unchanged at Grasp_Support until the slave robotic arm follows the given trajectory and the observation event Slave_Arrive occurs.

[0027] If the measured size of the target object is not less than the size threshold T co , with observed events>=T co Indicates that the upper-level task planner issues a control command to enter the action event Syn_Grasp and enter the synchronous master-slave collaborative grasping mode. In the synchronous master-slave collaborative grasping mode, the upper-level task planner requires the dual-soft robotic arm system to continuously trigger the observation event Grasp_Not_Complete, and the action event remains Syn_Grasp unchanged until the target object is successfully grasped and the observation event Grasp_Complete occurs;

[0028] After successfully grasping the target object, the upper-level task planner issues a control command to move the target object to the recycling bin, entering the action event Move_To_Box. The upper-level task planner requires the dual-soft robotic arm system to continuously trigger the observation event Not_Reach_Box, and the action event remains unchanged at Move_To_Box until the target object reaches the effective range of the recycling bin. The observation event Reach_Box occurs, and the upper-level task planner issues a control command to release the grasped target object, entering the action event Release, completing a grasp.

[0029] Compared with the existing technology, the dual-soft robotic arm collaboration method for underwater grasping tasks provided by the present invention formally describes the perception-action sequence based on the automaton model, transforms the upper-level task planning problem into a supervisory control problem to obtain a feasible task planner. The task planner takes underwater perception signals abstracted as discrete events as input and dual-arm action commands as output, thereby realizing the upper-level planning requirements of the underwater grasping task. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the technical solution in this embodiment, the following is a brief introduction to the drawings required for describing the embodiment. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0031] Figure 1 A schematic diagram of a dual-soft robotic arm collaboration method for underwater grasping tasks in an embodiment of the present invention;

[0032] Figure 2 is a schematic diagram of the automaton model;

[0033] Figure 3 Schematic diagram of the upper-level task planner. DETAILED DESCRIPTION

[0034] In order to make the technical means, creative features, objectives and effects achieved by the present invention easier to understand, the following further describes how the present invention is implemented in conjunction with the accompanying drawings and specific implementation methods.

[0035] Reference Figure 1 As shown, the present invention provides a dual soft robotic arm collaboration method for underwater grasping tasks, comprising the following steps:

[0036] S1. Establish an automaton model for the underwater collaborative grasping task of a dual soft robotic arm system.

[0037] S2. Based on the automaton model established in step S1, an upper-level task planner is established in combination with supervisory control theory.

[0038] S3. When the dual soft robotic arm system is working, the sensed underwater sensor signals are abstracted into discrete events, and the discrete events are used as the input of the upper-level task planner to trigger the upper-level task planner to generate action instructions, and the generated action instructions are sent to the independent motion controllers of the two robotic arms respectively.

[0039] In one specific embodiment, in step S1, when establishing an automaton model, a control strategy for a dual-soft robotic arm system performing underwater collaborative grasping is determined, along with controllable action events and uncontrollable observation events for the dual-soft robotic arm system performing underwater collaborative grasping. Automata, a primary model in discrete event systems, can describe dynamic systems composed of a series of discrete states and driven by events.

[0040] In this embodiment, the dual-soft robotic arm system includes a master soft robotic arm and a slave soft robotic arm. The control strategy for the dual-soft robotic arm system to perform underwater collaborative grasping includes two collaborative grasping modes: an asynchronous master-slave collaborative grasping mode and a synchronous master-slave collaborative grasping mode. In the asynchronous master-slave collaborative grasping mode, for target objects whose size is smaller than a preset threshold, the master soft robotic arm approaches the geometric center point of the target object and is responsible for grasping the target object, and the slave soft robotic arm follows the trajectory of the master soft robotic arm while maintaining a constant distance (e.g., the diameter of one robotic arm). The slave soft robotic arm is responsible for blocking the target object to prevent it from slipping. In the synchronous master-slave collaborative grasping mode, for target objects whose size is not smaller than a preset threshold, the two soft robotic arms jointly grasp the target object, specifically by synchronously grasping the target object from the shorter opposite sides. The asynchronous master-slave collaborative grasping mode solves the problem that smaller target objects may slip through the fingers of the soft robot arm with relatively poor fit; the synchronous master-slave collaborative grasping mode solves the problem that larger target objects may fall due to inappropriate grasping points when grasping and carrying them.

[0041] Furthermore, the set of action events is represented as , its mathematical expression is as follows:

[0042] ;

[0043] In the above formula, the action event Explore represents the dual soft robotic arm system exploring the target object in the underwater environment through the endoscope, the action event Carry_Out_Confirmation represents the dual soft robotic arm system using the information from the endoscope to confirm whether the target object has been detected, the action event Measure represents the dual soft robotic arm system measuring the size of the detected target object through the endoscope, the action event Grasp_Support represents the dual soft robotic arm system in asynchronous master-slave collaborative grasping mode, the action event Syn_Grasp represents the dual soft robotic arm system in synchronous master-slave collaborative grasping mode, the action event Move_To_Box represents the dual soft robotic arm system moving the grasped target object to the recycling box, and the action event Release represents the dual soft robotic arm system releasing the grasped target object.

[0044] Furthermore, the set of observed events is expressed as , its mathematical expression is as follows:

[0045] ;

[0046] In the above formula, the observation events Object_Detect and No_Object_Detect represent the dual soft robotic arm system detecting and failing to detect the suspected target object through the endoscope, respectively. <T co and>=T co They represent the dual soft robotic arm system confirming that the size of the detected target object is less than or not less than the size threshold T. co The observation events Target_Not_Confirmed and Target_Confirmed respectively represent that the dual soft robotic arm system cannot confirm and confirm that the suspected object is the target object. The observation events Slave_Arrive and Slave_Not_Arrive respectively represent that the slave robotic arm reaches the target position and has not yet reached the target position when the dual soft robotic arm system is in the master-slave asynchronous collaborative grasping mode. The observation events Grasp_Complete and Grasp_Not_Complete respectively represent that the dual soft robotic arm system is in the master-slave synchronous grasping mode and has successfully grasped and has not yet successfully grasped. The observation events Reach_Box and Not_Reach_Box respectively represent that the dual soft robotic arm system collaborates to carry the target object to the recycling box and has not yet reached the recycling box.

[0047] Determine the set of action events and the set of observed events Then the event set of the automaton model is expressed as Σ = Σ c Σ uc .

[0048] Reference Figure 2 As shown in FIG, a schematic diagram of the automaton model in this embodiment, all states are represented by numbers 1 to 15, and all states are assumed to be the expected states of the dual soft robotic arm system performing the underwater grasping task, for example: the state marked with number 14 corresponds to the situation where the dual soft robotic arm system has released the grasped target object into the recycling box.

[0049] In the automaton model, when the dual-soft robotic arm system starts working, the main robotic arm, which is equipped with an endoscope, begins exploring the environment (the action event Explore). If no object is detected (the observation event No_Object_Detect), the system continues exploring the environment (the action event Explore). Otherwise, if an object is detected (the observation event Object_Detect), the UBSM system can either continue exploring (the action event Explore) or execute the target object confirmation procedure (the action event Carry_Out_Confirmation) to check whether the object detected by the endoscope is the target object.

[0050] In the underwater target object confirmation procedure, there are usually three situations: 1) The detected object is not confirmed as the target (observation event Target_Not_Confirmed), the reason for which may be the low quality of the image captured by the endoscope; 2) The object is not detected again in the image collected by the endoscope in the current cycle (observation event No_Object_Detect), the reason for which may be that the soft robotic arm blocks the field of view of the endoscope; 3) The detected object has been confirmed as the target object (observation event Target_Confirmed).

[0051] In the third case, the UBSM system can perform the following two actions: 1) Repeat the confirmation procedure several times to ensure that the object detected by the endoscope is the target grasping object (action event Carry_Out_Confirmation); 2) The image processing PC of the dual soft robotic arm system measures the size of the target object in the endoscope image (action event Measure). If the measured size of the target object is smaller than the preset threshold T co (Observation event <T co ), the system enters the asynchronous master-slave collaborative grasping mode (action event Grasp_Support); otherwise, if the measured size of the target object is not less than the threshold T co (Observation event>=T co ), the system enters the synchronous master-slave collaborative grasping mode (action event Syn_Grasp).

[0052] In the asynchronous master-slave collaboration mode, as long as the slave robot has not reached the target position to assist the master robot (observation event Slave_Not_Arrive), the system's action event will remain Grasp_Support unchanged; once the slave robot reaches the target position, the system's observation event Slave_Arrive will be triggered, and the system's action event will be changed to moving the target object to the recycling box, that is, Move_To_Box.

[0053] Similarly, in the synchronous master-slave collaboration mode, as long as either the master or slave robot arm fails to grasp the target object (observation event Grasp_Not_Complete), the system's action event will remain Syn_Grasp unchanged; once both the master and slave robots grasp the target object, the system's observation event Grasp_Complete will be triggered, and the system's action event will also be transformed into moving the target object to the recycling box, that is, Move_To_Box.

[0054] While the system is moving the target object, if the master or slave arm does not reach the recycling bin, the system will continue to trigger the observation event Not_Reach_Box, and the system's action event will remain unchanged at Move_To_Box. If the system reaches the recycling bin, the system will trigger the observation event Reach_Box, and then execute the action event Release to release the grasped target object.

[0055] in addition, Figure 2 In the example, state 15 represents a "bad" state, which is intended to indicate the situation when the dual soft robot arm system performs some unreasonable operations. For example, it is reasonable for the system to perform the action event Explore when it starts (i.e., the initial state), but it is unreasonable to perform any other action event. Therefore, the system performs -Action events in {Explore} will cause the system to transition to the "bad" state.

[0056] Next, in step S2, an upper-level task planner is established to implement the desired perception-action strategy, such as Figure 3 The automaton model in step S1 and the upper-level task planner in step S3 differ only in the following ways: 1) The upper-level task planner removes all transitions that lead to "bad" states to prevent the system from failing to complete the collaborative grasping task; and 2) the upper-level task planner requires the system to execute the action event Carry_Out_Confirmation (confirmation of the target object) upon object detection, rather than continuing with the action event Explore.

[0057] Specifically, in each underwater grasping mission, the dual-soft robotic arm system repeatedly experiences the following events after startup:

[0058] First, the action event Explore is entered, and the target object is explored through the endoscope in the underwater environment. Then, after the observation events No_Object_Detect and Object_Detect, the suspected target object is detected.

[0059] Then enter the action event Carry_Out_Confirmation to confirm whether the suspected target object is the target object, until the observation event Target_Confirmed occurs, indicating that the object detected by the endoscope is the target object.

[0060] After the observation event Target_Confirmed occurs, the upper task planner issues a control command to enter the action event Measure to measure the size of the target object; the measurement can be based on the pixel value, and the predefined threshold T co , for example, it can be 10% of the pixel value of the endoscopic image.

[0061] If the measured size of the target object is smaller than the size threshold T co , using observed events <T co Indicates that the upper-level task planner issues a control command to enter the action event Grasp_Support and enter the asynchronous master-slave collaborative grasping mode. In the asynchronous master-slave collaborative grasping mode, the upper-level task planner requires the dual-soft robotic arm system to continuously trigger the observation event Slave_Not_Arrive, and the action event remains unchanged Grasp_Support until the slave robotic arm follows the given trajectory and the observation event Slave_Arrive occurs.

[0062] If the measured size of the target object is not less than the size threshold T co , with observed events>=T co Indicates that the upper-level task planner issues a control command to enter the action event Syn_Grasp and enter the synchronous master-slave collaborative grasping mode. In the synchronous master-slave collaborative grasping mode, the upper-level task planner requires the dual-soft robotic arm system to continuously trigger the observation event Grasp_Not_Complete, and the action event remains Syn_Grasp unchanged until the target object is successfully grasped and the observation event Grasp_Complete occurs.

[0063] After successfully grasping the target object, the upper-level task planner issues a control command to move the target object to the recycling bin, entering the action event Move_To_Box. The upper-level task planner requires the dual-soft robotic arm system to continuously trigger the observation event Not_Reach_Box, and the action event remains unchanged at Move_To_Box until the target object reaches the effective range of the recycling bin. The observation event Reach_Box occurs, and the upper-level task planner issues a control command to release the grasped target object, entering the action event Release, completing a grasp.

[0064] The upper-level planner needs to plan the system's desired motion instructions based on underwater continuous signals, artificially abstracted into discrete events. Therefore, in step S3, when the dual-soft robotic arm system is operating, it abstracts the perceived underwater sensor signals into discrete events. These discrete events serve as input to the upper-level task planner, triggering it to generate motion instructions. These instructions are then sent to the independent motion controllers of the two robotic arms.

[0065] In summary, the dual-soft robotic arm collaboration method for underwater grasping tasks provided by the present invention formally describes the perception-action sequence based on the automaton model, transforms the upper-level task planning problem into a supervisory control problem to obtain a feasible task planner. The task planner takes underwater perception signals abstracted as discrete events as input and dual-arm motion commands as output, thereby realizing the upper-level planning requirements of the underwater grasping task.

[0066] Finally, it should be noted that the above description is only an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A dual-soft robotic arm collaboration method for underwater grasping tasks, characterized by: The steps include: S1. Establish an automaton model for underwater collaborative grasping tasks of a dual soft robotic arm system; S2. Based on the automaton model established in step S1, an upper-level task planner is established in combination with supervisory control theory; S3. When the dual soft robotic arm system is working, the sensed underwater sensor signals are abstracted into discrete events, and the discrete events are used as the input of the upper-level task planner to trigger the upper-level task planner to generate action instructions, and the generated action instructions are sent to the independent motion controllers of the two robotic arms respectively.

2. The dual-soft robotic arm collaboration method for underwater grasping tasks according to claim 1 is characterized in that: In step S1, when establishing the automaton model, the control strategy of the dual soft robotic arm system for underwater collaborative grasping is determined, and the controllable action events and uncontrollable observation events of the dual soft robotic arm system for underwater collaborative grasping are determined.

3. The dual-soft robotic arm collaboration method for underwater grasping tasks according to claim 2 is characterized in that: In step S1, the dual-soft robotic arm system includes a master soft robotic arm and a slave soft robotic arm. The control strategy for the dual-soft robotic arm system to perform underwater collaborative grasping includes two collaborative grasping modes: asynchronous master-slave collaborative grasping mode and synchronous master-slave collaborative grasping mode. In the asynchronous master-slave collaborative grasping mode, for target objects whose size is smaller than a preset threshold, the master soft manipulator is responsible for grasping the target object, and the slave soft manipulator is responsible for blocking the target object to prevent it from slipping; In the synchronous master-slave collaborative grasping mode, two soft manipulators jointly grasp the target object for a size not less than a preset threshold.

4. The dual soft robotic arm collaboration method for underwater grasping tasks according to claim 3 is characterized in that: In step S1, the set of action events is represented as , its mathematical expression is as follows: ； In the above formula, the action event Explore represents the dual soft robotic arm system exploring the target object in the underwater environment through the endoscope, the action event Carry_Out_Confirmation represents the dual soft robotic arm system using the information from the endoscope to confirm whether the target object has been detected, the action event Measure represents the dual soft robotic arm system measuring the size of the detected target object through the endoscope, the action event Grasp_Support represents the dual soft robotic arm system in asynchronous master-slave collaborative grasping mode, the action event Syn_Grasp represents the dual soft robotic arm system in synchronous master-slave collaborative grasping mode, the action event Move_To_Box represents the dual soft robotic arm system moving the grasped target object to the recycling box, and the action event Release represents the dual soft robotic arm system releasing the grasped target object.

5. The dual soft robotic arm collaboration method for underwater grasping tasks according to claim 4 is characterized in that: In step S1, the set of observed events is represented as , its mathematical expression is as follows: ； In the above formula, the observation events Object_Detect and No_Object_Detect represent the dual soft robotic arm system detecting and failing to detect the suspected target object through the endoscope, respectively. <T co and>=T co They represent the dual soft robotic arm system confirming that the size of the detected target object is less than or not less than the size threshold T. co The observation events Target_Not_Confirmed and Target_Confirmed respectively represent that the dual soft robotic arm system cannot confirm and confirm that the suspected object is the target object. The observation events Slave_Arrive and Slave_Not_Arrive respectively represent that the slave robotic arm reaches the target position and has not yet reached the target position when the dual soft robotic arm system is in the master-slave asynchronous collaborative grasping mode. The observation events Grasp_Complete and Grasp_Not_Complete respectively represent that the dual soft robotic arm system is in the master-slave synchronous grasping mode and has successfully grasped and has not yet successfully grasped. The observation events Reach_Box and Not_Reach_Box respectively represent that the dual soft robotic arm system collaborates to carry the target object to the recycling box and has not yet reached the recycling box.

6. The dual soft robotic arm collaboration method for underwater grasping tasks according to claim 5, characterized in that: In step S1, the event set of the automaton model is represented as Σ = Σ c Σ uc .

7. The dual soft robotic arm collaboration method for underwater grasping tasks according to claim 6, characterized in that: In step S2, an upper-level task planner is established to implement the desired perception-action strategy; Specifically, in each underwater grasping mission, the dual-soft robotic arm system repeatedly experiences the following events after startup: First, enter the action event Explore, explore the target object through the endoscope in the underwater environment, and then go through the observation events No_Object_Detect and Object_Detect, from never detecting to detecting a suspected target object; Then enter the action event Carry_Out_Confirmation to confirm whether the suspected target object is the target object, until the observation event Target_Confirmed occurs, indicating that the object detected by the endoscope is the target object; After the observation event Target_Confirmed occurs, the upper-level task planner issues a control command to enter the action event Measure to measure the size of the target object; If the measured size of the target object is smaller than the size threshold T co , using observed events <T co Indicates that the upper-level task planner issues a control command to enter the action event Grasp_Support and enter the asynchronous master-slave collaborative grasping mode. In the asynchronous master-slave collaborative grasping mode, the upper-level task planner requires the dual-soft robotic arm system to continuously trigger the observation event Slave_Not_Arrive, and the action event remains unchanged at Grasp_Support until the slave robotic arm follows the given trajectory and the observation event Slave_Arrive occurs. If the measured size of the target object is not less than the size threshold T co , with observed events>=T co Indicates that the upper-level task planner issues a control command to enter the action event Syn_Grasp and enter the synchronous master-slave collaborative grasping mode. In the synchronous master-slave collaborative grasping mode, the upper-level task planner requires the dual-soft robotic arm system to continuously trigger the observation event Grasp_Not_Complete, and the action event remains Syn_Grasp unchanged until the target object is successfully grasped and the observation event Grasp_Complete occurs; After successfully grasping the target object, the upper-level task planner issues a control command to move the target object to the recycling bin, entering the action event Move_To_Box. The upper-level task planner requires the dual-soft robotic arm system to continuously trigger the observation event Not_Reach_Box, and the action event remains unchanged at Move_To_Box until the target object reaches the effective range of the recycling bin. The observation event Reach_Box occurs, and the upper-level task planner issues a control command to release the grasped target object, entering the action event Release, completing a grasp.

Images