Dual-arm robot master-slave teleoperation method and operation system

Abstract

The application belongs to the technical field of robot control, and particularly relates to a master-slave remote operation method and operation system of a dual-arm robot. The method comprises the following steps: building a master-slave remote operation system; controlling a master arm and a slave arm to perform synchronous micro-movement, and generating a pre-compensation parameter set; collecting a current master arm posture and a slave arm posture as corresponding baseline postures; obtaining a detection value of an operator intention detection module, and determining whether to enter a following mode or a standby mode; in the following mode, an offset of a current master arm posture relative to a master arm baseline posture is calculated in real time, a slave arm target posture is generated by combining the pre-compensation parameter set and a mapping relationship between the master arm and the slave arm, a slave arm target joint angle is corrected, and when the system is switched from the following mode to the standby mode, the current slave arm and master arm actual postures are updated as new corresponding baseline postures. The application does not need absolute zero position calibration, and integrates calibration-free deployment, baseline anti-drift and communication glitch suppression into a unified safety control closed loop.

Description

Technical Field

[0001] This invention belongs to the field of robot control technology, specifically relating to a master-slave teleoperation method and operating system for a dual-arm robot. Background Technology

[0002] Existing dual-arm robot master-slave teleoperation systems typically require absolute zero-position calibration of the master and slave arms, alignment of installation postures, or correction of external reference fixtures before power-on or deployment to establish a one-to-one mapping relationship between the master and slave. For isomorphic dual-arm systems, the common practice is to establish joint space mapping through homing motion, encoder zero-position calibration, and manual alignment; for heterogeneous dual-arm systems, Cartesian space mapping is often established using external vision, tooling bases, or preset transformation matrices.

[0003] The above solutions generally have the following problems in actual working conditions.

[0004] First, machining errors in the master and slave arms, reducer backlash, installation angle deviations, changes in the relative position of the base, and structural deformation after long-term operation can cause the original absolute zero-position relationship to drift. Once the zero position or installation reference deviates, a systematic error occurs in the master-slave mapping relationship, requiring manual recalibration or re-calibration using external measurement methods. In scenarios requiring frequent operation relocation, workstation changes, or rapid redeployment, each recalibration consumes a significant amount of time and manpower, resulting in low deployment efficiency and severely limiting the flexibility and practicality of the dual-arm teleoperation system.

[0005] Second, existing master-slave teleoperation links typically transmit master arm sampling data and slave arm control commands via network bus, serial bus, industrial Ethernet, or wireless links. Complex electromagnetic environments, network jitter, transient buffer backlogs, or communication errors can easily cause abnormal jumps in individual control frames. Without a mechanism to suppress these abnormal jumps, a single frame spike could be mistaken for a real operator action and transmitted to the slave arm, causing instantaneous impacts, jitter, or even collisions, posing a serious threat to equipment safety and operational accuracy. Some solutions employ low-pass filtering or signal smoothing algorithms to suppress anomalies, but these methods inevitably introduce response delays while filtering noise, sacrificing real-time tracking performance and failing to achieve a good balance between safety and sensitivity.

[0006] Third, some systems use a relative baseline instead of absolute zero for control to avoid frequent absolute zero calibration processes. However, without clearly defined baseline refresh conditions, abnormal factors such as communication glitches, no-load swaying, and sensor zero drift can all trigger erroneous baseline updates, causing the slave arm to gradually drift in attitude during periods of stillness when the operator has not applied any action. Simply relying on filters can only suppress high-frequency noise; it cannot logically block erroneous baseline update paths and cannot fundamentally solve the drift problem.

[0007] Fourth, existing technologies typically handle functions such as "calibration-free mapping," "anti-communication anomalies," and "human-computer interaction intent recognition" in a decentralized manner, lacking a unified methodology to collaboratively address these issues. The lack of a unified state coordination mechanism and signal interaction logic between functional modules makes it prone to mode switching conflicts or the circumvention of security constraints, hindering the formation of a complete and reliable security control loop at the system level.

[0008] Therefore, there is a need for a dual-arm robot master-slave teleoperation method that does not require absolute zero-point calibration, can simultaneously suppress communication glitches and baseline drift, and can collaboratively handle multiple issues, in order to meet the engineering requirements of rapid deployment, smooth following, and safe and reliable operation. Summary of the Invention

[0009] The technical problem to be solved by this invention is to provide a master-slave teleoperation method and operating system for a dual-arm robot. This invention does not require absolute zero-position calibration. It automatically generates pre-compensation parameters through synchronous micro-motion, updates the baseline posture in a controlled manner only when the operator releases the device, and performs saturation truncation on inter-frame command increments with the physical limits of the joints as the boundary. This integrates calibration-free deployment, baseline anti-drift, and communication glitch suppression into a unified safety control closed loop.

[0010] This invention provides a master-slave teleoperation method for a dual-arm robot, comprising the following steps: Step S1: Build a master-slave remote operating system. The system includes a master arm, a slave arm, a master controller, a slave controller, a communication link, a master arm status acquisition module for acquiring the status of the master arm, a slave arm status acquisition module for acquiring the status of the slave arm, and an operator intent detection module for detecting operator interaction signals. Step S2: During system initialization, control the main arm and the slave arm to perform synchronous micro-movements. Calculate the installation deviation and / or zero-position offset between the main arm and the slave arm based on the main arm status data and slave arm status data collected during the movement, and generate a pre-compensation parameter set. Step S3: After the system enters standby mode, the current master arm posture is collected as the master arm baseline posture, and the current slave arm posture is collected as the slave arm baseline posture and latched. Step S4: Obtain the detection value of the operator intention detection module in real time. When the detection value is greater than the trigger threshold, determine to enter the follow mode. When the detection value is less than the release threshold, determine to enter the standby mode. The release threshold is less than the trigger threshold. Step S5: In follow mode, calculate the offset of the current pose of the main arm relative to the baseline pose of the main arm in real time, and generate the target pose of the slave arm by combining the pre-compensation parameter set and the mapping relationship between the main arm and the slave arm. Step S6: For each joint of the slave arm, calculate the maximum allowable angle change based on the maximum allowable angular velocity of the joint, the current control cycle, and the preset safety factor; obtain the target increment of the slave arm target joint angle in the current frame relative to the actual joint angle of the slave arm in the previous frame; if the absolute value of the target increment exceeds the maximum allowable angle change, saturate and truncate the target increment to correct the slave arm target joint angle. Step S7: When the system switches from follow mode to standby mode, the current slave arm actual posture is updated to the new slave arm baseline posture, the current master arm actual posture is updated to the new master arm baseline posture and latched, and the baseline posture is prohibited from being refreshed until a valid operation intention is detected again.

[0011] Furthermore, in step S2, the synchronous micro-motions act on at least three non-collinear or non-coplanar joint degrees of freedom of the master arm and slave arm, and the amplitude of the motion is 0.5% to 5% of the working range of each target joint.

[0012] Furthermore, in step S2, the pre-compensation parameter set includes at least one of the following: joint offset vector, proportional gain matrix, and end-effector pose compensation transformation matrix.

[0013] Furthermore, in step S4, the operator intention detection module is a six-dimensional force / torque sensor installed on the main arm interaction end, and the detected value is the grip force; or, the operator intention detection module is a handle trigger switch or a pressure sensor; the trigger threshold and the release threshold are set to different values ​​to form a hysteresis band.

[0014] Furthermore, in step S7, the switch from follow mode to standby mode means that the detection value changes from being greater than the trigger threshold to being less than the release threshold, thereby determining that the operator has ended the effective operation; at this switching moment, the baseline attitude of the master arm and the baseline attitude of the slave arm are updated and locked in a unified manner.

[0015] Furthermore, in step S5, the mapping relationship between the master arm and the slave arm is as follows: If the master arm and slave arm are isomorphic, the target joint angle of the slave arm is directly calculated in the joint space. If the master arm and slave arm are heterogeneous dual arms, the pose increment of the master arm is first mapped to the pose increment of the slave arm end, and then the target joint angle of the slave arm is obtained through inverse kinematics.

[0016] Furthermore, in step S6, for the first... For each joint, the maximum allowable angular change is calculated using the following formula: ; in, For the j-th The maximum permissible angular velocity of each joint For the current control cycle, For safety factor and ; The target increment is calculated using the following formula: ; in, Target joint angles of the arm generated for the current frame The actual joint angle of the arm in the previous frame.

[0017] Furthermore, in step S6, when > At that time, the saturation cutoff is corrected for the target joint angle of the arm using the following formula: ; Otherwise .

[0018] Furthermore, the main controller has a built-in synchronous trajectory planner for generating a sequence of joint position commands for the synchronous micro-motion; the main controller sends the sequence of joint position commands to the joint actuators of the main arm and the slave controller simultaneously through the communication link; after receiving the command sequence, the slave controller controls the joint actuators of the slave arm to perform synchronous following, thereby ensuring that the main arm and the slave arm achieve time-synchronized micro-motions under the same control intention.

[0019] The present invention also provides a master-slave teleoperation system for a dual-arm robot, comprising: Main arm, slave arm, main controller, slave controller, communication link, main arm status acquisition module, slave arm status acquisition module, and operator intent detection module; The pre-compensation parameter generation module is used to control the main arm and the slave arm to perform synchronous micro-movements during system initialization, calculate the installation deviation and / or zero position offset between the main arm and the slave arm based on the main arm status data and slave arm status data collected during the movement, and generate a pre-compensation parameter set. The baseline attitude latching and updating module is used to acquire the current master arm attitude as the master arm baseline attitude and the current slave arm attitude as the slave arm baseline attitude and latch them after the system enters standby mode; and when the system switches from follow mode to standby mode, it updates the current slave arm actual attitude to the new slave arm baseline attitude and updates the current master arm actual attitude to the new master arm baseline attitude and latches them, and prohibits refreshing the baseline attitude until a valid operation intention is detected again; The mode switching control module is used to acquire the detection value of the operator intention detection module in real time. When the detection value is greater than the trigger threshold, it is determined to enter the follow mode. When the detection value is less than the release threshold, it is determined to enter the standby mode. The release threshold is less than the trigger threshold. The slave arm target pose generation module is used to calculate the offset of the current pose of the master arm relative to the baseline pose of the master arm in real time in follow mode, and generate the slave arm target pose by combining the pre-compensation parameter set and the mapping relationship between the master arm and the slave arm. The inter-frame velocity limiting module is used to calculate the maximum allowable angle change for each joint of the slave arm based on the maximum allowable angular velocity of the joint, the current control cycle, and a preset safety factor; obtain the target increment of the target joint angle of the slave arm in the current frame relative to the actual joint angle of the slave arm in the previous frame; if the absolute value of the target increment exceeds the maximum allowable angle change, the target increment is saturated and truncated to correct the target joint angle of the slave arm.

[0020] The beneficial effects of this invention are: 1. Eliminates absolute zero calibration, enabling rapid deployment upon startup.

[0021] Existing solutions typically require absolute zero-point calibration or homing of the main and slave arms before startup. If the mounting base of either the main or slave arm moves, is replaced, or undergoes structural deformation, the original calibration relationship becomes invalid, necessitating recalibration using external fixtures, resulting in low deployment efficiency. This invention controls the main and slave arms to perform synchronous micro-movements during the initialization phase, automatically identifying installation deviations and zero-point offsets and generating a pre-compensation parameter set. Without any external measuring equipment or manual alignment, it can quickly complete error compensation even after base position changes or equipment reinstallation, giving the system "deploy-ready" capability and significantly reducing the manpower and time costs of on-site maintenance and recalibration.

[0022] 2. Baseline latching and controlled updates fundamentally eliminate silent drift.

[0023] Existing solutions based on relative incremental control, if the reference baseline is continuously refreshed due to communication glitches, sensor zero drift, or idle fluctuations, can cause uncontrollable drift in the slave arm when no operator applies any action. Simply relying on signal smoothing methods such as low-pass filtering or Kalman filtering can only suppress high-frequency noise, but cannot logically block the baseline update path of erroneous data. This invention strictly limits the refresh of the baseline attitude to a single moment "switching from follow mode to standby mode," and latches the baseline in all other time periods. This cuts off the logical path for abnormal signals to propagate to the reference reference at the state machine level, eliminating cumulative drift during operational gaps or communication anomalies, and ensuring absolute stability of the slave arm when stationary.

[0024] 3. Based on the physical limits of the actuator, inter-frame limiting accurately identifies and suppresses communication glitches.

[0025] Existing technologies for handling communication glitches often rely on empirical thresholds or signal smoothing filters. The former lacks a physical basis for threshold setting; too lenient a threshold misses glitches, while too strict a threshold weakens the tracking sensitivity of normal operations. The latter introduces delay, affecting real-time performance. This invention directly constructs inter-frame physical limit thresholds based on the maximum permissible angular velocity, control cycle, and safety factor of each joint of the slave arm. These thresholds are entirely based on the actuator's actual driving capability. Any instruction increment exceeding the physical limit is identified as a communication glitch or abnormal jump and saturated and truncated. Normal, rapid operation instructions can pass without delay as long as they do not exceed the physical boundaries. This mechanism effectively protects the slave arm from impacts without introducing additional lag in the operator's intention response.

[0026] 4. Collaborative solutions to multiple problems form a unified security control closed loop.

[0027] Existing technologies typically handle issues such as "calibration-free mapping," "communication anomaly suppression," and "human-computer interaction intent recognition" in a fragmented manner, lacking a unified state coordination mechanism between modules, which easily leads to interaction logic conflicts. This invention organically integrates four mechanisms—installation deviation pre-compensation, baseline attitude latching, operator intent hysteresis discrimination, and inter-frame velocity limiting—into a single control architecture. This ensures that the system has clear safety constraints and state switching logic in each stage of initialization, following, and standby, forming a complete closed loop from calibration to execution, and from intent detection to instruction limiting, thereby improving the overall reliability and robustness of the system.

[0028] 5. Physical limiting and intention hysteresis work together to balance tracking sensitivity and operational safety.

[0029] This invention addresses two key issues. First, it creates a hysteresis band by differentiating the trigger and release thresholds in the operator intent detection module, eliminating boundary jitter during follow / standby mode switching. Second, it uses inter-frame speed limiting with the actuator's physical limits as a hard constraint, ensuring that even during mode switching or sudden communication anomalies, the slave arm's output commands remain within a safe and controllable range. The combined effect of these two mechanisms allows the slave arm to sensitively follow the operator's rapid, large-scale dragging of the master arm; and ensures a smooth, shock-free response even in the event of occasional communication glitches or accidental release by the operator, achieving a dynamic balance between sensitivity and safety.

[0030] 6. The unified approach is applicable to both isomorphic and heterogeneous two-arm systems.

[0031] In step S5, the master-slave mapping relationship of this invention can be flexibly adapted according to the system configuration: for homogeneous dual-arm systems, mapping is directly performed in joint space, resulting in high computational efficiency; for heterogeneous dual-arm systems, incremental mapping in pose space and inverse kinematics are used for solution. Combined with the unified correction of installation deviations using a pre-compensation parameter set, this method can be applied to both homogeneous and heterogeneous dual-arm robots without modifying the core control architecture, demonstrating good versatility and engineering applicability. Attached Figure Description

[0032] Figure 1 This is a flowchart illustrating the method of the present invention. Detailed Implementation

[0033] 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 a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0034] It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present invention are only used to explain the relative positional relationship and movement of each component in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indication will also change accordingly.

[0035] Furthermore, in this invention, descriptions involving "first," "second," etc., are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0036] In this invention, unless otherwise explicitly specified and limited, the terms "connection," "fixed," etc., should be interpreted broadly. For example, "fixed" can mean a fixed connection, a detachable connection, or an integral part; it can mean a mechanical connection, an electrical connection, a physical connection, or a wireless communication connection; it can mean a direct connection or an indirect connection through an intermediate medium; it can mean the internal communication of two elements or the interaction between two elements, unless otherwise explicitly limited. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0037] Furthermore, the technical solutions of the various embodiments of the present invention can be combined with each other, but only if they are feasible for those skilled in the art. If the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such combination of technical solutions does not exist and is not within the scope of protection claimed by the present invention.

[0038] like Figure 1 As shown, the present invention provides a master-slave teleoperation method for a dual-arm robot, comprising the following steps: Step S1: Establish a master-slave teleoperation system. The system includes a master arm, a slave arm, a master controller, a slave controller, a communication link, a master arm status acquisition module for acquiring the master arm's status, a slave arm status acquisition module for acquiring the slave arm's status, and an operator intent detection module for detecting operator interaction signals. Both the master arm and slave arm are multi-degree-of-freedom robotic arms, placed on the operating side and the work side respectively. The master controller and slave controller are connected via an industrial Ethernet network. The operator intent detection module is installed at the end handle of the master arm to sense the operator's gripping state. The master arm status acquisition module and the slave arm status acquisition module contain encoders for each joint.

[0039] Step S2: During system initialization, the main arm and the slave arm are controlled to perform synchronous micro-movements. Based on the main arm status data and slave arm status data collected during the movement, the installation deviation and / or zero-position offset between the main arm and the slave arm are calculated, and a pre-compensation parameter set is generated. This step is equivalent to power-on self-calibration. The operator does not need to manually align the main arm and the slave arm to a fixed zero position. The system automatically drives the two arms to complete a set of coordinated micro-movements and identifies the installation deviation and zero-position offset from the synchronously collected motion data, effectively shortening the deployment time.

[0040] Step S3: After the system enters standby mode, it collects the current main arm posture as the main arm baseline posture and collects the current slave arm posture as the slave arm baseline posture and locks it. At this time, the system records the current static posture of the two arms as the reference benchmark for relative motion and locks it. In subsequent operations, this benchmark will not change automatically due to signal noise or slight disturbances.

[0041] Step S4: The detection value of the operator intent detection module is acquired in real time. When the detection value is greater than the trigger threshold, the system determines to enter follow mode; when the detection value is less than the release threshold, the system determines to enter standby mode. The release threshold is less than the trigger threshold. When the operator grips the main arm handle, the detection value rises above the trigger threshold, and the system determines that the operator has begun active control and enters follow mode. After the operator releases the handle, the detection value drops below the release threshold, and the system returns to standby mode. Setting the release threshold below the trigger threshold creates a hysteresis interval, effectively avoiding frequent mode jumps in critical states.

[0042] Step S5: In follow mode, the offset of the current posture of the main arm relative to the baseline posture of the main arm is calculated in real time. Combined with the pre-compensation parameter set and the mapping relationship between the main arm and the slave arm, the target posture of the slave arm is generated. When the operator drags the main arm, the system calculates the change of the main arm posture relative to the baseline in real time, and uses the pre-compensation parameters to correct the installation deviation. After the master-slave mapping conversion, the target posture that the slave arm should reach is obtained to ensure the accuracy of following.

[0043] Step S6: For each joint of the slave arm, calculate the maximum allowable angle change based on the maximum permissible angular velocity of the joint, the current control cycle, and the preset safety factor; obtain the target increment of the target joint angle of the slave arm in the current frame relative to the actual joint angle of the slave arm in the previous frame; if the absolute value of the target increment exceeds the maximum allowable angle change, then saturate-truncate the target increment to correct the target joint angle of the slave arm; this step uses the physical capability limit of the slave arm joint as a safety boundary, and any instruction increment exceeding this boundary is considered an abnormal jump and is truncated and restricted, thereby effectively suppressing the impact of communication glitches on the slave arm, while not affecting the rapid response of normal operation instructions.

[0044] Step S7: When the system switches from follow mode to standby mode, the current slave arm's actual posture is updated to the new slave arm baseline posture, and the current master arm's actual posture is updated to the new master arm baseline posture and latched. Baseline posture updates are prohibited until a valid operational intent is detected again. After the operator releases the handle to end an operation, the system immediately captures the current postures of the master and slave arms, sets them as the new reference baseline, and latches them. The starting point of the next operation is the ending point of this operation, achieving continuous operation without cumulative drift. Baseline updates only occur when the operator explicitly stops control, fundamentally ensuring safety.

[0045] Compared with the prior art, the present invention has at least the following beneficial effects: 1. Eliminates absolute zero calibration, enabling rapid deployment upon startup.

[0046] Existing solutions typically require absolute zero-point calibration or homing of the main and slave arms before startup. If the mounting base of either the main or slave arm moves, is replaced, or undergoes structural deformation, the original calibration relationship becomes invalid, necessitating recalibration using external fixtures, resulting in low deployment efficiency. This invention controls the main and slave arms to perform synchronous micro-movements during the initialization phase, automatically identifying installation deviations and zero-point offsets and generating a pre-compensation parameter set. Without any external measuring equipment or manual alignment, it can quickly complete error compensation even after base position changes or equipment reinstallation, giving the system "deploy-ready" capability and significantly reducing the manpower and time costs of on-site maintenance and recalibration.

[0047] 2. Baseline latching and controlled updates fundamentally eliminate silent drift.

[0048] Existing solutions based on relative incremental control, if the reference baseline is continuously refreshed due to communication glitches, sensor zero drift, or idle fluctuations, can cause uncontrollable drift in the slave arm when no operator applies any action. Simply relying on signal smoothing methods such as low-pass filtering or Kalman filtering can only suppress high-frequency noise, but cannot logically block the baseline update path of erroneous data. This invention strictly limits the refresh of the baseline attitude to a single moment "switching from follow mode to standby mode," and latches the baseline in all other time periods. This cuts off the logical path for abnormal signals to propagate to the reference reference at the state machine level, eliminating cumulative drift during operational gaps or communication anomalies, and ensuring absolute stability of the slave arm when stationary.

[0049] 3. Based on the physical limits of the actuator, inter-frame limiting accurately identifies and suppresses communication glitches.

[0050] Existing technologies for handling communication glitches often rely on empirical thresholds or signal smoothing filters. The former lacks a physical basis for threshold setting; too lenient a threshold misses glitches, while too strict a threshold weakens the tracking sensitivity of normal operations. The latter introduces delay, affecting real-time performance. This invention directly constructs inter-frame physical limit thresholds based on the maximum permissible angular velocity, control cycle, and safety factor of each joint of the slave arm. These thresholds are entirely based on the actuator's actual driving capability. Any instruction increment exceeding the physical limit is identified as a communication glitch or abnormal jump and saturated and truncated. Normal, rapid operation instructions can pass without delay as long as they do not exceed the physical boundaries. This mechanism effectively protects the slave arm from impacts without introducing additional lag in the operator's intention response.

[0051] 4. Collaborative solutions to multiple problems form a unified security control closed loop.

[0052] Existing technologies typically handle issues such as "calibration-free mapping," "communication anomaly suppression," and "human-computer interaction intent recognition" in a fragmented manner, lacking a unified state coordination mechanism between modules, which easily leads to interaction logic conflicts. This invention organically integrates four mechanisms—installation deviation pre-compensation, baseline attitude latching, operator intent hysteresis discrimination, and inter-frame velocity limiting—into a single control architecture. This ensures that the system has clear safety constraints and state switching logic in each stage of initialization, following, and standby, forming a complete closed loop from calibration to execution, and from intent detection to instruction limiting, thereby improving the overall reliability and robustness of the system.

[0053] 5. Physical limiting and intention hysteresis work together to balance tracking sensitivity and operational safety.

[0054] This invention addresses two key issues. First, it creates a hysteresis band by differentiating the trigger and release thresholds in the operator intent detection module, eliminating boundary jitter during follow / standby mode switching. Second, it uses inter-frame speed limiting with the actuator's physical limits as a hard constraint, ensuring that even during mode switching or sudden communication anomalies, the slave arm's output commands remain within a safe and controllable range. The combined effect of these two mechanisms allows the slave arm to sensitively follow the operator's rapid, large-scale dragging of the master arm; and ensures a smooth, shock-free response even in the event of occasional communication glitches or accidental release by the operator, achieving a dynamic balance between sensitivity and safety.

[0055] 6. The unified approach is applicable to both isomorphic and heterogeneous two-arm systems.

[0056] In step S5, the master-slave mapping relationship of this invention can be flexibly adapted according to the system configuration: for homogeneous dual-arm systems, mapping is directly performed in joint space, resulting in high computational efficiency; for heterogeneous dual-arm systems, incremental mapping in pose space and inverse kinematics are used for solution. Combined with the unified correction of installation deviations using a pre-compensation parameter set, this method can be applied to both homogeneous and heterogeneous dual-arm robots without modifying the core control architecture, demonstrating good versatility and engineering applicability.

[0057] In one embodiment, in step S2, the synchronized micro-motion acts on at least three non-collinear or non-coplanar joint degrees of freedom of the master arm and slave arm, and the motion amplitude is 0.5% to 5% of the working range of each target joint. This setting ensures that effective data containing multi-dimensional spatial information can be collected within a sufficiently small motion range, without triggering collision risks or significantly prolonging the initialization time. The limitation of at least three non-collinear or non-coplanar joint degrees of freedom ensures that the motion trajectories of the master arm and slave arm have sufficient excitation dimensions in space, thereby accurately identifying the components in the installation deviation and zero-position offset, and avoiding singularity or coupling errors in the solution of the pre-compensation parameter set due to the overly simple motion trajectory.

[0058] In one embodiment, in step S2, the pre-compensation parameter set includes at least one of a joint offset vector, a proportional gain matrix, and an end-effector pose compensation transformation matrix. The joint offset vector is used to compensate for the angular zero-position deviation between the master arm and slave arm at corresponding joints, and is the most direct error correction in a isomorphic dual-arm system. The proportional gain matrix is ​​used to compensate for the motion scaling relationship caused by the difference in the length of the master and slave arm links. The end-effector pose compensation transformation matrix is ​​used to compensate for the installation pose deviation between the master arm base and the slave arm base, and includes position offset and attitude rotation information. Depending on the actual structural relationship between the master arm and slave arm, one or more combinations of these parameters can be flexibly selected to achieve error modeling at different accuracy levels, from simple joint offset compensation to full pose space compensation.

[0059] In one embodiment, in step S4, the operator intent detection module is a six-dimensional force / torque sensor installed on the main arm interaction end, and the detected value is grip force; or, the operator intent detection module is a handle trigger switch or a pressure sensor; the trigger threshold and the release threshold are set to different values ​​to form a hysteresis band. When using a six-dimensional force / torque sensor to detect grip force, no additional button action is required from the operator; gripping enters the follow mode and releasing exits, resulting in a natural and smooth operating experience. When using a handle trigger switch or pressure sensor, clear intent indications can be directly obtained through physical buttons or pressure signals, suitable for occasions where cost or signal simplicity is required. The hysteresis band formed by the trigger threshold being higher than the release threshold can prevent the system from frequently switching between follow mode and standby mode due to slight hand tremors or fluctuations in force near the critical grip force, effectively improving the stability of mode switching and the fault tolerance of operation.

[0060] In one embodiment, in step S7, the switch from follow mode to standby mode refers to the detection value changing from being greater than the trigger threshold to being less than the release threshold, thereby determining that the operator has ended the effective operation. At this switching moment, the baseline attitude of the main arm and the baseline attitude of the slave arm are updated and latched simultaneously. This mechanism strictly limits the timing of the baseline attitude refresh to the instant the operator actively releases the main arm, that is, the moment when an effective operation is completely completed, rather than continuously refreshing during the follow process or standby period. By using the exit condition that the detection value must cross the hysteresis band between the trigger threshold and the release threshold, it can be ensured that the operator has indeed clearly ended the operation and is not temporarily leaving the threshold boundary due to instantaneous force fluctuations. The unified update and immediate latching of the baseline attitude of the main arm and the slave arm at the switching moment ensures that the starting point of the relative offset calculation for the next operation is seamlessly connected with the ending point of the current operation, realizing the transmission of errors without cumulative error in continuous segmented operations. At the same time, it logically eliminates the possibility of communication glitches or sensor noise mistakenly refreshing the baseline during idle periods, thereby causing silent drift of the slave arm.

[0061] In one embodiment, in step S5, the mapping relationship between the master arm and the slave arm is as follows: If the master arm and slave arm are isomorphic, the target joint angle of the slave arm is directly calculated in the joint space. If the master arm and slave arm are heterogeneous dual arms, the master arm pose increment is first mapped to the slave arm end-effector pose increment, and then the target joint angle of the slave arm is obtained through inverse kinematics. For isomorphic dual-arm mapping, the joint layout and link length of the master and slave arms are consistent or in a fixed ratio. Pre-compensation can be directly applied to the offset of the master arm baseline in joint space and superimposed onto the slave arm baseline. The calculation process is a joint-by-joint linear operation, with high real-time performance, suitable for scenarios where the master and slave arm structures are completely identical. For heterogeneous dual-arm mapping, the master and slave arms have different configurations, sizes, or degree-of-freedom allocations. The master arm pose increment needs to be converted to the slave arm end-effector pose increment in Cartesian space first. During the conversion process, position scaling, coordinate axis mapping, and attitude constraints can be introduced. Then, the target angle of each joint is iteratively solved through the inverse kinematics model of the slave arm, suitable for scenarios where the master and slave arm structures differ significantly. The two mapping modes can be switched as needed within the same control framework, giving this method good system adaptability.

[0062] In one embodiment, in step S6, for the first... For each joint, the maximum allowable angular change is calculated using the following formula: ; in, For the first The maximum permissible angular velocity of each joint For the current control cycle, For safety factor and ; The target increment is calculated using the following formula: ; in, Target joint angles of the arm generated for the current frame The actual joint angle of the arm in the previous frame.

[0063] This calculation method has the following advantages: First, the limiting threshold... Directly from the maximum permissible angular velocity of the arm joint actuator As a calculation benchmark, this parameter is an inherent physical property of the actuator and can be directly obtained from the driver specification. It eliminates the need for trial and error or historical data statistics to determine the limiting boundary, making the objectivity and reliability of the threshold setting significantly superior to existing technologies that rely on manually set empirical thresholds. Secondly, the target increment... The method, obtained by subtracting the original target of the current frame from the actual execution result of the previous frame, can accurately capture abnormal jumps in single-frame instructions caused by communication errors, network jitter, or buffer backlog. The identification is direct, computationally efficient, and requires no complex statistical analysis or frequency domain transformation of historical data sequences. Furthermore, it offers a high level of security. The introduction of this feature allows the system to be flexibly adjusted between sensitivity and security: when the communication environment is good, it can be optimized. To maximize joint performance, a value close to 1 is preferable; a smaller value can be used in scenarios with strong electromagnetic interference or delicate operation. The value is increased to enhance the glitch suppression margin. This configurability makes the method significantly more adaptable to different operating conditions than the fixed threshold limiting scheme. In addition, since the limiting judgment is based on the physical limit of the actuator itself, any instruction increment exceeding the limit is judged as abnormal and truncated, while normal fast operation instructions can pass through without delay as long as they do not break the physical boundary. This achieves the effect of "resolutely intercepting false instructions and allowing true instructions to pass without hindrance", overcoming the inherent defect of traditional low-pass filtering or signal smoothing methods that inevitably introduce response delay while suppressing noise.

[0064] In one embodiment, in step S6, when > At that time, the saturation cutoff is corrected for the target joint angle of the arm using the following formula: ; Otherwise .

[0065] This saturation truncation correction method starts from the actual position of the arm in the previous frame and moves along the direction of the original target increment by the maximum physically permissible step. This ensures that the corrected command neither exceeds the instantaneous execution capability of the joint actuator nor fails to fully preserve the operator's original directional intention. Compared to methods that directly discard abnormal frames or keep the previous frame's command unchanged, this method does not cause motion interruption or loss of direction. The operator only perceives the arm continuously approaching the target at its maximum speed, resulting in a smooth and natural control experience. Compared to methods that smooth abnormal frames using low-pass filtering, this method does not introduce additional delays to normal, rapid operation commands because the target increment that does not exceed the limit will be sent as is, without any truncation. The use of this function allows for the preservation of direction information with only one sign determination, resulting in extremely low computational overhead and making it suitable for real-time execution within millisecond-level control cycles. Furthermore, this saturation truncation mechanism, combined with the physical limit-based calculations in the previous step... Together, they form a complete pipeline for identifying genuine and fake commands and mapping safety constraints: any target increment that exceeds the joint's capability boundary is identified as an untrusted command and remapped to the capability boundary; any target increment that does not exceed the boundary is identified as a trustworthy command and directly issued for execution, thereby ensuring the absolute safety of the slave arm while maximizing the teleoperation's following response performance.

[0066] In one embodiment, the main controller incorporates a synchronous trajectory planner to generate a sequence of joint position commands for the synchronous micro-motion. The main controller simultaneously sends this sequence of commands to the joint actuators of the main arm and the slave controller via the communication link. Upon receiving the command sequence, the slave controller controls the joint actuators of the slave arm to perform synchronous following, thereby ensuring that the main arm and slave arm achieve time-synchronized micro-motions under the same control intent. This control architecture centralizes the trajectory planning authority for synchronous micro-motions entirely in the main controller, which acts as the sole command source, simultaneously distributing the same command sequence to both the main arm and slave arm via the same communication link. The slave controller passively executes only the received commands, thus eliminating, at the system level, potential timing deviations between the main arm controller and slave arm controller caused by local clock differences, communication delay jitter, or differences in independent planning logic. Because the trajectory commands received by the master and slave arms are completely consistent in terms of timestamps and motion profiles, the movements executed by both within the same control cycle are strictly synchronized in the time dimension. The collected master and slave arm status data can form a precise time correspondence, providing a high-quality data foundation for the accurate identification of subsequent installation deviations and zero-position offsets. This solution fully utilizes the existing communication links and control hardware in the master-slave remote operating system, eliminating the need for additional hardware synchronization trigger lines or dedicated clock synchronization protocols. It is low-cost, simple to deploy, and has no additional requirements for the bandwidth and real-time performance of the communication link, making it easy to implement directly in existing industrial Ethernet or fieldbus architectures.

[0067] The present invention also provides a master-slave teleoperation system for a dual-arm robot, comprising: Main arm, slave arm, main controller, slave controller, communication link, main arm status acquisition module, slave arm status acquisition module, and operator intent detection module; The pre-compensation parameter generation module is used to control the main arm and the slave arm to perform synchronous micro-movements during system initialization. Based on the main arm status data and slave arm status data collected during the movement, it calculates the installation deviation and / or zero-position offset between the main arm and the slave arm, and generates a pre-compensation parameter set. This module enables the system to automatically complete error identification and compensation parameter generation by automatically performing a set of micro-synchronous movements after each power-on or redeployment, without the need for manual zeroing or absolute zero-position calibration with the help of external measuring equipment. This reduces deployment preparation time from minutes to seconds, significantly improving the system's practicality and deployment efficiency in scenarios such as mobile operations and frequent site changes.

[0068] The baseline attitude latching and updating module is used to acquire and latch the current master arm attitude as the master arm baseline attitude and the current slave arm attitude as the slave arm baseline attitude after the system enters standby mode. Furthermore, when the system switches from follow mode to standby mode, it updates the current slave arm actual attitude to the new slave arm baseline attitude and the current master arm actual attitude to the new master arm baseline attitude, and latches them. Baseline attitude updates are prohibited until a valid operational intent is detected again. This module strictly limits the timing of baseline attitude updates to the instant the operator explicitly ends an operation and the system completes the switch from follow to standby mode; the baseline remains latched and protected during all other time periods. Compared to existing solutions that continuously refresh the reference point or rely on filtering to suppress drift, this module, from a state management perspective, cuts off the logical path for abnormal signals such as communication glitches, sensor noise, or environmental vibrations to contaminate the reference benchmark during idle periods, achieving a fundamental elimination of silent drift.

[0069] The mode switching control module is used to acquire the detection value of the operator intention detection module in real time. When the detection value is greater than the trigger threshold, it determines to enter the follow mode; when the detection value is less than the release threshold, it determines to enter the standby mode. The release threshold is less than the trigger threshold. This module constructs a hysteresis interval by setting unequal trigger and release thresholds, which effectively avoids high-frequency oscillation switching between follow and standby modes caused by fluctuations in the operator's grip strength around a single threshold. This makes the system state transition smoother and more stable, and also provides a clear and reliable state switching trigger signal for the baseline attitude latching and update module.

[0070] The slave arm target pose generation module is used to calculate the offset of the main arm's current pose relative to the main arm's baseline pose in real time in follow mode. Combined with the pre-compensation parameter set and the mapping relationship between the main arm and the slave arm, the slave arm target pose is generated. This module embeds the pre-compensation parameter set into the target pose generation link, so that the compensation for installation deviation and zero position offset is integrated with the calculation of relative offset into a unified process, without the need for additional error correction in independent links, thus ensuring the real-time performance and accuracy of target pose generation.

[0071] The inter-frame velocity limiting module calculates the maximum permissible angle change for each joint of the slave arm based on the joint's maximum permissible angular velocity, the current control cycle, and a preset safety factor. It obtains the target increment of the slave arm's target joint angle in the current frame relative to the actual joint angle in the previous frame. If the absolute value of the target increment exceeds the maximum permissible angle change, the target increment is saturated and truncated to correct the slave arm's target joint angle. This module constructs a dynamic limiting boundary based on the inherent maximum permissible angular velocity of each joint actuator in the slave arm. It can instantly identify and truncate command jumps caused by communication errors or network anomalies within a single control cycle. Compared to existing solutions that rely on signal filtering or empirical thresholds, it has the advantages of clear physical judgment criteria and no additional delay, serving as a hardware capability constraint layer to ensure the safe and smooth movement of the slave arm.

[0072] The aforementioned modules work collaboratively under a unified control architecture: the pre-compensation parameter generation module eliminates deployment deviations; the baseline attitude latching and updating module provides a safe and reliable relative motion reference; the mode switching control module accurately identifies the operator's intent and triggers state transitions; the arm target attitude generation module fuses the relative offset with compensation parameters to generate the desired command; and the inter-frame velocity limiting module acts as the final safety barrier, ensuring that all issued commands do not exceed the physical capability boundaries of the actuator. This modular system architecture integrates calibration-free deployment, anti-communication interference, intent discrimination, and safe following functions into a single closed-loop circuit. The modules coordinate and link through clear interfaces and status signals. Compared to existing technologies that handle each problem separately, this architecture offers significant advantages in system integration, logical consistency, and overall reliability.

[0073] This embodiment provides a master-slave teleoperation method for a dual-arm robot applied in remote precision assembly scenarios. The system includes a master arm as the teaching end, a slave arm as the working end, a master controller, a slave controller, an industrial Ethernet communication link, master arm joint encoders, slave arm joint encoders, and a six-dimensional force / torque sensor installed in the interactive handle at the end of the master arm. In this embodiment, the master arm is a lightweight 6-DOF collaborative robotic arm placed on an operating table, which is dragged and taught by the operator by directly holding its end handle; the slave arm is a 6-DOF industrial robot installed at the assembly station, with a precision gripper at its end. The operator intent detection module is implemented by the six-dimensional force / torque sensor, which determines the operation intent by sensing the magnitude of the operator's gripping force on the master arm handle.

[0074] After the system powers on, it first enters the initialization phase. The synchronous trajectory planner built into the main controller generates a sequence of synchronous micro-motion commands for the shoulder, elbow, and wrist joints. These synchronous micro-motions act on at least three non-collinear joint degrees of freedom of the master arm and slave arm, with the amplitude of each joint motion set to 2% of its working range. This ensures that while fully exciting the multidimensional error components of space, the motion amplitude is far less than the safe collision boundary. The main controller simultaneously sends the command sequence to the master arm joint actuator and the slave controller via the communication link. The slave controller controls the slave arm joint actuator to execute synchronously, ensuring that both arms achieve time-synchronized micro-motions under the same control intent. During the motion, the master arm state acquisition module and the slave arm state acquisition module synchronously record their respective actual joint angle data. Since the master arm and slave arm receive the same command sequence, their movements are strictly synchronized in time, and the acquired data has a precise time correspondence. The pre-compensation parameter generation module calculates the joint offset vector between the master arm and the slave arm based on the acquired data. and end pose compensation transformation matrix Together they constitute the pre-compensation parameter set The joint offset vector is used to compensate for the zero-position deviation of the corresponding joint angle, and the end-effector pose compensation transformation matrix is ​​used to compensate for the spatial pose deviation between the main arm base and the slave arm base due to different installation positions.

[0075] After initialization, the system automatically enters standby mode. The baseline attitude latching and update module acquires the current main arm joint angles. As the baseline posture of the main arm, the current arm joint angle is collected. The slave arm is then positioned as the baseline attitude, and both are latched into a register. At this point, the slave arm remains stationary, awaiting operator intervention.

[0076] When the operator grips the handle at the end of the main arm, a six-dimensional force / torque sensor installed inside the handle detects the gripping force in real time. The mode switching control module compares the detected value with a preset trigger threshold. and release threshold A comparison is performed. When the operator grips the handle tightly, Rise above When the system determines that the operator has issued a valid operational intention, it enters follow mode. The threshold between the trigger threshold and the release threshold... The difference forms a hysteresis band. Even if the operator experiences a brief fluctuation in grip strength due to slight hand tremors during operation, as long as it does not fall below a certain level... Once the release threshold is reached, the system will stably maintain the follow mode and will not experience frequent state transitions.

[0077] In follow mode, the operator drags the end effector of the main arm to perform assembly actions. Taking the "downward insertion" action of the main arm being dragged as an example: the main arm status acquisition module acquires the current joint angle of the main arm in real time. The joint offset of the main arm relative to the baseline posture is calculated from the arm target posture generation module. And generate the original target joint angle of the arm by combining the pre-compensation parameter set: ; in The gain matrix is ​​a proportional matrix. In an isomorphic configuration where the main arm and slave arm have the same dimensions, an identity matrix can be used. In an isomorphic configuration where the main and slave arms have proportional dimensions, a diagonal matrix can be used to achieve motion scaling.

[0078] The inter-frame rate limiting module applies safety constraints to the generated raw target commands. Let the control period be... The maximum permissible angular velocity from the j-th joint of the arm Safety factor The maximum allowable angle change per frame. Read the previous frame from the actual joint angle of the arm. Then, calculate the current target increment. If a typical single-frame interference occurs in the communication link at this time, it will lead to... An abnormal jump to a value far exceeding the normal range makes... The inter-frame rate limiting module immediately performs saturation truncation: ; The truncated command only allows the arm to move 0.02 rad in the original operating direction within the current frame. This is the maximum physical rotation that the joint can achieve within a 10ms control cycle, effectively suppressing abnormal jumps. For normal, rapid dragging actions by the operator, as long as the target increment in a single frame does not exceed the physical boundary of 0.02 rad, the original target command will pass through as is without any truncation or smoothing, ensuring real-time tracking performance.

[0079] The controller sends the revised safety target command to the slave arm joint actuator, which then drives the gripper to smoothly perform the assembly action. The operator, through force feedback from the main arm or a visual monitoring screen, senses that the slave arm has accurately completed the insertion and assembly, and then releases the main arm handle. The six-dimensional force / torque sensor detects... Descending to If the release threshold is crossed, the mode switching control module determines that the operator has ended the current valid operation, and the system switches from follow mode to standby mode.

[0080] At the same moment of switching, the baseline attitude latching and updating module immediately performs a baseline update: reading the current actual joint angle of the arm as the new baseline. Read the current actual joint angle of the main arm as the new And then re-locked. The end point of this assembly action automatically becomes the starting point of the next assembly action. When the operator performs multi-station assembly continuously, the reference datum between each operation is seamlessly connected, and no cumulative error occurs. When the next gripping action occurs... Before the trigger threshold is exceeded again, the baseline attitude is in a latched protection state, which prohibits any automatic refresh. Residual noise on the communication link or weak vibration of the master arm on the table cannot contaminate the reference reference, and the slave arm remains absolutely still.

[0081] By repeating the above steps, the operator can continuously perform remote control operations on multiple assembly stations without having to perform any manual calibration or zeroing operations when switching stations.

[0082] The above embodiments achieve the following effects: First, the architecture of centralizing the control of synchronous micro-motion in the master controller and synchronously distributing instructions to the master and slave arms ensures the time consistency of calibration data acquisition and provides high-quality data for pre-compensation parameter identification; Second, the pre-compensation parameter set includes both joint offset vectors and end-effector pose compensation transformation matrices, which can completely compensate for error sources at two levels: joint zero-position deviation and base installation deviation; Third, the six-dimensional force / torque sensor, combined with the hysteresis threshold intention discrimination mechanism, allows the operator to control the start and stop of following simply by gripping and releasing the hand, making the operation intuitive; Finally, the inter-frame velocity limiting module constructs a limiting boundary of 0.02 rad using the slave arm joint's inherent physical parameter of 2.5 rad / s, instantly intercepting communication glitches within a 10ms control cycle, allowing normal operation to pass with zero delay, achieving the optimal balance between safety and following performance. Throughout the assembly process, the arm moves smoothly without impact, the assembly accuracy meets the process requirements, and the system can complete autonomous calibration and enter the work-ready state within seconds after power-on. The deployment efficiency is significantly better than the traditional master-slave remote operation scheme that requires manual zeroing.

[0083] The above description is merely an embodiment and does not constitute any limitation on the present invention. Any person skilled in the art can make many possible variations, modifications, or alterations to the technical solutions of the present invention without departing from the scope of the present invention. Therefore, any simple modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention, without departing from the scope of the present invention, should fall within the protection scope of the present invention.

Claims

1. A master-slave teleoperation method for a dual-arm robot, characterized in that, Includes the following steps: Step S1: Build a master-slave remote operating system. The system includes a master arm, a slave arm, a master controller, a slave controller, a communication link, a master arm status acquisition module for acquiring the status of the master arm, a slave arm status acquisition module for acquiring the status of the slave arm, and an operator intent detection module for detecting operator interaction signals. Step S2: During system initialization, control the main arm and the slave arm to perform synchronous micro-movements. Calculate the installation deviation and / or zero-position offset between the main arm and the slave arm based on the main arm status data and slave arm status data collected during the movement, and generate a pre-compensation parameter set. Step S3: After the system enters standby mode, the current master arm posture is collected as the master arm baseline posture, and the current slave arm posture is collected as the slave arm baseline posture and latched. Step S4: Obtain the detection value of the operator intention detection module in real time. When the detection value is greater than the trigger threshold, determine to enter the follow mode. When the detection value is less than the release threshold, determine to enter the standby mode. The release threshold is less than the trigger threshold. Step S5: In follow mode, calculate the offset of the current pose of the main arm relative to the baseline pose of the main arm in real time, and generate the target pose of the slave arm by combining the pre-compensation parameter set and the mapping relationship between the main arm and the slave arm. Step S6: For each joint of the slave arm, calculate the maximum allowable angle change based on the maximum allowable angular velocity of the joint, the current control cycle, and the preset safety factor; obtain the target increment of the slave arm target joint angle in the current frame relative to the actual joint angle of the slave arm in the previous frame; if the absolute value of the target increment exceeds the maximum allowable angle change, saturate and truncate the target increment to correct the slave arm target joint angle. Step S7: When the system switches from follow mode to standby mode, the current slave arm actual posture is updated to the new slave arm baseline posture, the current master arm actual posture is updated to the new master arm baseline posture and latched, and the baseline posture is prohibited from being refreshed until a valid operation intention is detected again.

2. The master-slave teleoperation method for a dual-arm robot according to claim 1, characterized in that, In step S2, the synchronous micro-motions act on at least three non-collinear or non-coplanar joint degrees of freedom of the master arm and slave arm, and the amplitude of the motion is 0.5% to 5% of the working range of each target joint.

3. The master-slave teleoperation method for a dual-arm robot according to claim 1, characterized in that, In step S2, the pre-compensation parameter set includes at least one of the following: joint offset vector, proportional gain matrix, and end-effector pose compensation transformation matrix.

4. The master-slave teleoperation method for a dual-arm robot according to claim 1, characterized in that, In step S4, the operator intention detection module is a six-dimensional force / torque sensor installed on the main arm interaction end, and the detection value is the grip force; or, the operator intention detection module is a handle trigger switch or a pressure sensor; the trigger threshold and the release threshold are set to different values ​​to form a hysteresis band.

5. The master-slave teleoperation method for a dual-arm robot according to claim 1, characterized in that, In step S7, the switch from follow mode to standby mode means that the detection value changes from being greater than the trigger threshold to being less than the release threshold, thereby determining that the operator has ended the effective operation; at this switching moment, the master arm baseline attitude and slave arm baseline attitude are updated and locked in a unified manner.

6. The master-slave teleoperation method for a dual-arm robot according to claim 1, characterized in that, In step S5, the mapping relationship between the master arm and the slave arm is as follows: If the master arm and slave arm are isomorphic, the target joint angle of the slave arm is directly calculated in the joint space. If the master arm and slave arm are heterogeneous dual arms, the pose increment of the master arm is first mapped to the pose increment of the slave arm end, and then the target joint angle of the slave arm is obtained through inverse kinematics.

7. The master-slave teleoperation method for a dual-arm robot according to claim 1, characterized in that, In step S6, for the j-th... For each joint, the maximum allowable angular change is calculated using the following formula: in, For the jth The maximum permissible angular velocity of each joint For the current control cycle, For safety factor and ; The target increment is calculated using the following formula: in, Target joint angles of the arm generated for the current frame The actual joint angle of the arm in the previous frame.

8. The master-slave teleoperation method for a dual-arm robot according to claim 7, characterized in that, In step S6, when > At that time, the saturation cutoff is corrected for the target joint angle of the arm using the following formula: Otherwise .

9. The master-slave teleoperation method for a dual-arm robot according to claim 1, characterized in that, The main controller has a built-in synchronous trajectory planner for generating a sequence of joint position commands for the synchronous micro-motions; the main controller sends the sequence of joint position commands to the joint actuators of the main arm and the slave controller simultaneously through the communication link. After receiving the instruction sequence, the slave controller controls the joint actuator of the slave arm to perform synchronous following, thereby ensuring that the master arm and the slave arm achieve time-synchronized micro-movements under the same control intention.

10. A master-slave teleoperation system for a dual-arm robot, characterized in that, include: Main arm, slave arm, main controller, slave controller, communication link, main arm status acquisition module, slave arm status acquisition module, and operator intent detection module; The pre-compensation parameter generation module is used to control the main arm and the slave arm to perform synchronous micro-movements during system initialization, calculate the installation deviation and / or zero position offset between the main arm and the slave arm based on the main arm status data and slave arm status data collected during the movement, and generate a pre-compensation parameter set. The baseline attitude latching and updating module is used to acquire the current master arm attitude as the master arm baseline attitude and the current slave arm attitude as the slave arm baseline attitude and latch them after the system enters standby mode; and when the system switches from follow mode to standby mode, it updates the current slave arm actual attitude to the new slave arm baseline attitude and updates the current master arm actual attitude to the new master arm baseline attitude and latches them, and prohibits refreshing the baseline attitude until a valid operation intention is detected again; The mode switching control module is used to acquire the detection value of the operator intention detection module in real time. When the detection value is greater than the trigger threshold, it is determined to enter the follow mode. When the detection value is less than the release threshold, it is determined to enter the standby mode. The release threshold is less than the trigger threshold. The slave arm target pose generation module is used to calculate the offset of the current pose of the master arm relative to the baseline pose of the master arm in real time in follow mode, and generate the slave arm target pose by combining the pre-compensation parameter set and the mapping relationship between the master arm and the slave arm. The inter-frame velocity limiting module is used to calculate the maximum allowable angle change for each joint of the slave arm based on the maximum allowable angular velocity of the joint, the current control cycle, and a preset safety factor; obtain the target increment of the target joint angle of the slave arm in the current frame relative to the actual joint angle of the slave arm in the previous frame; if the absolute value of the target increment exceeds the maximum allowable angle change, the target increment is saturated and truncated to correct the target joint angle of the slave arm.

Images