Precise positioning assisted load adaptive force balance control system for robot arm

Abstract

The present application relates to a kind of precision positioning assisted mechanical arm's load adaptive force balance control system.A kind of precision positioning assisted mechanical arm's load adaptive force balance control system, including main control unit and detection module, the control method of the main control unit includes the following steps: S1.By detection module acquisition empty load reference sensing data;S2.Workpiece is hung or clamped in end execution component, and it is positioned by dragging;S3.Detection module synchronous acquisition linear telescopic drive piece's real-time displacement signal, actual axial force signal and end real-time load force signal;S4.Develop torque-compensation force solution;S5.Adopt the composite control strategy of feedforward precontrol superimposed force feedback closed-loop correction, output corresponding balance compensation force;S6.End load and pitch attitude change are monitored in real time;S7.Accomplish workpiece precision alignment.The present application can realize load adaptive dynamic gravity balance, with the advantage of being convenient for realizing high-precision precision alignment.

Description

Technical Field

[0001] This invention relates to a load-adaptive force balance control system for a precision positioning-assisted robotic arm. Background Technology

[0002] In on-site operations such as parts assembly and equipment maintenance, the workpieces are quite heavy. Relying entirely on manual lifting and positioning for extended periods can easily lead to muscle strain for operators. Furthermore, manual operation poses safety hazards such as large positioning deviations and the risk of workpieces falling and impacting the workpiece. To reduce the intensity of manual labor, the industry is gradually promoting various types of power-assisted robotic arm equipment.

[0003] Traditional counterweight, spring, and gas spring-assisted robotic arms only have simple lifting and stopping switches and lack adaptive force control logic. When the workpiece weight changes or the robotic arm's pitch posture changes, the system cannot dynamically adjust the output compensation force, resulting in a continuous mismatch between the balancing torque and the workpiece's weight torque. Each time the workpiece is changed or the working height is adjusted, the counterweight and spring preload must be manually adjusted on-site, making the debugging process cumbersome, the load adaptability range narrow, and the dragging process exhibits obvious sags and jerks, significantly increasing the intensity of manual operation.

[0004] Existing electric servo-assisted robotic arms generally adopt position / speed single-loop control, relying on the servo motor winding current to roughly calculate the theoretical output thrust. When there is friction in the lead screw transmission, transmission backlash, preload deformation of the mechanism, or instantaneous load impact of the workpiece, there will be a large deviation between the motor current and the actual axial thrust of the servo cylinder. The control method lacks a real-time force correction link, the theoretical compensation force cannot match the actual output force, the gravity balance accuracy is poor, and the dragging process is severely shaking, which cannot meet the requirements of smooth operation in precision assembly. Summary of the Invention

[0005] The purpose of this invention is to provide a load adaptive force balance control system for a precision positioning assistive robotic arm that can provide dynamic balance compensation force and adapt to high-precision operation.

[0006] To achieve the above objectives, this invention employs a load-adaptive force balance control system for a precision positioning-assisted robotic arm, comprising a base and a pitch support arm. One end of the pitch support arm is connected to the base for vertical swinging and forms a rotation axis. An end effector is provided at the end of the pitch support arm. The pitch support arm is equipped with a linear telescopic drive component for swinging around the rotation axis. The system also includes a main control unit and a detection module. The control method of the main control unit includes the following steps: S1. Collect no-load reference sensor data at the end of the pitch support arm under no-load conditions through the detection module; S2. The workpiece is mounted or clamped on the end effector component, and the main control unit, in conjunction with the linear telescopic drive component, drives the pitch support arm to complete a large-range drag positioning. S3. During the towing and positioning process, the detection module synchronously collects the real-time displacement signal, actual axial force signal, and end real-time load force signal of the linear telescopic drive component; S4. Perform torque-compensation force calculation; The main control unit estimates the new load based on the no-load reference sensor data and the signal obtained in step S3, and calculates the combined gravitational torque generated by the new load, the end effector and the pitch support arm’s self-weight around the rotation axis by combining the pitch attitude angle, the pre-stored mass, center of gravity and installation geometry parameters of the end effector and the pitch support arm. Based on the combined gravitational torque and the real-time equivalent arm calculation, the desired compensation force is generated. S5. A composite control strategy of feedforward pre-control superimposed force feedback closed-loop correction is adopted to convert the desired compensation force into a drive command, input the drive command to the linear telescopic drive component, and output the corresponding balance compensation force. S6. Monitor the end load and pitch attitude changes in real time. If the operating conditions change, repeat steps S3 to S5 to dynamically update the output compensation force. If the operating conditions are stable, maintain the current force balance output state. S7. The workpiece is precisely aligned using the end effector.

[0007] This invention completely overcomes the limitations of traditional counterweights, spring-type fixed balance structures, and the single-ring rough estimation of thrust in existing electric servo robotic arms, and has the core advantages of fully automatic load adaptation, dynamic torque matching, and high-precision force balance compensation.

[0008] The system of this invention can rely on multi-source sensing detection and algorithm calculation to dynamically calculate and output a precisely matched compensation force based on the workpiece weight and the real-time pitch posture of the pitch support arm. This force completely offsets the combined gravitational torque generated by the self-weight of the pitch support arm and the load of the workpiece, significantly reducing the resistance of manual dragging and the intensity of operational intervention. It completely eliminates the cumbersome debugging steps of manually adding or removing counterweights and adjusting spring preload required by traditional equipment. While simplifying the operation process, it effectively broadens the load adaptability range of the equipment and can adapt to various unstructured assembly and maintenance operation sites.

[0009] The system abandons the crude method of traditional servo robotic arms that relies on motor current to indirectly estimate theoretical thrust. Instead, it uses a detection module to simultaneously collect three types of real physical signals: the extension and retraction displacement of the linear telescopic drive component, the actual axial force, and the real-time load force at the end effector. This directly obtains the true stress state of the mechanism, eliminating output estimation errors caused by lead screw transmission friction, mechanism backlash, component preload deformation, and instantaneous load impacts. Simultaneously, the system combines previously collected no-load baseline data, the robotic arm's inherent mass, center of gravity position, installation geometry parameters, and real-time pitch angles to accurately calculate the composite gravitational torque of the robotic arm around its rotation axis. Combined with real-time equivalent arm conversion, it obtains a high-precision expected compensation force. Compared to traditional current estimation methods, the accuracy of gravity balance calculation is significantly improved.

[0010] To further optimize the force balance control effect and suppress operational jitter, this invention adopts a composite control strategy of feedforward pre-control superimposed with force feedback closed-loop correction. Feedforward pre-control can issue basic drive commands in advance based on real-time torque calculation results, offsetting most of the gravitational torque in advance and avoiding jitter caused by control lag. Force feedback closed-loop correction uses the measured axial force of the drive component and the measured load force at the end as feedback basis to correct the drive output commands in real time, accurately compensating for output deviations caused by various transmission errors and load impacts. This completely solves the defects of traditional servo single-loop control, such as lack of real-time correction and mismatch between theoretical and actual output.

[0011] Meanwhile, the system has a real-time working condition monitoring and dynamic update mechanism, which can continuously monitor the end load and the pitch attitude of the robotic arm. Once the working condition changes, it immediately cycles through signal acquisition, torque calculation and compensation force update, and dynamically matches the real-time gravity torque throughout the process. This completely eliminates the problems of torque mismatch, dragging and falling, and movement jerking caused by traditional counterweight and spring equipment that can only adapt to fixed weight and fixed height working conditions. It greatly reduces the intensity of manual operation and improves the stability and safety of operation.

[0012] Preferably, in step S4, the equivalent force arm is obtained by dynamic geometry real-time solution, and the desired compensation force F^(t) is generated based on the combined gravitational torque M(t) and the real-time equivalent force arm r(t).

[0013] The equivalent lever arm can be calculated in real time by combining displacement and tilt angle sensing data with the geometric parameters of the mechanism, without the need for additional dedicated distance and angle sensing elements for lever arm length measurement.

[0014] This invention provides a real-time dynamic solution for the equivalent force arm r(t), updating the geometric length of the force arm in real time as the pitch bearing arm swings and the linear telescopic drive extends and retracts. Regardless of whether the pitch bearing arm is raised or lowered to any posture, the combined gravitational torque M(t) always corresponds one-to-one with the equivalent force arm. The calculated expected compensation force F^(t) does not have any inherent deviation in the geometric model, thus eliminating the problems of imbalance and drop jerking under different postures from the source.

[0015] Preferably, the equivalent force arm calculation process is as follows: the fixed geometric dimension parameters between the hinge points at both ends of the linear telescopic drive and the rotation axis are stored in the main control unit in advance; during calculation, the fixed geometric parameters are retrieved, and the current total length of the drive is solved in combination with the real-time telescopic displacement of the linear telescopic drive; based on the triangular geometric relationship formed by the rotation axis and the hinge points at both ends of the drive, the instantaneous vertical distance from the line of action of the linear telescopic drive force to the rotation axis is solved to obtain the dynamic real-time equivalent force arm r(t); and the desired compensation force F^(t) is generated based on the combined gravitational torque M(t) and the real-time equivalent force arm r(t).

[0016] This invention's system not only calculates thrust using conventional methods such as current detection of the drive components, but also performs real-time detection of the continuous changes in the drive component's lever arm after pitch attitude changes. Utilizing three-point hinged fixed geometry and real-time telescopic displacement, the system calculates the instantaneous vertical lever arm, i.e., the dynamic real-time equivalent lever arm r(t), through trigonometric geometry, and then calculates the desired compensation force F^(t) = M(t) / r(t).

[0017] The combined gravitational torque M(t) continuously changes with the load and pitch angle. If a constant value is used as an approximation for the lever arm, relying solely on force feedback closed-loop correction will generate a huge steady-state error, easily leading to drag jitter and response lag. The system of this invention calculates the theoretically desired compensation force using an equivalent lever arm during the feedforward calculation stage. The feedforward output value is infinitely close to the actual required force, significantly reducing the force closed-loop correction amount, suppressing fine-tuning jitter, controlling overshoot, and improving system stability and response speed.

[0018] This system utilizes only basic trigonometric geometric operations, without complex algorithms or the need for sophisticated sensors. It can quickly calculate r(t) and F^(t) using a conventional microcontroller or main controller, avoiding control lag. Furthermore, the calculation relies solely on the inherent geometric relationships of the mechanism and the displacement closed-loop signal of the telescopic drive component, unaffected by external environmental interference such as light, magnetic fields, or vibration. Compared to using tilt sensors, it offers higher signal stability in oily or vibrating workshop conditions. It also reduces the need for external sensing circuits and signal acquisition channels, lowering the probability of equipment failure due to circuit wear, signal interference, and sensor malfunction. The overall structure is simpler and more compact.

[0019] Preferably, in step S5, the strategy includes a feedforward branch and a feedback branch. The feedforward branch converts the desired compensation force F^(t) into a feedforward current If(t) based on the force-current mapping relationship. In the feedback branch, the force deviation e(t) is obtained by calculating the difference between the desired compensation force F^(t) and the actual axial force Fa(t). The main control unit includes a force PID controller. The force deviation e(t) is input to the force PID controller to calculate and output the current correction amount ΔI(t). The feedforward current If(t) and the current correction amount ΔI(t) are superimposed to generate the target current Ic(t). After safety limiting processing, Ic(t) is input into the current torque inner loop built into the linear telescopic drive to drive the linear telescopic drive to output the balance compensation force matching the working condition.

[0020] Feedforward can offset most of the self-weight and workpiece gravitational torque in advance, eliminating the need to rely on the force PID controller for adjustment after the force deviation occurs. This solves the inherent problems of lag, drop, and start-up jerking in pure feedback control; it also significantly reduces the correction amount that the force PID controller needs to output, reducing the closed-loop adjustment burden.

[0021] The expected compensation force obtained by the feedforward calculation is the theoretical value derived from the geometric model of the mechanism. Affected by factors such as lead screw transmission friction, assembly clearance, instantaneous load impact and force-current mapping drift, there is a difference between the actual axial thrust of the drive component and the expected compensation force. This difference is the force deviation. Using the force deviation as the input of the force PID controller for closed-loop correction can compensate for various undefined disturbance errors in the model in real time, eliminate hovering steady-state drop, and suppress jitter and impact during dragging.

[0022] Safety limiting can limit the maximum output current of the drive motor, which in turn limits the maximum output thrust, preventing overload damage to the workpiece or mechanism.

[0023] This invention has the advantages of directly correcting the output thrust in a closed loop, and improving response speed, balance accuracy, and smooth operation without adding complex external hardware.

[0024] Preferably, the detection module includes a displacement sensor, an axial force sensor, and a multi-dimensional force sensor. The displacement sensor is used to acquire real-time displacement signals of the linear telescopic drive component; the axial force sensor is used to acquire real-time axial force signals of the linear telescopic drive component; and the end-effector multi-dimensional force sensor is used to acquire real-time load torque information at the end-effector.

[0025] Preferably, step S5 also includes a safety judgment process: real-time comparison of target current, displacement signal, actual axial force signal and preset safety threshold; if the safety threshold is exceeded, the main control unit outputs a stop command and triggers an audible and visual alarm; if the safety threshold is not exceeded, the target current is output normally to drive the linear telescopic drive component.

[0026] The system simultaneously monitors three core parameters: target current (corresponding to the output thrust of the telescopic drive component), displacement of the telescopic drive component (stroke limit), and actual axial force (overload). If any of these parameters exceeds the threshold, the system immediately shuts down. This prevents the telescopic drive component from overtravel jamming or burning out due to overcurrent, significantly extending equipment lifespan and reducing maintenance costs.

[0027] Preferably, the end effector includes one or more modules selected from the XYZ axis fine-tuning slide module, angle tilting module, and rotation module.

[0028] It also has the ability to perform three-dimensional linear translation, tilt adjustment, and circumferential rotation, which can correct workpiece offset, tilt, and angular misalignment, and is suitable for precision docking scenarios of various workpieces.

[0029] Preferably, the multi-dimensional force sensor is used to determine the force and alignment status of the workpiece. If the force status of the workpiece does not meet the requirements or the alignment status is poor, the workpiece position is further adjusted by the module.

[0030] The force deviation signal output by the multi-dimensional force sensor can provide the adjustment basis for the main control unit, and link the XYZ fine-tuning slide module, tilt module, and rotation module to complete the three-dimensional translation, tilt angle, and rotation multi-angle micro-correction; it can automatically compensate for various assembly defects such as workpiece skewing, end face non-parallelism, circumferential angle misalignment, and uneven axial clearance, and solve the problems of assembly jamming and bump damage caused by slight workpiece offset and angle mismatch after coarse positioning, thus meeting the assembly requirements of high-precision parts.

[0031] This invention achieves adaptive dynamic gravity balance of load by dynamically solving the equivalent force arm based on geometric parameters, force feedforward plus PID composite closed loop, which has the advantage of facilitating high-precision alignment. Attached Figure Description

[0032] Figure 1 This is a schematic diagram of one structure of the present invention.

[0033] Figure 2 This is a schematic diagram of one structure of the end-effector component of the present invention.

[0034] Figure 3 This is a schematic diagram of the pitch support arm and swing frame of the present invention.

[0035] Figure 4 This is a flowchart of the control method of the present invention.

[0036] Figure 5 This is a block diagram of the servo electric cylinder load adaptive feedback control of the present invention. Detailed Implementation

[0037] The present invention will now be further described with reference to the accompanying drawings and specific embodiments.

[0038] Depend on Figures 1 to 3 As shown, this embodiment discloses a load adaptive force balance control system for a precision positioning assistive robotic arm, including a base 100, a pitch support arm 300, and an end effector 400. In this embodiment, the base 100 is equipped with pulleys at its bottom, and a locking mechanism is configured at the pulleys for transferring the whole machine between different working areas and locking and fixing it after it is in place.

[0039] The base 100 includes a column and a lifting platform 12 arranged along the column 11. The lifting platform 12 is driven to rise and fall by a motor 13 through a lifting transmission mechanism. A swing frame 200 is provided on the lifting platform 12. Vertical rods 21 are provided at both the upper and lower ends of the swing frame 200. The axis of the upper vertical rod is offset from the axis of the lower vertical rod. The end of the pitch support arm 300 away from the end effector 400 is rotatably connected to the upper vertical rod of the swing frame 200. The lower vertical rod of the swing frame 200 is rotatably connected to the lifting platform 12. Locking components 201 are provided at both vertical rods 21. In the locked state, the locking components 201 restrict the degree of rotation. The locking components can be any one of friction locking, pin locking, electromagnetic brake, pneumatic clamp, or servo brake.

[0040] The pitch support arm 300 also includes a first movable arm 32 and a second movable arm 33. Both ends of the first movable arm 32 and the second movable arm 33 are respectively hinged to an upper fixed part and a lower fixed part, forming a linkage mechanism. The hinge axis at the end of the first movable arm 32 is offset from the hinge axis at the same side end of the second movable arm 33. Both the upper and lower fixed parts include two oppositely arranged fixing plates 31. Both ends of the first movable arm 32 and the second movable arm 33 are hinged between corresponding fixing plates 31. The first movable arm 32 is located in front of the second movable arm 33. The lower hinge axis of the first movable arm 32 is located below and in front of the lower hinge axis of the second movable arm 33, and the upper hinge axis of the second movable arm 33 is located above and behind the upper hinge axis of the first movable arm 32.

[0041] The pitch support arm 300 is equipped with a linear telescopic drive component 5. When the linear telescopic drive component 5 extends or retracts, the pitch support arm 300 swings around the horizontal hinge axis hinged to the lower fixed part as the rotation center, thereby adjusting the pitch attitude of the pitch support arm. In this embodiment, the linear telescopic drive component 5 is a servo electric cylinder. One end of the servo electric cylinder is hinged and fixed to the first movable arm 32, and the other end of the servo electric cylinder is hinged and fixed to the middle of the second movable arm 33. The hinge point between the servo electric cylinder and the first movable arm 32 is coaxial with the hinge point between the first movable arm 32 away from the end effector and the fixed plate at the lower fixed part.

[0042] The pitch support arm 300 has an end effector 400 at its end. The end effector 400 includes a base 41, a first slide module 42, a second slide module 43, a third slide module 44, and an angle tilt module 45 mounted on the base. The multiple slide modules form intersecting linear fine-tuning axes. The angle tilt module 45 is used for fine-tuning the end effector's tilt. It also has a replaceable end effector 46, located at the output of one of the modules, including a rotating gripper that can function as a rotating module. The replaceable end effector is connected to a multi-dimensional force sensor via a quick-change structure and can be replaced with any of the following: gripper, clamp, bracket, hook, suction head, or tool mount, depending on the work object. In this embodiment, the first slide module 42, the second slide module 43, and the third slide module 44 are used for fine-tuning in the X, Y, and Z directions, respectively.

[0043] This system includes a main control unit and a detection module. The detection module includes a displacement sensor, an axial force sensor, and a multi-dimensional force sensor. In this embodiment, the displacement sensor is a built-in detection element integrated inside the servo cylinder. The axial force sensor is a single-axis strain gauge force sensor integrated inside the servo cylinder, and is also an integrated component of the servo cylinder. The displacement sensor is used to collect the real-time displacement signal of the linear telescopic drive component 5; the axial force sensor is used to collect the real-time axial force signal of the linear telescopic drive component 5; the multi-dimensional force sensor in this embodiment is located between the replaceable end effector and a module, and is used to collect the real-time load torque information at the end effector component.

[0044] When workpiece lifting and alignment is required, the robotic arm of this embodiment is first moved to the work area, and the lifting platform 12 is driven to rise and fall through the lifting transmission mechanism. The height of the pitch support arm 300 and the end effector 400 is adjusted, and the pulleys are locked.

[0045] Start the robotic arm of this embodiment, perform a power-on self-test, then adjust the working height of the lifting platform 12, and then start the control of the system of this embodiment.

[0046] Depend on Figures 1 to 5 As shown, the main control unit in this embodiment is used to perform load adaptive force balance control, and the control method of the main control unit includes the following steps S1-S7.

[0047] S1. System initialization: The system collects no-load reference data through the detection module. The no-load reference data includes the readings of the multi-dimensional force sensor and the initial displacement of the servo electric cylinder under no-load conditions.

[0048] S2. The workpiece to be operated is clamped or mounted on the replaceable end effector, the locking components are unlocked, and the operator drives the pitch support arm through the main control unit in conjunction with the servo electric cylinder to complete the large-range dragging coarse positioning.

[0049] S3. During the drag positioning process, the displacement feedback x(t) of the servo electric cylinder, the actual axial force Fa(t) detected by the axial force sensor, and the force information FL(t) of the multi-dimensional force sensor are read in real time.

[0050] S4. Perform torque-compensation force calculation; The main control unit estimates the new load based on the difference between the real-time force information FL(t) from the multi-dimensional force sensor and the no-load reference FL(0), and calculates the combined gravitational torque M(t) generated by the self-weight of the new load, the end effector and the pitch support arm around the rotation axis, in combination with the pitch attitude angle θ(t), the pre-stored mass, center of gravity and installation geometry parameters of the end effector and the pitch support arm. Based on the combined gravitational torque and the real-time equivalent arm calculation, the desired compensation force is generated.

[0051] That is, the hinge points at both ends of the linear telescopic drive component are pre-set ( Figure 3 The fixed geometric parameters between points a and b and the rotation axis (with point c on the axis) are stored in the main control unit. During calculation, these fixed geometric parameters are retrieved and combined with the real-time telescopic displacement of the linear telescopic drive to solve for the current total length of the drive. Based on the triangular geometric relationship formed by the rotation axis and the hinge points at both ends of the drive, the instantaneous vertical distance from the line of action of the linear telescopic drive to the rotation axis is solved to obtain the dynamic real-time equivalent force arm r(t). Based on the combined gravitational torque M(t) and the real-time equivalent force arm r(t), the desired compensation force F^(t) is generated.

[0052] S5. A composite control strategy of feedforward pre-control superimposed force feedback closed-loop correction is adopted to convert the desired compensation force F^(t) into a drive command, input the drive command to the servo electric cylinder, and output the corresponding balance compensation force.

[0053] S6. Monitor the end load and pitch attitude changes in real time. If the operating conditions change, repeat steps S3 to S5 to dynamically update the output compensation force. If the operating conditions are stable, maintain the current force balance output state.

[0054] S7. When the end effector 400 and the workpiece reach the vicinity of the target area, the locking component is tightened, and the end effector 400 completes the precise alignment of the workpiece.

[0055] In step S4, the calculation process of the real-time equivalent force arm r(t) is as follows: retrieve the fixed geometric dimension parameters between the hinge points at both ends of the servo cylinder and the rotation axis (the distance between point a and point c, and the distance between point c and point b), calculate the current actual length of the servo cylinder (the distance between point a and point b) in combination with the real-time extension and retraction displacement of the servo cylinder, and solve the perpendicular distance from the line of action of the servo cylinder force to the rotation axis through geometric trigonometric relationships to obtain the dynamic real-time equivalent force arm r(t).

[0056] In fact, the real-time equivalent arm r(t) is not a fixed constant. Its value changes dynamically with the pitch attitude of the pitch support arm. The main control unit recalculates the real-time equivalent arm r(t) based on the displacement feedback x(t) in each control cycle.

[0057] The calculation of the composite gravitational moment M(t) includes the gravitational moment components generated by the weight of the end effector 400, the workpiece, and the pitch support arm 300. Based on the composite gravitational moment M(t) and the real-time equivalent arm r(t), the desired compensation force F^(t) is generated, where the desired compensation force F^(t) = M(t) / r(t).

[0058] In this embodiment, without setting an independent pitch angle sensor, the real-time pitch attitude angle θ(t) of the pitch bearing arm is calculated based on the displacement feedback x(t) of the servo electric cylinder and the known geometric parameters between the hinge points at both ends and the rotation axis. The pitch attitude angle θ(t) is used to update the synthetic gravitational torque M(t) and the real-time equivalent arm r(t).

[0059] This scheme constructs a geometric triangle using three core points: the hinge axes at both ends of the servo cylinder and the rotation axis of the pitch support arm, all coplanar. The distance between the hinge points at both ends of the servo cylinder represents its real-time total length. The distances from the connection points between the servo cylinder and the second movable arm to the rotation axis, and from the connection points between the servo cylinder and the first movable arm to the rotation axis, are pre-measured and calibrated fixed geometric parameters. Based on these fixed parameters and the real-time extension and retraction displacement of the servo cylinder, the pitch angle θ(t) of the robotic arm can be calculated in real time. Furthermore, by solving for the area of ​​the geometric triangle formed by the three points using the law of cosines, and then combining this with the sine value of the pitch angle θ(t), the perpendicular distance of the line of action of the servo cylinder force relative to the rotation axis can be accurately calculated, ultimately yielding the dynamic real-time equivalent arm r(t).

[0060] The estimation method of the combined gravitational moment M(t) is as follows: the main control unit estimates the current equivalent load based on the difference between the force information FL(t) of the multi-dimensional force sensor and the initial force information FL(0) of the no-load reference data; combined with the pitch attitude angle θ(t) and the pre-stored mechanism geometric parameters, the combined gravitational moment M(t) generated by the equivalent load and the self-weight of the pitch bearing arm, workpiece, and end effector 400 around the rotation axis is calculated.

[0061] In step S5, the composite control strategy includes a feedforward branch and a feedback branch. The feedforward branch converts the desired compensation force F^(t) into a feedforward current If(t) through force-current feedforward conversion. The feedback branch subtracts the desired compensation force F^(t) from the actual axial force Fa(t) to obtain the force deviation e(t) = F^(t) - Fa(t). The force deviation e(t) is input to the force PID controller, which outputs a current correction ΔI(t). The feedforward current If(t) and the current correction ΔI(t) are positively superimposed to form the target current Ic(t) = If(t) + ΔI(t). After safety limiting, the target current Ic(t) is input to the current / torque inner loop of the servo cylinder's built-in servo driver, driving the servo cylinder to output the compensation force F(t), so that the actual axial force Fa(t) tracks the desired compensation force F^(t). The safety limiting method can adopt conventional technical solutions.

[0062] In step S5, the main control unit also executes a safety judgment process, comparing the target current Ic(t), displacement feedback x(t), and actual axial force Fa(t) with the preset safety thresholds of each parameter in real time. If the target current Ic(t) exceeds the current safety threshold, or the actual axial force Fa(t) exceeds the axial force overload threshold, or the displacement feedback x(t) exceeds the stroke limit of the servo cylinder, the main control unit outputs a stop command and triggers an audible and visual alarm. If the safety thresholds are not exceeded, the target current Ic(t) is output normally to drive the servo cylinder.

[0063] The safety judgment process in this embodiment also includes: current change rate limitation, PID integral anti-saturation, emergency stop braking, and locking interlock judgment at the locking assembly.

[0064] The multi-dimensional force sensor in this embodiment is also used to provide feedback on the end contact force and alignment status. If the alignment status does not meet the requirements, the position and orientation of the workpiece are further adjusted through the above module until the accuracy requirements of installation, testing or maintenance operations are met.

Claims

1. A load-adaptive force balance control system for a precision positioning-assisted robotic arm, comprising a base and a pitch support arm, wherein one end of the pitch support arm is connected to the base for vertical swinging and forms a rotation axis, the end of the pitch support arm is provided with an end effector, and the pitch support arm is equipped with a linear telescopic drive for realizing its own swinging about the rotation axis, characterized in that, It also includes a main control unit and a detection module. The control method of the main control unit includes the following steps: S1. Collect no-load reference sensor data at the end of the pitch support arm under no-load conditions through the detection module; S2. The workpiece is mounted or clamped on the end effector component, and the main control unit, in conjunction with the linear telescopic drive component, drives the pitch support arm to complete a large-range drag positioning. S3. During the towing and positioning process, the detection module synchronously collects the real-time displacement signal, actual axial force signal, and end real-time load force signal of the linear telescopic drive component; S4. Perform torque-compensation force calculation; The main control unit estimates the new load based on the no-load reference sensor data and the signal obtained in step S3, and calculates the combined gravitational torque generated by the new load, the end effector and the pitch support arm’s self-weight around the rotation axis by combining the pitch attitude angle, the pre-stored mass, center of gravity and installation geometry parameters of the end effector and the pitch support arm. Based on the combined gravitational torque and the real-time equivalent arm calculation, the desired compensation force is generated. S5. A composite control strategy of feedforward pre-control superimposed force feedback closed-loop correction is adopted to convert the desired compensation force into a drive command, input the drive command to the linear telescopic drive component, and output the corresponding balance compensation force. S6. Monitor the end load and pitch attitude changes in real time. If the operating conditions change, repeat steps S3 to S5 to dynamically update the output compensation force. If the operating conditions are stable, maintain the current force balance output state. S7. The workpiece is precisely aligned using the end effector.

2. The load adaptive force balance control system for the precision positioning-assisted robotic arm according to claim 1, characterized in that, In step S4, the equivalent force arm is obtained by dynamic geometry real-time solution, and the desired compensation force F^(t) is generated based on the combined gravitational torque M(t) and the real-time equivalent force arm r(t).

3. The load adaptive force balance control system for the precision positioning-assisted robotic arm according to claim 2, characterized in that, The equivalent force arm calculation process is as follows: the fixed geometric dimension parameters between the hinge points at both ends of the linear telescopic drive and the rotation axis are stored in the main control unit in advance; during calculation, the fixed geometric parameters are retrieved, and the current total length of the drive is solved in combination with the real-time telescopic displacement of the linear telescopic drive; based on the triangular geometric relationship formed by the rotation axis and the hinge points at both ends of the drive, the instantaneous vertical distance from the line of action of the linear telescopic drive force to the rotation axis is solved, and the dynamic real-time equivalent force arm r(t) is obtained.

4. The load adaptive force balance control system for the precision positioning-assisted robotic arm according to claim 1, characterized in that, In step S5, the strategy includes a feedforward branch and a feedback branch. The feedforward branch converts the desired compensation force F^(t) into a feedforward current If(t) based on the force-current mapping relationship. In the feedback branch, the force deviation e(t) is obtained by calculating the difference between the desired compensation force F^(t) and the actual axial force Fa(t). The main control unit includes a force PID controller. The force deviation e(t) is input to the force PID controller to calculate and output the current correction amount ΔI(t). The feedforward current If(t) and the current correction amount ΔI(t) are superimposed to generate the target current Ic(t). After safety limiting processing, Ic(t) is input into the current torque inner loop built into the linear telescopic drive to drive the linear telescopic drive to output the balance compensation force matching the working condition.

5. The load adaptive force balance control system for the precision positioning-assisted robotic arm according to claim 1, characterized in that, The detection module includes a displacement sensor, an axial force sensor, and a multi-dimensional force sensor. The displacement sensor is used to collect real-time displacement signals of the linear telescopic drive component; the axial force sensor is used to collect real-time axial force signals of the linear telescopic drive component; and the end-effector multi-dimensional force sensor is used to collect real-time load torque information at the end-effector.

6. The load adaptive force balance control system for the precision positioning-assisted robotic arm according to claim 4, characterized in that, Step S5 also includes a safety judgment process: real-time comparison of the target current, displacement signal, actual axial force signal and preset safety threshold; If the safety threshold is exceeded, the main control unit outputs a shutdown command and triggers an audible and visual alarm. When the safety threshold is not exceeded, the target current is normally output to drive the linear telescopic actuator.

7. The load adaptive force balance control system for the precision positioning-assisted robotic arm according to claim 2, characterized in that, The end effector includes one or more modules selected from the XYZ axis fine-tuning slide module, angle tilt module, and rotation module.

8. The load adaptive force balance control system for the precision positioning-assisted robotic arm according to claim 7, characterized in that, The multi-dimensional force sensor determines the force and alignment status of the workpiece. If the force status of the workpiece does not meet the requirements or the alignment status is poor, the module can be used to further adjust the position of the workpiece.

Images