A trajectory planning method of a manipulator in a vacuum isothermal forging environment

Abstract

The application relates to the field of industrial control and discloses a trajectory planning method of a manipulator under a vacuum isothermal forging environment. First, three-dimensional structure information of a vacuum isothermal forging furnace is acquired, and a three-dimensional model containing at least a forging chamber, a vacuum channel, a heating furnace and a die seat is constructed, and a kinematic model of the manipulator is established to determine the conversion relationship between joint space coordinates and three-dimensional model space coordinates; key work points of respective stages are determined according to a blank placing stage and a forged piece taking stage in a forging operation process, and then in the three-dimensional model, the blank placing trajectory and the forged piece taking trajectory are planned by using the key work points of the corresponding stages, with the minimum pose error of the blank placed in the die seat and the minimum pose error of the forged piece clamped from the die seat as targets. The trajectory planning of the application takes into account the long-stroke transfer efficiency of the vacuum isothermal forging manipulator and the high-precision control of the end pose of the blank placed and the forged piece clamped at the key points.

Description

Technical Field

[0001] This invention relates to the field of industrial control technology, specifically to a trajectory planning method for a robotic arm in a vacuum isothermal forging environment. Background Technology

[0002] Vacuum isothermal forging is an advanced process for precision forming of difficult-to-deform alloys under high vacuum and constant high temperature conditions. Its operating environment typically includes enclosed cavities such as a heating furnace, vacuum channel, and forging chamber. To achieve efficient transfer of the billet from the heating furnace to the die holder and reliable removal of the forging, a specialized robotic arm with multi-stage telescopic and rotary functions is usually employed. In the two critical stages of billet placement and forging removal, this robotic arm not only needs to traverse narrow vacuum channels and gates but also needs to achieve high-precision orientation alignment above the die holder. The quality of its trajectory planning directly affects the forming accuracy of the forging and the production cycle time.

[0003] Currently, a common method for robot trajectory planning in forging environments is to use a teach pendant to record several key position points, and then generate a continuous motion trajectory through polynomial interpolation or trapezoidal velocity curves in joint space. This method primarily uses collision avoidance and joint limit constraints as constraints, and is effective in terms of trajectory smoothness and motion efficiency. However, in vacuum isothermal forging scenarios, the blank placement and forging removal processes have strict but differentiated requirements for end-effector pose accuracy: blank placement requires accurately placing the high-temperature blank in a preset posture within the die holder, while forging removal requires precise alignment of the clamps with the clamping parts of the formed forging. Existing planning methods typically use a uniform interpolation strategy for all path segments, without differentiating between high-precision placement segments and long-distance transport planning segments. While task space planning can guarantee the geometric accuracy of the end-effector path, it leads to low efficiency in long-stroke joint motion. If joint space planning is used to improve efficiency, the shape of the end-effector path becomes uncontrollable, making it difficult to meet precise alignment pose constraints near the die holder.

[0004] Furthermore, existing methods typically prioritize collision-free reachability during the planning phase, without explicitly optimizing the minimization of placement or gripping pose errors. Consequently, although the generated trajectory may be collision-free, there may still be significant end-point deviations at the material placement or pick-up points. These deviations require subsequent manual adjustments or additional visual compensation to correct, reducing operational consistency and automation.

[0005] Therefore, the trajectory planning method for the robot in the vacuum isothermal forging environment failed to perform segmented optimization for the differentiated requirements of end-effector pose accuracy in the billet placement and forging removal processes, and did not take minimizing the pose error during placement or clamping as an explicit planning objective. As a result, while ensuring the efficiency of long-distance transfer, it is difficult to achieve high-precision control of the end-effector pose at key working points, which affects the forging forming accuracy and operational consistency. Summary of the Invention

[0006] The purpose of this invention is to provide a trajectory planning method for a robot in a vacuum isothermal forging environment, in order to solve the technical problem that the existing technology fails to perform segmented optimization for the differentiated requirements of end-effector pose accuracy in the billet placement and forging removal processes, and does not take minimizing the pose error during placement or clamping as an explicit planning objective, which makes it difficult to achieve high-precision control of end-effector pose at key working points while ensuring long-distance transfer efficiency.

[0007] To solve the above-mentioned technical problems, the present invention specifically provides the following technical solution: A trajectory planning method for a robotic arm under vacuum isothermal forging conditions includes the following steps: Obtain the three-dimensional structural information of the vacuum isothermal forging furnace and construct a three-dimensional model that includes at least the forging chamber, vacuum channel, heating furnace, and mold base; Establish a kinematic model of the robot to determine the transformation relationship between the joint space coordinates of the robot and the three-dimensional space coordinates of the three-dimensional model; Based on the billet feeding stage and the forging stage in the forging operation of the vacuum isothermal forging furnace, the key working points of the robot in the billet feeding stage and the key working points of the robot in the forging stage are determined respectively. In the three-dimensional model, with the goal of minimizing the pose error when the blank is placed on the mold base and minimizing the total operation time of the blank placement stage, the blank placement trajectory of the robot is planned using the key working points of the blank placement stage. In the three-dimensional model, with the goal of minimizing the pose error when picking up the forging from the mold base and minimizing the total operation time of the forging picking stage, the forging picking trajectory of the robot is planned using the key working points of the forging picking stage.

[0008] Furthermore, methods for establishing a kinematic model of the manipulator to determine the transformation relationship between the joint space coordinates of the manipulator and the three-dimensional space coordinates of the three-dimensional model include: Based on the mechanical structure of the manipulator, the homogeneous transformation matrix between the coordinates of each adjacent joint of the manipulator is established using the DH parameter method. The mechanical structure includes a traveling mechanism, a lifting mechanism, a rotating mechanism, a first-stage telescopic mechanism, a second-stage telescopic mechanism, and an end effector. The robot arm is represented in joint space to obtain the joint space coordinate vector of the robot arm. ,in This represents the displacement of the traveling mechanism along the track. This refers to the vertical displacement of the lifting mechanism. The rotation angle of the rotary mechanism. and These are the extension lengths of the first-stage telescopic mechanism and the second-stage telescopic mechanism, respectively. By multiplying the homogeneous transformation matrices of adjacent joint coordinates, a kinematic model representing the transformation relationship between joint spatial coordinates and Cartesian three-dimensional spatial coordinates is obtained. The kinematic model is ; in, Let be the homogeneous transformation matrix of the spatial coordinates of the traveling mechanism relative to the three-dimensional spatial coordinates. Let be the homogeneous transformation matrix of the spatial coordinates of the lifting mechanism relative to the spatial coordinates of the traveling mechanism. Let be the homogeneous transformation matrix of the spatial coordinates of the rotary mechanism relative to the spatial coordinates of the lifting mechanism. Let be the homogeneous transformation matrix of the spatial coordinates of the first-stage telescopic mechanism relative to the spatial coordinates of the rotary mechanism. Let be the homogeneous transformation matrix of the spatial coordinate system of the second-stage telescopic mechanism relative to the spatial coordinate system of the first-stage telescopic mechanism. Let be the homogeneous transformation matrix of the spatial coordinates of the end gripper relative to the spatial coordinates of the second-stage telescopic mechanism. is the homogeneous transformation matrix of the spatial coordinates of the end gripper relative to the three-dimensional spatial coordinates.

[0009] Furthermore, the key working points in the billet feeding stage include at least: material preparation point, material feeding point, material feeding exit point, transfer intermediate point, forging chamber entrance point, material feeding preparation point, material feeding point, material feeding exit point, and material feeding avoidance point.

[0010] Furthermore, the key working points in the forging stage include at least: the forging preparation point, the forging point, the forging exit point, the intermediate point of forging transfer, and the unloading point.

[0011] Furthermore, the method for planning the billet placement trajectory includes: Based on the key working points of the billet feeding stage, a trajectory framework for the billet feeding stage is established. In the trajectory framework of the billet feeding stage, each local trajectory from the material preparation point to the forging chamber entrance point and each local trajectory from the feeding point through the feeding exit point to the feeding avoidance point are marked as the transfer planning segment of the billet feeding stage, and each local trajectory from the forging chamber entrance point through the feeding preparation point to the feeding point is marked as the precise planning segment of the billet feeding stage. In the transfer planning section of the billet unloading stage, an S-shaped velocity curve is used for interpolation planning in the joint space to minimize the total operation time of the billet unloading stage, thereby obtaining the planning trajectory of the transfer planning section. In the precise planning segment of the blank placement stage, linear interpolation is used in Cartesian three-dimensional coordinate space for planning, with the optimization objective of minimizing the pose error when the blank is placed on the mold base, to obtain the planning trajectory of the precise planning segment. The planned trajectory of the transfer planning segment is transformed into a three-dimensional space through the kinematic model, so that the planned trajectory of the transformed transfer planning segment is unified with the planned trajectory of the precise planning segment in the same coordinate space; At the common point between the planned trajectory of the transfer planning segment and the planned trajectory of the precise planning segment, a speed smooth transition method is used to connect the planned trajectory of the transfer planning segment and the planned trajectory of the precise planning segment to obtain the billet release trajectory.

[0012] Furthermore, the method for planning the trajectory of the forging includes: Based on the key working points of the forging stage, a trajectory framework for the forging stage is established. In the trajectory framework of the forging stage, each local trajectory from the forging point through the forging exit point and the intermediate point of the forging transfer to the unloading point is marked as the transfer planning segment of the forging stage, and each local trajectory from the forging preparation point to the forging point is marked as the precise planning segment of the forging stage. In the transfer planning segment of the forging stage, an S-shaped velocity curve is used for interpolation planning in the joint space to minimize the total operation time of the forging stage, thereby obtaining the planned trajectory of the transfer planning segment. In the precise planning segment of the forging stage, linear interpolation is used in three-dimensional space to minimize the pose error when the forging is picked up from the mold base, and the planning trajectory of the precise planning segment is obtained. The planned trajectory of the transfer planning segment is transformed into a three-dimensional space through the kinematic model, so that the transformed planned trajectory of the transfer planning segment is unified with the planned trajectory of the precise planning segment in the same coordinate space; At the common point between the planned trajectory of the transfer planning segment and the planned trajectory of the precise planning segment, a speed smooth transition method is used to connect the planned trajectory of the transfer planning segment and the planned trajectory of the precise planning segment to obtain the forging trajectory.

[0013] Furthermore, the trajectory planning method for the precise planning segment in the billet feeding stage includes: In the three-dimensional model, a three-dimensional straight path is planned along the normal direction of the mold base table, so that the end gripper of the robot arm moves from the material preparation point to the material release point along the straight path; Constructing the pose error cost function for the precise planning segment in the billet laying stage ,in The position and orientation of the end clamp at the discharge point. This refers to the target position and orientation of the blank at the mold base. These are the weighting coefficients; With the goal of minimizing the cost function, the straight path is optimized under the constraints of zero maximum speed, maximum acceleration, joint displacement, and zero end speed and zero end acceleration at the unloading point. The optimal straight path that minimizes the cost function is then obtained. The optimal straight path is used as the planned trajectory of the precise planning segment.

[0014] Furthermore, the trajectory planning method for the precise planning segment in the forging stage includes: In the three-dimensional model, a straight path in three-dimensional space is planned so that the end effector of the robot moves from the part preparation point to the part picking point along the straight path, and the posture of the end gripper remains aligned with the normal of the forging clamping surface during the movement. Constructing the pose error cost function for the precise planning segment in the forging stage ,in The position and orientation of the clamping points on the forging. The position and orientation of the end gripper when it reaches the pick-up point. These are the weighting coefficients; With the goal of minimizing the cost function, the straight path is optimized under the constraints of maximum speed, maximum acceleration, joint displacement limit, and zero end velocity and zero end acceleration at the pick-up point. The optimal straight path that minimizes the cost function is then obtained. The optimal straight path is used as the planned trajectory of the precise planning segment.

[0015] Furthermore, the planning methods for the trajectory of the transfer planning segment include: Determine the start and end points of the planned transfer segment, and calculate the displacement of each joint of the robotic arm from the start to the end point; Based on the displacement of each joint and the constraints of maximum velocity, maximum acceleration, and maximum jerk, an S-shaped velocity curve is synchronously planned for each joint in joint space, wherein: For each joint of the robot, under the constraints, the duration of each stage of the S-shaped velocity curve is automatically calculated. The S-shaped velocity curve consists of an acceleration segment, a uniform acceleration segment, a deceleration segment, a uniform speed segment, an acceleration / deceleration segment, a uniform deceleration segment, and a deceleration / deceleration segment. When the displacement of the joint is insufficient to make the velocity reach the maximum velocity, the S-shaped velocity curve degenerates into an S-shaped velocity curve without a uniform velocity segment. When the displacement of the joint is shortened and the maximum acceleration cannot be achieved, the S-shaped velocity curve degenerates into a triangular acceleration curve. Using the joint with the longest required motion time as the main synchronization axis, the S-shaped velocity curves of the remaining joints are scaled in time to ensure the coordination of the movement of each joint of the robot. The planned trajectory of the transport planning segment is obtained by isochronous discrete sampling of the S-shaped velocity curves generated by each joint.

[0016] Compared with the prior art, the present invention has the following advantages: This invention divides the trajectory of billet placement and forging removal into a transfer planning segment and a precision planning segment, and adopts a differentiated planning strategy. In the transfer planning segment, an S-shaped velocity curve in joint space ensures efficient and stable long-distance movement. In the precision planning segment, Cartesian three-dimensional linear interpolation is performed with the goal of minimizing pose error, and a zero-speed and zero-acceleration constraint is applied at the end. At the same time, the planning results of the two spaces are uniformly converted to the three-dimensional model coordinate system to achieve smooth connection of speeds between segments. Thus, under the same planning framework, the long-stroke transfer efficiency of the vacuum isothermal forging robot and the high-precision control of the end pose of billet placement and forging removal at key points are taken into account, which effectively improves the forging forming accuracy and operation consistency. Attached Figure Description

[0017] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings in the following description are merely exemplary, and those skilled in the art can derive other embodiments based on the provided drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of the overall structure provided for an embodiment of this application.

[0019] Figure 2 This is a schematic diagram of the overall structure of the vacuum isothermal forging furnace provided in the embodiments of this application.

[0020] Figure 3 This is a schematic plan view of the overall structure of the robotic arm provided in an embodiment of this application.

[0021] Figure 4 This is a three-dimensional schematic diagram of the overall structure of the robotic arm provided in the embodiments of this application.

[0022] The reference numerals in the figure are as follows: 1. Traveling mechanism; 2. Lifting mechanism; 3. Rotating mechanism; 4. First-stage telescopic mechanism; 5. Second-stage telescopic mechanism; 6. End clamp; 7. Forging. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0024] like Figure 1 As shown, the present invention provides a trajectory planning method for a robot in a vacuum isothermal forging environment. First, the three-dimensional structural information of the vacuum isothermal forging furnace is obtained and a three-dimensional model including at least the forging chamber, vacuum channel, heating furnace and die base is constructed. At the same time, the kinematic model of the robot is established to determine the transformation relationship between the joint space coordinates and the three-dimensional model space coordinates. Based on this, the key working points of the billet placement stage and the forging removal stage in the forging operation are determined respectively. Then, in the three-dimensional model, the trajectory of billet placement and the trajectory of forging removal are planned by using the key working points of the corresponding stages, with the goal of minimizing the pose error when the billet is placed on the die base and minimizing the pose error when the forging is clamped from the die cavity.

[0025] The first step is to obtain the three-dimensional structural information of the vacuum isothermal forging furnace (such as...). Figure 2 As shown, construct a three-dimensional model that includes at least a forging chamber, a vacuum channel, a heating furnace, and a mold base. The three-dimensional model belongs to the Cartesian three-dimensional coordinate system.

[0026] The second step is to establish a kinematic model of the robot to determine the transformation relationship between the joint space coordinates of the robot and the three-dimensional space coordinates of the three-dimensional model.

[0027] like Figure 3 and Figure 4 As shown, the robotic arm of the present invention is a serial mechanism, consisting of a traveling mechanism 1, a lifting mechanism 2, a rotating mechanism 3, a first-stage telescopic mechanism 4, a second-stage telescopic mechanism 5, and an end effector 6 connected in sequence. The independent movements of each joint are ultimately synthesized into a composite movement of the end effector in three-dimensional space. Therefore, the present invention establishes a kinematic model to quantify this transmission relationship from multi-joint displacement to end effector pose.

[0028] Furthermore, methods for establishing a kinematic model of the manipulator to determine the transformation relationship between the joint space coordinates of the manipulator and the three-dimensional space coordinates of the three-dimensional model include: Based on the mechanical structure of the manipulator, the homogeneous transformation matrix between the coordinates of each adjacent joint of the manipulator is established using the DH parameter method. The mechanical structure includes a traveling mechanism, a lifting mechanism, a rotating mechanism, a first-stage telescopic mechanism, a second-stage telescopic mechanism, and an end effector.

[0029] The DH parameter method, by fixing a coordinate system to each joint and describing the relative pose relationship between adjacent joint coordinates using four parameters, can handle the forward kinematics problem of a series link in a standardized manner, and is particularly suitable for the hybrid configuration containing translational and rotational joints as described in this invention. Therefore, this invention utilizes the DH parameter method to establish the homogeneous transformation matrix between the coordinates of each adjacent joint.

[0030] The pose of the robot hand in joint space is represented to obtain the joint space coordinate vector of the robot hand. ,in This represents the displacement of the traveling mechanism along the track. This refers to the vertical displacement of the lifting mechanism. The rotation angle of the rotary mechanism. and These are the extension lengths of the first-stage telescopic mechanism and the second-stage telescopic mechanism, respectively. By multiplying the homogeneous transformation matrices of each adjacent joint coordinate system, a kinematic model representing the transformation relationship between joint spatial coordinates and Cartesian three-dimensional spatial coordinates is obtained. Kinematic model is ; in, Let be the homogeneous transformation matrix of the spatial coordinates of the traveling mechanism relative to the three-dimensional spatial coordinates. Let be the homogeneous transformation matrix of the spatial coordinates of the lifting mechanism relative to the spatial coordinates of the traveling mechanism. Let be the homogeneous transformation matrix of the spatial coordinates of the rotary mechanism relative to the spatial coordinates of the lifting mechanism. Let be the homogeneous transformation matrix of the spatial coordinates of the first-stage telescopic mechanism relative to the spatial coordinates of the rotary mechanism. Let be the homogeneous transformation matrix of the spatial coordinate system of the second-stage telescopic mechanism relative to the spatial coordinate system of the first-stage telescopic mechanism. Let be the homogeneous transformation matrix of the spatial coordinates of the end gripper relative to the spatial coordinates of the second-stage telescopic mechanism. is the homogeneous transformation matrix of the spatial coordinates of the end gripper relative to the three-dimensional spatial coordinates.

[0031] In this invention The translation of the traveling mechanism along the track is converted into a positional change in the three-dimensional model space; and Describe the contributions of lifting and rotating motions to the end effector pose, respectively; and Describe the radial extension of the two-stage telescopic mechanism; This is a fixed offset of the clamp relative to the end of the second-stage telescopic mechanism.

[0032] By performing forward calculations using this kinematic model, any set of joint coordinates can be mapped to the precise pose of the end effector clamp in the 3D model space. Through inverse kinematics, the planned end effector path in the 3D model can be transformed into motion commands for each joint. This bidirectional mapping capability enables segmented differentiated planning: the transport planning segment is planned in joint space to ensure motion efficiency, while the precision planning segment is planned in Cartesian 3D space to ensure the geometric accuracy of the end effector path. The two are seamlessly converted through this kinematic model, ultimately unifying the planning results from both spaces into the same 3D model coordinate system, ensuring the spatial consistency and physical executability of the entire trajectory.

[0033] The third step involves determining the critical working points of the robot arm during the billet feeding stage and the forging stage during the forging operation in the vacuum isothermal forging furnace, based on the billet feeding stage and the forging stage.

[0034] The complete cycle of vacuum isothermal forging is not a single round trip, but includes two independent processes that differ in terms of objectives, paths, and precision requirements: taking the cold billet out of the heating furnace and placing it into the die holder, and taking the formed forging 7 out of the die holder and sending it out of the forging chamber. Therefore, this invention clearly divides the operation process into the billet placement stage and the forging removal stage, and defines exclusive key working points for each stage, so that the trajectory planning of each stage revolves around its core tasks, namely the billet placement task and the forging removal task, rather than mixing them up.

[0035] Furthermore, the key work points in the billet feeding stage include at least: Material handling preparation point: The end gripper of the robot arm is located at a preset safe distance outside the material handling port of the heating furnace, and the end posture is aligned with the axis of the billet. This point is the positioning point before the robot arm moves from the standby position to the vicinity of the heating furnace and is ready to perform the gripping action. Picking point: The end effector of the robotic arm extends into the heating furnace, the center of the gripper coincides with the geometric center of the billet, the gripper closes and clamps the billet, this point is the actual position where the billet is picked up; Material removal exit point: After the robotic arm grips the billet, it exits the heating furnace in the opposite direction of the material entry direction and exits to a preset safe position outside the material removal port. This point ensures that the billet completely leaves the heating furnace and does not interfere with the furnace door. Transfer intermediate point: One or more intermediate path points set by the robot arm to bypass bends, gate valves or other fixed obstacles when moving in a vacuum channel; Forging chamber entrance point: The waiting point where the robotic arm holds the billet to the outside of the forging chamber gate and waits for the gate to open before entering the forging chamber; Material placement preparation point: After the robot enters the forging chamber, the positioning point is located at a preset height above the mold base. The end posture is adjusted to be aligned with the axis of the mold base. This point is the starting point of the precision placement section plan. Unloading point: When the end of the robot arm descends to contact the bottom surface of the mold base or reaches the preset placement position, the clamps release the blank. This point is the target position for blank placement. Material discharge exit point: After the robot releases the billet, it exits the forging chamber in the opposite direction of the material discharge entry direction and exits into the vacuum channel outside the forging chamber gate; Material discharge avoidance point: After the robot arm completely exits the forging chamber, it waits for the forging operation to be completed in the standby position in the vacuum channel. This point is kept at a safe distance from the gate of the forging chamber to avoid the high temperature and vibration generated during the forging process from affecting the robot arm.

[0036] During the billet feeding stage, key working points are sequentially arranged along the physical flow direction of the billet from the heating furnace to the die holder. The pre-retrieving point is located at a safe distance outside the heating furnace's retrieving port, with its end-effector pre-aligned with the billet's axis. This allows the robot to complete coarse posture adjustments before entering the high-temperature furnace cavity, reducing furnace dwell time and lowering the risk of heat exposure. The retrieving point uses the alignment of the clamp center with the billet's geometric center as a positioning condition, ensuring symmetrical distribution of clamping force and preventing slippage or deformation of the billet during transport due to clamping eccentricity. The retrieving exit point requires the robot to exit in the opposite direction of entry to a safe position. This constraint ensures the exit path is collinear with the entry path, simplifying obstacle avoidance complexity at narrow furnace openings.

[0037] Transfer intermediate points are one or more path points set by the robot arm in the vacuum channel to bypass fixed obstacles such as bends and gate valves. Their number and location are flexibly determined based on the channel geometry and obstacle distribution in the 3D model, providing spatial guidance for subsequent collision detection.

[0038] The forging chamber entrance is located outside the forging chamber gate, marking the boundary between the transfer planning section and the precision planning section. Before this point, the robot moves rapidly through the vacuum channel with efficiency as its goal; after this point, the robot enters the forging chamber and begins its high-precision approach to the die holder. The material unloading preparation point is located at a preset height above the die holder, with its end effector adjusted to align with the die holder's axis. This point is the starting point of the precision planning section, and its spatial location directly determines the length and direction of the final straight path, thus affecting the optimization space of the pose error cost function. The material unloading point terminates when the billet contacts the bottom surface of the die holder, representing the ultimate test of pose accuracy during the material unloading stage. The material unloading exit point and material unloading avoidance point ensure that the robot safely withdraws from the forging chamber after releasing the billet and remains in a safe position within the vacuum channel during the forging operation.

[0039] Furthermore, the key working points in the forging stage include at least: Retrieval preparation point: After forging is completed, the forging chamber gate is opened, and the robot arm moves from the standby position in the vacuum channel to the positioning point inside the forging chamber gate, located at a preset height above the mold base. The end posture is adjusted to be aligned with the normal of the forging clamping surface. This point is the starting point of the precision retrieval section. Pick-up point: The end effector of the robotic arm descends to align the clamp with the preset clamping part of the formed forging and closes the clamp. This point is the actual position for picking up the forging. Part removal exit point: After the robotic arm grips the forging, it exits the forging chamber in the opposite direction of the part removal entry direction and exits into the vacuum channel outside the forging chamber gate. This point ensures that the forging completely leaves the forging chamber and does not interfere with the gate. Intermediate points for picking up and transferring parts: one or more intermediate path points set by the robot arm to bypass bends, gate valves or other fixed obstacles when returning with the forging in the vacuum channel; Unloading point: The robotic arm clamps the forging to the designated unloading position, releases the forging, and completes the unloading operation of the forging process.

[0040] In the forging stage, key working points are sequentially arranged along the path of the robot arm from the standby position to the mold base and back. However, there are two important differences compared to the billet placement stage: First, in the forging stage, the robot arm enters unloaded and exits with the forging clamped, and the change in load affects the setting of motion parameters; second, the alignment target at the forging point changes from the billet to the formed forging. The forging may have slight positional deviations in the mold base due to forging deformation. Therefore, the alignment requirements for the end effector posture at the forging preparation point and the forging point are more stringent. The end effector posture must be aligned with the normal of the forging clamping surface, rather than simply coinciding with the axis. The forging exit point and the intermediate point of forging transfer take into account the changes in the overall outer envelope of the robot arm after clamping the forging, ensuring that the collision-free condition constraint is also met in the return path of the vacuum channel. The unloading point delivers the forging to the designated position, marking the closure of the entire forging operation cycle.

[0041] Through the establishment of the aforementioned key work points, a complete chain of points is formed in the billet unloading stage, including the material preparation point, material unloading point, material unloading exit point, intermediate transfer point, forging chamber entrance point, unloading preparation point, unloading point, unloading exit point, and material unloading avoidance point. Similarly, a complete chain of points is formed in the forging removal stage, including the part preparation point, part removal point, part removal exit point, intermediate transfer point, and unloading point. The setting of each point is based on the actual geometric relationships in the 3D model, and its spatial coordinates can be precisely quantified, providing reliable input for the subsequent establishment of the trajectory framework. More importantly, the segment between the forging chamber entrance point and the material preparation point, and the segment between the part preparation point and the part removal point, respectively constitute the spatial carriers of the precisely planned segments in billet unloading and forging removal; the segments between the remaining points constitute transfer planning segments guided by motion efficiency (shortest operation time). This segmented structure, naturally formed based on task attributes and spatial location, provides a clear physical basis and engineering operability for the subsequent proposal of differentiated planning strategies.

[0042] The fourth step is to plan the billet placement trajectory of the robot arm in the 3D model with the goal of minimizing the pose error when the billet is placed on the mold base and minimizing the total operation time of the billet placement stage.

[0043] This invention divides the billet placement trajectory into a transfer planning segment and a precision planning segment according to task characteristics, and uses different spatial coordinate systems and optimization objectives for planning. This allows for both the efficiency of long-distance transfer and the high precision requirements of placement operations above the mold base within the same trajectory framework.

[0044] Furthermore, the planning methods for the billet placement trajectory include: Based on the key work points in the billet feeding stage, a trajectory framework for the billet feeding stage is established.

[0045] Specifically, based on the key work points in the billet feeding stage, the following points are connected sequentially according to the work sequence: material preparation point, material collection point, material collection exit point, transfer intermediate point, forging chamber entrance point, material feeding preparation point, material feeding point, material feeding exit point, and material feeding avoidance point, thus establishing the trajectory framework for the billet feeding stage. This framework provides a complete spatial topology for the entire billet feeding trajectory, ensuring that the subsequent division and planning of local trajectories are carried out under clear point constraints.

[0046] In the trajectory framework, the local trajectories from the material preparation point to the forging chamber entrance and from the material unloading point through the material unloading exit point to the material unloading avoidance point are marked as the transfer planning segments of the billet unloading stage. The local trajectories from the forging chamber entrance through the material unloading preparation point to the material unloading point are marked as the precision planning segments of the billet unloading stage. This division method makes full use of the natural landmark of the forging chamber entrance. In the vacuum channel before this point, the robot arm only needs to move quickly without strict end-effector pose constraints. However, after entering the forging chamber, the robot arm end-effector needs to gradually approach the mold base and finally accurately place the billet in a preset posture, which has strict requirements on the path geometric accuracy and end-effector pose.

[0047] In the transfer planning section of the billet unloading stage, an S-shaped velocity curve is used for interpolation planning in the joint space to minimize the total operation time of the billet unloading stage, thereby obtaining the planning trajectory of the transfer planning section.

[0048] Joint space planning directly interpolates the displacements of each joint, eliminating the need for frequent inverse kinematics calculations. Furthermore, the acceleration and jerk of the joint trajectory are directly controllable, avoiding velocity fluctuations caused by singularities or joint limitations in Cartesian space planning. The S-shaped velocity curve, with its seven-segment design (acceleration, uniform acceleration, deceleration, uniform velocity, acceleration / deceleration, uniform deceleration, and deceleration), ensures the continuity of acceleration during joint movement, effectively suppressing impact and vibration during long-distance, rapid manipulator movements. This is crucial for maintaining the stability of the vacuum-sealed structure and extending the lifespan of the manipulator's transmission components.

[0049] In the precise planning segment of the billet placement stage, linear interpolation is used in Cartesian 3D coordinate space to minimize the pose error of the billet when placed on the mold base, resulting in the planned trajectory of the precise planning segment. The end of the precise planning segment must descend vertically along a straight path in the normal direction of the mold base surface to ensure the billet accurately falls into the mold base. If joint space planning is used, the geometry of the end path cannot be guaranteed, potentially causing lateral offset and placement deviation. By constructing a pose error cost function in Cartesian 3D space and using zero velocity and zero acceleration at the placement point as end constraints, the straight path is optimized. This ensures that the final precise planning segment trajectory is not only geometrically a definite straight line, but also that the velocity-time law of the manipulator's end gripper on the straight trajectory guarantees that the end gripper must be completely stationary at the moment of contact with the mold base, avoiding overshoot or billet bounce caused by inertial impact.

[0050] The planned trajectory of the transfer planning segment is transformed into a three-dimensional space through a kinematic model, so that the transformed planned trajectory of the transfer planning segment is unified with the planned trajectory of the precise planning segment in the same coordinate space.

[0051] At the common point between the planned trajectory of the transfer planning section and the planned trajectory of the precise planning section, a speed smooth transition method is used to connect the planned trajectory of the transfer planning section and the planned trajectory of the precise planning section, so as to obtain a complete billet release trajectory that is spatially continuous, speed continuous and without motion impact.

[0052] Specifically, the speed smooth transition method first determines the connection point between the transfer planning segment and the precise planning segment in the Cartesian 3D space of the 3D model. In the billet placement trajectory, this connection point is the forging chamber entrance point; in the forging removal trajectory, the connection point is the removal point. The connection point is both the end point of the transfer planning segment and the start point of the precise planning segment. At the connection point, or common point, the terminal Cartesian velocity vector obtained by mapping the transfer planning segment through the kinematic model is denoted as... The initial velocity vector at the end, determined by the precise planning segment, is denoted as... Since the transfer planning segment prioritizes motion efficiency, its end velocity at the connection point is usually high; while the precision planning segment aims to achieve high-precision pose control, its starting velocity is usually required to be low or zero. Therefore, there must be a velocity difference between the two.

[0053] The speed smooth transition method selects a preset length transition interval before and after the connection point, and re-plans a connection within the transition interval. and The continuous velocity curve ensures a smooth transition of the end velocity from the final velocity of the transfer planning section to the starting velocity of the precise planning section, while maintaining continuous acceleration without abrupt changes. In practical applications, a fifth-order polynomial or S-shaped velocity curve can be used to interpolate the velocity within the transition interval, using the displacement, velocity, and acceleration at the connection point as boundary conditions to solve for the velocity-time law of the transition interval. The length of the transition interval is adaptively determined based on the magnitude of the velocity difference and the maximum allowable acceleration and jerk constraints of the joint; the larger the velocity difference, the longer the transition interval, ensuring that the acceleration and jerk remain within preset limits. If the transition interval encroaches on the straight path of the precise planning section, the precise planning section is extended or the time is redistributed accordingly to ensure that the end velocity re-enters the original precise planning section at the specified starting velocity after the transition. After this processing, the transfer planning section and the precise planning section achieve a continuous transition of velocity and acceleration at the connection point, and the entire billet placement trajectory remains smooth in both spatial and temporal dimensions, eliminating the motion impact introduced by segmented planning.

[0054] It is important to note that trajectory planning in this invention is not merely the planning of path nodes. Path nodes only describe the sequence of geometric points traversed by the robot's end effector in three-dimensional space, indicating the robot's position at which it passes, without involving time information. Trajectory, on the other hand, is a combination of path and velocity-time patterns, indicating at what time the robot reaches which positions. The same straight path, paired with different velocity curves—for example, a rapid descent followed by a sudden stop, or a slow, uniform descent—will produce completely different motion effects and end-effector positioning accuracy. Therefore, the final result obtained by this invention is not simply a sequence of spatial path nodes, but a pose-time sequence, i.e., a complete description of the end effector's position and attitude at each discrete time point. This sequence can be directly mapped to the position commands of each joint servo controller within each control cycle through inverse kinematics.

[0055] The introduction of time variables stems from two needs: First, the joint movements of the robotic arm are strictly limited by the maximum speed, maximum acceleration, and maximum jerk. These limitations are essentially time-domain constraints, and cannot be defined or verified without the time dimension.

[0056] Secondly, the end-point constraint of the precise planning segment, where the end velocity and acceleration at the unloading point are zero, is a requirement regarding the velocity-time law. It stipulates that the end point must decelerate and come to a complete stop near the endpoint, rather than merely reaching a certain geometric position. Therefore, in both the S-shaped velocity curve planning of the transfer planning segment and the linear interpolation planning of the precise planning segment, the velocity-time curve must be explicitly generated and combined with the geometric path. After isochronous discrete sampling, a pose-time series is formed as the final output of the planning.

[0057] Furthermore, the trajectory planning method for the precise planning segment in the billet laying stage includes: In the 3D model, a straight path is planned along the normal direction of the mold base surface, so that the end effector of the robot moves from the material preparation point to the material placement point along the straight path. When placing the blank, it must fall vertically along the axis of the mold base. Any lateral deviation may cause the blank to scrape or get stuck with the side wall of the mold base. Therefore, the choice of the direction of the straight path directly determines the geometric feasibility of the placement action.

[0058] Constructing the pose error cost function for the precise planning segment in the billet laying stage ,in This refers to the position and orientation of the robotic arm's end effector gripper at the material unloading point. This refers to the target position and orientation of the blank at the mold base. These are the weighting coefficients.

[0059] The cost function of this invention quantifies placement accuracy as a weighted sum of squares of positional and orientation deviations, transforming placement accuracy, which was originally difficult to measure directly, into a calculable and optimizable mathematical indicator. The introduction of weighting coefficients allows for flexible adjustment based on the forging process's sensitivity to accuracy in different directions. For example, for thin-walled forgings, the orientation weight can be appropriately increased to ensure high-precision alignment between the billet axis and the die base axis.

[0060] With the goal of minimizing the cost function, under the constraints of maximum speed, maximum acceleration, joint displacement, and zero end speed and zero end acceleration at the unloading point, the straight path is optimized to find the optimal straight path that minimizes the cost function; the optimal straight path is then used as the planning trajectory of the precise planning segment.

[0061] The variables in the optimization process can include the spatial location of the pre-feeding point and the velocity-time relationship from the pre-feeding point to the feeding point. By iteratively adjusting these variables, the cost function gradually converges to its minimum value. Essentially, this optimization process seeks a descent path that minimizes the final placement deviation of the billet, while ensuring the physical reachability and motion safety of the robotic arm. The dual constraints of zero terminal velocity and zero acceleration ensure that the end effector is completely stationary upon contact with the mold base, eliminating the interference of inertial forces on placement accuracy and preventing the billet from bouncing or overshooting due to impact. Finally, the optimal straight path obtained from the optimization solution is used as the planned trajectory for the precision planning segment. This trajectory simultaneously satisfies the triple requirements of geometric straightness, motion smoothness, and high-precision terminal pose.

[0062] In some specific implementations, the specific process of solving the straight path optimization problem is as follows: First, define the optimization variables. This includes setting the spatial coordinates of the material preparation point. As the first set of optimization variables, their search range is limited to an allowable neighborhood near a preset height above the mold base. The boundary of this neighborhood is jointly determined by the workspace of the robot and the geometric constraints of the forging chamber. Simultaneously, the speed-time law parameter is used as the second set of optimization variables. When an S-shaped speed curve is adopted, the speed-time law parameter includes the maximum speed. Maximum acceleration and maximum jerk The adjustable range of the aforementioned velocity-time parameters is limited by the physical hardware driving capabilities of each joint. These parameters directly determine the shape of the velocity-time curve at the end of the straight path, thus affecting the overall trajectory length and dynamic quality.

[0063] Secondly, in each iteration, for the given pre-feeding point position and speed-time parameters, the calculation is as follows: The geometric equation of the straight path is determined by the pre-feeding point and the feeding point; the velocity-time curve of the end effector along the straight path is generated according to the velocity-time law parameters; the velocity-time curve is discretized to obtain the end effector pose-time sequence; the end effector pose-time sequence is mapped to the displacement-time sequence of each joint through the inverse kinematics model; it is checked whether the displacement of each joint exceeds the limit, whether the velocity of each joint exceeds the maximum velocity constraint, and whether the acceleration of each joint exceeds the maximum acceleration constraint; at the same time, it is checked whether the end effector velocity and acceleration at the feeding point meet the terminal constraint of zero; if any constraint is not met, the cost function value corresponding to the set of variables is set to the preset penalty value, so that it is automatically excluded in the optimization process.

[0064] Then, for a feasible solution that satisfies all constraints, the positional and attitude deviations between the end pose and the target pose at the release point are obtained by inverse kinematics of the kinematic model, and then substituted into the cost function to calculate the current generation value.

[0065] Finally, numerical optimization methods such as gradient descent or particle swarm optimization are used to iteratively update the pre-feeding point position and velocity-time behavior parameters within the search space of the optimization variables, gradually converging the cost function to its minimum. The convergence condition is that the change in the current generation value is below a preset threshold for a certain number of consecutive iterations, or the number of iterations reaches a preset upper limit. After convergence, the optimal pre-feeding point position and optimal velocity-time behavior parameters are output. The optimal straight-line path determined by these two parameters is the planned trajectory of the precise planning segment in the billet feeding stage.

[0066] The fifth step involves planning the robot's forging trajectory in the 3D model, with the goal of minimizing the pose error when picking up the forging from the mold base and minimizing the total operation time of the forging stage. This is done by utilizing the key working points of the forging stage.

[0067] The forging trajectory planning is symmetrical and complementary to the billet placement trajectory planning, both employing a segmented differentiated planning framework. However, due to the different action sequences and precision requirements of the forging task, the trajectory composition and optimization focus show significant differences. In the forging stage, the robot first enters the forging chamber from the standby position, aligns with the formed forging with high precision, and completes the clamping. Then, it clamps the forging, exits the forging chamber, and transfers it to the unloading point. This is the operational logic of precise clamping followed by rapid withdrawal, which is exactly the opposite of the billet placement stage's sequence of rapid transfer followed by precise placement. Therefore, the precise planning segment of the forging trajectory is set in the first segment, from the pre-pickup point to the pickup point, while the transfer planning segment is set in the second segment, from the pickup point through the pickup exit point and the intermediate pickup-transfer point to the unloading point.

[0068] Furthermore, the planning methods for obtaining the forging trajectory include: Based on the key work points in the forging stage, a trajectory framework for the forging stage is established.

[0069] This invention establishes a trajectory framework for the forging stage by sequentially connecting the key work points of the forging stage according to the operation sequence of the forging preparation point, the forging point, the forging exit point, the forging transfer intermediate point, and the unloading point. Within this trajectory framework, the local trajectory from the forging preparation point to the forging point is marked as the precise planning segment for the forging stage, with the core task of minimizing the positional deviation between the clamp and the forging holding part. The local trajectories from the forging point through the forging exit point and the forging transfer intermediate point to the unloading point are marked as transfer planning segments, with the goal of quickly and smoothly delivering the high-temperature forging.

[0070] In the trajectory framework, each local trajectory from the pick-up point through the pick-up exit point and the pick-up transfer intermediate point to the unloading point is marked as the transfer planning segment of the forging stage, and each local trajectory from the pick-up preparation point to the pick-up point is marked as the precise planning segment of the forging stage.

[0071] Compared to the billet placement stage, the forging removal stage experiences a significantly increased end-load due to the robotic arm already holding the forging, resulting in greater inertial forces during movement. Using the same acceleration and jerk parameters as in the no-load stage could lead to excessive joint drive torque or residual vibration, thus affecting the forging's surface quality. Therefore, when planning the S-shaped velocity curve for this stage, the maximum acceleration and jerk settings can be appropriately reduced based on load variations, resulting in a smoother velocity curve that minimizes load impact while maintaining motion efficiency.

[0072] In the transfer planning section of the forging stage, an S-shaped velocity curve is used for interpolation planning in the joint space to minimize the total operation time of the forging stage, and the planning trajectory of the transfer planning section is obtained. In the precise planning segment of the forging stage, linear interpolation planning is used in three-dimensional space to minimize the pose error when the forging is picked up from the mold base, and the planning trajectory of the precise planning segment is obtained. The planned trajectory of the transfer planning segment is transformed into a three-dimensional space through a kinematic model so that the transformed planned trajectory of the transfer planning segment is unified with the planned trajectory of the precise planning segment in the same coordinate space. At the common point between the planned trajectories of the transfer planning segment and the precise planning segment, a speed smooth transition method is used to connect the planned trajectories of the transfer planning segment and the precise planning segment to obtain the forging trajectory.

[0073] Furthermore, the trajectory planning method for the precise planning segment in the forging stage includes: In the 3D model, a straight path in 3D space is planned so that the end effector of the robot moves from the pre-picking point to the picking point along the straight path. During the movement, the posture of the end effector of the robot remains aligned with the normal of the forging clamping surface. Only when the clamping direction coincides with the normal of the forging clamping surface can the clamping force be uniformly applied to the surface of the forging, thus avoiding additional bending moment or point contact damage caused by posture deviation.

[0074] Constructing the pose error cost function for the precise planning segment in the forging stage ,in The position and orientation of the clamping points on the forging. This refers to the position and orientation of the end gripper when it reaches the pick-up point. These are the weighting coefficients.

[0075] This cost function quantifies the clamping accuracy into an optimizable mathematical form. By iteratively adjusting the position and velocity-time characteristics of the part-retrieving preparation point, and under the constraints of joint limits and zero end-effector velocity and zero acceleration at the part-retrieving point, the optimal straight-line path that minimizes the cost function is found. The zero-velocity and zero-acceleration constraint at the end-effector ensures that the clamp is completely stationary at the moment of contact with the forging, eliminating clamping misalignment caused by inertial slippage.

[0076] With the goal of minimizing the cost function, under the constraints of maximum speed, maximum acceleration, joint displacement limit, and zero end velocity and zero end acceleration at the pick-up point, the straight path is optimized to find the optimal straight path that minimizes the cost function. The optimal straight path is used as the planning trajectory for the precise planning segment.

[0077] Since the transfer planning segment is generated in joint space and the precise planning segment is generated in Cartesian 3D space, the two need to be unified into the 3D model space and connected. During the forging removal stage, the connection point is located at the removal point; the precise planning segment ends at this point, and the transfer planning segment starts at this point. After mapping the planned trajectory of the transfer planning segment in joint space to Cartesian 3D space through a kinematic model, a smooth velocity transition method is adopted at the removal point. A continuous velocity curve is planned for the transition interval between the end velocity (zero speed) at the end of the precise planning segment and the end velocity (acceleration from zero) at the beginning of the transfer planning segment, ensuring that the acceleration does not jump. This process ensures a smooth and shock-free start-up process for the robot arm after clamping the forging, avoiding instantaneous fluctuations in clamping force or slippage of the forging due to sudden acceleration.

[0078] Finally, the combined precise planning segment and the transfer planning segment are used to obtain a complete trajectory for retrieving the forging. Spatially, this trajectory precisely approaches the retrieving point in a straight line from the retrieving preparation point, then exits along a safe path and is quickly transferred to the unloading point. Temporally, it satisfies the continuity of speed and acceleration in each segment, and remains stationary at the moment of clamping, thus achieving efficient and reliable removal of the forging after forging while ensuring clamping accuracy.

[0079] The optimization process of the precise planning segment in the forging stage is similar to that in the billet laying stage. It also uses the spatial position and velocity-time law parameters of the forging preparation point as optimization variables. In each iteration, the pose-time sequence of the end along the straight path is generated, and each constraint condition is checked.

[0080] The specific process of optimizing the straight path of the forging trajectory is as follows: First, define the optimization variables. This includes the spatial coordinates of the pickup preparation point. As the first set of optimization variables, their search range is limited to an allowable neighborhood near a preset height above the mold base. The boundary of this neighborhood is jointly determined by the workspace of the robot and the geometric constraints of the forging chamber. Simultaneously, the speed-time law parameter is used as the second set of optimization variables; when an S-shaped speed curve is adopted, the speed-time law parameter includes the maximum speed. Maximum acceleration and maximum jerk The adjustable range of the above parameters is limited by the physical hardware driving capability of each joint. Furthermore, a low-speed approach segment is set at a preset distance near the pickup point, and the end speed of this segment is... It is also used as one of the optimization variables, and its value ranges from zero to normal movement speed.

[0081] Secondly, in each iteration, for the currently given pick-up preparation point position, velocity-time law parameters, and low-speed approximation segment parameters, the following calculations are performed: the geometric equation of the straight path is determined by the pick-up preparation point and the pick-up point, and the starting point of the low-speed approximation segment is marked on the straight path at a preset distance; the velocity-time curve of the end along the straight path is generated according to the velocity-time law parameters, wherein the end velocity in the low-speed approximation segment decreases from the normal value to the preset low value. The remaining stroke is completed at a constant speed or with gradual deceleration. Discrete sampling of the velocity-time curve yields the end-effector pose-time sequence, which maintains alignment between the end-effector pose and the normal to the forging clamping surface throughout the entire stroke. The end-effector pose-time sequence is mapped to the displacement-time sequence of each joint using an inverse kinematics model. The displacement of each joint is checked to ensure it does not exceed the limit, the velocity of each joint exceeds the maximum velocity constraint, and the acceleration of each joint exceeds the maximum acceleration constraint. Simultaneously, the end-effector velocity and acceleration at the pick-up point are checked to ensure they meet the zero end-effector constraint, and the clamp closing direction coincides with the normal to the forging clamping surface. If any constraint is not met, the cost function value corresponding to that set of variables is set to a preset penalty value, which is then automatically excluded during the optimization process.

[0082] Then, for a feasible solution that satisfies all constraints, the positional and attitude deviations between the actual pose of the clamp at the pick-up point and the target pose of the preset clamping point of the forging are obtained by inverse kinematics, and then substituted into the cost function to calculate the current generation value.

[0083] Finally, numerical optimization methods such as gradient descent or particle swarm optimization are used to iteratively update the pre-receiving point position, velocity-time law parameters, and low-speed approximation segment parameters within the search space of the optimization variables, so that the cost function gradually converges to the minimum value. The convergence condition is that the change in the current generation value is lower than a preset threshold for a certain number of consecutive iterations, or the number of iterations reaches a preset upper limit. After convergence, the optimal pre-receiving point position, optimal velocity-time law parameters, and optimal low-speed approximation segment parameters are output. The optimal straight-line path determined by these parameters is the planned trajectory of the precise planning segment in the forging stage.

[0084] This optimization process, while ensuring the safe and reachable movement of the robotic arm, seeks the approximation path that minimizes the deviation between the gripper and the forging clamping area by adjusting the position of the pre-grabbing point and the descent speed distribution. The low-speed approximation segment provides sufficient response time for visual perception and pose compensation, ensuring that the robotic arm's end effector is perfectly aligned with the target at the final gripping moment. The dual constraints of zero terminal velocity and zero acceleration guarantee that the robotic arm's end effector is absolutely stationary when the gripper closes, eliminating the interference of inertial slippage on gripping accuracy.

[0085] In some specific implementations, a velocity curve obtained through constraint optimization is superimposed on a three-dimensional straight path determined by key points.

[0086] For the precise planning segment of the blanking stage, the path is a straight line from the blanking preparation point to the blanking point, and the end posture gradually changes according to the alignment requirements of the mold base axis; its trajectory is expressed as follows: , where path parameters Due to the pose error cost function The solution is obtained by minimizing the objective and satisfying the constraints that the end velocity and acceleration at each joint limit and the unloading point are both zero.

[0087] For the precise planning segment of the forging removal stage, the path is a straight line from the pre-removal point to the removal point, and the end posture remains aligned with the normal of the forging clamping surface throughout the entire process; its trajectory is expressed as follows: Among them, the velocity curve By minimizing the cost function The results were obtained through optimization under joint motion constraints and zero-speed / zero-acceleration conditions at the pick-up point. Both outputs are ultimately isochronous discrete pose time series, which can be directly used for joint servo control.

[0088] Furthermore, the planning methods for the trajectory of the transfer planning segment include: Determine the start and end points of the planned transfer segment, and calculate the displacement of each joint of the robotic arm from the start to the end point; Based on the displacement of each joint and the constraints of maximum velocity, maximum acceleration, and maximum jerk, an S-shaped velocity curve is synchronously planned for each joint in joint space, where: For each joint of the robot, the duration of each stage of the S-shaped velocity curve is automatically solved based on the displacement of each joint and the constraints of maximum speed, maximum acceleration and maximum jerk. The S-shaped velocity curve consists of jerk segment, uniform acceleration segment, deceleration segment, uniform speed segment, acceleration and deceleration segment, uniform deceleration segment and deceleration and deceleration segment. When the displacement of the joint is insufficient to make the velocity reach the maximum speed, the S-shaped velocity curve degenerates into an S-shaped velocity curve without a uniform velocity segment. When the displacement of the joint is shortened and the maximum acceleration cannot be achieved, the S-shaped velocity curve degenerates into a triangular acceleration curve. Using the joint with the longest required motion time as the main synchronization axis, the S-shaped velocity curves of the remaining joints are scaled in time to ensure the coordination of the movement of each joint of the robot. The planned trajectory of the transport planning segment is obtained by isochronous discrete sampling of the S-shaped velocity curves generated by each joint.

[0089] The billet placement stage comprises two transfer planning sections. The first transfer planning section starts at the material preparation point and ends at the forging chamber entrance. The second transfer planning section starts at the material placement point and ends at the material placement avoidance point. The forging removal stage consists of a single transfer planning section, starting at the part removal point and ending at the unloading point. Although auxiliary path points such as exit points and intermediate points are interspersed within each transfer planning section, the entire transfer section is considered as a continuous motion interval with the starting and ending points as boundary conditions in the S-shaped velocity curve planning of the joint space, thereby ensuring the synchronicity and efficiency of the joint movements.

[0090] In some specific implementations, the specific solution process for the S-shaped velocity curve is as follows: Obtain the starting joint coordinate vector of the transfer planning segment. and endpoint joint coordinate vector and the preset maximum speed of each joint Maximum acceleration and maximum jerk subscript Indicates the first Each joint has a displacement from the start point to the end point. .

[0091] Plan the initial S-shaped velocity curve independently for each joint. For the first joint... Each joint's S-shaped velocity curve consists of seven stages: acceleration (constant positive Jerk), uniform acceleration (constant positive acceleration), deceleration (constant negative Jerk), uniform velocity (zero acceleration), acceleration / deceleration (constant negative Jerk), uniform deceleration (constant negative acceleration), and deceleration / deceleration (constant positive Jerk). The duration of each stage is calculated as follows: First, determine the displacement of the joint. Is it sufficient to make the speed curve reach the preset maximum speed? The judgment condition is: the joint starts from rest and accelerates to its maximum speed. and maximum jerk Accelerate to Required acceleration displacement Is the sum of the displacements required for the symmetrical deceleration phase less than or equal to... The displacement required to accelerate to maximum speed consists of three segments: an acceleration segment, a uniform acceleration segment, and a deceleration segment, with the durations of each segment being as follows: , , The sum of the three acceleration displacements is The displacement of the symmetrical deceleration section is equal to that of the symmetrical deceleration section.

[0092] like This indicates that the displacement is large enough that the velocity curve can reach the maximum velocity and include a uniform velocity segment, the duration of which is... .

[0093] like This indicates that the displacement is insufficient to make the velocity reach [the desired value]. The velocity curve degenerates into an S-shaped curve without a uniform velocity segment. At this point, the maximum speed... The solution needs to be recalculated to maintain maximum acceleration. and maximum jerk Keeping it unchanged, so that the acceleration and deceleration segments are symmetrically distributed and the total displacement is exactly equal to Solve for the actual peak speed And recalculate the corresponding time segments.

[0094] like It shortens further, so that the joint has not yet reached its full potential. When deceleration needs to begin, the velocity curve degenerates into a triangular acceleration curve. At this point, it is necessary to... As constraints, establish simultaneous jerk constraints. Solve for the actual maximum acceleration and actual peak speed And redetermine the time periods.

[0095] Because the displacement of each joint is different, the total motion time of each joint is obtained through independent planning. They also vary. To ensure that all joints start simultaneously and reach their respective endpoints simultaneously, the joint with the longest total motion time among all joints is selected as the primary synchronization axis, and its motion time is denoted as... For the remaining joints, maintain their respective displacements. Unchanged, with To maintain a uniform total motion time, the duration of each stage is synchronously scaled according to the shape proportions of the S-shaped velocity curve. During scaling, the maximum speed and maximum acceleration settings are reduced first, so that the time of each segment is proportionally lengthened until the total time equals the maximum speed. At the same time, it ensures that the speed and acceleration after stretching do not exceed the original preset maximum values. After scaling, the trajectories of all joints have the same total time, ensuring the coordination of the movements of each joint of the robotic arm.

[0096] With a fixed control cycle The synchronized S-shaped velocity curves of each joint are sampled discretely at equal intervals. For each sampling time... ,in Based on the specific stage of the S-curve at that moment, the instantaneous displacement values ​​of each joint are calculated using the displacement formulas for each stage. Finally, the position-time sequence of each joint is obtained, which constitutes the planned trajectory of the transport planning segment. This position-time sequence can be directly used as the input command for the joint servo controller.

[0097] This invention divides the trajectory of billet placement and forging pick-up into a transfer planning segment and a precision planning segment, and adopts a differentiated planning strategy. In the transfer planning segment, an S-shaped velocity curve in joint space ensures efficient and stable long-distance movement. In the precision planning segment, Cartesian linear interpolation is performed with the goal of minimizing pose error, and zero-speed and zero-acceleration constraints are applied at the end. At the same time, the planning results of the two spaces are uniformly converted to the three-dimensional model coordinate system and the velocity between segments is smoothly connected. Thus, under the same planning framework, the long-stroke transfer efficiency of the vacuum isothermal forging robot and the high-precision control of the end pose of billet placement and forging pick-up at key points are taken into account, which effectively improves the forging forming accuracy and operation consistency.

[0098] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and not to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. All such modifications or substitutions should be covered within the protection scope of this application, and should not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application. Therefore, the protection scope of this application should be determined by the protection scope of the claims.

Claims

1. A trajectory planning method for a robotic arm under vacuum isothermal forging conditions, characterized in that, Includes the following steps: Obtain the three-dimensional structural information of the vacuum isothermal forging furnace and construct a three-dimensional model that includes at least the forging chamber, vacuum channel, heating furnace, and mold base; Establish a kinematic model of the robot to determine the transformation relationship between the joint space coordinates of the robot and the three-dimensional space coordinates of the three-dimensional model; Based on the billet feeding stage and the forging stage in the forging operation of the vacuum isothermal forging furnace, the key working points of the robot in the billet feeding stage and the key working points of the robot in the forging stage are determined respectively. In the three-dimensional model, with the goal of minimizing the pose error when the blank is placed on the mold base and minimizing the total operation time of the blank placement stage, the blank placement trajectory of the robot is planned using the key working points of the blank placement stage. In the three-dimensional model, with the goal of minimizing the pose error when picking up the forging from the mold base and minimizing the total operation time of the forging picking stage, the forging picking trajectory of the robot is planned using the key working points of the forging picking stage.

2. The trajectory planning method for a robotic arm under vacuum isothermal forging environment according to claim 1, characterized in that, Methods for establishing a kinematic model of a robotic arm to determine the transformation relationship between the joint space coordinates of the robotic arm and the three-dimensional space coordinates of the three-dimensional model include: Based on the mechanical structure of the manipulator, the homogeneous transformation matrix between the coordinates of each adjacent joint of the manipulator is established using the DH parameter method. The mechanical structure includes a traveling mechanism, a lifting mechanism, a rotating mechanism, a first-stage telescopic mechanism, a second-stage telescopic mechanism, and an end effector. The robot arm is represented in joint space to obtain the joint space coordinate vector of the robot arm. ,in This represents the displacement of the traveling mechanism along the track. This refers to the vertical displacement of the lifting mechanism. The rotation angle of the rotary mechanism. and These are the extension lengths of the first-stage telescopic mechanism and the second-stage telescopic mechanism, respectively. By multiplying the homogeneous transformation matrices of adjacent joint coordinates, a kinematic model representing the transformation relationship between joint spatial coordinates and Cartesian three-dimensional spatial coordinates is obtained. The kinematic model is as follows: ; in, Let be the homogeneous transformation matrix of the spatial coordinates of the traveling mechanism relative to the three-dimensional spatial coordinates. Let be the homogeneous transformation matrix of the spatial coordinates of the lifting mechanism relative to the spatial coordinates of the traveling mechanism. Let be the homogeneous transformation matrix of the spatial coordinates of the rotary mechanism relative to the spatial coordinates of the lifting mechanism. Let be the homogeneous transformation matrix of the spatial coordinates of the first-stage telescopic mechanism relative to the spatial coordinates of the rotary mechanism. Let be the homogeneous transformation matrix of the spatial coordinate system of the second-stage telescopic mechanism relative to the spatial coordinate system of the first-stage telescopic mechanism. Let be the homogeneous transformation matrix of the spatial coordinates of the end gripper relative to the spatial coordinates of the second-stage telescopic mechanism. is the homogeneous transformation matrix of the spatial coordinates of the end gripper relative to the three-dimensional spatial coordinates.

3. The trajectory planning method for a robotic arm under vacuum isothermal forging environment according to claim 2, characterized in that, The key working points in the billet feeding stage include at least: material preparation point, material feeding point, material feeding exit point, transfer intermediate point, forging chamber entrance point, material feeding preparation point, material feeding point, material feeding exit point, and material feeding avoidance point.

4. The trajectory planning method for a robotic arm under vacuum isothermal forging environment according to claim 3, characterized in that, The key working points in the forging stage include at least: the forging preparation point, the forging point, the forging exit point, the intermediate point for forging transfer, and the unloading point.

5. The trajectory planning method for a robot arm in a vacuum isothermal forging environment according to claim 4, characterized in that, The method for planning the billet feeding trajectory includes: Based on the key working points of the billet feeding stage, a trajectory framework for the billet feeding stage is established. In the trajectory framework of the billet feeding stage, each local trajectory from the material preparation point to the forging chamber entrance point and each local trajectory from the feeding point through the feeding exit point to the feeding avoidance point are marked as the transfer planning segment of the billet feeding stage, and each local trajectory from the forging chamber entrance point through the feeding preparation point to the feeding point is marked as the precise planning segment of the billet feeding stage. In the transfer planning section of the billet unloading stage, an S-shaped velocity curve is used for interpolation planning in the joint space to minimize the total operation time of the billet unloading stage, thereby obtaining the planning trajectory of the transfer planning section. In the precise planning segment of the blank placement stage, linear interpolation is used in Cartesian three-dimensional coordinate space for planning, with the optimization objective of minimizing the pose error when the blank is placed on the mold base, to obtain the planning trajectory of the precise planning segment. The planned trajectory of the transfer planning segment is transformed into a three-dimensional space through the kinematic model, so that the planned trajectory of the transformed transfer planning segment is unified with the planned trajectory of the precise planning segment in the same coordinate space; At the common point between the planned trajectory of the transfer planning segment and the planned trajectory of the precise planning segment, a speed smooth transition method is used to connect the planned trajectory of the transfer planning segment and the planned trajectory of the precise planning segment to obtain the billet release trajectory.

6. The trajectory planning method for a robotic arm under vacuum isothermal forging environment according to claim 5, characterized in that, The method for planning the trajectory of the forging includes: Based on the key working points of the forging stage, a trajectory framework for the forging stage is established. In the trajectory framework of the forging stage, each local trajectory from the forging point through the forging exit point and the intermediate point of the forging transfer to the unloading point is marked as the transfer planning segment of the forging stage, and each local trajectory from the forging preparation point to the forging point is marked as the precise planning segment of the forging stage. In the transfer planning segment of the forging stage, an S-shaped velocity curve is used for interpolation planning in the joint space to minimize the total operation time of the forging stage, thereby obtaining the planned trajectory of the transfer planning segment. In the precise planning segment of the forging stage, linear interpolation is used in three-dimensional space to minimize the pose error when the forging is picked up from the mold base, and the planning trajectory of the precise planning segment is obtained. The planned trajectory of the transfer planning segment is transformed into a three-dimensional space through the kinematic model, so that the transformed planned trajectory of the transfer planning segment is unified with the planned trajectory of the precise planning segment in the same coordinate space; At the common point between the planned trajectory of the transfer planning segment and the planned trajectory of the precise planning segment, a speed smooth transition method is used to connect the planned trajectory of the transfer planning segment and the planned trajectory of the precise planning segment to obtain the forging trajectory.

7. The trajectory planning method for a robotic arm under vacuum isothermal forging environment according to claim 5, characterized in that, The trajectory planning method for the precise planning segment in the billet feeding stage includes: In the three-dimensional model, a three-dimensional straight path is planned along the normal direction of the mold base table, so that the end gripper of the robot arm moves from the material preparation point to the material release point along the straight path; Constructing the pose error cost function for the precise planning segment in the billet laying stage ,in The position and orientation of the end clamp at the discharge point. This refers to the target position and orientation of the blank at the mold base. These are the weighting coefficients; With the goal of minimizing the cost function, the straight path is optimized under the constraints of zero maximum speed, maximum acceleration, joint displacement, and zero end speed and zero end acceleration at the unloading point. The optimal straight path that minimizes the cost function is then obtained. The optimal straight path is used as the planned trajectory of the precise planning segment.

8. The trajectory planning method for a robot arm in a vacuum isothermal forging environment according to claim 6, characterized in that, The trajectory planning method for the precise planning segment in the forging stage includes: In the three-dimensional model, a straight path in three-dimensional space is planned so that the end effector of the robot moves from the part preparation point to the part picking point along the straight path, and the posture of the end gripper remains aligned with the normal of the forging clamping surface during the movement. Constructing the pose error cost function for the precise planning segment in the forging stage ,in The position and orientation of the clamping points on the forging. The position and orientation of the end gripper when it reaches the pick-up point. These are the weighting coefficients; With the goal of minimizing the cost function, the straight path is optimized under the constraints of maximum speed, maximum acceleration, joint displacement limit, and zero end velocity and zero end acceleration at the pick-up point. The optimal straight path that minimizes the cost function is then obtained. The optimal straight path is used as the planned trajectory of the precise planning segment.

9. The trajectory planning method for a robot arm in a vacuum isothermal forging environment according to claim 6, characterized in that, The planning methods for the trajectory of the transfer planning segment include: Determine the start and end points of the planned transfer segment, and calculate the displacement of each joint of the robotic arm from the start to the end point; Based on the displacement of each joint and the constraints of maximum velocity, maximum acceleration, and maximum jerk, an S-shaped velocity curve is synchronously planned for each joint in joint space, wherein: For each joint of the robot, under the constraints, the duration of each stage of the S-shaped velocity curve is automatically calculated. The S-shaped velocity curve consists of an acceleration segment, a uniform acceleration segment, a deceleration segment, a uniform speed segment, an acceleration / deceleration segment, a uniform deceleration segment, and a deceleration / deceleration segment. When the displacement of the joint is insufficient to make the velocity reach the maximum velocity, the S-shaped velocity curve degenerates into an S-shaped velocity curve without a uniform velocity segment. When the displacement of the joint is shortened and the maximum acceleration cannot be achieved, the S-shaped velocity curve degenerates into a triangular acceleration curve. Using the joint with the longest required motion time as the main synchronization axis, the S-shaped velocity curves of the remaining joints are scaled in time to ensure the coordination of the movement of each joint of the robot. The planned trajectory of the transport planning segment is obtained by isochronous discrete sampling of the S-shaped velocity curves generated by each joint.

Images