Industrial robot real-time obstacle avoidance trajectory planning method based on dynamic environment modeling

Abstract

The application relates to the technical field of robot trajectory planning, and particularly discloses a trajectory planning method based on dynamic environment modeling. The method is applied to a cooperative workstation and comprises the following steps: acquiring initial phase lag and spatial intersection anchor points; establishing an independent running logic domain to reconstruct cooperative orientation characteristics; performing reverse kinematics analysis along a time axis to obtain constraint orientation parameters; converting trajectory tangential deviation into initial time sequence compensation and performing flexible limiting regulation to output optimal phase delay compensation; extracting time sequence residual extreme points between the initial compensation and the optimal compensation and converting the time sequence residual extreme points into spatial trajectory reduction parameters; combining a safety tolerance corridor to reconstruct a spatial shortcut; and completing synchronization handover of master-slave control domains. The application converts time sequence errors generated by limiting regulation into shortcut adjustments in the spatial dimension, reduces pose dislocation caused by insufficient time sequence compensation, and realizes smooth cooperation recovery of master-slave robots.

Description

Technical Field

[0001] This invention relates to the field of robot trajectory planning technology, specifically to a method for real-time obstacle avoidance trajectory planning of industrial robots based on dynamic environment modeling. Background Technology

[0002] In a collaborative workstation with multiple robots, the master and slave robotic arms are bound by a kinematic base coordinate system. When the first industrial robot, which is in charge of the workpiece pose, is forced to perform online obstacle avoidance due to random environmental obstacles, it will inevitably deviate from the predetermined process trajectory, resulting in additional physical detour distance and time loss, thereby directly disrupting the original collaborative cycle of the system.

[0003] To restore collaborative operation, existing technologies attempt to compensate for the phase lag of the first robot by adjusting the running speed of the second industrial robot in real time. However, compensation strategies relying on time-domain control require the second industrial robot to make up for the time difference through drastic acceleration and deceleration when the obstacle avoidance lag of the first robot is too large. This can easily exceed the rated maximum allowable acceleration and impact limits of its underlying servo joints.

[0004] In actual industrial control logic, to prevent mechanical damage, the underlying servo system will inevitably trigger a hardware protection mechanism, forcibly limiting the speed compensation command. This limiting action results in a lack of time compensation, creating a timing residual extreme that cannot be absorbed in the time dimension. This uncompensated timing deviation will cause the second industrial robot to fail to accurately reach the preset absolute spatial intersection anchor point at the theoretical synchronization moment, leading to spatial position misalignment and dynamic rigidity impact during the moment of master-slave control rebinding, ultimately causing the collaborative reconfiguration after obstacle avoidance to fail.

[0005] To this end, the present invention provides a method for real-time obstacle avoidance trajectory planning of industrial robots based on dynamic environment modeling. Summary of the Invention

[0006] The purpose of this invention is to provide a real-time obstacle avoidance trajectory planning method for industrial robots based on dynamic environment modeling, which is beneficial to maintaining the dynamic response stability of the underlying control system and the service life of the physical mechanism.

[0007] The objective of this invention can be achieved through the following technical solutions: A method for real-time obstacle avoidance trajectory planning for industrial robots based on dynamic environment modeling includes the following steps: Collaborative marking is performed on industrial robots in the collaborative workstation. The collaborative running trajectory of the first industrial robot is extracted and path interference analysis is performed on the dynamic environment. The individual trajectory change sequence that deviates from the predetermined trajectory is extracted and the time loss is quantified to obtain the initial phase lag and spatial intersection anchor point. An independent operating logic domain for the second industrial robot is established; and based on the independent operating logic domain, combined with the initial phase lag and the absolute space intersection anchor point, master-slave synchronization reconstruction is performed to obtain the cooperative orientation characteristics of the second industrial robot. Inverse kinematic analysis along the time axis is performed on the cooperative orientation features to obtain the inverse trajectory segments that maintain the intersection condition; the inverse trajectory segments are mapped and aligned to the current system absolute clock to obtain the constraint orientation parameters of the second industrial robot theory. The trajectory tangential deviation between the theoretical constraint orientation parameters and the real-time position and attitude data is extracted and time-distance equivalent analysis is performed to obtain the initial timing compensation amount. Combined with the kinematic extreme boundary of the robotic arm, the initial timing compensation amount is subjected to compliant amplitude limiting control to obtain the optimal phase delay compensation amount.

[0008] Furthermore, the path interference analysis is performed as follows: Extract the kinematic master-slave dependency relationship of the collaborative workstation, mark the robot arm that dominates the global pose and has an independent base coordinate system as the first industrial robot, and mark the robot arm whose base coordinate system is dynamically subordinate to the first industrial robot as the second industrial robot. Construct a dynamic 3D model of the current workstation environment, combine it with the body geometry of the first industrial robot, generate a swept spatial point cloud along the cooperative running trajectory, and perform spatial collision and intersection comparison with the dynamic 3D model of the environment. If a geometrically overlapping region is detected on the expected travel path, path interference is determined to have occurred, and the operation of extracting the individual trajectory change sequence is triggered.

[0009] Furthermore, the method for extracting the single-unit orbital change sequence is as follows: Using the physical extension safety points at both ends of the determined path interference point as the reconstruction boundary, a detour trajectory is generated within the non-interference safety area of ​​the dynamic environment 3D model; Extract the extreme values ​​of the normal offset and the spatial curvature change data of the bypass trajectory relative to the cooperative trajectory, as spatial reconstruction features; Based on spatial reconstruction characteristics, discrete interpolation is performed on the bypass trajectory by combining the dynamic extreme value constraint parameters of the first industrial robot, generating a continuous data set containing the physical coordinates and transient motion associated states of each trajectory node, thus obtaining the single-unit trajectory change sequence.

[0010] Furthermore, the process of performing the master-slave synchronization reconstruction is as follows: Extract the absolute arrival timestamp contained in the absolute spatial intersection anchor point, and extract the logical synchronization time corresponding to the second industrial robot in the original process plan; Retrieve the theoretical planning node for the logical synchronization moment of the second industrial robot; Extract the three-dimensional spatial coordinate components and tool axis attitude components of the theoretical planning node in the independent operation logic domain of the second industrial robot, and encapsulate the attribute association between the three-dimensional spatial coordinate components and the tool axis attitude components. Based on the associated encapsulated three-dimensional spatial coordinate components and tool axial attitude components, an ideal spatial state benchmark that the second industrial robot must reach at the moment of future re-intersection is constructed, and the cooperative orientation characteristics of the second industrial robot target are established.

[0011] Furthermore, the mapping alignment is performed as follows: After generating the reverse trajectory segment, obtain the absolute system clock of the current moment inside the collaborative workstation; Perform a time matching search between the current absolute system clock and the absolute discrete timestamp sequence attached to the reverse trajectory segment; Retrieve and lock the target preceding space node that matches the absolute system clock at the current moment from the reverse trajectory segment; Extract the three-dimensional coordinate data and end-point orientation data contained within the target's preceding spatial node; The three-dimensional coordinate data and the end-effector orientation data are associated and encapsulated, and defined as the theoretical constraint orientation parameters of the second industrial robot.

[0012] Furthermore, the method for generating the reverse trajectory segment is as follows: Retrieve the absolute arrival timestamp contained in the absolute spatial intersection anchor point; Within the independent operating logical domain, a timestamp sequence is constructed based on the time attributes of the spatial intersection anchor points; based on the time stamp association between the timestamp sequence and the system real-time clock, the theoretical constraint orientation of the second industrial robot is determined; The cooperative orientation feature is set as the initial starting boundary of the negative discrete time series. The reverse trajectory generator is started to perform reverse discrete time slice backtracking processing to obtain the reverse trajectory segment that maintains the intersection condition.

[0013] Furthermore, the method for performing the time-distance equivalence analysis is as follows: Simultaneously collect real-time position and attitude data of the second industrial robot; Extract the three-dimensional coordinate data from the theoretical constraint orientation parameters, and extract the real-time position data from the real-time position and attitude data, and calculate the spatial position vector difference between the two. By projecting the spatial position vector difference onto the instantaneous tangent direction of the inverse reference trajectory, the trajectory tangential deviation reflecting the difference in master-slave coordination progress is separated. Based on the motion state characteristics of the second industrial robot, the trajectory tangential deviation is equivalently converted into the hysteresis duration in the time dimension; The hysteresis duration is used as the time base parameter for adjusting the local virtual clock tick, thus establishing the initial timing compensation amount.

[0014] Furthermore, it also includes: Extract the timing residual poles between the initial and optimal phase delay compensation amounts, and transform the timing residual poles into spatial trajectory reduction parameters; extract the safety tolerance corridor and combine it with the spatial trajectory reduction parameters to perform spatial shortcut reconstruction processing, and complete the synchronous handover of the master and slave control domains.

[0015] Furthermore, the spatial shortcut reconstruction process is performed as follows: Extract the intermediate transition nodes contained in the remaining reverse trajectory segment along the second industrial robot's journey towards the absolute space intersection anchor point. With the goal of compensating for the time-series residual poles, the remaining trajectory is geometrically reconstructed within the safety tolerance corridor boundary to generate a spatial shortcut trajectory. Drive the second industrial robot along the spatial shortcut trajectory to the absolute spatial intersection anchor point, and stop calling the second industrial robot's local virtual clock at the absolute spatial intersection anchor point; Switch the control drive source of the second industrial robot back to the global synchronization clock of the collaborative workstation to achieve synchronous handover of the master and slave control domains.

[0016] Furthermore, the three-dimensional boundary of the tolerance corridor is established as follows: Spatial distances between each intermediate transition node and the dynamic environment 3D model are measured and calculated to obtain the maximum safe redundancy radius around each intermediate transition node that does not physically interfere with obstacles. A safety tolerance corridor is established by spatially enveloping the maximum safety redundancy radii of each consecutive intermediate transition node.

[0017] The beneficial effects of this invention are as follows: 1. By comparing the collaborative running trajectory with the 3D model of the dynamic environment through spatial collision and intersection calculations, and generating a detour sequence with physical extreme value constraints along the safety points of the entity, the kinematic feasibility of the obstacle avoidance path is realized. Extracting the initial phase lag of the collaborative workstation and binding the absolute timestamp with the spatial regression coordinates to construct spatial intersection anchor points facilitates the identification of spatiotemporal deviations in the master-slave robot interference state. Establishing an independent operating logic domain for the second industrial robot helps to sever the rigid master-slave binding in abnormal states and reduces blind following and uncontrollable spatial displacement when the robot performs obstacle avoidance actions. Constructing collaborative orientation features facilitates the pre-anchoring of the spatial state benchmark that must be reached to resume collaborative operations in the virtual computing space.

[0018] 2. By using the cooperative orientation features of the target as the starting boundary and constructing a negative discrete-time series in the independent operating logic domain, the inversion operation of the physical nominal limit value is introduced to construct a symmetric physical constraint boundary. The reverse discrete-time slice backtracking process is performed to generate reverse trajectory segments, which is beneficial to improve the convergence efficiency of trajectory search in high-dimensional space and maintain the smooth transition conditions of the kinematic curve.

[0019] 3. By calculating the spatial position vector difference between the theoretical constraint orientation parameters and real-time physical data, and projecting it onto the instantaneous tangential direction of the inverse reference trajectory, the trajectory tangential deviation reflecting the difference in collaborative progress is separated, which helps to reduce the interference of non-tangential spatial displacement on the collaborative progress assessment. The allowable extreme values ​​of the underlying servo are established as constraint boundaries. When an over-limit is detected, a compliant protection mechanism is triggered to forcibly truncate and dynamically limit the initial timing compensation amount, outputting the optimal phase delay compensation amount within the safe range. This helps to mitigate the transient dynamic shock of joints caused by full compensation, maintaining the dynamic response stability of the underlying control system and the operational life of the physical mechanism. 5. By extracting intermediate transition nodes along the remaining trajectory and calculating the maximum safety redundancy radius using dynamic environment graphs, a non-interference safety tolerance corridor is established. Without exceeding the 3D boundary, the spatial fusion chamfer is enlarged with the goal of absorbing spatial trajectory reduction parameters, and corner optimization is performed to generate a shortcut trajectory with reduced path length. The physical reduction of the spatial redundancy path is used to offset the timing delay caused by servo limiting, enabling the second industrial robot to smoothly connect to the intersection anchor point and restore global synchronization clock drive, which is beneficial for completing the handover of the main control logic domain. Attached Figure Description

[0020] The invention will now be further described with reference to the accompanying drawings.

[0021] Figure 1 This is a flowchart of the real-time obstacle avoidance trajectory planning method for industrial robots based on dynamic environment modeling, as described in this invention. Figure 2 This is a flowchart of the process for evaluating whether an instantaneous dynamic response will exceed the kinematic extreme boundary in this invention; Figure 3 This is a functional module diagram of the industrial robot real-time obstacle avoidance trajectory planning system based on dynamic environment modeling in this invention. Detailed Implementation

[0022] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0023] Example 1

[0024] like Figure 1As shown, a real-time obstacle avoidance trajectory planning method for industrial robots based on dynamic environment modeling is applied to a collaborative workstation containing a first industrial robot and a second industrial robot, and includes the following steps: S10: Collaboratively mark the industrial robots in the collaborative workstation, extract the collaborative running trajectory of the first industrial robot and perform path interference analysis with the dynamic environment, extract the individual trajectory change sequence that deviates from the predetermined trajectory and quantify the time loss, and obtain the initial phase lag and spatial intersection anchor point. Among them, the collaborative marking of industrial robots in the collaborative workstation, the extraction of the collaborative operation trajectory of the first industrial robot and the path interference analysis of the dynamic environment, and the extraction of the individual trajectory change sequence deviating from the predetermined trajectory are carried out in the following way: By retrieving the process configuration master table in the controller of the collaborative workstation, the kinematic master-slave dependency relationship of each physical robotic arm in the workstation is read from the process configuration master table. The robotic arm responsible for controlling the global spatial position and attitude data of the workpiece and whose kinematic base coordinate system is independently bound is identified and marked as the first industrial robot. Simultaneously, the kinematic base coordinate system is dynamically subordinated to the end position and attitude of the first industrial robot, and the robotic arm performing local process following operations is identified and marked as the second industrial robot; It should be noted that the purpose of collaborative labeling of industrial robots in collaborative workstations is to establish the master-slave relationship in the process of trajectory interference and obstacle avoidance reconstruction. In some embodiments, the pre-set collaborative operation trajectory of the first industrial robot is read from the control system as a static comparison benchmark; the visual perception system in the workstation is activated in parallel to scan the physical space in real time and establish a dynamic three-dimensional model of the current workstation environment. The body geometry of the first industrial robot is read, and three-dimensional expansion is performed along the cooperative running trajectory to generate a swept spatial point cloud with actual physical volume. The sweeping spatial point cloud and the dynamic environment 3D model are overlapped in space coordinates to find the intersection. If a geometric overlap region consisting of shared coordinate points is detected, path interference is determined to have occurred. If no geometrically overlapping area with the 3D model of the dynamic environment is detected, spatial collision and intersection comparison processing will continue. After determining that path interference has occurred, the starting point of the interference is used as the reconstruction starting point, and the non-interference safe area of ​​the dynamic environment 3D model is read at the same time. Using the physical extension safety point along the original trajectory after the interference termination point as the reconstruction endpoint, the existing obstacle avoidance search algorithm is called to generate a bypass trajectory within the non-interference safety area of ​​the dynamic environment 3D model; Extract the extreme values ​​of the normal offset and spatial curvature changes of the bypass trajectory relative to the cooperative trajectory in the three-dimensional spatial coordinate system; The extreme values ​​of normal offset and the data of spatial curvature variation are organized and used as spatial reconstruction features that deviate from the predetermined trajectory. It should be noted that the purpose of extracting spatial reconstruction features is to analyze the spatial cost of avoidance actions so that the generated detour trajectory does not exceed the maximum reach limit of the first industrial robot. Based on the extracted spatial reconstruction features, and combined with the kinematic limit torque and maximum acceleration / deceleration limit parameters of each joint of the first industrial robot, the orbital trajectory is processed by kinematic discrete interpolation. Preferably, the kinematic discrete interpolation process is performed by using the trajectory interpolation period of the underlying control system as the reference time step. Extract spatial curvature change data for each segment on the detour trajectory. For flat road sections with small curvature, calculate the corresponding maximum running step distance based on the maximum acceleration / deceleration limit parameter. For road sections with significant curvature, the corresponding safe deceleration step distance is calculated based on the upper limit not exceeding the kinematic limit torque. Subsequently, based on the above running step distance and deceleration step distance, the entire bypass trajectory is discretized according to the reference time step to generate a continuous data set containing the physical coordinates of each trajectory node and the transient motion associated state. The interpolated trajectory nodes are arranged in chronological order to form a continuous data set containing the physical coordinates, instantaneous velocity and instantaneous acceleration of each node, thus obtaining the single-unit trajectory change sequence of the first industrial robot. The process of obtaining the initial phase lag and spatial intersection anchor point by quantizing the time loss of a single orbital sequence is as follows: Extract the total planned running time of a single orbit change sequence from the start point to the end point of reconstruction; At the same time, extract the standard collaborative reference time corresponding to the first industrial robot running from the same reconfiguration starting point to the reconfiguration endpoint at the original cycle speed in the original collaborative operation trajectory; The difference between the actual planned total operating time and the standard collaborative reference time is calculated to obtain the absolute additional time generated by the first industrial robot due to the execution of spatial navigation and speed interpolation actions. The absolute extra time is defined as the initial phase lag of the collaborative cycle in the collaborative workstation; After obtaining the initial phase lag, the three-dimensional physical coordinates of the last trajectory node of the first industrial robot in the single trajectory change sequence are extracted as spatial regression coordinates. Retrieve the global system clock of the collaborative workstation to obtain the absolute start timestamp when the first industrial robot initiates the avoidance action; The absolute start timestamp is added to the actual planned total running time to calculate the absolute arrival timestamp when the first industrial robot arrives at the spatial return coordinates. By binding and encapsulating the spatial regression coordinates with the absolute arrival timestamp, the spatial intersection anchor point of the first industrial robot is established.

[0025] It should be noted that during the execution of the single-unit trajectory change sequence by the first industrial robot, the visual perception system is continuously invoked to monitor and refresh the three-dimensional model of the dynamic environment in real time. If a secondary displacement of an environmental obstacle is detected, causing geometric interference to occur again in the single-unit trajectory change sequence, the current interference position is used as the new reconstruction starting point, and the above obstacle avoidance search and time loss quantification steps are re-executed to achieve dynamic iterative update of the absolute arrival timestamp and spatial intersection anchor point.

[0026] S20. Establish the independent operating logic domain of the second industrial robot; and based on the independent operating logic domain, combine the initial phase lag and the absolute space intersection anchor point to perform master-slave synchronization reconstruction processing to obtain the cooperative orientation characteristics of the second industrial robot. The process of performing time-sequential separation processing on the planned trajectory of the second industrial robot to establish an independent operating logic domain is as follows: In some embodiments, the real-time coordinate system refresh command sent from the main controller of the first industrial robot to the slave controller of the second industrial robot is intercepted, thereby blocking the dynamic update link of the base coordinate system of the second industrial robot as the spatial position and orientation of the end effector of the first industrial robot changes. Extract the local virtual clock from the control system of the second industrial robot; The task execution sequence of the second industrial robot is switched from the global synchronous clock of the collaborative workstation to the local virtual clock for driving, thus establishing an independent operating logic domain for the second industrial robot. The current physical position and attitude data of the second industrial robot are extracted synchronously, and the visual perception system is invoked in an independent operating logic domain to perform real-time local collision monitoring of the dynamic sweeping area of ​​the second industrial robot. It should be noted that the purpose of establishing an independent operating logic domain is to enable the second industrial robot to detach from the real-time motion traction of the first industrial robot at the control algorithm level, thereby realizing the pre-establishment of the future collaborative recovery state in the virtual computing space; the introduction of local collision monitoring is to prevent the second industrial robot from blindly moving and colliding when it loses the master control traction. The method for obtaining the collaborative orientation characteristics of the second industrial robot by performing master-slave synchronization reconstruction based on the independent operating logic domain combined with the initial phase lag and the absolute space intersection anchor point is as follows: Obtain the absolute spatial intersection anchor point and initial phase lag output in step S10, and extract the absolute arrival timestamp contained in the absolute spatial intersection anchor point. Subtract the initial phase lag from the absolute arrival timestamp to calculate the logical synchronization time of the second industrial robot in the original process plan. Retrieve the original collaborative task planning scheme of the second industrial robot before the path interference occurs, and search for the theoretical planning nodes corresponding to the logical synchronization moment in the original collaborative task planning scheme. Extract the three-dimensional spatial coordinate components and tool axis attitude components of the theoretical planning node in the independent operation logic domain of the second industrial robot, and encapsulate the attribute association between the three-dimensional spatial coordinate components and the tool axis attitude components. Based on the associated encapsulated three-dimensional spatial coordinate components and tool axial attitude components, an ideal spatial state benchmark that the second industrial robot must reach at the moment of future re-intersection is constructed, and the cooperative orientation characteristics of the second industrial robot target are established.

[0027] Example 2

[0028] Please see Figure 1 As shown, the real-time obstacle avoidance trajectory planning method for industrial robots based on dynamic environment modeling includes the following steps: S30: Perform inverse kinematic analysis along the time axis on the cooperative orientation features to obtain the inverse trajectory segment that maintains the intersection condition; map and align the inverse trajectory segment to the current system absolute clock to obtain the constraint orientation parameters of the second industrial robot theory; The process of performing inverse kinematic analysis along the time axis on the cooperative orientation features to obtain the inverse trajectory segment that maintains the intersection condition is as follows: Retrieve the cooperative orientation features of the target output in step S20, and the absolute arrival timestamp contained in the absolute spatial intersection anchor point extracted in step S10. Within the independent operating logic domain of the second industrial robot, a set of absolute timestamp sequences is constructed, with the absolute arrival timestamp as the starting reference value and the system interpolation period as the decreasing step size. By retrieving the time node in the sequence that is closest to the current absolute system clock, the corresponding reverse trajectory node is determined, thus achieving logical alignment between the clock system and the spatial node. The cooperative orientation features of the target are set as the initial starting boundary of the negative discrete time series. The reverse trajectory generator is started to perform reverse discrete time slice backtracking processing to obtain the reverse trajectory segment that maintains the intersection condition. Preferably, the method for performing reverse discrete-time slice backtracking is as follows: Extract the physical rated maximum speed and maximum acceleration of each joint of the second industrial robot; The physical rated maximum velocity and maximum acceleration are numerically inverted to construct a symmetric physical constraint boundary through reverse derivation; Along the negative discrete time series, without breaking the symmetric physical constraint boundary, the preceding spatial nodes that the second industrial robot must have to maintain smooth end-effector motion are deduced one by one from the interpolation cycle. All derived preceding spatial nodes are smoothly connected and serialized in reverse chronological order to generate inverse trajectory segments that maintain the intersection conditions. The process of mapping and aligning the inverse trajectory segment to the current system's absolute clock to obtain the constraint orientation parameters of the second industrial robot theory is as follows: After generating the reverse trajectory segment, obtain the absolute system clock of the current moment inside the collaborative workstation; Perform a time matching search between the current absolute system clock and the absolute discrete timestamp sequence attached to the reverse trajectory segment; Retrieve and lock the target preceding space node that matches the absolute system clock at the current moment from the reverse trajectory segment; Extract the three-dimensional coordinate data and end-point orientation data contained within the target's preceding spatial node; The three-dimensional coordinate data and the end-point orientation data are associated and encapsulated, and defined as the theoretical constraint orientation parameters that the second industrial robot must achieve at the current moment.

[0029] S40: Extract the trajectory tangential deviation between the theoretical constraint orientation parameters and the real-time position and attitude data, and perform time-distance equivalent analysis to obtain the initial timing compensation amount; combine the kinematic extreme boundary of the robotic arm to perform compliant amplitude limiting control on the initial timing compensation amount to obtain the optimal phase delay compensation amount; The method for extracting the trajectory tangential deviation between the theoretical constraint orientation parameters and the real-time position and attitude data, and performing time-distance equivalent analysis to obtain the initial timing compensation amount is as follows: Obtain the theoretical constraint orientation parameters output in step S30, and simultaneously collect the real-time position and attitude data of the second industrial robot; Extract the three-dimensional coordinate data from the theoretical constraint orientation parameters, and extract the real-time position data from the real-time position and attitude data, and calculate the spatial position vector difference between the two. By projecting the spatial position vector difference onto the instantaneous tangent direction of the inverse reference trajectory, the trajectory tangential deviation reflecting the difference in master-slave coordination progress is separated. Retrieve the current real-time operating speed of the second industrial robot; Based on the motion state characteristics of the second industrial robot, the trajectory tangential deviation is equivalently converted into the hysteresis duration in the time dimension; Preferred method for converting hysteresis duration into time dimension is as follows: when converting spatial deviation into time compensation, the real-time running speed of the second industrial robot is retrieved. To prevent the robot's speed from being zero during the start-up and stop phases, which could lead to computational overflow, a very small non-zero speed lower limit is preset. The real-time speed is compared with the lower limit, and the result of the comparison is used as the denominator for quotient calculation to ensure the numerical stability of the hysteresis duration calculation. The hysteresis duration is used as the time base parameter for the local virtual clock tick in adjustment step S20 to establish the initial timing compensation amount. The optimal phase delay compensation amount is obtained by performing compliant amplitude limiting control on the initial timing compensation amount in conjunction with the kinematic extreme boundary of the robotic arm: The hardware rated physical limit parameters preset by the underlying servo system of the second industrial robot are retrieved, and its maximum allowable acceleration and maximum allowable impact are extracted and established as the kinematic extreme boundary. like Figure 2 As shown, the evaluation assesses whether the target acceleration and impact required by each joint of the second industrial robot would exceed the kinematic extreme boundary if the initial timing compensation is fully absorbed within one interpolation cycle. If no overshoot is detected, it is determined that there is no risk of overshoot, and the initial timing compensation amount is output as the optimal phase delay compensation amount. If a breakthrough is detected, a compliant protection mechanism is triggered. This mechanism uses the kinematic extreme value as a constraint to calculate the maximum time compensation amount that can be absorbed in the current cycle in reverse. This is then used to limit the initial timing compensation amount to reduce the risk of the servo system triggering hardware protection due to excessive time domain compensation. The timing adjustment value that meets the safety tolerance range of the mechanical structure after amplitude limiting is defined as the final optimal phase delay compensation amount.

[0030] S50: Extract the timing residual poles between the initial and optimal phase delay compensation amounts, and transform the timing residual poles into spatial trajectory reduction parameters; extract the safety tolerance corridor and combine it with the spatial trajectory reduction parameters to perform spatial shortcut reconstruction processing, and complete the synchronous handover of the master and slave control domains; The method for extracting the temporal residual poles between the initial and optimal phase delay compensation amounts, and transforming these poles into spatial trajectory reduction parameters, is as follows: Obtain the initial timing compensation amount and the optimal phase delay compensation amount output in step S40; The difference between the initial timing compensation and the optimal phase delay compensation is calculated to obtain the uncompensated timing residual poles. Retrieve the current interpolation speed of the second industrial robot within its independent operating logic domain; Multiply the temporal residual poles with the current interpolation speed to convert the compensation deficit in the time dimension into a spatial trajectory reduction parameter in the distance dimension. The process of extracting the safety tolerance corridor and combining it with spatial trajectory reduction parameters to reconstruct the spatial shortcut, thereby completing the synchronous handover of the master and slave control domains, is as follows: Extract the intermediate transition nodes contained in the remaining reverse trajectory segment along the second industrial robot's journey towards the absolute space intersection anchor point. Spatial distance measurement and calculation are performed between each intermediate transition node and the dynamic environment 3D model constructed in step S10 to obtain the maximum safe redundancy radius around each intermediate transition node that does not physically interfere with obstacles. For example, the maximum safety redundancy radius can be obtained by calculating the shortest distance from each intermediate transition node to the surface of the dynamic environment 3D model, and subtracting the maximum physical outer envelope radius of the second industrial robotic arm from the distance. The remaining gap distance is defined as the maximum safety redundancy radius. The maximum safety redundancy radius of each consecutive intermediate transition node is spatially enveloped to establish a non-interference safety tolerance corridor. With the goal of compensating for the time-series residual poles, the remaining trajectory is geometrically reconstructed within the safety tolerance corridor boundary to generate a spatial shortcut trajectory. Preferably, the method for geometric feature reconstruction is as follows: the spatial trajectory reduction parameter of the equivalent transformation of temporal residual poles is set as the total target value of path shortening. During the optimization process, the spatial fusion chamfer of multiple intermediate transition nodes on the remaining trajectory is adjusted synchronously so that the sum of the path reduction caused by the chamfer amplification of all nodes is equal to the spatial trajectory reduction parameter. Meanwhile, after determining the spatial fusion chamfer of each intermediate transition node, the corner line segment with the transition node as the vertex in the original trajectory is replaced by a smooth curve (such as a circular arc or a high-order spline curve) with the radius of the spatial fusion chamfer; the replaced smooth curve segment is then tangentially spliced ​​with the remaining straight line segment in the original trajectory to synthesize a spatial shortcut trajectory with a reduced total length. Drive the second industrial robot along the spatial shortcut trajectory to the absolute spatial intersection anchor point, and stop calling the second industrial robot's local virtual clock at the absolute spatial intersection anchor point; Switch the control drive source of the second industrial robot back to the global synchronization clock of the collaborative workstation to achieve synchronous handover of the master and slave control domains.

[0031] Example 3

[0032] Please see Figure 3 As shown, the real-time obstacle avoidance trajectory planning system for industrial robots based on dynamic environment modeling includes the following modules: Loss Analysis Module: Used to collaboratively mark industrial robots in collaborative workstations, extract the collaborative running trajectory of the first industrial robot and perform path interference analysis with the dynamic environment, extract the individual trajectory change sequence that deviates from the predetermined trajectory and quantify the time loss, and obtain the initial phase lag and spatial intersection anchor point. Feature reconstruction module: used to establish the independent operating logic domain of the second industrial robot; and based on the independent operating logic domain, combined with the initial phase lag and the absolute space intersection anchor point, perform master-slave synchronous reconstruction processing to obtain the cooperative orientation features of the second industrial robot; Orientation constraint module: used to perform inverse kinematic analysis along the time axis on cooperative orientation features to obtain inverse trajectory segments that maintain the intersection condition; based on the inverse trajectory segments, it maps and aligns them to the current system absolute clock to obtain the constraint orientation parameters of the second industrial robot theory; Amplitude limiting control module: used to extract the trajectory tangential deviation between theoretical constraint orientation parameters and real-time position and attitude data, and perform time-distance equivalent analysis to obtain the initial timing compensation amount; combined with the kinematic extreme boundary of the robotic arm, perform compliant amplitude limiting control on the initial timing compensation amount to obtain the optimal phase delay compensation amount; Synchronous Reconstruction Module: Used to extract the timing residual poles between the initial and optimal phase delay compensation amounts, and transform the timing residual poles into spatial trajectory reduction parameters; extract the safety tolerance corridor and combine it with the spatial trajectory reduction parameters to perform spatial shortcut reconstruction processing, and complete the synchronous handover of the master and slave control domains.

[0033] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the present invention should still fall within the scope of the present invention.

Claims

1. A method for real-time obstacle avoidance trajectory planning of industrial robots based on dynamic environment modeling, characterized in that, Includes the following steps: Collaborative marking is performed on industrial robots in the collaborative workstation. The collaborative running trajectory of the first industrial robot is extracted and path interference analysis is performed on the dynamic environment. The individual trajectory change sequence that deviates from the predetermined trajectory is extracted and the time loss is quantified to obtain the initial phase lag and spatial intersection anchor point. An independent operating logic domain for the second industrial robot is established; and based on the independent operating logic domain, combined with the initial phase lag and the absolute space intersection anchor point, master-slave synchronization reconstruction is performed to obtain the cooperative orientation characteristics of the second industrial robot. Inverse kinematic analysis along the time axis is performed on the cooperative orientation features to obtain the inverse trajectory segments that maintain the intersection condition; the inverse trajectory segments are mapped and aligned to the current system absolute clock to obtain the constraint orientation parameters of the second industrial robot theory. The trajectory tangential deviation between the theoretical constraint orientation parameters and the real-time position and attitude data is extracted and time-distance equivalent analysis is performed to obtain the initial timing compensation amount. Combined with the kinematic extreme boundary of the robotic arm, the initial timing compensation amount is subjected to compliant amplitude limiting control to obtain the optimal phase delay compensation amount.

2. The method for real-time obstacle avoidance trajectory planning of industrial robots based on dynamic environment modeling according to claim 1, characterized in that: The path interference analysis is performed as follows: Extract the kinematic master-slave dependency relationship of the collaborative workstation, mark the robot arm that dominates the global pose and has an independent base coordinate system as the first industrial robot, and mark the robot arm whose base coordinate system is dynamically subordinate to the first industrial robot as the second industrial robot. Construct a dynamic 3D model of the current workstation environment, combine it with the body geometry of the first industrial robot, generate a swept spatial point cloud along the cooperative running trajectory, and perform spatial collision and intersection comparison with the dynamic 3D model of the environment. If a geometrically overlapping region is detected on the expected travel path, path interference is determined to have occurred, and the operation of extracting the individual trajectory change sequence is triggered.

3. The method for real-time obstacle avoidance trajectory planning of industrial robots based on dynamic environment modeling according to claim 2, characterized in that: The method for extracting the single-unit orbital change sequence is as follows: Using the physical extension safety points at both ends of the determined path interference point as the reconstruction boundary, a detour trajectory is generated within the non-interference safety area of ​​the dynamic environment 3D model; Extract the extreme values ​​of the normal offset and the spatial curvature change data of the bypass trajectory relative to the cooperative trajectory, as spatial reconstruction features; Based on spatial reconstruction characteristics, discrete interpolation is performed on the bypass trajectory by combining the dynamic extreme value constraint parameters of the first industrial robot, generating a continuous data set containing the physical coordinates and transient motion associated states of each trajectory node, thus obtaining the single-unit trajectory change sequence.

4. The method for real-time obstacle avoidance trajectory planning of industrial robots based on dynamic environment modeling according to claim 1, characterized in that: The process of performing the master-slave synchronization reconstruction is as follows: Extract the absolute arrival timestamp contained in the absolute spatial intersection anchor point, and extract the logical synchronization time corresponding to the second industrial robot in the original process plan; Retrieve the theoretical planning node for the logical synchronization moment of the second industrial robot; Extract the three-dimensional spatial coordinate components and tool axis attitude components of the theoretical planning node in the independent operation logic domain of the second industrial robot, and encapsulate the attribute association between the three-dimensional spatial coordinate components and the tool axis attitude components. Based on the associated encapsulated three-dimensional spatial coordinate components and tool axial attitude components, an ideal spatial state benchmark that the second industrial robot must reach at the moment of future re-intersection is constructed, and the cooperative orientation characteristics of the second industrial robot target are established.

5. The method for real-time obstacle avoidance trajectory planning of industrial robots based on dynamic environment modeling according to claim 1, characterized in that: The mapping alignment is performed as follows: After generating the reverse trajectory segment, obtain the absolute system clock of the current moment inside the collaborative workstation; Perform a time matching search between the current absolute system clock and the absolute discrete timestamp sequence attached to the reverse trajectory segment; Retrieve and lock the target preceding space node that matches the absolute system clock at the current moment from the reverse trajectory segment; Extract the three-dimensional coordinate data and end-point orientation data contained within the target's preceding spatial node; The three-dimensional coordinate data and the end-effector orientation data are associated and encapsulated, and defined as the theoretical constraint orientation parameters of the second industrial robot.

6. The method for real-time obstacle avoidance trajectory planning of industrial robots based on dynamic environment modeling according to claim 1, characterized in that: The method for generating the reverse trajectory segment is as follows: Retrieve the absolute arrival timestamp contained in the absolute spatial intersection anchor point; Within the independent operating logical domain, a timestamp sequence is constructed based on the time attributes of the spatial intersection anchor points; based on the time stamp association between the timestamp sequence and the system real-time clock, the theoretical constraint orientation of the second industrial robot is determined; The cooperative orientation feature is set as the initial starting boundary of the negative discrete time series. The reverse trajectory generator is started to perform reverse discrete time slice backtracking processing to obtain the reverse trajectory segment that maintains the intersection condition.

7. The method for real-time obstacle avoidance trajectory planning of industrial robots based on dynamic environment modeling according to claim 1, characterized in that: The method for performing the time-distance equivalent analysis is as follows: Simultaneously collect real-time position and attitude data of the second industrial robot; Extract the three-dimensional coordinate data from the theoretical constraint orientation parameters, and extract the real-time position data from the real-time position and attitude data, and calculate the spatial position vector difference between the two. By projecting the spatial position vector difference onto the instantaneous tangent direction of the inverse reference trajectory, the trajectory tangential deviation reflecting the difference in master-slave coordination progress is separated. Based on the motion state characteristics of the second industrial robot, the trajectory tangential deviation is equivalently converted into the hysteresis duration in the time dimension; The hysteresis duration is used as the time base parameter for adjusting the local virtual clock tick, thus establishing the initial timing compensation amount.

8. The method for real-time obstacle avoidance trajectory planning of industrial robots based on dynamic environment modeling according to claim 1, characterized in that: Also includes: Extract the timing residual poles between the initial and optimal phase delay compensation amounts, and transform the timing residual poles into spatial trajectory reduction parameters; extract the safety tolerance corridor and combine it with the spatial trajectory reduction parameters to perform spatial shortcut reconstruction processing, and complete the synchronous handover of the master and slave control domains.

9. The method for real-time obstacle avoidance trajectory planning of industrial robots based on dynamic environment modeling according to claim 8, characterized in that: The method for performing the aforementioned spatial shortcut reconstruction process is as follows: Extract the intermediate transition nodes contained in the remaining reverse trajectory segment along the second industrial robot's journey towards the absolute space intersection anchor point. With the goal of compensating for the time-series residual poles, the remaining trajectory is geometrically reconstructed within the safety tolerance corridor boundary to generate a spatial shortcut trajectory. Drive the second industrial robot along the spatial shortcut trajectory to the absolute spatial intersection anchor point, and stop calling the second industrial robot's local virtual clock at the absolute spatial intersection anchor point; Switch the control drive source of the second industrial robot back to the global synchronization clock of the collaborative workstation to achieve synchronous handover of the master and slave control domains.

10. The method for real-time obstacle avoidance trajectory planning of industrial robots based on dynamic environment modeling according to claim 9, characterized in that: The method for establishing the safety tolerance corridor boundary is as follows: Spatial distances between each intermediate transition node and the dynamic environment 3D model are measured and calculated to obtain the maximum safe redundancy radius around each intermediate transition node that does not physically interfere with obstacles. A safety tolerance corridor is established by spatially enveloping the maximum safety redundancy radii of each consecutive intermediate transition node.

Images