A method for dynamic scheduling of multiple workstations in the safety zone of an unmanned coffee machine robotic arm

Abstract

This invention relates to the field of robot motion control technology and discloses a multi-station dynamic scheduling method for the safety zone of an unmanned coffee machine robotic arm. The method first calculates a dedicated dynamic safety radius for each movement based on the motion sector of the scheduling task and the state of the arm's interaction devices. Next, it establishes a geometric envelope model for different load states of the robotic arm and corrects the boundary points of the path based on this model and the dynamic safety radius to ensure that the load entity remains within the safety zone during movement. Finally, by calculating the effective rotation angle between the corrected boundary points and judging based on angle thresholds, it decides to generate a three-segment or a four-segment scheduling path that introduces intermediate transition points. This invention solves the technical problems of path planning redundancy caused by static safety zones and interference caused by neglecting load volume. While ensuring safety, it improves the operating efficiency, reliability, and adaptability to complex paths of the robotic arm.

Description

Technical Field

[0001] This invention relates to the field of robot motion control technology, specifically to a method for dynamic scheduling of multiple workstations in the safety zone of an unmanned coffee machine robotic arm. Background Technology

[0002] In automated retail, especially in the field of unmanned coffee machines, the application of multi-joint robotic arms is becoming increasingly widespread. Their core task is to autonomously complete a series of precise material handling and placement actions, from retrieving cups and adding ingredients to dispensing, within a compact workspace filled with equipment. To ensure the efficient and reliable operation of the robotic arm, motion path planning is one of the key technologies, directly determining the preparation time of a single cup of beverage and the stability of the entire system.

[0003] However, existing robotic arm path planning methods still have shortcomings in balancing efficiency and safety. To prevent collisions between the robotic arm and surrounding equipment during movement, a common approach is to pre-define a global, fixed minimum safety zone. This zone is typically determined based on the location of the nearest device to the robotic arm's base, and all planned paths must be strictly confined within this fixed area. While this safety strategy logically avoids interference, its rigid boundaries limit the flexibility of path planning. When the robotic arm moves long distances in open sectors away from obstacles, it is still constrained by this minimum safety zone, forcing it to perform unnecessary detours, resulting in path redundancy and reduced overall operational efficiency.

[0004] Furthermore, existing methods often simplify the robotic arm's end effector and its load as a single geometric point mass during path planning, neglecting the actual physical volume of the load (such as the cup or lid). This simplification has minimal impact on linear motion, but it exposes significant safety hazards when performing rotational movements. When the robotic arm carrying the cup makes a large-angle turn, the actual trajectory of the cup's outer edge may exceed the planned point mass path. If this path happens to be close to the safety zone boundary, accidental collisions between the load and the equipment are highly likely, leading to task failure and compromising system reliability. Therefore, there is an urgent need for a scheduling method that can dynamically adapt to the environment and accurately consider the load volume. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a dynamic scheduling method for multiple workstations within the safety zone of an unmanned coffee machine robotic arm. This method solves the problems of path redundancy and low operating efficiency caused by the use of fixed static safety zones in existing robotic arm path planning, as well as the risk of collisions during rotational movement due to the neglect of the actual volume of the load.

[0006] To achieve the above objectives, the present invention provides a method for dynamic scheduling of multiple workstations within the safety zone of an unmanned coffee machine robotic arm, comprising the following steps:

[0007] S1. System Initialization and Spatial Modeling

[0008] Establish a two-dimensional Cartesian coordinate system with the center of the robotic arm base as the origin O, obtain the positions of all arm interaction devices within the coordinate system, and abstract the position of each device i as a position vector. Based on this, a safety zone is defined around the center of the robotic arm base.

[0009] S2, Task Breakdown

[0010] The system receives external beverage preparation instructions and decomposes these instructions into at least one sequentially executed scheduled task. Each scheduled task is defined as a data structure, which at least includes the robot arm's starting position vector S, ending position vector T, and the starting load state P associated with the starting point for this task. S and the target load state P associated with the endpoint. T .

[0011] S3. Path Key Point Calculation: For each scheduled task, path key points are calculated. These path key points include: a starting-side corrected boundary point S located on the boundary of the safety zone and associated with the starting point. ′ And the endpoint-side corrected boundary point T associated with the endpoint. ′ .

[0012] S4. Scheduling path generation and execution are based on the starting point S, the ending point T, and the starting point-side correction boundary point S. ′ and the endpoint side correction boundary point T ′ A scheduling path is generated. This scheduling path consists of at least three consecutive path segments. After the path is generated, the robotic arm is controlled to move along the scheduling path from the starting point to the ending point.

[0013] Preferably, before calculating the path key points in step S3, the method may further include the following steps:

[0014] First, based on the starting point S and ending point T of the scheduling task, and with the base origin O as the vertex, the motion sector for this movement is determined. Then, based on the position of the arm interaction devices within the motion sector, or the current task status of all arm interaction devices, a dedicated dynamic safety radius r is calculated for this scheduling task. dyn Ultimately, using this dynamic safety radius r dyn As the starting point for calculating the corrected boundary point S ′ and the endpoint side correction boundary point T′ The radius of the safety zone boundary on which the time is based.

[0015] In one specific embodiment, the dynamic safety radius r is calculated. dyn The methods may specifically include:

[0016] Identify the arm interaction device located within the motion sector, or identify the arm interaction device currently in a task-locked state among all arm interaction devices, and define these two types of devices together as the constraint device set. The dynamic safety radius r dyn The following formula is used to calculate:

[0017]

[0018] Where, r min As a preset global minimum safety radius, representing the lower limit of safety assurance, in a specific embodiment, the method for calculating the path key points in step S3 may specifically include:

[0019] First, geometric envelope models are established for the various load states of the robotic arm. Then, based on the initial load state P defined in the scheduling task... S Based on its corresponding geometric envelope model, the starting point-side corrected boundary point S is calculated. ′ .

[0020] Similarly, according to the target load state P defined in the scheduling task T Based on its corresponding geometric envelope model, the endpoint-side corrected boundary point T is calculated. ′ This calculation method ensures that the outer contour of the robotic arm, along with its load, remains within the safety zone defined by the current safety radius throughout its movement along the scheduling path.

[0021] Furthermore, specific methods for establishing and calculating the geometric envelope model may include defining a load offset distance d for different load states P of the robotic arm, such as no load, holding a cup, and holding a lid. offset (P). The starting point side correction boundary point S ′ and the endpoint correction boundary point T ′ The calculation is performed using a boundary point correction function F. crs_mod This is implemented. The function shrinks the load offset distance corresponding to each load state inwards from the safety zone boundary in both the starting and ending directions. The calculation formula is as follows:

[0022]

[0023] Where V is the input original position vector, P is the associated load state, and r is the currently used safety radius.

[0024] Preferably, in step S4, the method for generating the scheduling path may specifically include:

[0025] Using the center O of the robotic arm base as the vertex, calculate the corrected boundary point S from the starting point side. ′ Correct the boundary point T at the endpoint. ′ The effective rotation angle θ. Its calculation formula is:

[0026]

[0027] Determine whether the absolute value of the effective rotation angle θ is greater than a preset angle threshold (e.g.) If the value is not greater than 1, a scheduling path is generated that sequentially connects the starting point, the starting point-side correction boundary point, the ending point-side correction boundary point, and the ending point. If the value is greater than 1, a scheduling path is generated that includes four path segments with one additional intermediate transition point.

[0028] Furthermore, when the effective rotation angle is greater than the angle threshold, the calculation method for the intermediate transition point M may include: calculating the corrected boundary point S located on the starting point side of the safety zone boundary defined by the current safety radius. ′ With the endpoint side correction boundary point T ′ The point on the angle bisector between the points is designated as the intermediate transition point M. The scheduling paths of the four path segments sequentially connect the starting point S, the starting point-side correction boundary point S', and the starting point-side boundary point S'. ′ The intermediate transition point M, the endpoint correction boundary point T ′ , and the endpoint T.

[0029] In one embodiment, the method may further include a closed-loop process execution flow, specifically including: before the robotic arm moves from the starting point along the scheduling path, setting the arm interaction device corresponding to the endpoint to a task locked state; and after the robotic arm reaches the endpoint and completes the preset specified action, releasing the state of the arm interaction device to a task idle state.

[0030] In one specific embodiment, the safety zone defined in step S1 is specifically a global minimum safety zone, which is determined based on the shortest straight-line distance from the center of the robotic arm base to all arm interaction devices. Any dynamic safety radius calculated in the method is not less than the radius of the global minimum safety zone, and any generated scheduling path is limited to the range defined by the global minimum safety zone, or to a larger range defined by the dynamic safety radius.

[0031] This invention provides a method for dynamic scheduling of multiple workstations within the safety zone of an unmanned coffee machine's robotic arm. It offers the following advantages:

[0032] 1. This invention calculates a unique dynamic safety radius for each movement based on the motion sector of the scheduling task and the task status of the arm interaction device. This allows the safety zone to be adaptively adjusted according to the current task and device status. This setting provides greater optimization space for scheduling path planning. Under the premise of ensuring safety, it can generate shorter and more direct movement trajectories, thereby improving the overall operating efficiency of the robotic arm.

[0033] 2. This invention establishes a geometric envelope model for different load states of the robotic arm and calculates the corrected path boundary points based on the model. This makes the path planning no longer regard the end of the robotic arm as an abstract point, but accurately includes the actual physical volume of the load it carries (such as the cup body or cup lid). This effectively avoids physical interference caused by the load exceeding the safety boundary during rotational motion, and enhances the reliability and operational safety of the entire scheduling process.

[0034] 3. This invention generates three-segment or four-segment scheduling paths by calculating the effective rotation angle and judging based on the angle threshold. It provides a structured and deterministic path generation strategy for scheduling tasks with different spans. In particular, for large-angle rotations, the introduction of intermediate transition points ensures that the robotic arm body always stays within the safe zone during the movement, solving potential safety problems under complex paths and improving the adaptability and robustness of the method to various scheduling scenarios. Attached Figure Description

[0035] Figure 1 This is a schematic diagram of the workspace layout of the unmanned coffee machine of the present invention;

[0036] Figure 2 This is a schematic diagram illustrating the breakdown of the beverage preparation instructions according to the present invention;

[0037] Figure 3 This is a flowchart of the dynamic path planning process for a single scheduling task according to the present invention;

[0038] Figure 4 This is a schematic diagram illustrating the geometric relationship between the path key points and the scheduling path in this invention;

[0039] Figure 5 This is a flowchart of the scheduling execution and status management process of the present invention. Detailed Implementation

[0040] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0041] See attached document Figure 1 This invention provides a method for dynamic scheduling of multiple workstations in the safety zone of an unmanned coffee machine robotic arm. In the initialization phase of this method, a step of parametric modeling of the physical environment and the execution state of the robotic arm may be included.

[0042] In one specific embodiment, a spatial modeling process is first performed. This process establishes a two-dimensional Cartesian coordinate system with the center of the robotic arm base as the origin O. Subsequently, the positions of all arm-interacting devices interacting with the robotic arm within the workspace are obtained. These arm-interacting devices may include coffee machines, ice makers, cup dispensers, cup retrieving stations, etc. The position of each arm-interacting device i in this coordinate system is abstracted as a point on its circumscribed contour closest to the origin O, and its position is determined by a position vector. express.

[0043] After completing the spatial modeling process, a global safety parameter calculation process is performed. This process is used to determine a global minimum safety zone, which is the absolute safety boundary that no subsequent scheduling path can cross.

[0044] Specifically, calculate the position vector from the origin O to each arm interaction device. The Euclidean norm between the given and given values ​​is taken, and the minimum value among them is taken as the global minimum safe radius r. min The calculation formula is as follows:

[0045]

[0046] Where, r min Indicates the global minimum safe radius; Let |||||| represent the position vector of the i-th arm interactive device; ||·|| represents the Euclidean norm of the vector.

[0047] To accurately describe the changes in the physical dimensions of the robotic arm at different stages of operation, this method also includes the establishment of a load state model. This model defines a load state set P, which contains all preset load conditions of the robotic arm during operation. For example, this set may include: an unloaded state P. empty Holding the cup P cup Holding the cover P lid wait.

[0048] Corresponding to the load state model, this method also includes a load geometry modeling process. This process establishes a geometric envelope model for each load state in the load state set P. To facilitate subsequent calculations, this three-dimensional geometric envelope model is simplified and extracted into a single key parameter: the load offset distance d. offset (P). The physical meaning of this parameter is the maximum radial distance from the center point of the end effector of the robotic arm and its load under this load state, in the radial direction with the origin O as the observation point.

[0049] For example, for the no-load state P empty Its load offset distance d offset (P empty This can be set to zero, or a small value representing the radius of the end effector's own structure. For the cup-holding state P... cup Its load offset distance d offset (P cup This corresponds to the sum of the cup's radius and the portion exceeding the end gripper structure. In this way, the space occupancy of the robotic arm under different load conditions is precisely parameterized, providing a data foundation for subsequent path correction calculations.

[0050] See attached document Figure 2 After completing system initialization and static parameter modeling, this method performs a process of converting macroscopic instructions into a series of executable robotic arm scheduling tasks.

[0051] In one specific embodiment, the method first performs an instruction receiving and parsing process. This process receives a structured beverage preparation instruction from an external system. This instruction contains all the information needed to prepare a specific beverage, such as a beverage recipe identifier and specifications such as temperature or sweetness level.

[0052] Upon receiving a beverage preparation instruction, a process breakdown procedure is initiated. This procedure retrieves and calls up the corresponding process flow from a pre-stored process recipe library based on the recipe identifier in the instruction. Each process flow is defined as an ordered sequence of multiple process steps. For example, the process flow for an iced Americano can be defined as: 1. Take a cup; 2. Take ice; 3. Move to the coffee machine; 4. Prepare espresso; 5. Move to the cup dispensing station.

[0053] Subsequently, the core of this method lies in executing a scheduling task generation process. This process traverses each process step in the process flow sequence and transforms it into a precisely defined scheduling task that can be used for subsequent path planning. Each scheduling task is a tuple containing multiple data elements, with the structure defined as follows:

[0054]

[0055] The definitions of each data element are as follows:

[0056] S: The starting position vector of the robotic arm for this scheduled task. This vector is usually the ending position vector of the previous scheduled task. For the first task in the sequence, this starting point can be a preset system initial position or standby position; T: The ending position vector of the robotic arm for this scheduled task. This vector corresponds to the position vector of the arm interaction device required for the current process step. A: The physical action that the robotic arm needs to perform after reaching the endpoint T. This action is a preset action instruction code, for example, used to control the robotic arm's end effector to perform operations such as grasping, releasing, or pressing buttons; P S The initial load state before executing this scheduled task corresponds to the physical state of the robotic arm at the starting point S and is directly inherited from the target load state P of the previous scheduled task. T ;P T After completing physical action A at endpoint T, the robotic arm will enter the target load state. This state is determined based on a prediction of the outcome of action A. For example, if action A is grasping a cup, then regardless of the initial load state P... S Why, target load state P T All will be set to cup-holding state P cup .

[0057] By sequentially executing this scheduling task generation process for each step in the process flow, a macroscopic beverage preparation instruction is ultimately transformed completely and unambiguously into a time-series of scheduling tasks T. This sequence provides accurate and complete input for subsequent dynamic path planning and execution.

[0058] See attached document Figure 3 and attached Figure 4 After the process flow is decomposed into a series of scheduling tasks, this method performs a dynamic path planning process independently for each scheduling task in the sequence.

[0059] In one specific embodiment, the dynamic path planning process receives a scheduling task. As input, proceed with the following steps:

[0060] First, a context-aware and dynamic safety radius calculation process is executed. The first step of this process is to determine the movement sector for this move.

[0061] Specifically, taking the origin O at the center of the robotic arm base as the vertex, and connecting the starting point S and the ending point T respectively, a vector is formed. and The region enclosed by these two vectors is defined as the motion sector for this task.

[0062] Subsequently, based on this motion sector, the constraint devices are identified. The identification process includes two parallel conditions: first, traversing all arm interaction devices and assigning their position vectors... Devices located within the motion sector are identified as constraint devices; secondly, all arm interaction devices are traversed, and devices currently in a task-locked state are also identified as constraint devices. Devices meeting any of the above conditions are added to the constraint device set for this task. middle.

[0063] Determining the set of constraint devices Then, calculate the dynamic safety radius r specific to this scheduling task. dyn This calculation process ensures that the generated safety radius is neither less than the system-defined global minimum safety radius r. min It can also be adapted to actual constraints. The calculation formula is as follows:

[0064]

[0065] Where, r min This is the global minimum safety radius determined during system initialization. If the set of constraint devices... If r is empty, it means there are no task conflicts or nearby devices around the current movement path. dyn A preset maximum operating radius for the system can be selected.

[0066] Next, a path critical point correction calculation process is performed. This process uses the dynamic safety radius r obtained in the previous step. dyn Based on the load state of the robotic arm, the starting point correction boundary point S is calculated to ensure that the load entity is completely within the safe zone. ′ and the endpoint correction boundary point T ′ The calculation is performed using a boundary point correction function F. crs_mod accomplish:

[0067]

[0068] Where V is the input original position vector (i.e., the starting point S or the ending point T), P is the load state associated with that point, and r is the currently used dynamic safety radius r dyn d offset (P) represents the load offset distance corresponding to load state P. It is a unit vector in the direction from O to V.

[0069] The load offset distance d here offset(P), in physical terms, represents the maximum radial dimension of the overall geometric envelope (including the actuator itself and the load it holds) of the robotic arm end effector in a specific load state P, in the direction away from the origin O of the base. This parameter is crucial to ensuring that the load entity is always completely enveloped by the dynamic safety zone during movement.

[0070] In one specific embodiment, this offset distance can be pre-calculated and stored through offline modeling. For example, for an unloaded state P empty Its offset distance d offset (P empty This refers to the maximum radial dimension of the end effector itself; for the cup-holding state P cup By retrieving the 3D CAD models of the cup body and the actuator, the radial distance from the farthest point from the actuator center on the overall contour of the combined structure can be calculated, and this distance can be used as d. offset (P cup In this way, the system can establish a precise library of offset distance parameters for every possible load state.

[0071] When applying this function, calculate the corrected boundary point S on the starting side. ′ The initial load state P before the task begins is used. S :

[0072] S ′ =F crs_mod (S,P S ,r dyn );

[0073] Calculate the corrected boundary point T on the end side ′ The target load state P used after the task is completed is... T :

[0074] T′=F crs_mod (T,P T ,r dyn );

[0075] To obtain the precise corrected boundary point S ′ and T ′ Then, a path decision and generation process is executed. This process first calculates the path from the origin O as the vertex, using the vector OS as the starting point. ′ To vector OT ′ Effective rotation angle θ:

[0076]

[0077] Then, the absolute value of the effective rotation angle θ is compared with a preset angle threshold (e.g., The angle threshold value can be compared. The specific value can be empirically set or experimentally calibrated based on the robot arm model, arm span, and the compactness of the workspace. The basic principle for setting it is:

[0078] When the robotic arm moves only along S without adding a midpoint... ′ To T ′ During circular motion, the arm (such as the elbow joint) must not exceed the dynamic safety radius r during rotation. dyn The defined cylindrical safety space. A conservative threshold (such as 90 degrees) can be suitable for most application scenarios. However, in cases with shorter arm spans or more open workspaces, the threshold can be increased to reduce unnecessary four-segment path generation and further improve efficiency.

[0079] If |θ| is not greater than the threshold, it is determined to be a small-angle or straight-line movement, generating a three-segment scheduling path consisting of three path segments. The node sequence of this path is (S→S ′ →T ′ →T), its trajectory is: moving in a straight line from the starting point S to the starting side correction boundary point S. ′ Then along the radius of or The safe zone boundary arc is moved to the end point side to correct the boundary point T. ′ Finally from T ′ Move in a straight line to the endpoint T °

[0080] If |θ| is greater than this threshold, it is determined to be a large-angle rotation. To prevent the robotic arm from exceeding the safe zone during rotation, an intermediate transition point M needs to be added to the path. This intermediate transition point M is located on the boundary of the dynamic safe zone and is situated within... and Along the angle bisector direction, generate a four-segment scheduling path consisting of four path segments, with the node sequence being (S→S). ′ →M→T ′ →T). The trajectory of this path is: moving from the starting point S to S... ′ Move along the circular arc of the safe zone boundary to the intermediate transition point M, and then continue moving along the circular arc to T. ′ Finally from T ′ Move to the destination T.

[0081] See attached document Figure 5 After generating a deterministic scheduling path for a single scheduled task, this method performs a scheduling execution and state closed-loop management process to ensure the orderly and conflict-free execution of tasks and to provide the correct initial state for subsequent tasks.

[0082] In one specific embodiment, for each scheduled task with a generated scheduling path, the method is performed according to the following steps:

[0083] First, before sending the motion commands for the scheduled path to the robotic arm controller, a resource locking process is performed. This process extracts the endpoint position vector T from the current scheduled task T and uses it as an index to look up the corresponding device entry in the system-maintained list of arm interaction device statuses.

[0084] Subsequently, the status of the device entry is changed from task idle to task locked. This status change ensures that no other parallel scheduling process will use the device as a target point for path planning before the current task is completed, thereby avoiding access conflicts to device resources.

[0085] After completing the resource locking process, the method executes the path execution process, which takes the sequence of scheduling path nodes generated in the previous step (e.g., S→S) ′ →T ′ →T ′ →T) is converted into motion commands that the robotic arm controller can recognize, and the robotic arm is controlled to move physically according to the path sequence.

[0086] In a typical implementation environment, the scheduling method described in this invention runs on a host computer or industrial computer. This host computer is responsible for performing all calculations from instruction decomposition to path planning. When the path execution process starts, the host computer generates a sequence of path nodes (e.g., S→S). ′ →T ′ →T) is sent to the dedicated controller of the robotic arm through the communication interface. After receiving these target points, the robotic arm controller uses its own inverse kinematics solver and servo drive system to complete the calculation of joint angles and motor drive, thereby enabling the end effector to move accurately along a predetermined path in Cartesian space.

[0087] Once the robotic arm's end effector reaches its destination T, it first executes the physical action A defined in the scheduled task T. After a confirmation signal indicating the completion of action A is returned, the system immediately performs two parallel state update steps:

[0088] First, the resource release process is executed. This process corresponds to the resource locking process. The arm interaction device status list is accessed again, and the task status identifier of the device corresponding to the endpoint T is restored from the task locked state to the task idle state, so that the device can be used for other subsequent scheduling tasks.

[0089] Second, the robotic arm state update process is executed. This process updates the robotic arm's own state recorded internally by the system. Specifically, the current load state of the robotic arm is updated to the target load state P defined in the current scheduled task T. T The updated load state will serve as the starting load state P for the next scheduled task. S This ensures the continuity and accuracy of the robotic arm's load status throughout the entire process.

[0090] After completing all the above steps, the current scheduled task is marked as completed. The method then executes an iterative process, checking if there are any subsequent tasks in the scheduled task sequence. If so, the next scheduled task is extracted, and the process returns to the aforementioned dynamic path planning, initiating a new round of planning and execution. If no task exists, it indicates that the entire beverage preparation instruction has been executed. This iterative execution method constitutes the closed-loop management core of the scheduling method of this invention.

Claims

1. A method for dynamic scheduling of multiple workstations within the safety zone of an unmanned coffee machine robotic arm, characterized in that, Includes the following steps: S1. Establish a coordinate system with the center of the robotic arm base as the origin, and define a safety zone around the center of the robotic arm base based on the positional relationship between the center of the robotic arm base and multiple arm interaction devices. S2. Receive a beverage preparation instruction and decompose the beverage preparation instruction into at least one scheduling task. The scheduling task includes the starting point and ending point of the robotic arm, as well as the starting load state and target load state associated with the starting point and ending point, respectively. S3. For the scheduling task, calculate the path key points, including the start-side correction boundary point located on the boundary of the safety zone and associated with the start point, and the end-side correction boundary point associated with the end point. S4. Based on the starting point, the ending point, the starting point-side correction boundary point, and the ending point-side correction boundary point, generate a scheduling path containing at least three path segments, specifically including: Using the center of the robotic arm base as the vertex, calculate the effective rotation angle from the starting point correction boundary point to the ending point correction boundary point; Determine whether the effective rotation angle is greater than a preset angle threshold; If the effective rotation angle is not greater than the angle threshold, a scheduling path is generated that sequentially connects the starting point, the starting point-side correction boundary point, the ending point-side correction boundary point, and the ending point. If the effective rotation angle is greater than the angle threshold, then on the boundary of the safe zone, the intermediate transition point located on the angle bisector between the starting point correction boundary point and the ending point correction boundary point is calculated, and a scheduling path is generated that sequentially connects the starting point, the starting point correction boundary point, the intermediate transition point, the ending point correction boundary point, and the ending point. Control the robotic arm to move from the starting point to the ending point along the scheduling path.

2. The method for dynamic scheduling of multiple workstations in the safety zone of an unmanned coffee machine robotic arm according to claim 1, characterized in that, In step S3, before calculating the path key points, the following steps are also included: Based on the start and end points of the scheduling task, determine the movement sector for this move; Based on the position of the arm interaction devices within the motion sector or the task status of all arm interaction devices, a unique dynamic safety radius is calculated for this scheduling task. The starting point correction boundary point and the ending point correction boundary point are calculated using the dynamic safety radius.

3. The method for dynamic scheduling of multiple workstations in the safety zone of an unmanned coffee machine robotic arm according to claim 1, characterized in that, Step S3, the step of calculating the key points of the path, specifically includes: Establish geometric envelope models for the various load states of the robotic arm; Based on the initial load state defined in the scheduling task and combined with the corresponding geometric envelope model, the starting point side correction boundary point is calculated. Based on the target load state defined in the scheduling task and combined with the corresponding geometric envelope model, the endpoint correction boundary point is calculated.

4. The method for dynamic scheduling of multiple workstations in the safety zone of an unmanned coffee machine robotic arm according to claim 2, characterized in that, The steps for calculating the dynamic safety radius specifically include: Identify the arm interaction device located within the motion sector, or identify the arm interaction device currently in a task-locked state among all arm interaction devices, and determine both as the constraint device. The shortest distance from the center of the robotic arm base to all the constraint devices is taken as the dynamic safety radius.

5. The method for dynamic scheduling of multiple workstations in the safety zone of an unmanned coffee machine robotic arm according to claim 3, characterized in that, The steps for establishing the geometric envelope model specifically include: Define load offset distances for different load states of the robotic arm: no load, holding a cup, and holding a lid. When calculating the starting point and ending point correction boundary points, the load offset distance corresponding to each load state is contracted inward from the boundary of the safety zone in the starting direction and ending direction, respectively, to obtain the final correction boundary points.

6. The method for dynamic scheduling of multiple workstations in the safety zone of an unmanned coffee machine robotic arm according to claim 1, characterized in that, The steps for generating and executing the scheduling path also include: Before the robotic arm moves along the scheduling path, the arm interaction device corresponding to the endpoint is set to a task locked state; After the robotic arm reaches the endpoint and completes the designated action, the arm interaction device is released to a task idle state.

7. The method for dynamic scheduling of multiple workstations in the safety zone of an unmanned coffee machine robotic arm according to claim 4, characterized in that, Before calculating the effective rotation angle, the following steps are also included: A dynamic safety radius is determined for the scheduled task; Based on the starting load state and target load state associated with the scheduling task, the starting point correction boundary point and the ending point correction boundary point are calculated respectively. The calculation of the effective rotation angle is based on the safety zone boundary determined by the dynamic safety radius, as well as the positions of the starting point correction boundary point and the ending point correction boundary point.

8. The method for dynamic scheduling of multiple workstations in the safety zone of an unmanned coffee machine robotic arm according to claim 1, characterized in that, In step S1, the safety zone is a global minimum safety zone determined based on the shortest straight-line distance from the center of the robotic arm base to all arm interaction devices, and the scheduling paths generated in the method are all limited to the range defined by the global minimum safety zone.

Images