Test chamber robot control system based on adaptive trajectory planning

Abstract

The application belongs to the technical field of intelligent control, and discloses a test cabin manipulator control system based on adaptive trajectory planning, which comprises a topological constraint module, which is used for acquiring the load physical properties of a carried object and the physical positions of movable components in the test cabin, constructing a dynamic motion limited domain in a preset manipulator motion space, and performing regional division on the manipulator motion space to generate a spatial topological field; a trajectory adaptive module, which is used for constructing a scalar potential field according to the spatial topological field, generating a motion trajectory meeting a preset fluid stability condition in combination with the dynamic motion limited domain, with the scalar potential field as a guiding direction; and a conflict arbitration module, which is used for detecting the spatiotemporal collision risk of the motion trajectory in the spatial topological field, and constructing a cooperative avoidance trajectory on the premise of keeping the end pose of the manipulator unchanged when the spatiotemporal collision risk exists, so as to improve the motion stability and operation reliability of the manipulator in a complex limited space.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology, and more specifically, to a control system for a laboratory robot based on adaptive trajectory planning. Background Technology

[0002] With the continuous improvement of laboratory automation, robotic arms are gradually replacing manual labor in handling samples and reagent containers within the testing chamber. The rationality of the robotic arm's motion trajectory planning directly affects handling efficiency and the safety of test samples. Existing robotic arm trajectory planning methods in testing chambers typically generate motion trajectories based on preset fixed paths or simple obstacle avoidance algorithms. These methods fail to fully consider the physical properties of the load on the object being handled (such as liquid volume and container sealing status) as well as the real-time position status of other movable components within the testing chamber to create restricted areas. This results in the robotic arm being prone to causing liquid sloshing due to excessive acceleration and deceleration when handling liquid sample containers, potentially leading to container tipping or leakage. Furthermore, it struggles to adapt to the collision risks arising from the dynamic changes in the position of movable components within the testing chamber.

[0003] Furthermore, existing robotic arms, upon detecting potential collision risks, typically employ either a complete trajectory replanning or a complete halt to avoid collisions. On one hand, complete replanning is time-consuming, impacting operational continuity; on the other hand, directly adjusting the end effector pose can easily cause the transported container to tilt, compromising end effector pose stability during transport. Simultaneously, the existing trajectory execution process lacks a continuous monitoring and online correction mechanism for deviations between the actual execution state and the theoretically planned state. When abnormal forces are applied to the robotic arm joints or the end effector pose deviates, it is difficult to promptly correct unexecuted trajectories, leading to accumulated deviations that affect the robotic arm's handling accuracy and the safety of laboratory samples.

[0004] In view of this, the present invention proposes a laboratory robot control system based on adaptive trajectory planning to solve the above problems. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art and to achieve the above objectives, the present invention provides the following technical solution: a laboratory chamber robotic arm control system based on adaptive trajectory planning, comprising:

[0006] The topology constraint module is used to obtain the load physical properties of the object being transported and the physical positions of the movable components in the test chamber, construct a dynamic motion-constrained domain within the preset manipulator motion space, and divide the manipulator motion space into regions to generate a spatial topology field.

[0007] The trajectory adaptive module is used to construct a scalar potential energy field based on the spatial topology field. Using the scalar potential energy field as the guiding direction, combined with the dynamic motion confined domain, it generates a motion trajectory that meets the preset fluid stability conditions.

[0008] The conflict arbitration module is used to detect the spatiotemporal collision risk of the motion trajectory in the spatial topological field. When there is a spatiotemporal collision risk, a cooperative avoidance trajectory is constructed on the premise of keeping the pose of the robot end effector unchanged.

[0009] The deviation perception module is used to collect the actual driving torque and end pose when executing the cooperative avoidance trajectory at a preset period, and compare them with the theoretical driving torque and theoretical pose respectively to obtain the trajectory correction increment vector.

[0010] The online correction module is used to correct the unexecuted part of the cooperative avoidance trajectory online when the trajectory correction increment vector exceeds a preset threshold, and generate corresponding control commands to be output to the robot arm for execution.

[0011] The load physical properties of the transported object include the mass, dimensions, center of gravity, container type, liquid load, and container sealing status of the transported object; the physical location of the movable components inside the laboratory chamber includes the spatial coordinates, external dimensions, and motion status of each movable component;

[0012] The mass and container type of the transported object are obtained from the laboratory task database, the external dimensions and center of gravity are determined by the pre-stored container mold or corrected by the visual measurement device, the liquid loading is obtained by the liquid level detection device, and the container sealing status is confirmed by the sealing detection device.

[0013] Preferably, the method for generating a spatial topological field includes:

[0014] The equivalent envelope of the object being transported is calculated based on the load physical properties of the object being transported. The dynamic envelope of each movable component is calculated based on the physical position and motion state of the movable components in the laboratory chamber. Based on the equivalent envelope and the dynamic envelope, the boundary of each envelope is expanded according to a preset expansion coefficient to obtain the dynamic motion confined domain.

[0015] The robot's motion space is divided into several spatial grid cells according to a preset grid size. The positional relationship between each spatial grid cell and the dynamic motion restricted area is determined, and the spatial grid cells are marked as safe areas, buffer areas, or restricted areas. Based on the marking results of each spatial grid cell and the connectivity between adjacent grid cells, a spatial topology field containing regional attributes and connectivity topology is constructed.

[0016] Preferably, the method for constructing a scalar potential energy field based on a spatial topological field includes:

[0017] For spatial grid cells marked as restricted regions in the spatial topology field, a potential energy value positively correlated with the distance to the target location is assigned; for spatial grid cells marked as safe regions, a potential energy value negatively correlated with the distance to the target location is assigned; and for spatial grid cells marked as buffer regions, a transition potential energy value decreasing from the restricted region to the safe region is assigned.

[0018] Based on the spatial grid cells corresponding to the current pose of the robot and the spatial grid cells corresponding to the target pose, the potential energy gradient vector is calculated; the potential energy value of each spatial grid cell is associated with and stored with the corresponding potential energy gradient vector to obtain a scalar potential energy field covering the robot's motion space.

[0019] Preferably, the method for generating a motion trajectory that satisfies preset fluid stability conditions includes:

[0020] Candidate path points are generated step by step along the descent direction of the potential energy gradient vector in the scalar potential energy field to obtain the initial motion trajectory; based on the container type, liquid loading, and container sealing status of the object being transported, the acceleration threshold and angular acceleration threshold of the robot end effector corresponding to each path point on the initial motion trajectory are calculated.

[0021] Determine whether the acceleration and angular acceleration of each path point on the initial motion trajectory exceed the corresponding acceleration threshold and angular acceleration threshold. If they do, adjust the motion speed and direction of the path point to obtain a motion trajectory that meets the preset fluid stability conditions.

[0022] Preferably, the method for detecting the risk of spatiotemporal collision of a motion trajectory in a spatial topological field includes:

[0023] Obtain the timestamps corresponding to each path point on the motion trajectory to obtain the spatiotemporal trajectory sequence of the motion trajectory; obtain the position state of the dynamic motion confined domain of the movable component in the test chamber at the corresponding timestamp to obtain the spatiotemporal occupancy sequence of the movable component;

[0024] The spatiotemporal trajectory sequence of the motion trajectory is spatiotemporally matched with the spatiotemporal occupancy sequence of the movable component. It is determined whether there is an overlap between the spatial grid cell corresponding to the motion trajectory of the robot and the spatial grid cell corresponding to the dynamic motion restricted domain of the movable component under the same timestamp. If there is an overlap, it is determined that there is a spatiotemporal collision risk in the motion trajectory, and the time interval of the collision risk and the corresponding spatial grid cell are recorded.

[0025] Preferably, the method for constructing a cooperative avoidance trajectory includes:

[0026] Based on the time interval of the collision risk and the corresponding spatial grid cell, determine the trajectory segment that needs to be avoided; while keeping the end effector pose unchanged, use the redundant degrees of freedom of the manipulator base joint and intermediate joint as the adjustment object, and search for alternative path points to bypass the corresponding spatial grid cell in the spatial topology field.

[0027] A cooperative avoidance trajectory is generated based on alternative path points and path points in the original trajectory that are not affected by collision risk. The cooperative avoidance trajectory is then reviewed for spatiotemporal collision risk. When the review result meets the no-collision condition, the cooperative avoidance trajectory is used as the trajectory to be executed.

[0028] Preferably, the method for collecting the actual driving torque and end-effector pose during the execution of the cooperative avoidance trajectory at a preset period includes:

[0029] Torque sensors are installed at each joint of the robotic arm, and a pose detection device is installed at the end of the robotic arm. According to a preset acquisition cycle, the actual driving torque of each joint of the robotic arm is collected by the torque sensors, and the actual position and actual attitude angle of the end of the robotic arm are collected by the pose detection device to obtain the actual pose of the end.

[0030] Based on the kinematic and dynamic models corresponding to the cooperative avoidance trajectory, the theoretical driving torque of each joint and the theoretical pose of the manipulator end effector are calculated at the same moment. The actual driving torque and actual end effector pose within the same acquisition cycle are synchronized with the theoretical driving torque and theoretical pose, respectively.

[0031] Preferably, the method for obtaining the trajectory correction increment vector includes:

[0032] The difference between the actual driving torque and the theoretical driving torque within the same acquisition cycle is calculated to obtain the torque deviation; the difference between the actual pose and the theoretical pose of the end effector within the same acquisition cycle is calculated to obtain the pose deviation, which includes the position deviation component and the attitude angle deviation component.

[0033] The torque deviation, position deviation component, and attitude angle deviation component are normalized, and then weighted according to preset weighting coefficients to obtain the trajectory correction increment vector.

[0034] Preferably, the method for online correction of the unexecuted portion of the cooperative avoidance trajectory includes:

[0035] Based on the actual pose of the robot arm's end effector at the current moment, determine the set of unexecuted path points in the cooperative avoidance trajectory; based on the trajectory correction increment vector, compensate and correct the position coordinates and attitude angles of each path point in the set of unexecuted path points to obtain the corrected set of path points.

[0036] The modified path point set is subjected to spatiotemporal collision risk verification and fluid stability condition verification. When the verification and validation results meet the preset conditions, the corresponding control command is generated based on the modified path point set and output to the robot for execution; when the verification and validation results do not meet the preset conditions, the cooperative avoidance trajectory is reconstructed.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] This invention constructs a dynamic motion-constrained domain by acquiring the load physical properties of the object being transported and the physical positions of movable components within the laboratory chamber. It then divides the manipulator's motion space into regions to generate a spatial topological field. This allows the constrained region to dynamically change according to the object's own properties and the real-time positions of the movable components, improving the accuracy and adaptability of the manipulator's motion space constraints compared to fixed path planning methods. Based on the spatial topological field, a scalar potential energy field is constructed, using the potential energy gradient as the trajectory guidance direction. Simultaneously, acceleration and angular acceleration thresholds are set based on the container type, liquid load, and sealing status of the object being transported. This ensures that the generated motion trajectory avoids constrained areas while satisfying fluid stability conditions, reducing the risk of shaking, tipping, or leakage during the transport of liquid containers.

[0039] Secondly, upon detecting a risk of spatiotemporal collision, while maintaining the end effector's pose unchanged, the redundant degrees of freedom of the base joints and intermediate joints are used to search for alternative path points to construct a cooperative avoidance trajectory. This avoids the problems of time-consuming overall trajectory replanning or container tilting caused by adjusting the end effector pose, ensuring the stability of the transported object's posture during the avoidance process. Furthermore, by collecting the actual driving torque and the actual end effector pose at preset intervals and comparing them with theoretical values ​​to obtain the trajectory correction increment vector, only the unexecuted portion of the cooperative avoidance trajectory is corrected online when the increment vector exceeds a preset threshold. This method is faster than overall replanning, can promptly eliminate accumulated deviations, and improves the robot's handling accuracy and the safety and stability of the sample handling process. Attached Figure Description

[0040] Figure 1 This is a schematic diagram of the control system for the laboratory robot based on adaptive trajectory planning according to the present invention.

[0041] Figure 2 This is a schematic diagram of the control method for the robotic arm of the laboratory chamber based on adaptive trajectory planning according to the present invention. Detailed Implementation

[0042] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only 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.

[0043] Example 1

[0044] Please see Figure 1 As shown, this embodiment provides a control system for a laboratory chamber robotic arm based on adaptive trajectory planning, specifically including the following steps:

[0045] The topology constraint module is used to obtain the load physical properties of the object being transported and the physical positions of the movable components in the test chamber, construct a dynamic motion-constrained domain within the preset manipulator motion space, and divide the manipulator motion space into regions to generate a spatial topology field.

[0046] The trajectory adaptive module is used to construct a scalar potential energy field based on the spatial topology field. Using the scalar potential energy field as the guiding direction, combined with the dynamic motion confined domain, it generates a motion trajectory that meets the preset fluid stability conditions.

[0047] The conflict arbitration module is used to detect the spatiotemporal collision risk of the motion trajectory in the spatial topological field. When there is a spatiotemporal collision risk, a cooperative avoidance trajectory is constructed on the premise of keeping the pose of the robot end effector unchanged.

[0048] The deviation perception module is used to collect the actual driving torque and end pose when executing the cooperative avoidance trajectory at a preset period, and compare them with the theoretical driving torque and theoretical pose respectively to obtain the trajectory correction increment vector.

[0049] The online correction module is used to correct the unexecuted portion of the cooperative avoidance trajectory online when the trajectory correction increment vector exceeds a preset threshold, and generates corresponding control commands to be output to the robot arm for execution. The various modules are connected via wired and / or wireless means to achieve data transmission between them.

[0050] The physical properties of the load on the transported object include the object's mass, dimensions, center of mass, container type, liquid load, and container sealing status; the physical location of movable components inside the laboratory chamber includes the spatial coordinates, external dimensions, and motion status of each movable component.

[0051] In this embodiment, the information of the object to be transported corresponding to the test task to be performed is first obtained. A corresponding load information file is established according to the test task number to be performed, which is used to record the load physical properties of the object to be transported. The mass and container type of the object to be transported can be directly obtained from the test task database. The external dimensions and center of gravity of the object to be transported can be determined based on the pre-stored container mold, or can be corrected in real time by a visual measurement device. The liquid loading volume is obtained by the liquid level detection device, and the container sealing status is confirmed by the sealing detection device.

[0052] For example, when the laboratory receives a test task with the number T-20250601, the system creates a load information file based on this task number and queries the test task database to find that the object to be transported for this task is a 250 ml conical flask, the container type is recorded as a conical flask, and the mass is recorded as 0.18 kg. Based on the mold data corresponding to the conical flask in the pre-stored container mold library, the system determines the container's external dimensions to be a bottom diameter of 0.085 meters, a height of 0.155 meters, and a center of gravity located 0.06 meters above the bottom of the container. Simultaneously, the visual measurement device takes a real-time picture of the conical flask and compares it with the pre-stored mold data. After correction, it confirms that the external dimensions and center of gravity are consistent with the pre-stored data. The liquid level detection device detects that the liquid load in the conical flask is 180 ml, accounting for 72% of the container volume. The sealing detection device detects that the conical flask mouth is not fitted with a sealing cap, confirming that the container is in an open state. The system writes the above information into the payload information file and records it as follows: mission number T-20250601, container type conical flask, mass 0.18 kg, bottom diameter 0.085 m, height 0.155 m, center of gravity height 0.06 m, liquid loading 180 ml, and the sealed state is open.

[0053] It should be noted that the movable components within the testing chamber include all parts capable of positional changes, excluding the robotic arm itself, such as movable racks, rotating platforms, and opening / closing doors. The spatial coordinates of these movable components can be obtained through positioning devices installed within the testing chamber or through position sensors carried by each component itself. Their external dimensions are determined based on pre-stored models of the movable components, and their motion status includes their current speed, direction, and whether they are in motion. For example, a rotating sampling platform within a testing chamber has its spatial coordinates, external radius, and current rotational angular velocity all uploaded in real-time by the platform's own controller to the topology constraint module.

[0054] Methods for generating spatial topological fields include:

[0055] The equivalent envelope of the object being transported is calculated based on the load physical properties of the object being transported. The dynamic envelope of each movable component is calculated based on the physical position and motion state of the movable components in the laboratory chamber. Based on the equivalent envelope and the dynamic envelope, the boundary of each envelope is expanded according to a preset expansion coefficient to obtain the dynamic motion confined domain.

[0056] It should be noted that the expansion distance corresponding to the preset expansion coefficient is determined by adding the values ​​of three parts: the positioning repeatability of the robot arm's end effector, the measurement error of the object being transported and its movable components, and the reserved safety margin. For example, if the positioning repeatability of the robot arm's end effector is 0.02 meters, the measurement error of the visual measurement device in the external dimensions of the object being transported is 0.01 meters, and a safety margin of 0.02 meters is reserved, then the three are added together, and the expansion distance corresponding to the preset expansion coefficient is set to 0.05 meters.

[0057] The robot's motion space is divided into several spatial grid cells according to a preset grid size. The positional relationship between each spatial grid cell and the dynamic motion restricted area is determined, and the spatial grid cells are marked as safe areas, buffer areas, or restricted areas. Based on the marking results of each spatial grid cell and the connectivity between adjacent grid cells, a spatial topology field containing regional attributes and connectivity topology is constructed.

[0058] The preset grid size is determined based on the overall dimensions of the robot's motion space and the minimum controllable step length of the robot's end effector. The grid size is set to be no smaller than the minimum step length of the robot's end effector to ensure that the trajectory planning accuracy matches the computational efficiency. For example, if the robot's motion space is 1.5m × 1.2m × 1.0m and the minimum controllable step length of the robot's end effector is 0.05m, then the preset grid size is set to 0.05m × 0.05m × 0.05m.

[0059] It should be noted that the equivalent envelope is generated based on the external dimensions and center of mass of the object being transported, and is used to characterize the actual geometric contour occupied by the object in the robot's motion space; the dynamic envelope is generated based on the current position, external contour dimensions, and motion state of the movable component, and is used to characterize the spatial range that the movable component may reach during its motion. Expansion processing refers to using the original boundaries of the equivalent and dynamic envelopes as a reference, expanding outwards by a preset expansion coefficient by a certain distance, thereby reserving a safety margin on the basis of the original geometric contour, and obtaining the dynamic motion confined domain.

[0060] For example, if the object being transported is a container, its equivalent envelope is a cylinder with a radius of 0.08 meters, and the movable component is a moving rack, its dynamic envelope is a rectangle with a length of 0.30 meters and a width of 0.20 meters, and the expansion distance corresponding to the preset expansion coefficient is set to 0.05 meters, then after expanding the equivalent envelope of the object being transported outward by 0.05 meters, a dynamic motion confined region with a radius of 0.13 meters is obtained, and after expanding the dynamic envelope of the moving rack outward by 0.05 meters, a dynamic motion confined region with a length of 0.40 meters and a width of 0.30 meters is obtained.

[0061] Furthermore, the robot's motion space is divided into several spatial grid cells according to a preset grid size (e.g., 0.05 m × 0.05 m × 0.05 m). The positional relationship between each spatial grid cell and the aforementioned dynamic motion-constrained domain is determined one by one: when a spatial grid cell is completely within the dynamic motion-constrained domain, it is marked as a constrained region; when a spatial grid cell is within a preset range outside the boundary of the dynamic motion-constrained domain, it is marked as a buffer region; when a spatial grid cell is outside the dynamic motion-constrained domain and its buffer range, it is marked as a safe region. Based on the marking results of each spatial grid cell, and combined with the spatial connectivity between adjacent grid cells (i.e., whether adjacent grid cells can be directly traversed), a spatial topology field containing regional attributes and connectivity topology is constructed, providing a spatial constraint basis for subsequent trajectory planning.

[0062] Methods for constructing scalar potential energy fields based on spatial topological fields include:

[0063] For spatial grid cells marked as restricted areas in the spatial topological field, a higher potential energy value is assigned according to the positional relationship between the grid cell and the center of the corresponding restricted area, so that the overall potential energy value inside the restricted area is at a preset high level, which is used to characterize that the robot should avoid the area.

[0064] For spatial grid cells marked as safe areas, potential energy values ​​are assigned based on the distance between the grid cell and the target position. The closer to the target position, the lower the potential energy value, and the farther from the target position, the higher the potential energy value, which is used to guide the robot arm to move towards the target position.

[0065] For a spatial grid cell marked as a buffer region, an interpolation is performed between the higher potential energy value corresponding to the restricted region and the potential energy value corresponding to the safe region, based on the relative position of the grid cell from the boundary of the restricted region and the boundary of the safe region. This results in the potential energy value in the buffer region gradually decreasing from the side closer to the restricted region to the side closer to the safe region, forming a smooth transition.

[0066] It should be noted that the higher potential energy value corresponding to the restricted area is preset to a fixed value based on the unified range of potential energy values ​​in the system; the potential energy ratio coefficient of the safe area is determined based on the distance between the target position and the farthest boundary within the robot's movement space, and the potential energy value corresponding to the farthest point of the desired safe area, i.e., the ratio coefficient is obtained by dividing the potential energy value at the farthest point of the desired safe area by the corresponding distance; the width of the buffer area is determined based on the maximum movement speed of the robot and the response time of the control system, ensuring that the robot has sufficient distance to complete deceleration after entering the buffer area. For example, if the potential energy value of the restricted area is set to 100; the maximum span of the robot's movement space is 2 meters; and the potential energy value at the farthest point of the desired safe area is 15, then the ratio coefficient is 15 ÷ 2 = 7.5 (unit: per meter); the maximum movement speed of the robot is 0.5 m / s; the response time of the control system is 0.2 seconds; and a safety factor of 2 is set, then the width of the buffer area is 0.5 × 0.2 × 2 = 0.2 meters.

[0067] Furthermore, based on the potential energy difference between the spatial grid cell corresponding to the current pose of the robot and its adjacent spatial grid cells, the adjacent direction with the largest decrease in potential energy value is determined, and this direction is defined as the potential energy gradient vector corresponding to that spatial grid cell. The process is repeated to traverse each spatial grid cell within the robot's motion space, obtaining the potential energy gradient vector corresponding to each spatial grid cell. The potential energy value of each spatial grid cell is associated with and stored with its corresponding potential energy gradient vector, resulting in a scalar potential energy field covering the entire robot's motion space and reflecting the potential energy level and direction of decrease at each position.

[0068] For example, a spatial grid cell corresponding to a restricted area is assigned a preset potential energy value of 100. In a buffer area adjacent to the restricted area and with a width of 0.1 meters, the spatial grid cell closer to the restricted area is assigned a potential energy value of 80, the spatial grid cell closer to the safe area is assigned a potential energy value of 20, and the intermediate grid cells are assigned transitional potential energy values ​​decreasing from the inside to the outside, such as 60 and 40. In the safe area, the spatial grid cell 0.5 meters from the target position is assigned a potential energy value of 15, the spatial grid cell 0.2 meters from the target position is assigned a potential energy value of 6, and the spatial grid cell 0.05 meters from the target position is assigned a potential energy value of 1.5. After comparing the spatial grid cell corresponding to the current pose of the robot with its adjacent grid cells, if the adjacent grid cell with the largest decrease in potential energy value is determined to be located on the side of the target position, then the direction pointing to that adjacent grid cell is determined as the potential energy gradient vector corresponding to the current grid cell.

[0069] Methods for generating motion trajectories that satisfy preset fluid stability conditions include:

[0070] Candidate path points are generated step by step along the descent direction of the potential energy gradient vector in the scalar potential energy field to obtain the initial motion trajectory; based on the container type, liquid loading, and container sealing status of the object being transported, the acceleration threshold and angular acceleration threshold of the robot end effector corresponding to each path point on the initial motion trajectory are calculated.

[0071] Determine whether the acceleration and angular acceleration of each path point on the initial motion trajectory exceed the corresponding acceleration threshold and angular acceleration threshold. If they do, adjust the motion speed and direction of the path point to obtain a motion trajectory that meets the preset fluid stability conditions.

[0072] In this embodiment, it should be noted that the acceleration threshold and angular acceleration threshold are determined in advance through liquid surface sloshing experiments conducted on different container types, different liquid loading ratios, and different sealing conditions. The critical values ​​at which the container will overflow or tilt under different motion acceleration and angular acceleration conditions are recorded, and the measurement results are compiled into a lookup table for the system to access. For example, experimentally determined, the maximum allowable acceleration for an open container to avoid overflow when the liquid loading is 80% of its volume is 0.5, and the maximum allowable angular acceleration is 0.3; for a sealed container with the same loading, the corresponding thresholds can be set to 1.2 and 0.8.

[0073] In this embodiment, it should be noted that generating candidate path points step by step along the descent direction of the potential energy gradient vector means starting from the spatial grid cell corresponding to the current pose of the robot, sequentially selecting the grid cell with the largest decrease in potential energy value among adjacent grid cells as the next candidate path point, until reaching the spatial grid cell corresponding to the target pose, thereby obtaining the initial motion trajectory that passes through the safe area and buffer area and avoids the restricted area.

[0074] Furthermore, the acceleration and angular acceleration thresholds for each path point on the initial motion trajectory are determined based on the container type, liquid load, and sealing status of the object being transported. For example, when the object being transported is an open container with a large liquid load, a smaller acceleration and angular acceleration threshold is set accordingly; when the object being transported is a sealed container or has a small liquid load, a relatively larger acceleration and angular acceleration threshold is set accordingly. For instance, if a container is open and the liquid load is 80% of its volume, the system sets the acceleration threshold for that path segment to 0.5 and the angular acceleration threshold to 0.3. When the calculated acceleration at a certain path point on the initial motion trajectory is 0.8, exceeding the preset acceleration threshold, the motion speed at that path point is reduced and the motion direction is adjusted appropriately to make the acceleration change more gradual until the acceleration and angular acceleration at that path point do not exceed the corresponding thresholds, thus obtaining a motion trajectory that meets the preset fluid stability conditions.

[0075] Methods for detecting the risk of spatiotemporal collisions of motion trajectories in a spatial topological field include:

[0076] Obtain the timestamps corresponding to each path point on the motion trajectory to obtain the spatiotemporal trajectory sequence of the motion trajectory; obtain the position state of the dynamic motion confined domain of the movable component in the test chamber at the corresponding timestamp to obtain the spatiotemporal occupancy sequence of the movable component;

[0077] The spatiotemporal trajectory sequence of the motion trajectory is spatiotemporally matched with the spatiotemporal occupancy sequence of the movable component. It is determined whether there is an overlap between the spatial grid cell corresponding to the motion trajectory of the robot and the spatial grid cell corresponding to the dynamic motion restricted domain of the movable component under the same timestamp. If there is an overlap, it is determined that there is a spatiotemporal collision risk in the motion trajectory, and the time interval of the collision risk and the corresponding spatial grid cell are recorded.

[0078] For example, the movement trajectory of the robotic arm corresponds to spatial grid cells G1, G2, G3, G4, and G5 sequentially from time t1 to t5, forming a spatiotemporal trajectory sequence. Within the same time interval, the dynamic motion confined domain of a moving material rack in the laboratory sequentially occupies spatial grid cells G6, G6, G3, G3, and G7, forming a spatiotemporal occupancy sequence of the movable component. By matching and comparing the two sequences with the same timestamp, it can be seen that at time t3, the spatial grid cell G3 corresponding to the robotic arm's movement trajectory overlaps with the spatial grid cell G3 corresponding to the dynamic motion confined domain of the moving material rack. Therefore, it is determined that there is a spatiotemporal collision risk for this movement trajectory near time t3, and the time interval of the collision risk is recorded as a preset time range before and after t3, corresponding to spatial grid cell G3.

[0079] Methods for constructing cooperative avoidance trajectories include:

[0080] Based on the time interval of the collision risk and the corresponding spatial grid cell, determine the trajectory segment that needs to be avoided; while keeping the end effector pose unchanged, use the redundant degrees of freedom of the manipulator base joint and intermediate joint as the adjustment object, and search for alternative path points to bypass the corresponding spatial grid cell in the spatial topology field.

[0081] A cooperative avoidance trajectory is generated based on alternative path points and path points in the original trajectory that are not affected by collision risk. The cooperative avoidance trajectory is then reviewed for spatiotemporal collision risk. When the review result meets the no-collision condition, the cooperative avoidance trajectory is used as the trajectory to be executed.

[0082] In this embodiment, it should be noted that the redundant degrees of freedom of the manipulator's base joints and intermediate joints refer to the degrees of freedom that allow the manipulator to change its overall configuration by adjusting the angle combinations of the base joints and intermediate joints, while ensuring that the manipulator's end-effector pose (end-effector position and attitude angle) remains unchanged. Taking the collision risk of bearing spatial grid unit G3 as an example, under the constraint of keeping the end-effector pose unchanged, the system traverses the feasible angle combinations of the manipulator's base joints and intermediate joints within the range of redundant degrees of freedom, and searches in the spatial topology field for alternative path points that allow the overall configuration of the manipulator to bypass spatial grid unit G3 while still allowing the end-effector to reach the corresponding position of the original path point.

[0083] The alternative path points obtained from the search are concatenated with unaffected path points in the original trajectory that are outside the collision risk time interval to generate a cooperative avoidance trajectory. Further, the spatiotemporal collision risk detection process is re-executed on the generated cooperative avoidance trajectory, i.e., the aforementioned spatiotemporal matching steps are repeated. When the verification result shows that the cooperative avoidance trajectory does not overlap with the spatiotemporal occupancy sequence of the movable component throughout the entire execution time period, the cooperative avoidance trajectory is determined to meet the collision-free condition and is identified as the trajectory to be executed. If the verification still indicates a collision risk, other alternative path points are searched within the redundant degrees of freedom until a cooperative avoidance trajectory that meets the collision-free condition is generated.

[0084] The methods for collecting the actual driving torque and actual end-effector pose during the execution of the cooperative avoidance trajectory at a preset period include:

[0085] Torque sensors are installed at each joint of the robotic arm, and a pose detection device is installed at the end of the robotic arm. According to a preset acquisition cycle, the actual driving torque of each joint of the robotic arm is collected by the torque sensors, and the actual position and actual attitude angle of the end of the robotic arm are collected by the pose detection device to obtain the actual pose of the end.

[0086] It should be noted that the preset acquisition period is determined based on the control period of the robot arm servo control system and is set to an integer multiple of the servo control period to ensure that the acquired data matches the control command period. For example, if the robot arm servo control period is 10 milliseconds, the preset acquisition period is set to 20 milliseconds, meaning that the actual driving torque and the actual end effector pose are acquired every 2 servo control cycles.

[0087] Based on the kinematic and dynamic models corresponding to the cooperative avoidance trajectory, the theoretical driving torque of each joint and the theoretical pose of the manipulator end effector are calculated at the same moment. The actual driving torque and actual end effector pose within the same acquisition cycle are synchronized with the theoretical driving torque and theoretical pose, respectively.

[0088] It should be noted that the kinematic and dynamic models in this invention are pre-established using existing modeling methods in the field of robotic arms and stored in the system model library for use in calculating theoretical driving torque and theoretical pose. The specific establishment method is as follows:

[0089] The kinematic model is established using the Denavit-Hartenberg (DH) parameter method. This involves determining four parameters—link length, link torsion angle, link offset, and joint angle—for each adjacent joint coordinate system based on the structural parameters of each link. Following the homogeneous transformation relationship between adjacent coordinate systems, transformation matrices are sequentially established between each joint coordinate system. These transformation matrices are then multiplied sequentially according to the link connection order to obtain the total transformation matrix of the manipulator's end-effector coordinate system relative to the base coordinate system. This matrix constitutes the kinematic model, used to calculate the position and attitude angles of the manipulator's end-effector based on the joint angles.

[0090] For example, the DH parameters corresponding to the second link of a six-joint manipulator are set as follows: link length 0.3 meters, link twist angle 0 degrees, link offset 0 meters, and joint angle is the current rotation angle of the joint. After multiplying the transformation matrices corresponding to the six joints in sequence, when the input joint angles are 10 degrees, -20 degrees, 30 degrees, 0 degrees, 45 degrees, and 0 degrees respectively, the kinematic model calculates the end effector position coordinates as (0.42 meters, 0.15 meters, 0.68 meters) and the end effector attitude angles as (0 degrees, 30 degrees, 10 degrees).

[0091] The dynamic model is established using the Lagrange equation method. Based on the structural parameters of each link in the manipulator, such as mass, moment of inertia, and center of mass position, expressions for the total kinetic energy and total potential energy of the manipulator system are established. The difference between the kinetic and potential energy expressions is used to construct a Lagrange function. Partial derivatives of the Lagrange function with respect to joint angles and angular velocities are then calculated and rearranged according to the standard form of the Lagrange equations. This yields the functional relationships between the driving torque of each joint and the joint angle, angular velocity, and angular acceleration. This dynamic model is used to calculate the corresponding theoretical driving torque based on the angle, angular velocity, and angular acceleration of each joint.

[0092] For example, the third link of a robotic arm has a mass of 2 kg, a moment of inertia of 0.05 kg·m² about its joint axis, and a distance of 0.15 m from its center of mass to the joint axis. When the joint's current angle is 30 degrees, its angular velocity is 0.2 radians / second, and its angular acceleration is 0.1 radians / second², substituting these values ​​into the established dynamic model yields a theoretical driving torque of 3.6 N·m. Both the kinematic and dynamic models are established once the robotic arm's structural parameters are determined. During actual operation, these models are repeatedly called to calculate the theoretical pose and driving torque at each moment based on the joint angle, angular velocity, and angular acceleration data corresponding to the cooperative avoidance trajectory.

[0093] It should be noted that torque sensors are installed on the rotating parts of each joint of the robotic arm to detect the actual torque borne by each joint during the driving process in real time; the pose detection device is installed on the end effector of the robotic arm to detect the actual position coordinates and attitude angle of the end effector in the laboratory chamber coordinate system in real time. The preset acquisition period is set by the operator according to the response frequency of the robotic arm control system, for example, the acquisition period is set to 20 milliseconds.

[0094] Within each acquisition cycle, the system acquires the actual driving torque and end-effector pose using torque sensors and pose detection devices. Simultaneously, based on the kinematic model (used to determine the mapping relationship between joint angles and end-effector pose) and dynamic model (used to determine the mapping relationship between joint angles, angular velocity, angular acceleration, and joint driving torque) corresponding to the cooperative avoidance trajectory, it calculates the theoretical driving torque that each joint should output and the theoretical pose that the robot's end-effector should reach at the corresponding moment in that acquisition cycle. The actual driving torque and end-effector pose acquired within the same acquisition cycle are timestamped and aligned with the theoretical driving torque and theoretical pose calculated at the corresponding moment, providing a comparison benchmark at the same moment for subsequent deviation calculations.

[0095] The method for obtaining the trajectory correction increment vector includes:

[0096] The actual driving torque of each joint of the manipulator acquired within the same acquisition cycle is compared with the calculated theoretical driving torque one by one to obtain the torque deviation of each joint; the actual position and actual attitude angle of the manipulator end acquired within the same acquisition cycle are compared with the calculated theoretical position and theoretical attitude angle to obtain the pose deviation, which includes a position deviation component representing the position difference and an attitude angle deviation component representing the attitude angle difference.

[0097] Furthermore, since the torque deviation, position deviation, and attitude angle deviation have different dimensions and numerical ranges, the three types of deviations are normalized according to their respective preset value ranges, converting them into dimensionless values ​​that can be directly compared and superimposed. Based on preset weighting coefficients, corresponding weight ratios are assigned to the normalized torque deviation, position deviation, and attitude angle deviation, and the three are weighted and summed according to their corresponding weight ratios to obtain a trajectory correction increment vector that can comprehensively reflect the overall deviation of the robot's actual execution state from the theoretical planning state at the current moment.

[0098] It should be noted that the preset weighting coefficients are determined based on the degree of influence of torque deviation, position deviation component, and attitude angle deviation component on the safety of the handling process. These coefficients are assigned after statistically analyzing the correlation between each deviation and trajectory execution anomalies using historical execution data, and the sum of the three weighting coefficients is 1. For example, statistical analysis shows that the position deviation component has the most significant impact on handling safety, so its weighting coefficient is set to 0.4, while the weighting coefficients for torque deviation and attitude angle deviation are each set to 0.3. The normalized value range for each deviation is determined based on the historical statistical maximum value of that deviation under normal system operation, and this historical maximum value is used as the reference denominator for normalization. For example, if the historical maximum value of torque deviation under normal system operation is 2 N·m, then 2 N·m is used as the reference value for normalization of torque deviation in subsequent calculations. The current torque deviation is divided by this reference value to obtain the normalized torque deviation.

[0099] For example, in a certain acquisition cycle, the actual driving torque of joint 1 is detected to be 12.5 N·m, corresponding to a theoretical driving torque of 12.0 N·m, resulting in a torque deviation of 0.5 N·m. The deviations of the actual and theoretical positions of the end effector in the three coordinate directions are 0.002 m, 0.001 m, and 0.003 m, respectively, yielding the position deviation component. The deviation of the actual and theoretical attitude angles of the end effector is 0.01 radians, yielding the attitude angle deviation component. After normalizing the torque deviation, position deviation, and attitude angle deviation components according to preset value ranges, they are weighted and summed according to preset weighting coefficients (e.g., torque deviation weight is 0.3, position deviation component weight is 0.4, and attitude angle deviation component weight is 0.3) to obtain the trajectory correction increment vector corresponding to that acquisition cycle.

[0100] Methods for online correction of the unexecuted portion of the cooperative avoidance trajectory include:

[0101] Based on the actual pose of the robot arm's end effector at the current moment, determine the set of unexecuted path points in the cooperative avoidance trajectory; based on the trajectory correction increment vector, compensate and correct the position coordinates and attitude angles of each path point in the set of unexecuted path points to obtain the corrected set of path points.

[0102] The modified path point set is subjected to spatiotemporal collision risk verification and fluid stability condition verification. When the verification and validation results meet the preset conditions, the corresponding control command is generated based on the modified path point set and output to the robot for execution. When the verification and validation results do not meet the preset conditions, the cooperative avoidance trajectory is reconstructed.

[0103] For example, when the trajectory correction increment vector exceeds a preset threshold, the system determines the corresponding position in the cooperative avoidance trajectory based on the actual pose of the robot's end effector at the current moment, and identifies the path points after that position that have not yet been executed as the set of unexecuted path points. Further, based on the position deviation component and attitude angle deviation component contained in the trajectory correction increment vector ΔV, reverse compensation amounts are applied to the position coordinates and attitude angles of each path point in the set of unexecuted path points, so that the corrected path points can theoretically offset the currently generated execution deviation, resulting in the corrected set of path points.

[0104] The revised path point set is re-performed with spatiotemporal collision risk detection to determine whether the revised path still overlaps with the spatiotemporal occupancy sequence of movable components within the laboratory chamber. Simultaneously, based on the container type, liquid load, and sealing status of the transported object, the acceleration and angular acceleration corresponding to the revised path points are re-verified to ensure they do not exceed preset thresholds. When both the collision risk verification and fluid stability condition verification results meet the preset conditions, corresponding joint angle control commands are generated based on the revised path point set and output to the robotic arm actuator for execution. If the verification or validation results do not meet the preset conditions—that is, if the revised path still poses a collision risk or does not meet the fluid stability conditions—the conflict arbitration module is triggered to reconstruct the cooperative avoidance trajectory, repeating the aforementioned avoidance trajectory generation and verification process until an executable trajectory that meets the conditions is obtained.

[0105] In this embodiment, it should be noted that the preset threshold corresponding to the trajectory correction increment vector is determined by conducting multiple sets of trajectory execution deviation experiments in advance, recording the actual position offset of the robot's end effector under different trajectory correction increment vector values, and combining this with the requirements of the testing task for the handling position accuracy. For example, it was found through experiments that when the trajectory correction increment vector value reaches 0.15, the corresponding position offset of the robot's end effector is approximately 3 mm, exceeding the maximum allowable offset of 2 mm required by the testing task. Therefore, the preset threshold is set to 0.12 to ensure that online correction is triggered in time before the end effector position offset reaches 2 mm.

[0106] Example 2

[0107] Please see Figure 2 As shown, the parts not described in detail in this embodiment are described in Embodiment 1. A control method for a laboratory chamber robotic arm based on adaptive trajectory planning is provided, including:

[0108] S1. Obtain the load physical properties of the object being transported and the physical positions of the movable components inside the testing chamber. Construct a dynamic motion-restricted domain within the preset robotic arm motion space, and divide the robotic arm motion space into regions to generate a spatial topological field.

[0109] S2. Construct a scalar potential energy field based on the spatial topological field. Using the scalar potential energy field as the guiding direction, and combined with the dynamic motion confined domain, generate a motion trajectory that satisfies the preset fluid stability conditions.

[0110] S3. Detect the spatiotemporal collision risk of the motion trajectory in the spatial topological field. When there is a spatiotemporal collision risk, construct a cooperative avoidance trajectory on the premise of keeping the end pose of the robot unchanged.

[0111] S4. Collect the actual driving torque and end pose when executing the cooperative avoidance trajectory according to the preset cycle, and compare them with the theoretical driving torque and theoretical pose respectively to obtain the trajectory correction increment vector.

[0112] S5. When the trajectory correction increment vector exceeds the preset threshold, the unexecuted part of the cooperative avoidance trajectory is corrected online, and the corresponding control command is generated and output to the robot arm for execution.

[0113] Example 3

[0114] This embodiment discloses an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the operation mode of the laboratory chamber robotic arm control system based on adaptive trajectory planning provided above.

[0115] Since the electronic device described in this embodiment is the electronic device used in implementing the laboratory chamber robotic arm control system based on adaptive trajectory planning in the embodiments of this application, those skilled in the art can understand the specific implementation methods and various variations of the electronic device in this embodiment based on the laboratory chamber robotic arm control system based on adaptive trajectory planning described in the embodiments of this application. Therefore, how the electronic device implements the method in the embodiments of this application will not be described in detail here. As long as those skilled in the art implement the electronic device used in the laboratory chamber robotic arm control system based on adaptive trajectory planning in the embodiments of this application, it falls within the scope of protection of this application.

[0116] It should be noted that all formulas in this manual are calculated by removing dimensions and taking their numerical values. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0117] The above description is merely a preferred embodiment of the present invention, and the scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for users of ordinary technical skills, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A control system for a laboratory robot based on adaptive trajectory planning, characterized in that, include: The topology constraint module is used to obtain the load physical properties of the object being transported and the physical positions of the movable components in the test chamber, construct a dynamic motion-constrained domain within the preset manipulator motion space, and divide the manipulator motion space into regions to generate a spatial topology field. The trajectory adaptive module is used to construct a scalar potential energy field based on the spatial topology field. Using the scalar potential energy field as the guiding direction, combined with the dynamic motion confined domain, it generates a motion trajectory that meets the preset fluid stability conditions. The method for generating motion trajectories that satisfy preset fluid stability conditions includes: Candidate path points are generated step by step along the descent direction of the potential energy gradient vector in the scalar potential energy field to obtain the initial motion trajectory; based on the container type, liquid loading, and container sealing status of the object being transported, the acceleration threshold and angular acceleration threshold of the robot end effector corresponding to each path point on the initial motion trajectory are calculated. Determine whether the acceleration and angular acceleration of each path point on the initial motion trajectory exceed the corresponding acceleration threshold and angular acceleration threshold. If they exceed, adjust the motion speed and direction of the path point to obtain a motion trajectory that meets the preset fluid stability conditions. The conflict arbitration module is used to detect the spatiotemporal collision risk of the motion trajectory in the spatial topological field. When there is a spatiotemporal collision risk, a cooperative avoidance trajectory is constructed on the premise of keeping the pose of the robot end effector unchanged. The deviation perception module is used to collect the actual driving torque and end pose when executing the cooperative avoidance trajectory at a preset period, and compare them with the theoretical driving torque and theoretical pose respectively to obtain the trajectory correction increment vector. The online correction module is used to correct the unexecuted part of the cooperative avoidance trajectory online when the trajectory correction increment vector exceeds a preset threshold, and generate corresponding control commands to be output to the robot arm for execution.

2. The control system for the laboratory robot based on adaptive trajectory planning according to claim 1, characterized in that, The load physical properties of the transported object include the mass, dimensions, center of gravity, container type, liquid load, and container sealing status of the transported object; the physical location of the movable components inside the laboratory chamber includes the spatial coordinates, external dimensions, and motion status of each movable component; The mass and container type of the transported object are obtained from the laboratory task database, the external dimensions and center of gravity are determined by the pre-stored container mold or corrected by the visual measurement device, the liquid loading is obtained by the liquid level detection device, and the container sealing status is confirmed by the sealing detection device.

3. The control system for the laboratory robot based on adaptive trajectory planning according to claim 2, characterized in that, The method for generating spatial topological fields includes: The equivalent envelope of the object being transported is calculated based on the load physical properties of the object being transported. The dynamic envelope of each movable component is calculated based on the physical position and motion state of the movable components in the laboratory chamber. Based on the equivalent envelope and the dynamic envelope, the boundary of each envelope is expanded according to a preset expansion coefficient to obtain the dynamic motion confined domain. The robot's motion space is divided into several spatial grid cells according to a preset grid size. The positional relationship between each spatial grid cell and the dynamic motion restricted area is determined, and the spatial grid cells are marked as safe areas, buffer areas, or restricted areas. Based on the marking results of each spatial grid cell and the connectivity between adjacent grid cells, a spatial topology field containing regional attributes and connectivity topology is constructed.

4. The control system for the laboratory robot based on adaptive trajectory planning according to claim 3, characterized in that, The method for constructing a scalar potential energy field based on a spatial topological field includes: For spatial grid cells marked as restricted regions in the spatial topology field, a potential energy value positively correlated with the distance to the target location is assigned; for spatial grid cells marked as safe regions, a potential energy value negatively correlated with the distance to the target location is assigned; and for spatial grid cells marked as buffer regions, a transition potential energy value decreasing from the restricted region to the safe region is assigned. Based on the spatial grid cells corresponding to the current pose of the robot and the spatial grid cells corresponding to the target pose, the potential energy gradient vector is calculated; the potential energy value of each spatial grid cell is associated with and stored with the corresponding potential energy gradient vector to obtain a scalar potential energy field covering the robot's motion space.

5. The control system for the laboratory robot based on adaptive trajectory planning according to claim 4, characterized in that, The method for detecting the risk of spatiotemporal collisions of motion trajectories in a spatial topological field includes: Obtain the timestamps corresponding to each path point on the motion trajectory to obtain the spatiotemporal trajectory sequence of the motion trajectory; obtain the position state of the dynamic motion confined domain of the movable component in the test chamber at the corresponding timestamp to obtain the spatiotemporal occupancy sequence of the movable component; The spatiotemporal trajectory sequence of the motion trajectory is spatiotemporally matched with the spatiotemporal occupancy sequence of the movable component. It is determined whether there is an overlap between the spatial grid cell corresponding to the motion trajectory of the robot and the spatial grid cell corresponding to the dynamic motion restricted domain of the movable component under the same timestamp. If there is an overlap, it is determined that there is a spatiotemporal collision risk in the motion trajectory, and the time interval of the collision risk and the corresponding spatial grid cell are recorded.

6. The control system for the laboratory robot based on adaptive trajectory planning according to claim 5, characterized in that, The method for constructing cooperative avoidance trajectories includes: Based on the time interval of the collision risk and the corresponding spatial grid cell, determine the trajectory segment that needs to be avoided; while keeping the end effector pose unchanged, use the redundant degrees of freedom of the manipulator base joint and intermediate joint as the adjustment object, and search for alternative path points to bypass the corresponding spatial grid cell in the spatial topology field. A cooperative avoidance trajectory is generated based on alternative path points and path points in the original trajectory that are not affected by collision risk. The cooperative avoidance trajectory is then reviewed for spatiotemporal collision risk. When the review result meets the no-collision condition, the cooperative avoidance trajectory is used as the trajectory to be executed.

7. The control system for the laboratory robot based on adaptive trajectory planning according to claim 6, characterized in that, The method for collecting the actual driving torque and end-effector pose during the execution of the cooperative avoidance trajectory at a preset period includes: Torque sensors are installed at each joint of the robotic arm, and a pose detection device is installed at the end of the robotic arm. According to a preset acquisition cycle, the actual driving torque of each joint of the robotic arm is collected by the torque sensors, and the actual position and actual attitude angle of the end of the robotic arm are collected by the pose detection device to obtain the actual pose of the end. Based on the kinematic and dynamic models corresponding to the cooperative avoidance trajectory, the theoretical driving torque of each joint and the theoretical pose of the manipulator end effector are calculated at the same moment. The actual driving torque and actual end effector pose within the same acquisition cycle are synchronized with the theoretical driving torque and theoretical pose, respectively.

8. The control system for the laboratory robot based on adaptive trajectory planning according to claim 7, characterized in that, The method for obtaining the trajectory correction increment vector includes: The difference between the actual driving torque and the theoretical driving torque within the same acquisition cycle is calculated to obtain the torque deviation; the difference between the actual pose and the theoretical pose of the end effector within the same acquisition cycle is calculated to obtain the pose deviation, which includes the position deviation component and the attitude angle deviation component. The torque deviation, position deviation component, and attitude angle deviation component are normalized, and then weighted according to preset weighting coefficients to obtain the trajectory correction increment vector.

9. The control system for the laboratory robot based on adaptive trajectory planning according to claim 8, characterized in that, The method for online correction of the unexecuted portion of the cooperative avoidance trajectory includes: Based on the actual pose of the robot arm's end effector at the current moment, determine the set of unexecuted path points in the cooperative avoidance trajectory; based on the trajectory correction increment vector, compensate and correct the position coordinates and attitude angles of each path point in the set of unexecuted path points to obtain the corrected set of path points. The modified path point set is subjected to spatiotemporal collision risk verification and fluid stability condition verification. When the verification and validation results meet the preset conditions, the corresponding control command is generated based on the modified path point set and output to the robot for execution; when the verification and validation results do not meet the preset conditions, the cooperative avoidance trajectory is reconstructed.

Images