Robot tool switching method and system based on multi-modal recognition

Abstract

The invention relates to the technical field of robot switching, and discloses a robot tool switching method and system based on multi-modal recognition, and the system comprises an acquisition module which is configured to determine the initial tool release speed and the initial tool grabbing speed of an operation robot based on the task information of the current stage and the task information of the next stage; the judging module is configured to judge whether the initial tool release speed is adjusted or not based on the tool grabbing state information; the controller is further configured to judge whether the initial tool grabbing speed is adjusted or not according to the matching degree; the multi-modal adjustment module is configured to determine a release adjustment coefficient of the initial tool release speed based on the historical release record and obtain a final tool release speed; the grabber is also configured to determine a grabber adjustment coefficient of the initial tool grabber speed based on the contact state information and the grabber force information, and obtain a final tool grabber speed; and an execution module. The tool switching efficiency and accuracy of the operation robot can be effectively improved.

Description

Technical Field

[0001] This invention relates to the field of robot switching technology, and more specifically, to a robot tool switching method and system based on multimodal recognition. Background Technology

[0002] Traditional robotic tool switching technologies mostly employ a single recognition method, such as relying solely on visual recognition or sensor data for tool switching operations. This single-modal recognition approach has numerous limitations, failing to comprehensively and accurately perceive the working environment and tool status. For example, visual recognition is prone to errors in poor lighting or obstructed conditions; while sensor data alone may not provide sufficient environmental information, leading to collisions and unstable gripping during tool switching. Furthermore, existing robotic tool switching methods often lack dynamic adjustment mechanisms for tool release and gripping speeds. Under different work scenarios and task requirements, fixed tool release and gripping speeds may not meet actual operational requirements, thus affecting work efficiency and quality. For instance, in tasks with high precision requirements, excessively fast tool release speeds may lead to inaccurate tool placement; while excessively slow gripping speeds reduce overall work efficiency.

[0003] Therefore, it is necessary to design a robot tool switching method and system based on multimodal recognition to solve the problems existing in the current technology. Summary of the Invention

[0004] In view of this, the present invention proposes a robot tool switching method and system based on multimodal recognition, aiming to solve the problem that the current single-modal recognition method has many limitations and cannot fully and accurately perceive the working environment and tool status.

[0005] In one aspect, the present invention proposes a robot tool switching system based on multimodal recognition, comprising: The data acquisition module is configured to identify the working robot, acquire the current stage task information and the next stage task information of the working robot, and determine the initial tool release speed and the initial tool grasping speed of the working robot based on the current stage task information and the next stage task information. The judgment module is configured to collect the tool grasping status information of the robot when the robot completes the current stage task and moves to the target tool library, and determine whether to adjust the initial tool release speed based on the tool grasping status information. It is also configured to determine the matching degree between the position of the end effector of the work robot and the position of the target tool, and determine whether to adjust the initial tool grasping speed based on the matching degree; The multimodal adjustment module is configured to, when it is determined that the initial tool release speed should be adjusted, collect the historical release records of the task tools used in the current stage of the task, determine the release adjustment coefficient of the initial tool release speed based on the historical release records, and obtain the final tool release speed. It is also configured to collect contact state information between the end effector of the work robot and the target tool and the grasping force information of the end effector, determine the grasping adjustment coefficient of the initial tool grasping speed based on the contact state information and the grasping force information, and obtain the final tool grasping speed; The execution module is configured to control the work robot to perform tool release and grasping operations based on the final tool release speed and the final tool grasping speed.

[0006] Furthermore, when determining the initial tool release speed and initial tool grasping speed of the robot based on the current stage task information and the next stage task information, the process includes: The task information at the current stage is parsed to obtain the task tool type, task tool weight, and task tool usage time; The next stage task information is analyzed to obtain the target tool type, target tool weight, and the required installation accuracy of the target tool; The initial tool release speed of the work robot is determined based on the task tool type, task tool weight, and task tool usage time; The initial tool-grabbing speed of the robot is determined based on the target tool type, target tool weight, and required installation accuracy of the target tool.

[0007] Furthermore, when determining the initial tool release speed of the work robot based on the task tool type, task tool weight, and task tool usage duration, the following is included: Construct a release vector group based on the task tool type, task tool weight, and task tool usage duration; The release vector group is compared with the historical release group. If there is a historical release vector group in the historical release group that is the same as the release vector group, then the historical tool release speed corresponding to the historical release vector group is taken as the initial tool release speed. If there is no historical release vector group in the historical release group that is the same as the release vector group, then all historical release vector groups that are the same as the release vector group in terms of task tool type and task tool weight are selected and constructed as the first filtering group. Obtain the average and maximum historical tool usage durations in the first filter group, and record them as the historical average tool usage duration and the historical maximum tool usage duration, respectively. The initial tool release speed is determined by comparing the usage time of the task tools in the release vector group with the historical average tool usage time and the historical maximum tool usage time. When the usage time of the task tool is less than or equal to the historical average usage time of the tool, the initial tool release speed is determined as the first release speed; When the duration of the task tool usage is greater than the historical average tool usage duration, and less than or equal to the historical maximum tool usage duration, the initial tool release speed is determined as the second release speed. When the duration of the task tool usage exceeds the historical maximum tool usage duration, the initial tool release speed is determined to be the third release speed.

[0008] Furthermore, when determining the initial tool grasping speed of the robot based on the target tool type, target tool weight, and required installation accuracy of the target tool, the process includes: The basic tool grasping speed of the robot is determined based on the target tool type. The optimization coefficient of the basic tool's grasping speed is determined based on the target tool's weight and the required installation accuracy of the target tool, and the product of the optimization coefficient and the basic tool's grasping speed is used as the initial tool's grasping speed. When determining the optimization coefficient for the gripping speed of the basic tool based on the weight of the target tool and the required installation accuracy of the target tool, the following steps are included: The weight of the target tool is compared with the weight of the standard tool, and the required installation accuracy of the target tool is compared with the standard installation accuracy. The optimization coefficient is determined based on the comparison results. When the weight of the target tool is greater than or equal to the weight of the standard tool, and the required installation accuracy of the target tool is greater than or equal to the standard installation accuracy, the optimization coefficient is determined as the first optimization coefficient. When the weight of the target tool is greater than or equal to the weight of the standard tool, and the required installation accuracy of the target tool is less than the standard installation accuracy, the optimization coefficient is determined to be the second optimization coefficient. When the weight of the target tool is less than the weight of the standard tool, and the required installation accuracy of the target tool is greater than or equal to the standard installation accuracy, the optimization coefficient is determined to be the third optimization coefficient. When the weight of the target tool is less than the weight of the standard tool, and the required installation accuracy of the target tool is less than the standard installation accuracy, the optimization coefficient is determined to be the fourth optimization coefficient.

[0009] Furthermore, when determining whether to adjust the initial tool release speed based on the tool's grasping state information, the process includes: Feature extraction is performed on the tool's grasping state information to obtain tool grasping feature values; The tool's grasping feature value is compared with the tool's grasping threshold, and the initial tool release speed is adjusted based on the comparison result. When the tool's grasping feature value is within the tool's grasping threshold, it is determined that the initial tool release speed will not be adjusted. Otherwise, it is determined that the release speed of the initial tool should be adjusted.

[0010] Further, determining the matching degree between the end effector position of the robot and the target tool position, and deciding whether to adjust the initial tool grasping speed based on the matching degree, includes: A three-dimensional Cartesian coordinate system is established with the geometric center of the target tool as the origin, and the position coordinates of the end effector of the robot and the position coordinates of the target tool are obtained respectively. Calculate the distance deviation and angle deviation between the end effector and the target tool based on the position coordinates of the end effector and the target tool. The degree of matching between the end effector position and the target tool position is calculated based on the distance deviation value and the angle deviation value; The matching degree is compared with the matching degree threshold. If the matching degree is greater than or equal to the matching degree threshold, it is determined that the initial tool's crawling speed will not be adjusted. Otherwise, it is determined that the initial tool's grasping speed should be adjusted.

[0011] Further, when determining the release adjustment coefficient for the initial tool release speed based on the historical release records and obtaining the final tool release speed, the process includes: The historical release records are parsed to obtain the maximum impact force value of the end effector corresponding to the release process of each task tool, and a maximum impact force value data set is constructed. The historical release deviation values ​​of the mission tool are obtained based on the maximum impact force value data set. The historical release deviation value is compared with the first historical release deviation value and the second historical release deviation value, and the release adjustment coefficient is determined based on the comparison result; wherein, the first historical release deviation value is less than the second historical release deviation value; When the historical release deviation value is less than or equal to the first historical release deviation value, the release adjustment coefficient is determined to be the first release adjustment coefficient. When the historical release deviation value is greater than the first historical release deviation value and less than or equal to the second historical release deviation value, the release adjustment coefficient is determined to be the second release adjustment coefficient. When the historical release deviation value is greater than the second historical release deviation value, the release adjustment coefficient is determined to be the third release adjustment coefficient; The final tool release speed is obtained by multiplying the release adjustment coefficient by the initial tool release speed.

[0012] Further, when determining the initial tool grasping speed adjustment coefficient based on the contact state information and grasping force information, and obtaining the final tool grasping speed, the process includes: The contact state information is analyzed to obtain the contact angle and contact area between the end effector and the target tool; The grasping force information is parsed to obtain the actual grasping force of the end effector; The gripping influence index is determined based on the contact angle, contact area, and actual gripping force. The crawling impact index is compared with historical data, and the crawling adjustment coefficient is determined based on the comparison results; When there is a historical crawling impact index in the historical data that is the same as the crawling impact index, the historical crawling adjustment coefficient corresponding to the historical crawling impact index shall be used as the crawling adjustment coefficient. When there is a historical crawling influence index in the historical data that is the same as the crawling influence index, calculate the difference between the crawling influence index and each of the historical crawling influence indices one by one, and record it as the influence difference. Obtain the absolute value of the impact difference, and denot it as the absolute impact difference; The capture adjustment coefficient is determined based on the absolute impact difference. The final tool grasping speed is the product of the grasping adjustment coefficient and the initial tool grasping speed.

[0013] Further, when determining the capture adjustment coefficient based on the absolute impact difference, it includes: When the historical crawling adjustment coefficient corresponding to the absolute impact difference is unique, the historical crawling adjustment coefficient shall be used as the crawling adjustment coefficient. When the historical crawling adjustment coefficient corresponding to the absolute impact difference is not unique, the average value of all the historical crawling adjustment coefficients is calculated, and the average value is used as the crawling adjustment coefficient.

[0014] Compared with existing technologies, the beneficial effects of this invention are as follows: The robot tool switching system based on multimodal recognition provided by this invention can effectively improve the efficiency and accuracy of tool switching for operational robots. Through multimodal recognition and comprehensive consideration of multiple parameters, the tool release speed and gripping speed are scientifically and rationally adjusted, reducing problems such as tool damage and operational errors caused by inappropriate speeds. Regarding tool release, a release vector group is constructed based on the weight and usage time of the task tool, and compared with historical release groups. Adjustments are made in conjunction with tool gripping status information and historical release records, making the release speed more closely match actual needs. This not only avoids tool drops and collisions caused by excessively fast release speeds but also prevents wasted time due to excessively slow release speeds, improving the safety and efficiency of tool release. Regarding tool gripping, the initial gripping speed is determined based on the type, weight, and required installation accuracy of the target tool, and then adjusted in conjunction with the matching degree between the end effector position and the target tool position, contact status information, and gripping force information. This ensures that the robot can grasp tools accurately without being too fast or too slow, thus improving the success rate and accuracy of tool grasping.

[0015] In another aspect, this invention also proposes a robot tool switching method based on multimodal recognition, comprising the following steps: Identify the work robot, collect the current stage task information and the next stage task information of the work robot, and determine the initial tool release speed and the initial tool grasping speed of the work robot based on the current stage task information and the next stage task information. When the robot completes the current stage task and moves to the target tool library, the tool grasping status information of the robot is collected, and it is determined whether to adjust the initial tool release speed based on the tool grasping status information. Determine the matching degree between the position of the end effector of the robot and the position of the target tool, and determine whether to adjust the initial tool grasping speed based on the matching degree; When it is determined that the initial tool release speed needs to be adjusted, the historical release records of the task tools used in the current stage of the task are collected, the release adjustment coefficient of the initial tool release speed is determined based on the historical release records, and the final tool release speed is obtained. The contact state information between the end effector of the work robot and the target tool and the grasping force information of the end effector are collected. Based on the contact state information and the grasping force information, the grasping adjustment coefficient of the initial tool grasping speed is determined, and the final tool grasping speed is obtained. The final tool release speed and the final tool grasping speed are used to control the working robot to perform tool release and grasping operations.

[0016] It is understandable that the above-mentioned robot tool switching methods and systems based on multimodal recognition have the same beneficial effects, and will not be elaborated further here. Attached Figure Description

[0017] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 This is a structural block diagram of a robot tool switching system based on multimodal recognition provided in an embodiment of the present invention; Figure 2 A flowchart of a robot tool switching method based on multimodal recognition provided in an embodiment of the present invention. Detailed Implementation

[0018] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the disclosure to those skilled in the art. It should be noted that, unless otherwise specified, embodiments and features in the embodiments of the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0019] See Figure 1 As shown in some embodiments of this application, this embodiment provides a robot tool switching system based on multimodal recognition, including: The data acquisition module is configured to identify the working robot, acquire the current stage task information and the next stage task information of the working robot, and determine the initial tool release speed and the initial tool grasping speed of the working robot based on the current stage task information and the next stage task information. The judgment module is configured to collect the tool grasping status information of the robot when the robot completes the current stage task and moves to the target tool library, and determine whether to adjust the initial tool release speed based on the tool grasping status information. It is also configured to determine the matching degree between the position of the end effector of the work robot and the position of the target tool, and determine whether to adjust the initial tool grasping speed based on the matching degree; The multimodal adjustment module is configured to, when it is determined that the initial tool release speed should be adjusted, collect the historical release records of the task tools used in the current stage of the task, determine the release adjustment coefficient of the initial tool release speed based on the historical release records, and obtain the final tool release speed. It is also configured to collect contact state information between the end effector of the work robot and the target tool and the grasping force information of the end effector, determine the grasping adjustment coefficient of the initial tool grasping speed based on the contact state information and the grasping force information, and obtain the final tool grasping speed; The execution module is configured to control the work robot to perform tool release and grasping operations based on the final tool release speed and the final tool grasping speed.

[0020] It is understood that the multimodal recognition-based robot tool switching system provided in this embodiment can effectively improve the efficiency and accuracy of tool switching for operational robots. Through multimodal recognition and comprehensive consideration of multiple parameters, the tool release speed and gripping speed are scientifically and rationally adjusted, reducing problems such as tool damage and operational errors caused by inappropriate speeds. Regarding tool release, a release vector group is constructed based on the weight and usage time of the task tool, and compared with historical release groups. Adjustments are made in conjunction with tool gripping status information and historical release records, making the release speed more closely match actual needs. This not only avoids tool drops and collisions due to excessively fast release speeds but also prevents wasted time due to excessively slow release speeds, improving the safety and efficiency of tool release. Regarding tool gripping, the initial gripping speed is determined based on the type, weight, and required installation accuracy of the target tool, and then adjusted in conjunction with the matching degree between the end effector position and the target tool position, contact status information, and gripping force information. This ensures that the operational robot will not fail to grip the tool accurately due to excessive speed, nor will it reduce work efficiency due to excessively slow speeds, improving the success rate and accuracy of tool gripping.

[0021] Specifically, determining the initial tool release speed and initial tool grasping speed of the robot based on the current stage task information and the next stage task information includes: The task information at the current stage is parsed to obtain the task tool type, task tool weight, and task tool usage time; The next stage task information is analyzed to obtain the target tool type, target tool weight, and the required installation accuracy of the target tool; The initial tool release speed of the work robot is determined based on the task tool type, task tool weight, and task tool usage time; The initial tool-grabbing speed of the robot is determined based on the target tool type, target tool weight, and required installation accuracy of the target tool.

[0022] Understandably, analyzing current and next-stage task information allows for a precise understanding of the characteristics of both the task tool and the target tool. Different types of task tools, with their varying structures and materials, affect release speed; for example, irregularly shaped or fragile tools require slow release to prevent damage. The weight and usage duration of the task tool are also crucial; heavy tools may be damaged with rapid release, while older tools require speed adjustments for safety. The type, weight, and installation precision of the target tool are equally important. Different types require different grasping methods and levels of difficulty; for example, small, precision tools require slow grasping to ensure accuracy. The weight of the target tool affects the grasping force and speed, while heavy tools require fast grasping to ensure stability. Installation precision determines grasping accuracy; higher precision results in slower, more stable grasping. Determining the initial speed based on task information provides a foundation for multimodal adjustments. Subsequent adjustments, incorporating factors such as tool grasping status, historical release records, contact status, and grasping force, can optimize tool release and grasping speeds, achieving efficient, accurate, and safe tool switching for the robot.

[0023] Specifically, determining the initial tool release speed of the work robot based on the task tool type, task tool weight, and task tool usage duration includes: Construct a release vector group based on the task tool type, task tool weight, and task tool usage duration; The release vector group is compared with the historical release group. If there is a historical release vector group in the historical release group that is the same as the release vector group, then the historical tool release speed corresponding to the historical release vector group is taken as the initial tool release speed. If there is no historical release vector group in the historical release group that is the same as the release vector group, then all historical release vector groups that are the same as the release vector group in terms of task tool type and task tool weight are selected and constructed as the first filtering group. Obtain the average and maximum historical tool usage durations in the first filter group, and record them as the historical average tool usage duration and the historical maximum tool usage duration, respectively. The initial tool release speed is determined by comparing the usage time of the task tools in the release vector group with the historical average tool usage time and the historical maximum tool usage time. When the usage time of the task tool is less than or equal to the historical average usage time of the tool, the initial tool release speed is determined as the first release speed; When the duration of the task tool usage is greater than the historical average tool usage duration, and less than or equal to the historical maximum tool usage duration, the initial tool release speed is determined as the second release speed. When the duration of the task tool usage exceeds the historical maximum tool usage duration, the initial tool release speed is determined to be the third release speed.

[0024] Understandably, constructing a release vector group and comparing it with historical release groups allows us to determine the initial tool release speed based on historical experience. If a matching historical release vector group exists, directly using the corresponding historical tool release speed can quickly and accurately provide a suitable release speed for the current task. If no such group exists, a first filtering group is constructed, and the initial release speed is determined by combining historical averages and maximum tool usage time, taking into account the tool usage time factor. The first release speed is used when the tool usage time is short, at which point the tool wear is minimal and the structure is stable; releasing at this speed improves efficiency and avoids damaging the tool. The second release speed is for when the tool usage time is moderate, at which point the tool has some wear, and the speed needs to be appropriately reduced to ensure safety and stability. The third release speed is for when the tool usage time is long and the wear is severe; to avoid tool damage or dropping, a slower release speed is required.

[0025] It should also be noted that when only one set of historical release vectors is selected, the difference between the current task tool usage time and the historical task tool usage time in the unique set of historical release vectors is calculated to obtain the duration difference. This duration difference is then compared with a preset tool release speed mapping table to determine the initial tool release speed. The tool release speed mapping table is the basis for precisely adjusting the tool release speed based on changes in task tool usage time. It records the tool release speed corresponding to different duration differences. For example, when the duration difference is 3 (in hours), the corresponding initial tool release speed is 10 cm / s; when the duration difference is 5 (in hours), the corresponding initial tool release speed is 8 cm / s; and when the duration difference is 8 (in hours), the corresponding initial tool release speed is 5 cm / s. This method of determining the initial tool release speed based on the duration difference and the tool release speed mapping table can more accurately take into account the specific differences in task tool usage time and further optimize the setting of the initial tool release speed.

[0026] Specifically, when determining the initial tool-grabbing speed of the robot based on the target tool type, target tool weight, and required installation accuracy of the target tool, the following steps are included: The basic tool grasping speed of the robot is determined based on the target tool type. The optimization coefficient of the basic tool's grasping speed is determined based on the target tool's weight and the required installation accuracy of the target tool, and the product of the optimization coefficient and the basic tool's grasping speed is used as the initial tool's grasping speed. When determining the optimization coefficient for the gripping speed of the basic tool based on the weight of the target tool and the required installation accuracy of the target tool, the following steps are included: The weight of the target tool is compared with the weight of the standard tool, and the required installation accuracy of the target tool is compared with the standard installation accuracy. The optimization coefficient is determined based on the comparison results. When the weight of the target tool is greater than or equal to the weight of the standard tool, and the required installation accuracy of the target tool is greater than or equal to the standard installation accuracy, the optimization coefficient is determined as the first optimization coefficient. When the weight of the target tool is greater than or equal to the weight of the standard tool, and the required installation accuracy of the target tool is less than the standard installation accuracy, the optimization coefficient is determined to be the second optimization coefficient. When the weight of the target tool is less than the weight of the standard tool, and the required installation accuracy of the target tool is greater than or equal to the standard installation accuracy, the optimization coefficient is determined to be the third optimization coefficient. When the weight of the target tool is less than the weight of the standard tool, and the required installation accuracy of the target tool is less than the standard installation accuracy, the optimization coefficient is determined to be the fourth optimization coefficient.

[0027] Understandably, the optimization coefficients are ranked as follows: First optimization coefficient < Second optimization coefficient < Third optimization coefficient < Fourth optimization coefficient. When the target tool is heavy and requires high installation precision (corresponding to the first optimization coefficient), the gripping speed needs to be reduced to ensure stability and installation accuracy, so the first optimization coefficient is the smallest. When the target tool is heavy but the installation precision requirement is low (corresponding to the second optimization coefficient), the gripping speed can be appropriately increased while ensuring stable tool gripping, so the second optimization coefficient is greater than the first. If the target tool is light but the installation precision requirement is high (corresponding to the third optimization coefficient), although the tool is lighter and easier to grip, the gripping speed cannot be too fast to achieve high installation precision, although it can be slightly faster than the previous two cases, so the third optimization coefficient is greater than the second. When the target tool is light and the installation precision requirement is also low (corresponding to the fourth optimization coefficient), the restriction on gripping speed is less, and a relatively faster gripping speed can be used, so the fourth optimization coefficient is the largest. By setting such optimization coefficients, the initial tool grasping speed of the robot can be determined more scientifically and rationally based on the different characteristics of the target tool, providing a more accurate basis for subsequent multimodal adjustments and further improving the efficiency and accuracy of tool switching of the robot.

[0028] Specifically, when determining whether to adjust the initial tool release speed based on the tool's grasping state information, the process includes: Feature extraction is performed on the tool's grasping state information to obtain tool grasping feature values; The tool's grasping feature value is compared with the tool's grasping threshold, and the initial tool release speed is adjusted based on the comparison result. When the tool's grasping feature value is within the tool's grasping threshold, it is determined that the initial tool release speed will not be adjusted. Otherwise, it is determined that the release speed of the initial tool should be adjusted.

[0029] It is understandable that extracting features from the tool grasping status information and comparing it with a threshold is a scientific and effective method of judgment. The tool grasping feature value reflects the actual state of the robot during tool grasping, while the tool grasping threshold is a reasonable range determined based on extensive experiments and experience. When the tool grasping feature value is within the threshold, it indicates that the robot's grasping state is normal, and operating at the initial tool release speed is appropriate, requiring no adjustment. This avoids unnecessary adjustments that waste time and resources, ensuring a smooth tool switching process. Conversely, when the tool grasping feature value exceeds the tool grasping threshold, it indicates an abnormal grasping state of the robot. This may be due to inaccurate tool placement, end effector malfunction, or interference from the external environment. In this case, if the initial tool release speed is not adjusted in time, problems may occur during tool release, such as tool falling or collisions, affecting the safety and efficiency of the operation. Therefore, it is necessary to adjust the initial tool release speed. Subsequent multimodal adjustment modules collect relevant information and determine release adjustment coefficients to obtain a final tool release speed more suitable for the current situation, ensuring smooth tool release operations.

[0030] In this embodiment, the preferred tool grasping characteristic values ​​are the tool's wobbling amplitude and the tightness of the fit between the end effector and the tool when the robot grasps the tool. The tool's wobbling amplitude can be obtained using an accelerometer mounted on the tool or end effector. Excessive wobbling amplitude may cause the tool to collide with surrounding objects during release or grasping. The tightness of the fit between the end effector and the tool can be measured by a pressure sensor detecting the contact pressure between them. Insufficient fit may cause the tool to fall off during grasping or release.

[0031] Specifically, determining the matching degree between the end effector position of the robot and the target tool position, and deciding whether to adjust the initial tool grasping speed based on the matching degree, includes: A three-dimensional Cartesian coordinate system is established with the geometric center of the target tool as the origin, and the position coordinates of the end effector of the robot and the position coordinates of the target tool are obtained respectively. Calculate the distance deviation and angle deviation between the end effector and the target tool based on the position coordinates of the end effector and the target tool. The degree of matching between the end effector position and the target tool position is calculated based on the distance deviation value and the angle deviation value; The matching degree is compared with the matching degree threshold. If the matching degree is greater than or equal to the matching degree threshold, it is determined that the initial tool's crawling speed will not be adjusted. Otherwise, it is determined that the initial tool's grasping speed should be adjusted.

[0032] Understandingly, establishing a three-dimensional Cartesian coordinate system and calculating the matching degree is a method to accurately quantify the positional relationship between the end effector and the target tool. Distance deviation reflects the spatial proximity of the two, while angular deviation reflects their relative posture differences. The matching degree threshold is a standard determined based on actual operational requirements and experience; it represents the degree of positional matching that ensures successful tool grasping. When the matching degree is greater than or equal to the threshold, it indicates a good positional relationship between the end effector and the target tool, and operating at the initial tool grasping speed is feasible without adjustment. This maintains the stability and efficiency of the grasping process, avoiding unnecessary adjustments that could affect the work progress. Conversely, when the matching degree is less than the threshold, it means there is a significant positional deviation between the end effector and the target tool. This could be due to improper target tool placement, robot positioning errors, or external interference. In this case, if the initial tool grasping speed is not adjusted promptly, it may lead to grasping failure, such as tool slippage or inaccurate grasping, reducing the success rate and efficiency of the operation. Therefore, it is reasonable to adjust the initial tool grasping speed. By further collecting relevant information and determining the grasping adjustment coefficient through the subsequent multimodal adjustment module, a final tool grasping speed that is more adapted to the actual position matching situation can be obtained, ensuring that the robot can accurately and stably grasp the target tool.

[0033] In this embodiment, the matching degree calculation process is as follows: First, the distance deviation value and angle deviation value are normalized and mapped to the [0,1] interval to eliminate the influence of different units on the calculation results, since their units and magnitudes may be different, normalization makes them comparable. Then, according to the importance of distance and angle to the grasping operation in actual operation, different weights are assigned to the normalized distance deviation value and angle deviation value respectively. For example, if angle accuracy is more critical to grasping, a higher weight is assigned to the angle deviation value, and vice versa. Next, the normalized distance deviation value and angle deviation value are multiplied by their corresponding weights and then added together to obtain the matching degree between the end effector position and the target tool position. This comprehensively considers distance and angle factors, making the matching degree calculation result more accurately reflect the actual positional relationship. Based on this accurate matching degree and the comparison with the matching degree threshold, it is possible to more scientifically determine whether to adjust the initial tool grasping speed, thereby improving the accuracy and efficiency of tool switching of the robot.

[0034] Specifically, when determining the release adjustment coefficient for the initial tool release speed based on the historical release records and obtaining the final tool release speed, the process includes: The historical release records are parsed to obtain the maximum impact force value of the end effector corresponding to the release process of each task tool, and a maximum impact force value data set is constructed. The historical release deviation values ​​of the mission tool are obtained based on the maximum impact force value data set. The historical release deviation value is compared with the first historical release deviation value and the second historical release deviation value, and the release adjustment coefficient is determined based on the comparison result; wherein, the first historical release deviation value is less than the second historical release deviation value; When the historical release deviation value is less than or equal to the first historical release deviation value, the release adjustment coefficient is determined to be the first release adjustment coefficient. When the historical release deviation value is greater than the first historical release deviation value and less than or equal to the second historical release deviation value, the release adjustment coefficient is determined to be the second release adjustment coefficient. When the historical release deviation value is greater than the second historical release deviation value, the release adjustment coefficient is determined to be the third release adjustment coefficient; The final tool release speed is obtained by multiplying the release adjustment coefficient by the initial tool release speed.

[0035] Understandably, the release adjustment coefficients are ranked as follows: First Release Adjustment Coefficient < Second Release Adjustment Coefficient < Third Release Adjustment Coefficient. When the historical release deviation is small (less than or equal to the first historical release deviation, corresponding to the first release adjustment coefficient), it indicates that the task tool's deviation during the historical release process is minimal, and the release process is relatively stable. In this case, to improve release efficiency, the release speed can be appropriately increased, hence the first release adjustment coefficient is small. When the historical release deviation is in a moderate range (greater than the first historical release deviation but less than or equal to the second historical release deviation, corresponding to the second release adjustment coefficient), the task tool's release deviation has increased, but remains within an acceptable range. To ensure relative stability in the release process, the release speed needs to be adjusted to some extent, hence the second release adjustment coefficient is greater than the first. If the historical release deviation is large (greater than the second historical release deviation, corresponding to the third release adjustment coefficient), it indicates that the task tool's deviation during the historical release process is significant, and the release process is unstable. To ensure the safety and accuracy of the tool's release, the release speed needs to be significantly reduced, hence the third release adjustment coefficient is the largest. By setting such a release adjustment coefficient, the initial tool release speed can be adjusted more scientifically and reasonably based on the historical deviation of the task tool release, thereby obtaining a final tool release speed that is more suitable for the current task and further improving the safety and efficiency of tool switching for the work robot.

[0036] In this embodiment, the calculation process for the historical release deviation value is as follows: First, statistical analysis is performed on the maximum impact force data set to calculate the mean; then, the absolute value of the difference between each maximum impact force value and the mean is calculated to measure the degree of deviation between the impact force borne by the end effector and the average impact force during each release process. The larger the absolute value of the difference, the greater the difference; next, these absolute values ​​are summed to obtain a total, reflecting the overall deviation; finally, the sum is divided by the number of data points, and the result is the historical release deviation value. This calculation method can comprehensively consider impact force fluctuations and more accurately reflect the degree of deviation of the task tool's historical release. Based on this accurate historical release deviation value and comparison with the first and second historical release deviation values, the release adjustment coefficient can be scientifically and reasonably determined to obtain a final tool release speed that is more suitable for the current task, ensuring the safety and efficiency of tool switching for the work robot.

[0037] Specifically, when determining the initial tool gripping speed adjustment coefficient based on the contact state information and gripping force information, and obtaining the final tool gripping speed, the process includes: The contact state information is analyzed to obtain the contact angle and contact area between the end effector and the target tool; The grasping force information is parsed to obtain the actual grasping force of the end effector; The gripping influence index is determined based on the contact angle, contact area, and actual gripping force. The crawling impact index is compared with historical data, and the crawling adjustment coefficient is determined based on the comparison results; When there is a historical crawling impact index in the historical data that is the same as the crawling impact index, the historical crawling adjustment coefficient corresponding to the historical crawling impact index shall be used as the crawling adjustment coefficient. When there is a historical crawling influence index in the historical data that is the same as the crawling influence index, calculate the difference between the crawling influence index and each of the historical crawling influence indices one by one, and record it as the influence difference. Obtain the absolute value of the impact difference, and denot it as the absolute impact difference; The capture adjustment coefficient is determined based on the absolute impact difference. The final tool grasping speed is the product of the grasping adjustment coefficient and the initial tool grasping speed.

[0038] Understandably, analyzing contact state and grasping force information allows for a comprehensive understanding of the robot's actual performance when grasping target tools. Contact angle and contact area reflect the contact method and tightness between the end effector and the target tool, while the actual grasping force reflects the intensity of the grasp. These three factors together constitute the grasping influence index, which comprehensively reflects the actual effect of the grasping operation. Comparing the grasping influence index with historical data allows for determining adjustment strategies based on past experience. If identical grasping influence indices exist in historical data, directly using the corresponding historical grasping adjustment coefficient can quickly and accurately determine the direction and magnitude of adjustments, improving decision-making efficiency. When identical grasping influence indices do not exist, calculating the influence difference and absolute influence difference quantifies the degree of difference between the current grasping situation and historical situations. Determining the grasping adjustment coefficient based on this degree of difference allows for more precise adjustments to adapt to the actual situation. The final tool grasping speed determined in this way considers both the current actual grasping operation and historical experience, better meeting the robot's needs for accurate and stable grasping of target tools in different scenarios. This helps to further improve the efficiency and quality of tool switching for operational robots, ensuring the smooth operation of the entire process, reducing problems such as tool damage and operational errors caused by improper grasping, and providing more reliable support for automated operations in industrial production and other fields. At the same time, this adjustment method, based on comprehensive judgment of multiple factors and reference to historical data, also reflects the system's intelligence and flexibility, enabling it to continuously adapt to changes in different target tools and operating environments.

[0039] In this embodiment, the calculation process of the grasping influence index is as follows: First, different weights are assigned to the contact angle, contact area, and actual grasping force. These weights are determined based on their importance in the grasping operation. For example, if the contact angle is crucial for successfully grasping the tool, it can be given a higher weight; while the weights of the contact area and actual grasping force are reasonably allocated according to their relative importance. Next, the contact angle, contact area, and actual grasping force are standardized. The contact angle can be quantified based on its deviation from the ideal contact angle; the smaller the deviation, the higher the standardized value. The contact area is standardized by the ratio to the theoretically optimal contact area; the actual grasping force can be compared with a preset appropriate grasping force range and standardized. Then, the standardized contact angle, contact area, and actual grasping force are multiplied by their respective weights. Finally, the three products are added together, and the result is the grasping influence index. This calculation method comprehensively considers contact state information and grasping force information, which can accurately reflect the actual effect of the robot when grasping the target tool. It provides a reliable basis for determining the grasping adjustment coefficient, making the final tool grasping speed adjustment more scientific and reasonable, and further improving the performance and quality of the robot's tool switching.

[0040] Specifically, when determining the capture adjustment coefficient based on the absolute impact difference, it includes: When the historical crawling adjustment coefficient corresponding to the absolute impact difference is unique, the historical crawling adjustment coefficient shall be used as the crawling adjustment coefficient. When the historical crawling adjustment coefficient corresponding to the absolute impact difference is not unique, the average value of all the historical crawling adjustment coefficients is calculated, and the average value is used as the crawling adjustment coefficient.

[0041] Understandably, when the historical adjustment coefficient corresponding to the absolute impact difference is unique, directly using that historical adjustment coefficient is appropriate because it indicates a high degree of similarity between the current capture situation and a specific historical situation. Using this historical adjustment coefficient maximizes the use of past experience, ensuring the accuracy and effectiveness of the adjustment. However, when the historical adjustment coefficient corresponding to the absolute impact difference is not unique, it indicates a correlation between the current capture situation and multiple historical situations, but not identical ones. In this case, calculating the average of all historical adjustment coefficients as the capture adjustment coefficient is a comprehensive approach. This balances the influence of different historical situations, avoiding over- or under-adjustment due to relying on only a single historical case. The capture adjustment coefficient determined in this way better adapts to the actual capture situation by comprehensively considering historical experience. This method of determining the capture adjustment coefficient based on the absolute impact difference further demonstrates the system's intelligence and flexibility. It can not only make accurate adjustments quickly when there are clear historical counterparts, but also derive reasonable adjustment plans by comprehensively analyzing historical data when facing complex situations. This helps the robot to precisely adjust its grasping speed according to the actual situation in various grasping scenarios, thereby improving the efficiency and quality of tool switching. At the same time, this adjustment mechanism based on historical data and comprehensive judgment of multiple factors also enables the robot to better cope with various complex and changing working environments, reduce various problems caused by improper grasping, and provide more stable and reliable support for automated operations in industrial production and other fields.

[0042] See Figure 2 As shown in some embodiments of this application, this embodiment provides a robot tool switching method based on multimodal recognition, including the following steps: S100: Determine the working robot, collect the current stage task information and the next stage task information of the working robot, and determine the initial tool release speed and the initial tool grasping speed of the working robot based on the current stage task information and the next stage task information. S200: When the robot completes the current stage task and moves to the target tool library, the tool grasping status information of the robot is collected, and the initial tool release speed is adjusted based on the tool grasping status information. S300: Determine the matching degree between the position of the end effector of the work robot and the position of the target tool, and determine whether to adjust the initial tool grasping speed based on the matching degree; S400: When it is determined that the initial tool release speed needs to be adjusted, the historical release records of the task tools used in the current stage of the task are collected, the release adjustment coefficient of the initial tool release speed is determined based on the historical release records, and the final tool release speed is obtained. S500: Collect the contact state information between the end effector of the work robot and the target tool and the grasping force information of the end effector; determine the grasping adjustment coefficient of the initial tool grasping speed based on the contact state information and the grasping force information; and obtain the final tool grasping speed. S600: Control the working robot to perform tool release and grasping operations based on the final tool release speed and the final tool grasping speed.

[0043] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program goods. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program goods embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0044] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program goods according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0045] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0046] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0047] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A robot tool switching system based on multimodal recognition, characterized in that, include: The data acquisition module is configured to identify the working robot, acquire the current stage task information and the next stage task information of the working robot, and determine the initial tool release speed and the initial tool grasping speed of the working robot based on the current stage task information and the next stage task information. The judgment module is configured to collect the tool grasping status information of the robot when the robot completes the current stage task and moves to the target tool library, and determine whether to adjust the initial tool release speed based on the tool grasping status information. It is also configured to determine the matching degree between the position of the end effector of the work robot and the position of the target tool, and determine whether to adjust the initial tool grasping speed based on the matching degree; The multimodal adjustment module is configured to, when it is determined that the initial tool release speed should be adjusted, collect the historical release records of the task tools used in the current stage of the task, determine the release adjustment coefficient of the initial tool release speed based on the historical release records, and obtain the final tool release speed. It is also configured to collect contact state information between the end effector of the work robot and the target tool and the grasping force information of the end effector, determine the grasping adjustment coefficient of the initial tool grasping speed based on the contact state information and the grasping force information, and obtain the final tool grasping speed; The execution module is configured to control the work robot to perform tool release and grasping operations based on the final tool release speed and the final tool grasping speed.

2. The robot tool switching system based on multimodal recognition according to claim 1, characterized in that, When determining the initial tool release speed and initial tool grasping speed of the robot based on the current stage task information and the next stage task information, the following are included: The task information at the current stage is parsed to obtain the task tool type, task tool weight, and task tool usage time; The next stage task information is analyzed to obtain the target tool type, target tool weight, and the required installation accuracy of the target tool; The initial tool release speed of the work robot is determined based on the task tool type, task tool weight, and task tool usage duration; The initial tool-grabbing speed of the robot is determined based on the target tool type, target tool weight, and required installation accuracy of the target tool.

3. The robot tool switching system based on multimodal recognition according to claim 2, characterized in that, When determining the initial tool release speed of the work robot based on the task tool type, task tool weight, and task tool usage duration, the following are included: Construct a release vector group based on the task tool type, task tool weight, and task tool usage duration; The release vector group is compared with the historical release group. If there is a historical release vector group in the historical release group that is the same as the release vector group, then the historical tool release speed corresponding to the historical release vector group is taken as the initial tool release speed. If there is no historical release vector group in the historical release group that is the same as the release vector group, then all historical release vector groups that are the same as the release vector group in terms of task tool type and task tool weight are selected and constructed as the first filtering group. Obtain the average and maximum historical tool usage durations in the first filter group, and record them as the historical average tool usage duration and the historical maximum tool usage duration, respectively. The initial tool release speed is determined by comparing the usage time of the task tools in the release vector group with the historical average tool usage time and the historical maximum tool usage time. When the usage time of the task tool is less than or equal to the historical average usage time of the tool, the initial tool release speed is determined as the first release speed; When the duration of the task tool usage is greater than the historical average tool usage duration, and less than or equal to the historical maximum tool usage duration, the initial tool release speed is determined as the second release speed. When the duration of the task tool usage exceeds the historical maximum tool usage duration, the initial tool release speed is determined to be the third release speed.

4. The robot tool switching system based on multimodal recognition according to claim 2, characterized in that, When determining the initial tool-grabbing speed of the robot based on the target tool type, target tool weight, and required installation accuracy of the target tool, the following steps are included: The basic tool grasping speed of the robot is determined based on the target tool type. The optimization coefficient of the basic tool's grasping speed is determined based on the target tool's weight and the required installation accuracy of the target tool, and the product of the optimization coefficient and the basic tool's grasping speed is used as the initial tool's grasping speed. When determining the optimization coefficient for the gripping speed of the basic tool based on the weight of the target tool and the required installation accuracy of the target tool, the following steps are included: The weight of the target tool is compared with the weight of the standard tool, and the required installation accuracy of the target tool is compared with the standard installation accuracy. The optimization coefficient is determined based on the comparison results. When the weight of the target tool is greater than or equal to the weight of the standard tool, and the required installation accuracy of the target tool is greater than or equal to the standard installation accuracy, the optimization coefficient is determined as the first optimization coefficient. When the weight of the target tool is greater than or equal to the weight of the standard tool, and the required installation accuracy of the target tool is less than the standard installation accuracy, the optimization coefficient is determined to be the second optimization coefficient. When the weight of the target tool is less than the weight of the standard tool, and the required installation accuracy of the target tool is greater than or equal to the standard installation accuracy, the optimization coefficient is determined to be the third optimization coefficient. When the weight of the target tool is less than the weight of the standard tool, and the required installation accuracy of the target tool is less than the standard installation accuracy, the optimization coefficient is determined to be the fourth optimization coefficient.

5. The robot tool switching system based on multimodal recognition according to claim 3, characterized in that, When determining whether to adjust the initial tool release speed based on the tool's grasping status information, the following is included: Feature extraction is performed on the tool's grasping state information to obtain tool grasping feature values; The tool's grasping feature value is compared with the tool's grasping threshold, and the initial tool release speed is adjusted based on the comparison result. When the tool's grasping feature value is within the tool's grasping threshold, it is determined that the initial tool release speed will not be adjusted. Otherwise, it is determined that the release speed of the initial tool should be adjusted.

6. The robot tool switching system based on multimodal recognition according to claim 4, characterized in that, When determining the matching degree between the end effector position of the robot and the target tool position, and deciding whether to adjust the initial tool grasping speed based on the matching degree, the process includes: A three-dimensional Cartesian coordinate system is established with the geometric center of the target tool as the origin, and the position coordinates of the end effector of the robot and the position coordinates of the target tool are obtained respectively. Calculate the distance deviation and angle deviation between the end effector and the target tool based on the position coordinates of the end effector and the target tool. The degree of matching between the end effector position and the target tool position is calculated based on the distance deviation value and the angle deviation value; The matching degree is compared with the matching degree threshold. If the matching degree is greater than or equal to the matching degree threshold, it is determined that the initial tool's crawling speed will not be adjusted. Otherwise, it is determined that the initial tool's grasping speed should be adjusted.

7. The robot tool switching system based on multimodal recognition according to claim 5, characterized in that, When determining the release adjustment coefficient for the initial tool release rate based on the historical release records, and obtaining the final tool release rate, the process includes: The historical release records are parsed to obtain the maximum impact force value of the end effector corresponding to the release process of each task tool, and a maximum impact force value data set is constructed. The historical release deviation values ​​of the mission tool are obtained based on the maximum impact force value data set. The historical release deviation value is compared with the first historical release deviation value and the second historical release deviation value, and the release adjustment coefficient is determined based on the comparison result; wherein, the first historical release deviation value is less than the second historical release deviation value; When the historical release deviation value is less than or equal to the first historical release deviation value, the release adjustment coefficient is determined to be the first release adjustment coefficient. When the historical release deviation value is greater than the first historical release deviation value and less than or equal to the second historical release deviation value, the release adjustment coefficient is determined to be the second release adjustment coefficient. When the historical release deviation value is greater than the second historical release deviation value, the release adjustment coefficient is determined to be the third release adjustment coefficient; The final tool release speed is obtained by multiplying the release adjustment coefficient by the initial tool release speed.

8. The robot tool switching system based on multimodal recognition according to claim 6, characterized in that, When determining the initial tool gripping speed adjustment coefficient based on the contact state information and gripping force information, and obtaining the final tool gripping speed, the process includes: The contact state information is analyzed to obtain the contact angle and contact area between the end effector and the target tool; The grasping force information is parsed to obtain the actual grasping force of the end effector; The gripping influence index is determined based on the contact angle, contact area, and actual gripping force. The crawling impact index is compared with historical data, and the crawling adjustment coefficient is determined based on the comparison results; When there is a historical crawling impact index in the historical data that is the same as the crawling impact index, the historical crawling adjustment coefficient corresponding to the historical crawling impact index shall be used as the crawling adjustment coefficient. When there is a historical crawling influence index in the historical data that is the same as the crawling influence index, calculate the difference between the crawling influence index and each of the historical crawling influence indices one by one, and record it as the influence difference. Obtain the absolute value of the impact difference, and denot it as the absolute impact difference; The capture adjustment coefficient is determined based on the absolute impact difference. The final tool grasping speed is the product of the grasping adjustment coefficient and the initial tool grasping speed.

9. The robot tool switching system based on multimodal recognition according to claim 8, characterized in that, When determining the capture adjustment coefficient based on the absolute impact difference, the following are included: When the historical crawling adjustment coefficient corresponding to the absolute impact difference is unique, the historical crawling adjustment coefficient shall be used as the crawling adjustment coefficient. When the historical crawling adjustment coefficient corresponding to the absolute impact difference is not unique, the average value of all the historical crawling adjustment coefficients is calculated, and the average value is used as the crawling adjustment coefficient.

10. A robot tool switching method based on multimodal recognition, applied to the robot tool switching system based on multimodal recognition as described in any one of claims 1-9, characterized in that, include: Identify the work robot, collect the current stage task information and the next stage task information of the work robot, and determine the initial tool release speed and the initial tool grasping speed of the work robot based on the current stage task information and the next stage task information. When the robot completes the current stage task and moves to the target tool library, the tool grasping status information of the robot is collected, and it is determined whether to adjust the initial tool release speed based on the tool grasping status information. Determine the matching degree between the position of the end effector of the robot and the position of the target tool, and determine whether to adjust the initial tool grasping speed based on the matching degree; When it is determined that the initial tool release speed needs to be adjusted, the historical release records of the task tools used in the current stage of the task are collected, the release adjustment coefficient of the initial tool release speed is determined based on the historical release records, and the final tool release speed is obtained. The contact state information between the end effector of the work robot and the target tool and the grasping force information of the end effector are collected. Based on the contact state information and the grasping force information, the grasping adjustment coefficient of the initial tool grasping speed is determined, and the final tool grasping speed is obtained. The final tool release speed and the final tool grasping speed are used to control the working robot to perform tool release and grasping operations.

Images