JP4984576B2 - Grasping control method by robot hand - Google Patents

Description

  The present invention relates to a gripping control method when gripping an object with a robot hand, and more particularly to a gripping control method when an object exists around the gripping object and becomes an obstacle to an object gripping operation.

Development of a robot that acquires information on the outside world using a visual sensor and performs complicated work is underway. For example, the technique disclosed in Patent Document 1 is an environment recognition system that actively recognizes an external environment in an intelligent robot or the like. The basic idea of this technology is that if you cannot recognize an object from a certain point of view, you can move the viewpoint to a more visible position, or an object that is concealed and difficult to recognize It is intended to create an environment where the visual system is easy to function by actively changing itself and the external environment, such as moving to a place that does not hide the obstacles, or changing the lighting conditions appropriately. It is described.
JP-A-8-30327

  In this technique, for example, when another robot is visually recognized, examples such as movement of itself, prompting movement of another robot, and moving an obstacle in between are disclosed. By the way, in a robot that operates in a human living environment, there may be an application in which a certain object is searched for and gripped to perform a predetermined motion. There is no disclosure on how to set the gripping operation including the selection of the obstacle to be removed when there is another object in the object and it is an obstacle to gripping the object .

  Therefore, an object of the present invention is to provide a grip control method using a robot hand that can appropriately control a gripping operation of a target object when another object exists near the target object.

In order to solve the above problems, a grip control method using a robot hand according to the present invention is a grip control method using a robot hand that grasps a target object and its surroundings using a visual sensor and grips the target object. (1) a step of setting a first type path, which is a movement path of the hand from the initial position to the gripping position of the object by the path search algorithm, in consideration of the object, the operated portion, and the operation posture ; (2) A step of resetting the route in consideration of the object, the operated part, and the operation posture when the first type of route setting is impossible due to an obstacle, and (a) the first type of route is set. and setting the second type of route by connecting the presence or constraints of the robot hand of routing when the obstacle from the final point of the candidate outside the route to the gripping position, (b) the second type Robo along Path when moving the Tohando, a step of extracting an obstacle any part of the robot hand arm with possibility of contact or collision, it is removed by grasping by (c) extracted obstacle robot hand And a step of resetting the movement path of the hand up to the gripping position including.

  In the route search algorithm in the step (1), as one embodiment, when a search operation for extending the route in a random direction from the initial position is executed and the robot hand hits a constraint condition or an obstacle, The search from that position is stopped and the route that finally reaches the gripping position is selected as the route. The route when the route is found in this way is referred to as a first type route. In the present invention, when the first type route is not found, the route is set by connecting the position where the continuation of the search is stopped and the last point to the gripping position ((a) in (2) above) Process). In this route (second type route) calculation, a route search algorithm may be used. In this search, the constraint condition of the robot hand is considered, but the obstacle condition is ignored. Then, in the following step (b), obstacles on the set route are extracted, and in step (c), the route including the route for removing each obstacle is obtained, and the route to the gripping position is obtained. Reset the settings.

  Note that the above route search algorithm does not need to be a random search algorithm as described above. For example, a potential field from the initial point to the final point is set and the route is obtained by the steepest descent method, Any route search algorithm or the like may be used.

  A plurality of the second type routes may be set, the route priority may be set according to the obstacle present on the route, and the route may be reset. By setting a plurality of routes when setting the second type route, the number of route options increases.

Alternatively, in the grip control method by the robot hand according to the present invention, instead of the above (2), (2 ′) there is an obstacle, and therefore, when the first type of route setting is impossible , A step of resetting the path in consideration of the operation unit and the operation posture , (a) detecting a proximity obstacle that is close to the grasped object and obstructs the grasping operation; and (b) the detected proximity obstacle A step of setting the second type route to the gripping object when there is no object, and (c) when the second type of route setting is impossible, from the initial position on the condition that the movement of the robot hand is limited A step of setting the third type path to the gripping position; and (d) when moving the robot hand along the third type path , any part of the robot hand arm may contact or collide. A step of extracting sexual obstacles, and (e) these obstacles. And a step of resetting the moving path of the hand to the gripping position including the path for gripping and removing the harmful object by the robot hand .

  In the present invention, when the first type path in the step (1) does not exist, a proximity obstacle that may interfere with the gripping operation itself is first detected around the gripping target ((a ) Process). If the path to the gripping object (second type path) can be set if this obstacle does not exist, the second type path is set as a path candidate (step (b)). As a method for setting the second type of route, for example, a random search algorithm may be used. If this route cannot be set, a third type route from the initial position to the gripping position is further set under conditions that limit the movement of the robot hand (step (c)). By limiting the movement of the robot hand, the route range to be searched is limited. Obstacles present in the setting of the third type of route are ignored. In step (d), an obstacle on the set route is extracted, and in step (e), a route including the route for removing each obstacle / obstacle is obtained to obtain a route to the gripping position. Reset the settings.

  It is preferable to set a plurality of the third type routes, set the priority order to the routes according to the obstacles present on the route, and execute the route resetting. By setting a plurality of routes when setting the third type route, there are more route options.

  In the step (a), when the obstacle is detected, if it is impossible to reset the route by the presumed gripping operation, the preconditioning gripping operation may be changed to perform the resetting step. . Specifically, the path setting options are increased by changing the gripping posture and the gripping position and redoing the resetting process.

  The route priority is given priority to routes that have obstacles close to the robot hand, and if the distance is approximately the same, priority is given to routes that have obstacles on the outside as viewed from the robot hand. May prioritize routes with tall obstacles. This is because the closer to the robot hand, the closer to the outside, the taller the obstacle, the more likely to be an obstacle when gripping an object.

  According to the above step (2), since there is an object around the gripping object, if the gripping object cannot be directly touched as it is, the obstacle object is removed and the gripping object is gripped. It is possible to reliably set the route until. At this time, for example, by using the result of the random search algorithm, the amount of calculation required for the search can be reduced, and the route can be searched in a shorter time.

  Even in the step (2 ′), since an object exists around the gripping object, if the gripping object cannot be directly touched as it is, the obstacle object is removed and the gripping object is removed. It is possible to reliably set a route until gripping. In this case, it is possible to reduce the amount of calculation required for the search in a shorter time by preferentially removing an object that obstructs the gripping operation and performing a route search under a condition that restricts the movement of the robot hand. The route can be searched with.

  By setting a plurality of second type routes of the step (2) and third type routes of the step (3), route options increase. As to which of the set routes is to be selected, for example, a route with few obstacles and easy to remove may be selected. By determining based on the distance from the robot hand, the position to the outside, and the height as this order, the determination criteria are simplified, the processing is easy, and the selection can be performed at high speed.

  DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, preferred embodiments of the invention will be described in detail with reference to the accompanying drawings. In order to facilitate the understanding of the description, the same reference numerals are given to the same components in the drawings as much as possible, and duplicate descriptions are omitted.

  FIG. 1 is a block diagram of a robot that executes a grip control method according to the present invention. In this robot, a multi-finger multi-joint hand is arranged at the tip of left and right multi-joint arms, and a predetermined operation such as gripping an object is performed by driving an actuator 31 arranged in each joint. be able to. A stereo camera 20 is disposed on the head of the robot, acquires a stereo image of the object and the surrounding environment, and transfers it to the control unit 1.

  The control unit 1 includes a CPU, a ROM, a RAM, and the like, and causes the robot hand / arm to execute a predetermined operation in accordance with a built-in program. Although not shown, a predetermined operation is performed in response to an instruction from an input means (keyboard, voice input, wired / wireless LAN, etc.), or an object searched from an image acquired by the stereo camera 20 is displayed. The corresponding operation may be performed. The predetermined operation includes an operation of holding an object and moving it to a predetermined position, an operation of executing a series of programmed operations while holding the object, and the like.

  The control unit 1 includes an image processing unit 10 that processes an image acquired by the stereo camera 20, a peripheral environment recognition unit 11 that acquires the shape and position information of an object in the robot's peripheral environment from the processing result, and each joint of the robot. The position / orientation recognition unit 12 that recognizes the position and orientation of the robot from the output of the arranged angle sensor 30, the path search unit 13 that searches for the movement path of the robot hand for performing a desired operation, and the actuator 31 are driven to search. A drive control unit 14 for supporting movement on the route. Each of these units 10 to 14 may be configured by separate hardware, but some or all may share part or all of the hardware, and even in this case, the software is completely independent. It is not necessary to be configured, and a part of the configuration may be shared.

  Before describing the gripping control of an object by the robot, the concept of an object operated portion will be briefly described. FIG. 2 is a diagram illustrating a relationship among an object, an operated unit, and an operation posture in gripping path setting. An object to be gripped (target object) includes one or more operated parts (EfM: Effective for Manipulation part) that are grippable places. One or more operation postures exist for each operated part. Taking the case of gripping as an example, this operation posture corresponds to the position / posture of the hand to be gripped.

  If the object is a vertical cylinder, the operated portion that is the position to be gripped may include at least a side wall portion and an upper portion. For example, even when the upper part is gripped, various postures are conceivable as gripping positions. If the posture is different, the positional relationship between the object around the object and the robot is also affected, and the object = obstacle that becomes an obstacle to the gripping operation can be changed. In the grip control method according to the present invention, the route is set in consideration of the object, the operated portion, and the operation posture.

  FIG. 3 is a flowchart showing the processing of the first embodiment of the grip control method according to the present invention. This process is executed in the control unit 1 when a gripping command is issued from the outside or from an internal program.

First, a route search process to an object is performed by a random search algorithm (step S1). This random search algorithm is a well-known technique for various route setting techniques. For example, as shown in FIG. 3, when searching for a route from the start position q ₁ to the destination target position q _{G, TO} , the route is set as a combination of short nodes, and the direction of the node is randomly determined from the start position. This is a method of searching for a correct route by setting nodes at random in the same way for branches from each node. When a node touches or collides with another object, or when the end position of the node exceeds a boundary position set based on the constraint condition of the robot hand, the setting of the node or branching from the node is stopped. In addition, branching with overlapping nodes is also restricted.

  Specifically, based on the image acquired by the stereo camera 20, the image processing unit 10 and the surrounding environment recognition unit 11 acquire the shape and position information of the object and an object that can be an obstacle around the object. On the other hand, the position and orientation recognition unit 12 grasps the state of each joint from the output of the angle sensor 30 and grasps the position of the hand based on this. The route search unit 13 sets the final target position of the hand by setting the target unit to be operated and its operation posture based on each position information and posture information, and sets the final target position of the hand from the current position. Perform a random search for the route to

  In step S2, it is evaluated whether or not the route to the target position (first type route) has been successfully set by this random search algorithm. If the setting is successful, the gripping operation is possible without touching other obstacles, etc., so that the process proceeds to step S3, the set search route is output, and the process ends. Thereafter, the drive control unit 14 drives the actuator 31 to move the hand along the set path. At this time, the output of the angle sensor 30 may be read by the position / orientation recognition unit 12 and feedback control may be performed to determine whether movement along the route is performed. This control is not limited to feedback control, and various control methods may be appropriately combined.

If the first type of route setting fails, that is, if there is an obstacle up to the target position, the process proceeds to obstacle removal determination. Specifically, first, in step S4, route candidates obtained by the random search algorithm in step S2 are extracted, and the end positions are processed. The route candidate here is a route in which node setting / branching is stopped due to a constraint condition or contact with or colliding with an obstacle in the node connection described above. As the first route candidate, a route in which node setting / branching is stopped at a position closest to the target position is selected. The termination process is a process in which the position physically closest to the target position among the nodes constituting the extracted route candidate is the termination position. For example, in the path shown in Figure _4, q B, even node extend to _{q C, q B, q C} q from the target position _{G, TO} close _{q A} is set to the end position.

  Next, a connection path from the terminal position of the path candidate to the target position is derived (step S5). Here, an obstacle on the route is ignored, and a route connecting the end position and the target position is obtained based on the dynamic characteristics of the robot hand. Specifically, a route (second type route) that can be reached by a combination of simple joint operations is set from the end position.

When the route is set, an object = obstacle existing on the route is detected (step S6). In the example of FIG. _{4, O 10} corresponds to this. Here, the obstacle to be detected is not simply an object existing on the locus of a specific point on the robot hand, but when moving the hand along the set route, either the robot hand or the arm is detected. All objects that are likely to touch or collide are detected.

  Next, the obstacles are ordered (step S7). Specifically, when there are a plurality of detected obstacles, the obstacle removal priority is set based on a preset criterion. For example, as a reference, first, the obstacle closest to the robot is prioritized. When both are at approximately the same distance, the one near the boundary where the robot's hand can reach, when grasping with the left hand, the leftmost object within the reach of the robot hand, when grasping with the right hand, the robot Priority is given to the rightmost object in the reach of the hand. These are used when the obstacles are apart from each other by a certain distance, and when the distance between the obstacles is short, those having the operated portion at a high position are removed preferentially.

  After the obstacle removal priority order is determined, a route for removing the obstacle itself is set. This is performed by setting a route by the same method as the processing of step S1 with an obstacle as an object. At this time, in order to remove the obstacle, if another obstacle needs to be removed, a path for removing the other obstacle is searched before removing the obstacle.

For example, in the path shown in FIG. 4, since the arrival position q _G10 for removing obstacles O ₁₀ is blocked by O _20, it is necessary to remove in advance the O _20. Since the reaching position q _G20 for removing O ₂₀ is blocked by the obstacles O ₃₀ and O ₃₁ , it is necessary to remove these O ₃₀ and O ₃₁ . FIG. 5 shows the relationship between these obstacles as a tree diagram. In this tree diagram, the highest position represents the final target position (q _{G, TO in} this figure), and each of the arrival positions or obstacles to the object are branched downward. O ₃₀ and O ₃₁ are written in parallel because both are independent obstacles to O ₂₀ . Then, O ₃₀ close to the robot is removed with priority.

  Next, in step S9, it is checked whether the route search is successful. If the route cannot be set, the process proceeds to step S10, another route candidate is selected, and the route setting process from step S4 is performed again. Note that the alternative route candidates also include routes in which the operated portion of the gripping object and the gripping posture are different. If the route search is successful, the searched route and the obstacle are output (step S11), and the process ends. Thereafter, as described above, the drive control unit 14 performs the control of moving the hand along the path set by driving the actuator 31.

  In a rare case when there are a plurality of obstacles on the route, the object B becomes an obstacle to remove the object A, and the object A becomes an obstacle to remove the object B. A deadlock relationship may be established. In this control, even in an environment where such a situation may occur, selecting a route from a plurality of route candidates set by random search increases the possibility of selecting a route that does not interfere with these objects. The angle for setting a route to an object is increased, and route generation is facilitated.

  In addition, when a route that reaches the target arrival position without touching or colliding with an obstacle is preferentially set by the random search algorithm and the route cannot be set, the route candidate used in this random search is set. Since a route for which an obstacle needs to be removed is set by using the route setting process at the time of detecting an obstacle, the processing time can be increased.

  Note that the processing in steps S4 to S8 may be executed simultaneously for a plurality of route candidates, and a route with few removed obstacles or a route with a simple route may be preferentially set. In this case, there is an advantage that a more appropriate route can be set.

  The grip control method according to the present invention is not limited to the above embodiment. FIG. 6 is a flowchart showing the processing of the second embodiment of the grip control method according to the present invention. Similar to the first embodiment, this process is also executed in the control unit 1 when a gripping command is issued from the outside or from an internal program.

  First, a route search process to the object is performed by a random search algorithm (step S21), the success or failure of the search is determined (step S22), and if successful, the search route (first type route) is output. (Step S23) The process is terminated in the same way as the processes in steps S1 to S3 in the first embodiment.

  If it is determined in step S22 that the route cannot be set, the process proceeds to step S24, and an obstacle close to the object is determined. Here, the obstacle close to the object refers to an obstacle near the object that may affect the gripping posture. FIG. 7 shows the relationship between such an obstacle, the object, the operated part, and the operation posture. The “obstacle close to the object” (hereinafter referred to as a proximity obstacle) detected here is determined by the operation posture. One or more operation postures can be determined by the operated portion EfM. Furthermore, one or more operated parts EfM can be set depending on the object. Therefore, there can be one or a plurality of operated parts EfM for one object, and these are in a selective relationship. Each operated part EfM can have one or a plurality of gripping postures, which are also in a selective relationship. Furthermore, there can be one or more proximity obstacles for each gripping posture. A plurality of proximity obstacles for a specific gripping posture is not a selective relationship, and all of them are equivalent obstacles.

This proximity obstacle determination process will be described more specifically. FIG. 8 shows an example of the surrounding environment. Here, a case where the teapot 50 on the table 70 is gripped by the robot 40 is taken as an example. An obstacle around the teapot 50 is denoted by reference numeral 60. 9, as shown in FIG. 10, the handle 51 is a grip of the teapot 50 is approximated by a cylinder and its length w _h, the diameter and d _c. The operation area 55 for actually performing gripping is a space having a diameter (d _c + 2 × d _h ) surrounding the handle 51 and a length (w _h + w _m ).

  FIG. 11 is a flowchart showing this proximity obstacle determination process. First, each object shape is projected on a horizontal plane (step S41). For example, the projection is performed on the table 70 in the environment as shown in FIG. This projection is performed on an xy plane with the front-rear direction as the x-axis and the left-right direction as the y-axis, as viewed from the robot. Next, conversion is performed to a predetermined coordinate system having a predetermined point near the gripping position as an origin (step S42). In the example shown in FIG. 10, the projection point at the base of the handle on the central axis of the handle 51 is set as the origin O ′, the X ′ axis is the central axis of the handle 51, and the Y ′ axis is set in a direction perpendicular thereto. Convert system to X'Y 'coordinate system.

After the coordinate conversion, an object that overlaps the operation area 55 and the projection plane is extracted as an obstacle candidate (step S43). By converting the coordinate as described above, the operation area 55, X 'is 0 or higher _(w h + w _m) or less, Y' absolute value of _(r c + _{d h),} and the this operating region 55 It becomes easy to determine the overlap. In the example shown in FIG. 10, horizontal cylinders 60 _3,1 , cubes 60 _1,2 , vertical cylinders 60 _2,1 and spheres 60 _4,1 are recognized as obstacle candidates with overlapping projection planes.

Next, it is determined from the extracted obstacle candidate and the three-dimensional shape and positional relationship of the target object whether or not there is actually a proximity relationship, and a candidate satisfying the proximity relationship is extracted as an obstacle (step S44). For example, as shown in FIG. 12, the presence or absence of contact may be determined from the height positional relationship between the obstacle candidates 60 _i and 60 _j and the operation area 55. Under the conditions shown in the figure, the obstacle candidate 60 _i is determined not to be an obstacle, and the obstacle candidate 60 _j is determined to be an obstacle. Finally, the extracted obstacles are ordered as the priority of removal (step S45).

  When the obstacle determination process ends, the process proceeds to step S25 in the flowchart shown in FIG. 6 to determine whether there is a nearby obstacle that satisfies the condition. If it is determined that there is a nearby obstacle, the process proceeds to step S26 and a route to the target object (second type route) is searched on the assumption that there is no extracted nearby obstacle. This route search may be performed based on the same random search as in step S21, or may be performed by extending the route candidates whose node connection is stopped by the obstacle.

  After the route search, the success or failure is determined (step S27). When the route setting is successful, the process proceeds to step S30 described later. On the other hand, if it is determined that the route setting is impossible, and if it is determined in step S25 described above that there is no proximity obstacle, the process proceeds to step S28 to simplify the movement of the robot joint. Thus, a route from the initial position to the target arrival position (simple route = third type route) is set, and an obstacle on the simple route is detected.

  Here, two examples of this simple path in an environment as shown in FIG. 8 will be described. A first example is shown in FIG. First, extend the robot's right arm straight to the shoulder to the right. From that position, the hand is lowered to the lower left so as to draw an arc toward the object 50. At this time, the shoulder joint 40s and the elbow joint 40e are bent simultaneously. Each joint makes an equiangular velocity motion. At this time, the trajectory drawn by the wrist 40w becomes an Archimedean spiral, and the trajectory drawn by the elbow joint 40e corresponds to the arc of the oblique cross section of the cone formed by the Archimedean spiral surface. These can be obtained geometrically from the shape of the robot. At this time, an object whose part is located in an area through which the robot hand arm passes is recognized as an obstacle.

  A second example is shown in FIGS. In this case, first, the elbow joint 40e is driven and the forearm is bent toward the upper arm. Next, extend the upper arm to the right and raise it to the height of the shoulder. At this time, the forearm is in a posture folded on the upper arm. From here, the shoulder joint 40 s is rotated to rotate the forearm portion to the left, and is moved to a position where the plane formed by the central axes of the forearm portion and the upper arm portion is located on the object 50. Next, the tip of the hand is moved straight to the object. At this time, the elbow joint 40e is moved to the right side. Even in this case, the movement trajectory of each joint or the like can be obtained geometrically. For this reason, the region through which the robot hand arm passes and the obstacle located there can also be obtained by geometric calculation.

  If an obstacle is detected, a route to the object is searched assuming that all the detected obstacles do not exist (step S29). In this case, since the simple route described above can be set at least, the simple route described above may be used as the route, but other routes may be searched. Although such a path is complicated as the operation of the drive system, the path of the path itself can be shorter than a simple path.

  In step S30, when there are a plurality of obstacles / proximity obstacles, the removal is ranked. This order setting may be performed according to the same criteria as in the first embodiment described above. When the ranking is completed, a route for obstacle removal is searched (step S31). In this route search, the route may be set by a random search algorithm similar to step S21 with individual obstacles as objects.

  When the route setting is completed, the process proceeds to step S32 to determine whether or not the search is successful. If the route cannot be set, the process proceeds to step S33, another route or another nearby obstacle is selected, and the route setting process from step S24 is performed again. Note that the alternative route candidates also include routes in which the operated portion of the gripping object and the gripping posture are different. If the route search is successful, the searched route and the obstacle are output (step S34), and the process ends. Thereafter, as described above, the drive control unit 14 performs the control of moving the hand along the path set by driving the actuator 31.

  In the control of the present embodiment, a route for preferentially removing the nearby obstacle is set, so that the grip posture can be surely reached. In addition, when detecting an obstacle, the determination is performed using a simple path that restricts the movement of the hand and arm. Therefore, the obstacle determination process is easy and simple, and the processing speed can be increased. . In addition, the possibility of setting other routes with respect to the deadlock relationship is increased as in the first embodiment.

  The simple route described above is not limited to this, and various route settings can be performed. Also in the present embodiment, the processes in steps S24 to S32 may be performed concurrently for a plurality of gripping units and gripping postures.

It is a block block diagram of the robot which performs the holding | grip control method concerning this invention. It is a figure which shows the relationship between the target object, to-be-operated part, and operation attitude | position in grip path | route setting. It is a flowchart which shows the process of 1st Embodiment of the holding | grip control method concerning this invention. It is a figure which shows an example of the object and path | route candidate in the process of FIG. It is a figure which shows the structure of the path | route setting result in the state of FIG. It is a flowchart which shows the process of 2nd Embodiment of the holding | grip control method concerning this invention. It is a figure which shows the relationship between the obstruction which adjoins the target object used as a detection target in the process of FIG. It is a figure which shows an example of the surrounding environment explaining the determination process of a proximity | contact obstacle. It is a figure which shows the teapot which is a holding | grip target object in the environment of FIG. FIG. 10 shows an obstacle close to the teapot in FIG. 9. It is a flowchart which shows the determination process of the proximity | contact obstacle in the process of FIG. It is a figure which shows the positional relationship with the obstacle which becomes object by the process of FIG. 11 for the teapot shown in FIG. It is a figure which shows the 1st example of the simple path | route in the process of FIG. It is a figure which shows the 2nd example of the simple path | route in the process of FIG. It is the figure which looked at the simple path | route of FIG. 14 from the top.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Control part, 10 ... Image processing part, 11 ... Ambient environment recognition part, 12 ... Position and orientation recognition part, 13 ... Path search part, 14 ... Drive control part, 20 ... Stereo camera, 30 ... Angle sensor, 31 ... Actuator , 40 ... robot, 40 s ... shoulder joint, 40 w ... wrist, 40e ... elbow joint, 50 ... teapot (object), 51 ... grip, 55 ... operation region, ₆₀ i, 60 j _... obstacle candidate, 70 ... table , EfM: operated part.

Claims (6)

In a grip control method by a robot hand that grasps a target object and its surroundings using a visual sensor and grips the target object,
A step of setting a first type path that is a movement path of a hand from an initial position to a gripping position of an object by a path search algorithm in consideration of the object, an operated part, and an operation posture;
A step of resetting the path in consideration of the object, the operated part, and the operation posture when the first type of path setting is impossible because an obstacle exists.
A step of setting a second type of route by connecting from the final point of the route that is not a candidate due to the presence of an obstacle or the constraint condition of the robot hand during the first type of route setting to a gripping position;
Extracting an obstacle with which any part of the robot hand arm may contact or collide when moving the robot hand along the second type path;
Resetting the movement path of the hand to the gripping position including the path for gripping and removing the extracted obstacle by the robot hand; and
A gripping control method using a robot hand, comprising: The robot hand gripping according to claim 1, wherein a plurality of the second type routes are set, a route priority is set according to an obstacle existing on the route, and the route is reset. Control method. In a grip control method by a robot hand that grasps a target object and its surroundings using a visual sensor and grips the target object,
A step of setting a first type path that is a movement path of a hand from an initial position to a gripping position of an object by a path search algorithm in consideration of the object, an operated part, and an operation posture;
A step of resetting the path in consideration of the object, the operated part, and the operation posture when the first type of path setting is impossible because an obstacle exists.
Detecting a proximity obstacle that is close to the object to be grasped and obstructs the grasping operation;
A step of setting a second type route to the grasped object when there is no detected proximity obstacle;
A step of setting a third type path from an initial position to a gripping position under a condition in which the movement of the robot hand is limited when the second type path setting is impossible;
Extracting an obstacle that may contact or collide with any part of the robot hand arm when moving the robot hand along the third type path;
A step of resetting a movement path of the hand to a gripping position including a path for gripping and removing an obstacle by the robot hand, and a gripping control by the robot hand Method. 4. The robot hand according to claim 3, wherein a plurality of the third type routes are set, a priority is set for the route according to an obstacle existing on the route, and the route is reset. Gripping control method. 5. When the obstacle is detected, if resetting of the path is impossible with the presumed gripping operation, the resetting process is performed by changing the presumed gripping operation. The grip control method by the robot hand described in 1. The route priority is given priority to routes that have obstacles close to the robot hand, and if the distance is approximately the same, priority is given to routes that have obstacles on the outside as viewed from the robot hand. 5. The grip control method using a robot hand according to claim 2, wherein priority is given to a route in which a tall obstacle exists.

Images