JP7125745B2 - ENVIRONMENTAL ADAPTABILITY REINFORCEMENT SYSTEM OF AUTONOMOUS WORK SUPPORT ROBOT, OPERATION SIMULATION DEVICE, AND THEREOF PROGRAM - Google Patents

Description

In the present invention, when a robot capable of autonomous operation is used to respond to a disaster, etc., various tasks performed by the robot in the field can be improved by confirming the operation of the robot in advance in a virtual space that matches the actual environment. The present invention relates to an environment adaptability enhancement system for an autonomous work support robot, a motion simulation device, and a program therefor, which enable the robot to perform work quickly and reliably.

Introduction of robots for recovery work after a disaster is in progress, and such disaster response robots are required to investigate on-site conditions and carry out urgent recovery work. Here, when performing disaster response work in a work environment where secondary disasters may occur, or in a harsh environment where humans cannot enter, such as radiation, fire, and toxic gas, remote control from a safe remote location is required. autonomous robots will be introduced.

In recent years, next-generation disaster response robots have been researched for the purpose of responding to disasters with a higher degree of difficulty. A high level of work performance is required to remove debris, etc. In order to meet these demands, disaster response robots must have not only environmental applicability but also traversability and workability. Here, treadability is the ability to move safely and efficiently over uneven terrain with steps and slopes, and workability is the ability to handle even complex objects, and furthermore, the ability to cut into the object. It is the ability to perform various actions such as tearing and peeling.

Therefore, the inventors of the present invention have already proposed a multi-arm mobile robot having a plurality of working arms as a next-generation disaster response robot (see Patent Document 1). This multi-arm mobile robot has an arm unit including four arms that can move within a predetermined range, two main crawlers that operate while contacting the ground so as to move a vehicle body to which the arm unit is attached, and four main crawlers. It has a crawler unit consisting of two flippers, and a control device for controlling the operation of each arm unit and the crawler unit according to the operation of the operator. In this multi-arm mobile robot, the operator can select two types of functions: a manual operation function in which the arm unit and crawler unit are all manually operated by remote control, and an automation function in which the arm unit and crawler unit are all automatically operated. It's becoming The automation function is provided with the intention that the robot automatically executes the task in response to a task command from an operator who is away from the site.

JP 2017-52013 A

However, when the robot moves automatically, it may not be possible to perform a desired task when the actual environmental conditions around the robot differ from those assumed in advance. In other words, in the automation function, the robot operates to execute the task instructed by the operator under any circumstances, and problems such as failure of the task and the robot becoming immobile occur. In other words, there are cases where unpredicted robot control conditions are required due to the actual situation at the site, and there are situations where the robot does not actually come into contact with the site, such as the friction and stability of the ground or obstacles that adversely affect the robot's operation. The above problem occurs when there is a hindrance factor in the actual environment that is unknown. For example, when climbing a step with two steps, if only one-step step-climbing operation control is prepared, the actual step cannot be climbed. In addition, if the ground on which the robot runs has unevenness, it is expected that the left-right balance will be lost. Furthermore, if the ground is a slippery step, the slippage can affect task performance when acting like a normal step. In addition, when the actual step height exceeds the robot's performance, it is often difficult for the operator to judge the robot's motion performance even if the operator monitors the robot from a remote location with a camera or the like. If the robot tries to forcefully climb a step beyond the operating performance, there is a risk that it will fall over.

When performing disaster response work using an autonomous robot at an actual site, it is required that the work should not fail and the robot should not be unable to escape. However, with conventional automation functions, due to the nature of disaster sites, it is not only difficult to actually use robots in a real space for trial and error, but also work in conditions where many of the physical properties of the work target at the site are unknown. Be strong. For this reason, there is a limit to how well a robot can perform a desired task operation at a disaster site, and this is a factor that hinders the promotion of automation of disaster response work.

SUMMARY OF THE INVENTION The present invention has been devised to solve such problems. An object of the present invention is to provide an environment adaptability enhancement system for an autonomous work support robot, a motion simulation device, and a program therefor, which enable tasks to be performed more reliably with appropriate task motions.

In order to achieve the above object, the present invention mainly provides a robot having a function of automatically performing a task action corresponding to a task command from an operator in a real space, and an appropriate robot according to the environment of the real space. An environmental adaptability enhancement system for executing a task motion, comprising motion confirmation means for performing a motion check of the robot in advance by simulating the task motion in a virtual space, and motion confirmation by the motion confirmation means. a motion command means for commanding the task motion to the robot through processing according to the result, wherein the motion confirmation means transforms the environment of the real space into the virtual space based on topographical information around the robot; and a virtual motion generator for generating a virtual motion of the robot corresponding to the task command in the virtual space, wherein the motion command means is based on motion data of the virtual motion. a reliability calculation unit that calculates a reliability that serves as an index for judging the success or failure of the commanded task in the real space; and a correspondence determination unit that determines the correspondence of the robot.

According to the present invention, before a robot executes a task action corresponding to a predetermined task command in a real space, the task action is simulated in a virtual space, and the success or failure of the task by the robot can be estimated in advance. Further, when it is determined that the task is highly likely to fail, it is possible to adjust the motion of the robot in advance in a virtual space simulation in which the robot's control conditions and the like are changed. Furthermore, if it is determined by the simulation that the task cannot be performed, the robot will not be able to perform the task in the real space. can be avoided. As described above, according to the present invention, when the robot is caused to automatically perform a task operation corresponding to a task command, the task can be performed more reliably with an appropriate task operation according to the actual environment.

In addition, before the robot executes a task operation in real space, the robot performs a search operation in which it directly contacts an object such as the ground or an obstacle in real space, and obtains property information such as the physical properties and stability of the object. can be actively obtained. Accordingly, by including the property information in the simulation in the virtual space, it is possible to further improve the certainty of the task motion by the robot in the real space.

1 is a block diagram showing the configuration of an environment adaptability enhancement system according to an embodiment; FIG. FIG. 2 is a diagram mainly showing a schematic perspective view of an autonomous work support robot applied to the environmental adaptability enhancement system; (A) to (C) are conceptual diagrams for exemplifying an operation when the robot moves up a step. 4 is a flow chart for explaining a procedure for issuing a task execution command to a robot by a motion simulation device;

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 shows a block diagram showing the configuration of an environment adaptability enhancement system according to this embodiment. In this figure, an environment adaptability enhancement system 10 according to the present embodiment is an autonomous work support robot (hereinafter simply referred to as This system is applied to a robot 11 (referred to as a “robot”) and causes the robot 11 to perform an appropriate task operation according to the environment of the real space in which the robot 11 exists.

This environmental adaptability enhancement system 10 includes an environment information acquisition device 13 that is provided integrally with the robot 11 and acquires various data related to environment information in the real space, and A motion simulation device 14 confirms the motion of the robot 11 in a virtual space that reproduces the space environment, and determines the response of the robot 11 in the real space.

By the way, the application of the present invention is not particularly limited, but the robot 11 in this embodiment is, as shown in FIG. , the robot 11 can automatically perform a task operation corresponding to a task command from an operator who is away from the robot 11 . Here, the tasks instructed by the operator are roughly classified into a locomotion task for moving to a destination and a manipulation task for performing a desired operation while grasping an object. As the locomotion task, for example, assuming movement at a disaster site, in addition to running on flat ground, there are motions such as running on uneven ground, stepping over obstacles on the ground, and going up and down steps. Examples of the manipulation task include work such as removal and collection of objects such as rubble and obstacles at the site, cutting work of the object, and operation of valves, doors, etc. existing at the site.

An operator uses an input device 18 consisting of a known interface to issue a desired task command while viewing a display device 17 on which images obtained by a camera 16 mounted on the robot 11 are displayed. Then, an electric signal corresponding to the task command is transmitted from the input device 18 to the motion simulation device 14 via wired communication or wireless communication, and after being processed by the motion simulation device 14, the task motion is automatically executed by the robot 11. It has become so.

The robot 11 has a movable part 20 which is powered by a driving device (not shown) such as a motor and which can perform a predetermined movement. and an operation control unit 21 .

The movable section 20 includes a body 24 including a plurality of arms 23 operable within a predetermined range, and a traveling body 25 that operates while contacting the ground so as to move the body 24 . For example, as shown in FIG. It has a structure that enables various actions for disaster response work, such as enabling actions to grasp and remove objects such as rubble. In addition, since the detailed structure of the movable portion 20 is not an essential part of the present invention, the description thereof is omitted.

The environment information acquisition device 13 is integrally attached to the robot 11, and as shown in FIG. It comprises a shape detection sensor 27 for detecting data for detecting physical properties, and a physical property detection sensor 28 for detecting physical property data such as frictional force, reaction force, and weight of an object.

Various sensors can be used as the shape detection sensor 27 as long as they can measure data for specifying the position and shape of the object. In this embodiment, as the shape detection sensor 27, a range sensor or a depth sensor that acquires point cloud data consisting of position data of each point along the surface shape of the object is used.

The physical property detection sensor 28 is composed of a force sensor for measuring the friction force and reaction force on the surface of the object with respect to the robot 11 .

Note that the sensors 27 and 28 described above have known structures and are not essential parts of the present invention, so detailed description of the configuration and the like will be omitted. Further, these sensors 27 and 28 may be arranged separately from the robot 11 as long as the environmental information around the robot 11 can be obtained.

The detection result of the environment information acquisition device 13 is transmitted to the motion simulation device 14 at a predetermined timing using wireless communication or the like.

The operation simulation device 14 is composed of a computer including an arithmetic processing unit such as a CPU and a storage device such as a memory and a hard disk, and programs are installed to make the computer function as the following means.

The motion simulation device 14 in this embodiment functions as a server that is arranged at a location separate from the robot 11 , but it can also be arranged integrally with the robot 11 .

As shown in FIG. 1, the motion simulation device 14 is motion confirmation means for performing motion confirmation of the robot 11 in advance by simulating task motions of the robot 11 corresponding to task commands from the operator in a virtual space. 30, and an operation command means 31 for issuing an operation command to the operation control section 21 through processing according to the result of the operation confirmation by the operation confirmation means 30. FIG.

The operation confirmation means 30 includes a real environment reproduction section 33 that reproduces the environment around the robot in the real space (hereinafter referred to as "real environment") in the virtual space based on the detection result of the shape detection sensor 27, and a virtual A virtual motion generator 34 generates a virtual motion of the robot 11 corresponding to the task command from the operator in the space, and outputs the result of the virtual motion of the robot 11 simulated in the virtual space to the motion command means 31. and an output unit 35 .

The real environment reproduction unit 33 uses a physical simulator composed of known software having a dynamics simulation function for robots. Further, here, from the point cloud data of the object in the real space acquired by the shape detection sensor 27, the occupied area of the object in the real space is specified, and the terrain consisting of the object and the ground existing in the real space is determined. Information is reproduced in the virtual space as shape information. In this virtual space, a terrain model that reproduces a real environment and a robot model corresponding to the robot 11 are arranged, and a coefficient of friction between the robot 11 and the ground is set. It should be noted that the details of the processing performed by the real environment reproducing unit described above are based on known technology and are not essential to the present invention, and thus detailed description thereof will be omitted.

In the virtual motion generator 34, many patterns of basic motions of the robot 11 are stored in advance. Based on the terrain information, the basic motions required to perform the task are extracted, and a series of virtual motions are generated by combining them in chronological order. Although not particularly limited, for example, in the case of a locomotion task, various basic motions such as flat ground running, step climbing, step descending, rough terrain running, etc. are prepared. As for the step up and down, a plurality of basic motion patterns are prepared so as to be able to cope with changes in the height of the step. Therefore, for example, when the operator issues a task command to move over a step, the basic motion corresponding to the terrain information of the virtual space is extracted from among the basic motions used in the locomotion task, and each motion is executed. Divided into phases and combined in chronological order. Further, for example, in the case of a manipulation task such as removal of an object, various basic operations such as reaching, grasping, lifting, and carrying an object by the arm 23 are prepared. Further, in the case of door opening work, various basic operations such as reaching, gripping, rotating the doorknob by the arm 23, and opening the door while moving the door are prepared. As described above, in each phase of the virtual motion, the target joint angle, joint angle, and the like of the movable portion 20 of the robot 11 are set in advance according to the content of the motion. The motion generation described above is performed by a well-known algorithm, and further detailed description is omitted.

The output unit 35 outputs the position and orientation of the robot model 24 in the robot model, the time taken in each phase, the motion of the movable unit 20 as a result of executing the virtual motion of the robot model corresponding to the task command from the operator in the virtual space. Motion data of virtual motions such as joint angles, angular velocities, and torques (hereinafter referred to as “virtual motion data”) are output.

The motion commanding means 31 includes a reliability calculation unit 37 that calculates a reliability that serves as an index for judging the success or failure of a commanded task in the real space based on the virtual motion data, and a robot 11 based on the reliability. When the response determination unit 38 that determines the response of the robot 11 including execution or non-execution of the task in the real space and the response determination unit 38 determines that the reliability is such that execution or non-execution of the task cannot be determined, A motion adjustment unit 39 is provided for adjusting the motion of the robot model by searching for a motion that is more appropriate than the virtual motion.

From the virtual motion data, the reliability calculation unit 37 calculates a _first evaluation value P1 relating to task stability, a _second evaluation value P2 relating to task efficiency, and a real environment After obtaining a _third evaluation value P3 relating to the reproducibility of , the reliability _C is obtained using these evaluation values P1, P2 _, and _P3 . The reliability C is set such that the greater the value, the higher the probability of success of the task instructed by the operator. Also, the reliability C is obtained for each phase that constitutes the virtual motion generated by the virtual motion generation unit 34 .

Each of the evaluation values P ₁ , P ₂ , and P ₃ takes a value from 0 to 1, and the larger the value, the worse the performance of each, and the more likely it is to hinder the execution of the task. is set to

Specifically, in the case of the locomotion task, the _first evaluation value P1 is the overturn probability P1 representing the possibility of the robot ₁₁ overturning during movement. As shown in the above formula ( ₂ ), the possibility of overturning P1 is determined by the maximum angle θ _max (deg) representing the maximum inclination angle of the body 24 with respect to the ground during movement, and the maximum is obtained from the limit angle θ _lim (deg), which is the inclination angle of . Here, the limit angle θ _lim is preset to a default value based on the actual performance of the robot 11 . That is, the overturn possibility P1 takes _a value of "0" when the maximum body angle θ _max is 0 (deg), and is set to increase as the maximum body angle θ _max increases. In the above formula (2), when the _maximum aircraft angle θ _max becomes the same as the _limit angle θ _lim , the value becomes "1". In this case, the value "1" is adopted as the movement of the robot 11 becomes impossible.

Here, the overturn possibility P1 is calculated for _each of the roll angle and the pitch angle of the robot ₁₁ , and the greater overturn possibility P1 is adopted. Although not particularly limited, the limit angle θ _lim is set to different values depending on the roll angle (eg, 20 (deg)) and the pitch angle (eg, 45 (deg)).

Further, in the case of the object lifting work in the manipulation task, as the _first evaluation value P1, for example, the grasping angle (roll, pitch) of the object when lifting the object to be grasped is calculated in the same manner as described above. Desired. Further, in the case of the door opening operation, as the _first evaluation value P1, the change in the gripping angle at the tip portion of the arm 23 is obtained in the same manner as described above.

The _second evaluation value P2 is an evaluation value representing the possibility of task stalemate. and a predetermined reference time T set to . That is, the _second evaluation value P2 takes a value of "0" when the operation time t is the same as the reference time T, indicating that the working efficiency is good. On the other hand, the _second evaluation value P2 is set to increase and approach the value "1" because the longer the operation time t takes than the reference time T, the lower the working efficiency. In addition, in the above equation (3), when the operating time t is shorter than the reference time T, P ₂ =0.

The _third evaluation value P3 is an evaluation value corresponding to the possibility of a measurement error. A positive value of 1 or less is taken from the distance d between the position of 11 and an environmental object (for example, a step) related to the operation in the target phase, and the known error E of the shape detection sensor 27. You are asked to An example of the error E is 5% (E=0.05).

The correspondence determination unit 38 determines the following correspondence based on a preset threshold based on the reliability calculated in each phase that constitutes the virtual action generated in the virtual space in response to the operator's task command. done.

In all phases of the virtual operation, if the reliability exceeds the upper limit threshold (for example, 0.7), that is, if the reliability of task execution is high (hereinafter referred to as “reliability “high” case”) , it is determined that the robot 11 can reliably execute the task in the real space, and the virtual motion generated by the virtual motion generator 34 is commanded to the motion controller 21 of the robot 11 .

On the other hand, in any phase of the virtual operation, if the reliability is less than the lower threshold (for example, 0.3) lower than the upper threshold, that is, if the reliability of task execution is low (hereinafter referred to as "reliability" In the case of "low"), it is determined that the robot 11 cannot reliably execute the task in the real space, and a command is given to the robot 11 not to execute the task operation in the real space.

In the case of "medium" reliability that does not correspond to either the "high" reliability or the "low" reliability, that is, the reliability in all phases of the virtual operation exceeds the lower threshold, but at least one If the reliability in the phase does not exceed the upper threshold, the motion adjusting unit 39 adjusts the virtual motion in that phase as follows.

In the motion adjusting unit 39, for each phase in which the reliability is equal to or higher than the lower limit threshold and less than the upper limit threshold, the physical properties and states of objects in the real environment that affect the success or failure of the task of the robot 11 are also obtained. A process of searching for an optimum motion whose reliability is equal to or higher than the upper limit threshold is performed by performing motion learning in consideration thereof.

This operation adjustment unit 39 includes a property information acquisition unit 41 that acquires property information about the physical properties and states of objects in the real environment, and a robot robot through machine learning in a virtual space that adds property information to terrain information as environment information. and a motion search unit 42 for searching for 11 optimal motions.

The property information acquisition unit 41 causes the robot 11 in the real space to perform a search operation in which a portion of the robot 11 moves while being in direct contact with the object, thereby detecting property data with the property detection sensor 28 and detecting the object. Get the property information of This property information includes friction information for estimating the frictional state of the surface of the object, weight information for estimating the mass and center of gravity of the object, deformation state due to the hardness of the object, and the position of the object. and stability information for estimating the state. Therefore, the property information acquisition unit 41 has a friction information acquisition function for acquiring friction information, a weight information acquisition function for acquiring weight information, and a stability information acquisition function for acquiring stability information. Here, in each phase of the task operation of the robot 11, necessary information in the property information is stored in advance, and the property information acquisition unit 41 acquires necessary information for each phase in which motion adjustment is performed. It is designed to For example, in the locomotion task for moving the robot 11, in the step-climbing phase (see FIG. 3) in which the robot body 24 is raised while the arm 23 in front of the movement direction touches the step, the friction information on the ground that the arm 23 contacts is and stability information such as whether the ground is firm or not, and the property information acquisition unit 41 functions to acquire such information.

In the friction information acquisition function, the tip of the arm 23 is brought into contact with the surface of the object whose friction information is to be acquired, and the arm 23 is moved along the surface at a predetermined speed and contact force. Then, an operation command is issued to the operation control unit 21 of the robot 11 . Furthermore, here, the friction coefficient of the object is obtained as the friction information from the measurement value of the physical property detection sensor 28 accompanying the movement of the arm 23, that is, the measurement value of the frictional force that is the resistance force in the movement direction of the arm 23. be done.

The weight information acquisition function instructs the motion control unit 21 of the robot 11 to execute a search motion in which an object whose weight information is to be acquired is gripped by the tips of the two arms 23 and lifted. is done. Furthermore, here, the weight and the center of gravity of the object can be obtained as weight information by specifying the frictional force from the measured value by the physical property detection sensor 28 when the object is lifted.

In the stability information acquisition function, the motion control unit 23 of the robot 11 is caused to perform a search motion in which a predetermined pressing force is applied while the tip of the arm 23 is in contact with the surface of the object whose property information is to be acquired. An operation command is issued. Furthermore, here, the stability information, which is a numerical value corresponding to the elasticity and stability representing the ease with which the object is displaced, is taken as the stability information from the relationship between the amount of displacement of the pressed surface and the reaction force.

The property information acquisition unit 41 described above, together with the physical property detection sensor 28, causes the robot 11 to physically contact an object in the environment in real space, so that the tactile information determines the success or failure of the task of the robot 11. An active information collecting means is configured to collect property information of an object based on physical data.

The motion search unit 42 uses a motion learning method such as deep reinforcement learning or genetic algorithm, and changes a predetermined number of learning parameters and control rules while performing all phases in which motion adjustment needs to be performed in parallel. , the optimum operation is searched for in the number of times of learning or operation time. In the motion search process, the reliability calculation unit 37 calculates the reliability of each of the candidate motions. Therefore, when the reliability of all the phases requiring motion adjustment becomes equal to or higher than the upper limit threshold, the response determining unit 38 determines that the robot 11 can reliably execute the task in the real space, All virtual motions whose reliability is equal to or higher than the upper limit threshold, including the adjusted virtual motion (hereinafter referred to as “adjusted virtual motion”), are instructed to the motion control unit 21 of the robot 11 . On the other hand, if it is not possible to search for a motion whose reliability is equal to or higher than the upper threshold within a predetermined number of times or within a predetermined period of time in all phases by the motion search here, the correspondence determination unit 38 determines the task motion of the robot 11. is determined to be non-executable.

In addition, the motion pattern of the robot 11 for which the reliability of the task success probability has been obtained by the motion search unit 42 is added as a basic motion pre-stored in the virtual motion generation unit 34 each time. , the virtual motion is generated when the next motion verification by the motion verification means 30 is performed.

In the motion search unit 42, the collection of the property information by the motion adjustment unit 39 is not essential, and only shape information including terrain information in the actual environment is used for motion learning by arbitrary selection of learning parameters and control rules. It is also possible to

Next, a procedure for issuing a task execution command to the robot 11 by the motion simulation device 14 will be described below with reference to the flow chart of FIG.

First, before the robot 11 operates at the site, an operator positioned away from the robot 11 uses the input device 18 while checking the image obtained from the camera 16 attached to the robot 11 on the display device 17. Then, a task command is issued, and the task command is input to the motion simulation device 14 .

Then, in the state before the task operation where the robot 11 has stopped its operation, the shape detection sensor 27 attached to the robot 11 detects the position and position of an object such as the ground or an object existing around the robot 11 in the real space. Data for specifying the shape is measured and transmitted to the motion simulation device 14 .

After that, in the motion simulation device 14, based on the task command from the operator and the data from the shape detection sensor 27, the motion confirmation means 30 determines the motion of the robot 11 in the virtual space before actually moving the robot 11 in the real space. An operation check is performed to simulate the operation. That is, in the real environment reproduction unit 33, the shape information of the object is acquired from the measurement data of the shape detection sensor 27 (step S101), and the real environment around the robot 11 in the real space is reproduced in the virtual space. (step S102). Further, in the virtual motion generator 34, corresponding motions are extracted from the pre-stored basic motions for each phase according to the task command, and virtual motions are generated by combining them (step S103).

Next, based on the result of executing a series of virtual actions generated in the virtual space, the reliability calculation unit 37 calculates the reliability for each phase (step S104). Then, the reliability in all phases is evaluated using the upper threshold and the lower threshold (step S105).

Here, when the reliability in all the phases exceeds the upper limit threshold and the reliability is "high", the possibility of task success in the virtual action generated by the virtual action generating unit 34 is considered high. is instructed to the robot 11, and the task operation is executed by the robot 11 (step S106).

In addition, when the reliability in any phase is less than the lower limit threshold and the reliability is "low", the probability of task success in the virtual motion generated by the virtual motion generating unit 34 is low. The task operation of the robot 11 is made non-executable (step S107).

Furthermore, when the reliability is "medium", which does not correspond to any of the above cases, the virtual action generated by the virtual action generating unit 34 is evaluated as having a possibility of task success, although it cannot be said to be certain. be. Therefore, in order to further increase the certainty of task success, motion learning is performed a predetermined number of times or over a predetermined period of time (step S108) for phases where the reliability is equal to or higher than the lower limit threshold and less than the upper limit threshold. adjustments are made. That is, first, the property information acquisition unit 41 issues a search operation command to the robot 11 to obtain tactile information while a part of the robot 11 in the real space is in contact with the object, and obtains tactile information related to the physical properties, state, etc. of the object in the real environment. Property information is acquired (step S109). After that, the motion searching unit 42 adds property information to the shape information of the object as environment information, and then searches for the optimum motion for task success by motion learning (step S110). As a result, the reliability of the adjusted virtual motion, which is the searched motion, is calculated again in the same manner as described above, and evaluation is performed based on the reliability (steps S104 and S105). As a result, when the reliability of the adjusted virtual motion becomes "high" in all phases, the robot is instructed to execute the adjusted virtual motion (step S106). , the task motion of the robot 11 is not executed (step S107).

In the above-described embodiment, the multi-arm mobile robot was illustrated as an example of the autonomous work support robot 11, but the present invention is not limited to this, and can be applied to autonomous work support robots of various structures. .

In addition, the configuration of each part of the device in the present invention is not limited to the illustrated configuration example, and various modifications are possible as long as substantially the same action is exhibited.

10 Environmental adaptability enhancement system 11 Autonomous work support robot 14 Motion simulation device 28 Physical property detection sensor (active information gathering means)
30 action confirmation means 31 action command means 33 real environment reproduction part 34 virtual action generation part 37 reliability calculation part 38 correspondence determination part 39 action adjustment part 41 property information acquisition part (active information gathering means)

Claims (6)

An environmental adaptability enhancing system for making a robot, which has a function to automatically perform a task operation corresponding to a task command from an operator in a real space, perform the task operation appropriately according to the environment of the real space. There is
By simulating the task motion in a virtual space, an operation confirmation means for confirming the operation of the robot in advance, and the task operation is commanded to the robot through processing according to the result of the operation confirmation by the operation confirmation means. and an operation command means,
The action confirmation means includes a real environment reproduction unit that reproduces the environment of the real space in the virtual space based on topographical information around the robot, and a robot controller that reproduces the environment of the real space in the virtual space in response to the task command. a virtual motion generator that generates a virtual motion;
The motion command means includes a reliability calculation unit that calculates a reliability that serves as an index for judging the success or failure of the commanded task in the real space based on the motion data of the virtual motion, and a robot based on the reliability. a response determination unit that determines a response of the robot including execution or non-execution of the task in the real space ; and a motion adjusting unit that adjusts the motion of the robot by searching for a motion that is more appropriate than the virtual motion . The reliability is determined by a first evaluation value regarding task stability, a second evaluation value regarding task efficiency, and a third evaluation value regarding reproducibility of the real environment in the virtual space. 2. The system for enhancing environmental adaptability of an autonomous work support robot according to claim 1. Further comprising active information collecting means for collecting property information of the object based on physical property data that determines the success or failure of the task by bringing the robot into physical contact with the object in the real space,
said active information collecting means instructing said robot to perform a search operation for collecting said property information when said reliability is such that execution or non-execution of said task cannot be determined by said correspondence determination unit;
2. The system for enhancing environmental adaptability of an autonomous work support robot according to claim 1 , wherein said motion adjusting unit searches for said motion in consideration of said property information. The virtual motion generation unit extracts a basic motion corresponding to the task command from among a plurality of preset basic motions to generate the virtual motion,
The motion adjustment unit conducts motion learning to search for an optimum motion while appropriately adjusting a plurality of learning parameters and control rules,
In the virtual action generation unit, as a result of the action learning, the action for which the reliability is obtained with a high probability of task success is added to the basic action and used when generating the next virtual action. 2. The system for enhancing environmental adaptability of an autonomous work support robot according to claim 1 . A simulation device for causing a robot having a function to automatically perform a task operation corresponding to a task command from an operator in a real space to perform the appropriate task operation according to the environment of the real space,
By simulating the task motion in a virtual space, an operation confirmation means for confirming the operation of the robot in advance, and the task operation is commanded to the robot through processing according to the result of the operation confirmation by the operation confirmation means. and an operation command means,
The action confirmation means includes a real environment reproduction unit that reproduces the environment of the real space in the virtual space based on topographical information around the robot, and a robot controller that reproduces the environment of the real space in the virtual space in response to the task command. a virtual motion generator that generates a virtual motion;
The motion command means includes a reliability calculation unit that calculates a reliability that serves as an index for judging the success or failure of the commanded task in the real space based on the motion data of the virtual motion, and a robot based on the reliability. a response determination unit that determines a response of the robot including execution or non-execution of the task in the real space ; and a motion adjusting unit that adjusts the motion of the robot by searching for a motion that is more appropriate than the virtual motion . A program for a simulation apparatus for causing a robot having a function to automatically perform a task operation corresponding to a task command from an operator in a real space to perform the appropriate task operation according to the environment of the real space. hand,
By simulating the task motion in a virtual space, an operation confirmation means for confirming the operation of the robot in advance, and the task operation is commanded to the robot through processing according to the result of the operation confirmation by the operation confirmation means. functioning a computer as an operation command means,
The action confirmation means includes a real environment reproduction unit that reproduces the environment of the real space in the virtual space based on topographical information around the robot, and a robot controller that reproduces the environment of the real space in the virtual space in response to the task command. a virtual motion generator that generates a virtual motion;
The motion command means includes a reliability calculation unit that calculates a reliability that serves as an index for judging the success or failure of the commanded task in the real space based on the motion data of the virtual motion, and a robot based on the reliability. a response determination unit that determines a response of the robot including execution or non-execution of the task in the real space ; and a motion adjusting unit that adjusts the motion of the robot by searching for a motion that is more appropriate than the virtual motion .

Images