JP4878462B2 - Communication robot - Google Patents

Description

  The present invention relates to a communication robot, and more particularly to a communication robot that includes a touch sensor and responds to human tactile behavior.

An example of this type of communication robot (a robot that can communicate with a person) is disclosed in Patent Document 1.
JP2002355578 [B25J 13/00 A63H 3/33 11/00 B25J 5/00]

  In the communication robot in Patent Document 1, when a person touches the touch sensor, the person reacts to the human tactile action to notify the person that he / she has perceived the tactile action, thereby further facilitating communication with the person. ing.

  On the other hand, it is the same with other animals as well as humans, but before other people actually touch, it is detected in advance by some means, and the reaction behavior that avoids or responds to that touch behavior Sometimes However, in the prior art of Patent Document 1, the communication robot cannot detect the tactile behavior until a human actually touches the robot, and in that sense, it is said that it is somewhat lacking in “biological character”. I don't get it.

  Therefore, the main object of the present invention is to provide a novel communication robot system.

  Another object of the present invention is to provide a communication robot system capable of a more biological reaction.

The invention according to claim 1 is a communication robot that detects human tactile behavior based on sensor inputs from a plurality of touch sensors provided in a plurality of parts, and determines whether or not a human has actually touched. Means for detecting whether or not a person is within the predetermined range when the touch determination means does not determine that the person is actually touching, and the presence detection means for detecting that the person is within the predetermined range. When detected, it comprises first specifying means for specifying a part where a human hand is approaching within a predetermined distance as an estimated contact part, and the first specifying means is a part where the human hand is approaching within a predetermined distance Including first determination means for determining whether or not there are a plurality of parts, wherein the first determination means detects that there are a plurality of parts where a human hand approaches within a predetermined distance. Identifying a satisfying one site as the estimated contact site, a communication robot.

According to the first aspect of the present invention, the communication robot (the reference numeral exemplifying the corresponding or corresponding part in the embodiment is “10”, the same applies hereinafter) includes the CPU (50). Acts as a detection means. The touch determination means determines whether or not a human has touched any part of the communication robot based on the presence or absence of sensor input from touch sensors provided in a plurality of parts. Then, when such a touch determination means does not detect a touch, the presence detection means, for example, based on the three-dimensional coordinates of the respective parts of the robot and the human from the motion capture (76), the human is predetermined from the robot. Determine if it is within range. Specifically, the CPU determines whether or not the human is within a predetermined distance range based on the length of the human hand (S37, S39). The CPU also functions as the first specifying means. When the CPU determines that the human is within the predetermined range, the CPU determines whether there is a portion where the human hand is approaching within the predetermined distance. For example, the part is specified as an estimated contact part (S49-S59).
The CPU also functions as first determination means, and determines whether there are a plurality of parts where the human hand approaches within a predetermined distance (S49). Then, when the first determining unit detects that there are a plurality of parts where the human hand approaches within a predetermined distance, the first specifying part specifies a part that satisfies the predetermined condition as an estimated contact part (S57).
Similarly, in the invention of claim 2, the first determining means determines whether there are a plurality of parts where the human hand approaches within a predetermined distance, and the first specifying means determines whether the human hand is the first determining means. When it is detected that there are a plurality of parts approaching within a predetermined distance, a part satisfying the predetermined condition is specified as the estimated contact part.

According to the first aspect of the present invention, it is possible to detect a human tactile behavior without actually touching a human, and to obtain a communication robot that is more biological. Furthermore, even if there are a plurality of parts close to a human hand, any one of them can be specified as an estimated contact part, so that even when a reaction action to a tactile action is caused, a correct reaction action can be caused. it can.

A second aspect of the present invention, the predetermined condition is whether the human gaze is directed, it is a communication robot according to claim 1.

In the invention of claim 2 , the CPU estimates the direction in which the human line of sight is directed, for example, by calculating a human line-of-sight vector, and the part to which the line of sight is directed is the part that the person is trying to contact Since it specifies (S51), the part which a human is interested can be correctly specified as an estimated contact part.

The invention according to claim 3 is the communication according to claim 1 or 2 , wherein the first specifying means specifies the part as an estimated contact part when there is only one part where a human hand approaches within a predetermined distance. It is a robot.

In the invention of claim 3 , when there is only one part where the human hand approaches within the predetermined distance, the CPU specifies that part as the estimated contact part (S55).

The invention according to claim 4 is the communication robot according to any one of claims 1 to 3 , further comprising a reaction action occurrence predetermined means for causing a reaction action in accordance with the estimated contact site specified by the first specifying means.

In the invention of claim 4 , the CPU also functions as a reaction behavior generating unit, and the reaction behavior generating unit generates a different reaction behavior (Table 1) depending on the estimated contact site (S7). Therefore, even if a human does not actually touch, the reaction behavior corresponding to the estimated contact site can be executed.

The invention of claim 5 further includes a plurality of touch sensors provided in the plurality of parts, and second specifying means for specifying an actual contact part actually touched by a human in response to a sensor input from the touch sensor. Item 5. The communication robot according to any one of Items 1 to 4 .

According to the fifth aspect of the present invention, the CPU can also specify the actual contact site actually touched by a human by sensor input from a plurality of touch sensors (721-728). Can detect behavior.

In the invention of claim 6 , the second specifying means includes second determining means for determining whether or not there is a sensor input from a plurality of touch sensors, and the second determining means indicates whether or not there is a sensor input from the plurality of touch sensors. The communication robot according to claim 5 , wherein when detected, one actual contact site is identified in accordance with a human gaze direction.

In the invention of claim 6, CPU (second specifying means), when it detects that one of the plurality of touch sensors in the second determining means (S25) has sensor inputs, only one in accordance with the human eye direction A contact site is specified (S27-S31). It is possible to correctly identify a part that a person is interested in as an estimated contact part.

  According to the present invention, even if a human does not actually touch, it is possible to detect a human's tactile behavior, and in some cases, a reaction behavior to the tactile behavior can be generated. A communication robot that is more biological is obtained.

  The above object, other objects, features and advantages of the present invention will become more apparent from the following detailed description of embodiments with reference to the drawings.

  A communication robot (hereinafter, simply referred to as a “robot”) 10 according to an embodiment of the present invention shown in FIG. 1 includes a carriage 12, and wheels on which the robot 10 autonomously moves are provided on the lower surface of the carriage 12. 14 is provided. The wheel 14 is driven by a wheel motor (indicated by reference numeral “70” in FIG. 2), and the carriage 12, that is, the robot 10 can be moved in any direction, front, back, left, and right. Although not shown, a collision sensor (indicated by reference numeral “74” in FIG. 2) is attached to the front surface of the carriage 12, and this collision sensor is used to detect a person or other obstacle to the carriage 12. Detect contact. When contact with an obstacle is detected during the movement of the robot 10, the driving of the wheels 14 is immediately stopped to suddenly stop the movement of the robot 10 to prevent a collision.

  In this embodiment, the height of the robot 10 is set to about 100 cm so as not to intimidate people, particularly children. However, this height can be arbitrarily changed.

  A polygonal column sensor mounting panel 16 is provided on the carriage 12, and an ultrasonic distance sensor 18 is mounted on each surface of the sensor mounting panel 16. The ultrasonic distance sensor 18 measures a distance between the mounting panel 16, that is, a person mainly around the robot 10.

  On the carriage 12, the lower part is surrounded by the above-described mounting panel 16, and the body of the robot 10 is attached so as to stand upright. This body is composed of a lower body 20 and an upper body 22, and the lower body 20 and the upper body 22 are connected by a connecting portion 24. Although not shown, the connecting portion 24 has a built-in lifting mechanism, and the height of the upper body 22, that is, the height of the robot 10 can be changed by using the lifting mechanism. As will be described later, the elevating mechanism is driven by a waist motor (indicated by reference numeral “68” in FIG. 2). The height 100 cm of the robot 10 described above is a value when the upper body 22 is at its lowest position. Therefore, the height of the robot 10 can be 100 cm or more.

  However, it is also possible to integrate the lower body 20 and the upper body 22 into a single body whose height cannot be adjusted.

  One omnidirectional camera 26 and one microphone 28 are provided in the approximate center of the upper body 22. The omnidirectional camera 26 photographs the surroundings of the robot 10 and is distinguished from an eye camera 46 described later. The microphone 28 captures ambient sounds, particularly human voice.

  Upper arms 32R and 32L are attached to both shoulders of the upper body 22 by shoulder joints 30R and 30L, respectively. The shoulder joints 30R and 30L each have three degrees of freedom. That is, the shoulder joint 30R can control the angle of the upper arm 32R around each of the X axis, the Y axis, and the Z axis. The Y axis is an axis parallel to the longitudinal direction (or axis) of the upper arm 32R, and the X axis and the Z axis are axes orthogonal to the Y axis from different directions. The shoulder joint 30L can control the angle of the upper arm 32L around each of the A, B, and C axes. The B axis is an axis parallel to the longitudinal direction (or axis) of the upper arm 32L, and the A axis and the C axis are axes orthogonal to the B axis from different directions.

  Forearms 36R and 36L are attached to the respective distal ends of upper arms 32R and 32L via elbow joints 34R and 34L. The elbow joints 34R and 34L can control the angles of the forearms 36R and 36L around the axes of the W axis and the D axis, respectively.

  In the X, Y, X, W axes and the A, B, C, D axes that control the displacement of the upper arms 32R and 32L and the forearms 36R and 36L (FIG. 1), “0 degree” is the home position. In this home position, the upper arms 32R and 32L and the forearms 36R and 36L are directed downward.

  Further, as shown in FIG. 3, a touch sensor is provided on the shoulder portion of the upper body 22 including the shoulder joints 30R and 30L, the upper arms 32R and 32L, the forearms 36R and 36L, and the head 42 and the abdomen. 721-728 are provided, and these touch sensors 721-728 detect whether a person has touched these parts of the robot 10 or not. These touch sensors are generally indicated by reference numeral 72 in FIG.

  Spheres 38R and 38L corresponding to hands are fixedly attached to the tips of the forearms 36R and 36L, respectively. Instead of the spheres 38R and 38L, a “hand” in the shape of a human hand can be used when a finger function is required unlike the robot 10 of this embodiment.

  A head portion 42 is attached to the upper center of the upper body 22 via a neck joint 40. The neck joint 40 has three degrees of freedom and can be angle-controlled around each of the S, T, and U axes. The S-axis is an axis that goes directly from the neck, and the T-axis and the U-axis are axes that are orthogonal to the S-axis in different directions. The head 42 is provided with a speaker 44 at a position corresponding to a human mouth, and an eye camera 46 at a position corresponding to the eyes. The speaker 44 is used by the robot 10 to communicate with a person around it by voice or voice. The eye camera 46 captures the video signal by photographing the face and other parts of the person approaching the robot 10. However, the speaker 44 may be provided in another part of the robot 10, for example, the trunk.

  Note that each of the omnidirectional camera 26 and the eye camera 46 described above may be a camera using a solid-state imaging device such as a CCD or a CMOS.

  The configuration of the control system of the robot 10 shown in FIG. 1 is shown in the block diagram of FIG. As shown in FIG. 2, the robot 10 includes a microcomputer or a CPU 50 for overall control. The CPU 50 is connected to a memory 54, a motor control board 56, a sensor input / output board 58 and a voice through a bus 52. An input / output board 60 is connected.

  Although not shown, the memory 54 includes a ROM and a RAM, in which a control program for the robot 10 is written in advance and voice or voice data to be generated from the speaker 44 is stored. The RAM is used as a temporary storage memory and can be used as a working memory.

  The motor control board 56 is constituted by, for example, a DSP (Digital Signal Processor), and controls each axis motor of each arm and head. That is, the motor control board 56 receives the control data from the CPU 50, and controls the angles of the three motors for controlling the X, Y, and Z axes of the right shoulder joint 30R and the axis W of the right elbow joint 34R. The rotation angle of a total of four motors (one in FIG. 2 and collectively shown as “right arm motor”) 62 is adjusted. The motor control board 56 includes a total of four motors including three motors that control the angles of the A, B, and C axes of the left shoulder joint 30L and one motor that controls the angle of the D axis of the left elbow joint 34L. (They are collectively shown as “left arm motor” in FIG. 2). The rotation angle of 64 is adjusted. The motor control board 56 also adjusts the rotation angle of three motors 66 (collectively shown as “head motors” in FIG. 2) that control the angles of the S, T, and U axes of the head 42. To do. The motor control board 56 also controls a waist motor 68 and two motors 70 (referred to collectively as “wheel motors” in FIG. 2) that drive the wheels 14.

  In addition, the above-described motors of this embodiment are stepping motors or pulse motors for simplifying the control, except for the wheel motors 70, but may be DC motors similarly to the wheel motors 70.

  Similarly, the sensor input / output board 58 is configured by a DSP, and takes in signals from each sensor and camera and gives them to the CPU 50. That is, data relating to the reflection time from each of the ultrasonic distance sensors 18 is input to the CPU 50 through the sensor input / output board 58. The video signal from the omnidirectional camera 26 is input to the CPU 50 after being subjected to predetermined processing by the sensor input / output board 58 as required. The video signal from the eye camera 46 is also given to the CPU 50 in the same manner. In FIG. 2, the touch sensors described in FIG. 1 are collectively represented as “touch sensors 72”, and signals from those touch sensors 72 are given to the CPU 50 via the sensor input / output board 58. .

  The speaker 44 is provided with synthesized voice data from the CPU 50 via the voice input / output board 60, and in response to this, the speaker 44 outputs voice or voice according to the data. Then, the voice input from the microphone 28 is taken into the CPU 50 via the voice input / output board 60.

  Here, the touch sensor 72 shown comprehensively in FIG. 2 will be described in detail. In the robot 10 of this embodiment, a total of eight touch sensors 721-728 are provided as shown in FIG. The touch sensor 721 is provided on the upper surface (the top of the head) of the head 42, and the touch sensors 722 and 725 are provided on the left shoulder and the right shoulder, respectively. Touch sensors 723 and 724 are disposed on the left upper arm 32L and the left forearm 34L, and touch sensors 7256 and 727 are disposed on the right upper arm 32R and the left forearm 34R. A touch sensor 728 is provided on the chest and / or belly. Note that numbers (1) to (8) are assigned to the touch sensors 721 to 728, respectively, and their center positions are represented by P1 to P8.

  However, the number of touch sensors and the place or part where they are arranged are not limited to this embodiment, and can be changed as appropriate.

  Further, the communication robot system of the embodiment uses an optical motion capture 76 shown in FIG. As is well known, motion capture is a system that digitally records the movements and postures of real people and objects. Markers that are attached to human bodies and objects and trackers that detect this are used. Capture in combination. In the optical system, for example, a reflective seal is used as a marker, and a plurality of cameras are used as trackers.

  Explaining in the embodiment, 19 reflective stickers are attached as markers 78 to the communication robot 10 shown in FIG. On the other hand, as shown by a black circle in FIG. 5, 19 reflective stickers are similarly attached as markers 78 to a person 80 who communicates such as touching the robot 10. Then, the robot 10 and the human 80 are simultaneously photographed by a plurality of MC (motion capture) cameras 821-82n shown in FIG. 2, and image signals from these cameras 821-82n are input to the MC computer 84. The MC computer 84 calculates the three-dimensional position (coordinates) of each marker 78 (FIGS. 4 and 5) in real time, and supplies it to the robot computer, that is, the CPU 50 through the bus 52. However, the MC computer 84 and the robot computer 50 may be linked by a wireless LAN instead of wired.

  As the motion capture system (three-dimensional motion measurement device) 76, a known motion capture system is applied. However, in the experiments by the inventors, the optical system of VICON (http://www./vicon.com/) is used. The motion capture system of was used. In this case, 12 MC cameras 82 are used.

  The movement and posture of the robot 10 can be grasped without necessarily using motion capture. For example, the movement and posture of the robot 10 can be calculated from the rotation angles of the respective axis motors of the arms and head, the right arm motor 62, the left arm motor 64, and the head motor 66 described above. However, in the embodiment, the motion capture 76 is used to reduce the real-time calculation amount of the CPU 50.

  FIG. 6 is a flowchart for explaining the operation of the CPU 50 of the embodiment shown in FIGS. 1 and 2. In the first step S <b> 1, the CPU 50 determines whether or not a stop command is input from the operator of the robot 10. If “YES” is determined in the step S1, the processing is naturally ended.

  If "NO" is determined in the step S1, the CPU 50 is contacted in the next step S3 whether or not the human 80 has made contact with the robot 10, and if there has been a contact action, any position or part of the robot is contacted. It is determined.

  Here, the subroutine of step S3 will be described in detail with reference to FIGS.

  In the first step S <b> 21 in FIG. 7, the CPU 50 checks the presence or absence of a contact reaction of the touch sensor for each part of the robot 10. Specifically, the CPU 50 determines whether there is a sensor input for the touch sensor 72, that is, the touch sensors 721-728 shown in FIG. 3, through the sensor input / output board 58 (FIG. 2). If there is a sensor input, the CPU 50 sets the touch sensor flag Ts_flag (n) formed in the memory 54 of FIG. 2 to “1”. However, “n” in parentheses represents the number of the touch sensor, which is “1-8” in the embodiment. When there is no sensor input, the CPU 50 writes “0” in the corresponding flag Ts_flag (n). Note that the number of touch sensors with sensor input is also counted at the same time, and each time “1” is written in the touch sensor flag, the CPU 50 increments the touch sensor counter Ts_count formed in the same memory 54 (+1).

  In subsequent step S23, CPU 50 determines whether or not the count value of touch sensor counter Ts_count is “1” or more. If “YES” is determined in this step S23, it means that there is a sensor input from at least one touch sensor in the previous step S21, that is, the human 80 (FIG. 5) is actually somewhere in the robot 10. Means that the human 80 has not actually touched (contacted) the robot 10.

  If “YES” in the step S23, in the subsequent step S25, the CPU 50 determines whether or not the count value of the counter Ts_count is “2 or more” (Ts_count> 1). This step S25 functions as a second determination means, and that “YES” is determined in step S25 means that there are sensor inputs from two or more touch sensors in step S21, that is, a plurality of robots 10 Means that a place or part has been touched by a human. In this embodiment, even if there are touches at a plurality of locations on the robot 10, the location is specified as one location. For this purpose, the steps after step S27 are executed.

  In step S27 constituting the second specifying means, simply speaking, the CPU 50 specifies the contact location according to the line-of-sight direction of the human 80 at that time.

  The line-of-sight direction of the other person is calculated as a line-of-sight vector from the coordinates of the marker attached to the head as shown in FIG. Specifically, as shown in FIG. 9, two marker positions attached to the front of the head are shown as Right_head_front on the right side, Pref_head_front on the left side, and Pcenter_head_back as the marker position attached on the rear side. The line-of-sight vector is calculated from these marker positions (coordinates). The line-of-sight vector EyeVector is a vector from the Center_head_back to the center position of the line connecting the Rights_head_front and the Left_head_front. The line-of-sight vector can be obtained by the following equation (1).

  Then, the touch sensor Te in the direction in which the line-of-sight vector EyeVector faces is specified. That is, it is determined which of the eight touch sensors 721-728 (FIG. 3) corresponds to the touch sensor in the direction of the line-of-sight vector EyeVector.

  Here, first, the center coordinates P1-P8 of the touch sensors 721-728 of the body of the robot are obtained as follows.

  The center point P1-P8 of each touch sensor 721-728 of the robot can be easily calculated if the vector from the coordinates of the marker attached in the vicinity of the touch sensor 721-728 is known in advance. However, in this case, as shown in FIGS. 3 and 4, the marker near the touch sensor 721-728 has a relative position with respect to the center point P1-P8 of the touch sensor 721-728 even when the robot operates. It must be installed where it will not change.

  In the following specific example, as an example, a method of calculating the coordinates Ptouchsensor_shoulder of the center positions P1 to P8 of the shoulder touch sensor 722 (725) will be described.

  A vector Vshoulder (= Ptouchsensor_shoulder-Psoulder) from the coordinate Pshoulder of the marker attached to the shoulder to the center coordinate Ptouchsensor_shoulder of the shoulder touch sensor 722 (725) is measured in advance. By using this vector Vshoulder, the coordinates of the center position P1 (P5) of the shoulder touch sensor 722 (725) in which the robot is operating can be obtained by Ptouchsensor_shoulder = Pshoulder + Vsoulder.

  In addition, the center of each part (touch sensor center) point of the robot 10 can be calculated by solving the forward kinematic equation from each information of the current robot joints if the body center coordinates of the robot are known. is there.

  Next, in FIG. 10, the distance d between the point and the straight line is obtained by Equation 2. However, in FIG. 10 and Equation 2, the vector symbols are omitted for convenience.

  That is, a straight line indicating the line-of-sight direction is calculated from the normalized line-of-sight vector and the coordinates of the marker attached to the back position of the opponent (human) head, and the center coordinate positions P1-P8 of each touch sensor (FIG. 3). Distance D (n) is calculated from the above formula of the distance d between the point and the straight line. However, (n) is the number of the touch sensor 721-728, and is “1-8” in the embodiment.

  Considering the correspondence with Equation 2 in the specific example, the point p1 corresponds to the marker position Pcenter_head_back attached to the back of the opponent's head, the point p2 corresponds to the center position Ptouchsensor_n of each touch sensor 721-728, The normalized direction vector v (however, vector symbols are omitted for convenience) corresponds to (EyeVector / | EyeVector |). Then, the distance d is calculated according to Equation 2 and is set to D (n).

  In this way, the distance D (n) between the straight line in the line-of-sight direction and the center of each touch sensor 721-728 is obtained. Then, as shown in FIG. 11, a touch sensor whose distance D (n) is equal to or less than an arbitrary threshold value L is estimated as a touch sensor Te that actually has a contact act.

  Thereafter, in step S29, the CPU 50 determines whether or not the touch sensor Te estimated from the line-of-sight direction matches the contact reaction location detected in step S21. That is, it is determined whether the flag Ts_flag (Te) is “1”. However, (Te) is one of the numbers “1-8” of the estimated touch sensors 721-728.

  If “YES” is determined in the step S29, the CPU 50 identifies the touch sensor Te estimated from the line-of-sight direction as a touch sensor actually touched in a step S31, and returns.

  If “NO” in the step S29, that is, if the touch sensor Te estimated from the line-of-sight direction and the contact reaction portion detected in the step S21 do not coincide with each other, it is recognized that the determination is impossible in the step S33, and the process returns.

  If “NO” is determined in step S25, that is, if only one touch sensor has detected a contact reaction in step S21, the touch sensor actually detected by the touch sensor detected at that time is touched. It recognizes that it is a sensor (step S35), and returns.

  As described above, steps S25 to S31 and S35 function as second specifying means for specifying the actual contact site when the opponent person 80 actually touches the robot 10 somewhere.

  However, in this embodiment, even if a human does not actually touch, it senses the “sign” that the human is trying to touch, and causes a reaction behavior similar to the touch behavior when there is an actual touch. By doing so, we will try to obtain a communication robot system that is more biological. Therefore, even when there is no actual touch, the CPU 50 executes step S37 and subsequent steps to recognize whether or not there has been a human tactile action.

  In step S37 constituting the first determination means, the CPU 50 calculates the distance Db between the opponent person.

  First, the center coordinates of the bodies of the robot 10 and the human 80 are obtained. Specifically, the center coordinate Pbodycenter of the body is represented by the center position of each coordinate of the marker attached to both shoulders, as shown in the following equation (3).

  According to Equation 3, the center coordinates Probe_bodycenter and Human_bodycenter of the body of the robot 10 and the body of the person 80 are obtained. If the distance between the two points is Db, the distance Db is given by the following equation (4).

  Thereafter, in step S39 constituting the first determination means, the CPU 50 checks whether or not the robot 10 is clearly within the reach of the opponent's hand based on the length La of the opponent's human hand. As an example, if the robot 10 is within the range of twice the hand length La of the opponent 80, the opponent can touch the robot, and it is determined whether the distance Db is within the range of La * 2. However, such a threshold value (La * 2) can be changed as appropriate.

  Here, an example of how to obtain the length La of the opponent's human hand will be described. The length of the opponent's arm is calculated from the positions of the markers attached to the shoulder, elbow and hand of the opponent's body. That is, as shown in Equation 5, the distance from the elbow to the shoulder + the distance from the elbow to the hand = the arm length La.

  If “NO” is determined in the first determination means, that is, step S39, in the next step S41, it is recognized that there is no tactile action, and the process returns.

  If “YES” in the step S39, the process proceeds to a step S43 shown in FIG. Steps subsequent to step S43 constitute first specifying means. First, in step S43, the CPU 50 follows the equation (6) between the distance between the opponent's left hand and the center point Pn of each touch sensor of the robot body (left hand distance). ) Dlh (n) is calculated (n = 1-8), and the distance (right hand distance) Drh (n) between the opponent's right hand and the center point Pn of each touch sensor of the robot body is calculated (n = 1− 8) Do it.

  In the next step S45, the CPU 50 sets the smallest n among the left hand distances Dlh (n) smaller than the approach distance L (this is a variable threshold) as the left hand contact estimation part Tlh. 2 is written in the left-hand register Tlh formed in the memory 54 (If min {Dlh (n)} <L → Tlh = n). However, if there is no Dlh (n) closer than the approach distance L, “0” is written in the left-hand register Tlh (If min {Dlh (n)}> L → Tlh = 0).

  Furthermore, in the same step S45, the CPU 50 is formed in the memory 54 of FIG. 2 with n that is the smallest among the right hand distances Drh (n) smaller than the approach distance L (variable threshold) as the right hand contact estimation part Trh. Is written in the right-hand register Trh (If min {Drh (n)} <L → Trh = n). However, if there is no Drh (n) closer than the approach distance L, “0” is written in the right-hand register Trh (If min {Drh (n)}> L → Trh = 0).

  In the subsequent step S47, the CPU 50 determines whether or not there is a part of the robot that is likely to come into contact with the opponent's right hand and left hand. Specifically, it is checked whether both the left hand register Tlh and the right hand register Trh are “0”. The reason is that, as described above, the registers Tlh and Trh are set to “0” when there is no Dlh (n) and Drh (n) that are closer than the approach distance L, respectively. This is because the presence or absence of a part that is likely to come into contact is known.

  The determination of “NO” in step S47 means that at least the opponent's right hand or left hand is approaching somewhere in the robot. Therefore, in the next step S49, it is determined whether or not both Dlh and Drh are “0” (Dlh ≠ 0, Drh ≠ 0), that is, whether the other person's left hand and right hand are closer to the approach distance L. To do.

  If “YES” is determined in step S49, that is, if both the left hand and the right hand of the other person are closer than the approach distance L, the CPU 50 is about to touch the person (closer than the distance L). It is necessary to specify only one part of the robot. As the identification method, the human gaze direction is used as a clue for estimation as in step S27 described above. That is, in step S51, the CPU 50 calculates the direction of the other party's line of sight from the direction of the other person's face and the position of the eyes, and the person is not actually touching the robot. Based on this, the proximity part is estimated as the estimated contact touch sensor Te.

  As described above with reference to FIG. 5, the line-of-sight vector EyeVector representing the partner's line-of-sight direction can be calculated according to Equation 1 from the coordinates of the three markers Pride_head_front, Pleft_head_front, and Center_head_back attached to the head. Then, the touch sensor Te in the direction in which the line-of-sight vector EyeVector faces is specified. That is, it is determined which of the eight touch sensors 721-728 (FIG. 3) corresponds to the touch sensor in the direction of the line-of-sight vector EyeVector.

  Thereafter, in step S53, the CPU 50 determines whether or not the touch sensor Te estimated from the line-of-sight direction matches the touch sensor numbers registered in the left hand register Tlh and the right hand register Trh detected in step S45. That is, it is determined whether Te = Tlh or Te = Trh is satisfied.

  When they match, that is, when either Te = Tlh or Te = Trh is satisfied, in step S55, the CPU 50 determines the position of the touch sensor Te of the robot 10 to be touched by the human 80. Identify the location or location and return.

  If they do not match, the CPU 50 specifies a human hand closer to the robot 10 as the estimated contact portion in step S57. Specifically, the left hand distance Dlh and the right hand distance Drh obtained in step S43 are compared. When the left hand distance Dlh is smaller than the right hand distance Drh (Dlh <Drh), the CPU 50 sets the left hand contact part Tlh as the estimated contact part T. Identify and return. Conversely, when the left hand distance Dlh is larger than the right hand distance Drh (Dlh> Drh), the CPU 50 specifies the right hand contact part Trh as the estimated contact part T and returns.

  Furthermore, when “NO” is determined in the previous step S49, that is, when either the estimated contact portion Tlh or Trh is “0”, there is a possibility that only one of the hands may come into contact with the CPU 50. In step S59, one of the estimated contact sites Tlh or Trh is specified as the estimated contact site T, and the process returns.

  Finally, when “NO” is determined in step S47, there is no human hand to contact, and the estimated contact sites Tlh and Trh are both “0”. In this case, the CPU 50 determines in step S61. If there is no contact act, return.

  In this way, in step S3 of FIG. 6, it is possible to detect whether the human 80 is actually touching or trying to touch the robot 10 (presence or absence of tactile behavior).

  Then, as a result of executing the subroutine shown in FIG. 7 and FIG. 8 in step S3, when the tactile action by the human 80 is confirmed, “YES” is determined in step S5, and the CPU 50 performs the processing shown in FIG. In step S7, the robot 10 is commanded to react differently depending on the contact location or the estimated contact location. That is, step S7 functions as a reaction behavior generating unit. Table 1 shows an example of such reaction behavior. However, the reaction behaviors illustrated in Table 1 are basically reaction behaviors that do not like being touched by humans, but conversely, reaction behaviors that express a favorable feeling to humans or reactions in which they are mixed. Naturally, it is possible to set an action.

  As shown in Table 1, for example, when the contact part or the estimated contact part is the head, that is, when Te = T (1), the CPU 50 performs the motor control shown in FIG. The wheel motor 70 is controlled through the board 56.

  For example, when the contact site or the estimated contact site is the abdomen, that is, Te = T (8), the CPU 50 moves the robot 10 backward while guarding the abdomen with a hand so as not to touch the abdomen. In addition, the right arm motor 62, the left arm motor 64 and the wheel motor 70 are controlled through the motor control board 56.

  For example, when the contact site or the estimated contact site is the left shoulder, that is, Te = T (2), the CPU 50 passes through the motor control board 56 so as to rotate the robot 10 so that the left shoulder is kept away from the human. The waist motor 68 is controlled. When the contact part or the estimated contact part is the right shoulder, that is, when Te = T (5), the CPU 50 controls the waist motor 68 to rotate the robot 10 so that the right shoulder is kept away from the human. To do.

  For example, when the contact site or the estimated contact site is the left upper arm, that is, Te = T (3), the CPU 50 turns the left upper arm backward while rotating the robot 10 so as to keep the left upper arm away from the human. The lower motor 68 and the left arm motor 64 are controlled through the motor control board 56 so as to lower. When the contact part or the estimated contact part is the upper right arm, that is, Te = T (6), the CPU 50 lowers the left upper arm backward while rotating the robot 10 so as to keep the upper right arm away from the human. In addition, the waist motor 68 and the right arm motor 62 are controlled.

  For example, when the contact part or the estimated contact part is the left forearm, that is, Te = T (4), the CPU 50 turns the left forearm backward while rotating the robot 10 so as to keep the left forearm away from the human. The lower motor 68 and the left arm motor 64 are controlled through the motor control board 56 so as to lower. When the contact site or the estimated contact site is the upper right arm, that is, Te = T (7), the CPU 50 lowers the left forearm while rotating the robot 10 so as to keep the right forearm away from the human. In addition, the waist motor 68 and the right arm motor 62 are controlled.

  In any reaction behavior, the CPU 50 controls the head motor 66 and the like so that the eye camera 46 (FIG. 2) of the robot 10 approaches a human hand.

  As described above, in the passive mode, when it is detected whether the hand of the human 80 actually touches the body of the robot 10 or comes close enough to be touched, a predetermined reaction behavior is performed. However, in FIG. 6, when the human tactile action is not detected in step S5, the CPU 50 determines in the next step S9 whether or not it is in the standby state for a predetermined time or more. However, the standby state refers to a state where there is no sensor input from the sensor board 58 (FIG. 2), for example. When “NO” is determined in step S9, that is, when there is any sensor input within a predetermined time, the CPU 50 returns to the previous step S3 and detects the presence or absence of tactile behavior by the method described above. To do. If “YES” is determined in the step S9, the CPU 50 enters a standby state in a step S11. In the standby state, the CPU 50 performs predetermined processing such as maintaining the icon contact or maintaining the interpersonal distance with respect to the human 80 by the robot 10 (Idling Mode).

  However, if the standby state continues for a certain period of time or longer, the robot 10 is set to the active mode (Agent Mode), and for example, a self-acting action such as touching a human hand with its own hand (arm) is executed. You can also.

  In the embodiment described above, in order to detect the three-dimensional position coordinates of each part of a human or a robot, a motion capture method in which the marker 78 is photographed by the cameras 801-80n (FIG. 2) is employed. However, motion capture using other methods (for example, magnetism) may be used, and further, the state and position of a human or robot arm may be detected by general image processing instead of motion capture. . In the case of the robot 10, the three-dimensional position coordinates of each part may be specified by detecting each joint angle.

  In addition, in the embodiment, when there are two or more parts approaching within a predetermined distance by the first specifying means, is a human line of sight directed as a condition for specifying one part as an estimated contact part from the two parts? The condition of how was used. However, the first specifying means may adopt another condition, for example, specifying one estimated contact site on condition that a human hand is closer.

FIG. 1 is an illustrative view showing one example of a configuration of a communication robot according to one embodiment of the present invention. FIG. 2 is a block diagram showing an electrical configuration of the communication robot shown in FIG. FIG. 3 is an illustrative view illustrating the arrangement of touch sensors provided in the communication robot shown in FIG. FIG. 4 is an illustrative view illustrating the arrangement of markers attached to the communication robot shown in FIG. FIG. 5 is an illustrative view exemplifying an arrangement of markers attached to a person who cooperates with a communication robot. FIG. 6 is a flowchart showing the operation of this embodiment. FIG. 7 is a flowchart showing a part of the subroutine of FIG. FIG. 8 is a flowchart showing a part of a subroutine subsequent to FIG. FIG. 9 is an illustrative view for explaining a method of detecting a human gaze direction in the embodiment. FIG. 10 is an illustrative view for explaining a method of detecting a distance between a straight line in a human visual line direction and the center of the touch sensor in the embodiment. FIG. 11 is an illustrative view for explaining a method of specifying or estimating a touch location or part in the embodiment.

Explanation of symbols

10 ... Communication robot 50 ... CPU
72, 721-728 ... touch sensor 76 ... motion capture 821-82n ... motion capture (MC) camera 84 ... MC computer

Claims (6)

A communication robot that detects human tactile behavior based on sensor inputs from a plurality of touch sensors provided in a plurality of parts,
Touch determination means for determining whether a human has actually touched,
Presence detection means for detecting whether or not a human is within a predetermined range when the touch determination means does not determine that the user is actually touching; and that the human is within the predetermined range by the presence detection means A first specifying means for specifying a part where the human hand is approaching within a predetermined distance as an estimated contact part,
The first specifying means includes first determination means for determining whether or not there are a plurality of parts where the human hand approaches within the predetermined distance, and the first determination means allows the human hand to move to the predetermined distance. A communication robot that identifies one part that satisfies a predetermined condition as the estimated contact part when it is detected that there are a plurality of parts that are approaching within. Wherein the predetermined condition is whether the human gaze is directed, according to claim 1, wherein the communication robot. It said first specifying means, the human hand to identify the site as the estimated contact site when only one site approaching within the predetermined distance, according to claim 1 or 2, wherein the communication robot. The communication robot according to any one of claims 1 to 3 , further comprising reaction behavior generating means for generating a reaction behavior in accordance with the estimated contact site specified by the first specifying means. Second, further comprising a specifying means, communication robot according to any one of claims 1 to 4 to identify the actual contact portion of actual touch humans in response to sensor inputs from the touch sensor. The second specifying means includes second determination means for determining whether there is a sensor input from a plurality of touch sensors, and when the second determination means detects whether there is a sensor input from the plurality of touch sensors, the human being The communication robot according to claim 5 , wherein one actual contact site is specified according to the line-of-sight direction.

Images