robot

Abstract

To provide a multi-leg walking robot capable of causing the robot to advance (walk) by controlling the robot so as to alternatively perform grounded movement operation whin moves each of three pairs of legs mobile with respect to a body (first leg, second leg and third leg from the rear) and free leg movement operation which moves the tip in a state of isolating from the ground. and adaptively switching a gait.SOLUTION: A robot implements a first gait in which such a state that the number of legs performing grounded movement operation is three appears in a period and, in other periods, implements a second gait in which such a state that the number of legs performing grounded movement operation is four or more is maintained by lessening the rate of time during which a second leg and a third leg perform the free leg movement operation to time during which a first leg performs the free leg movement operation by comparing the case of the first gait.SELECTED DRAWING: Figure 6

Description

[Technical field]

[0001] The present invention relates to a robot, for example, a walking robot (legged mobile robot) that has multiple legs that are movable relative to a body. [Background technology]

[0002] A walking robot moves forward by controlling the positions of the tips of multiple legs relative to the torso. In general, a robot moves forward by alternately executing a contact movement motion and a swing leg movement motion. In a contact movement motion, the tip of a leg that is in contact with the ground (contact leg) is moved backward relative to the torso. In a swing leg movement motion, the tip of a leg whose tip is not in contact with the ground (swing leg) is moved forward relative to the torso and then touched down on the ground.

[0003] This type of walking robot divides the timing pattern for switching between the grounded leg and the swing leg of each of the multiple legs into multiple states, and can switch between each divided state. In other words, it is possible to adaptively switch the walking mode (gait) (for example, see Patent Document 1).

[0004] However, the robot in Patent Document 1 is premised on four-legged walking, and therefore does not disclose gait control for a walking robot having more than four legs, for example, six legs. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] JP 2001-246579 A Summary of the Invention [Problem to be solved by the invention]

[0006] Therefore, there has been a need to adaptively switch gait patterns in multi-legged robots with six or more legs. [Means for solving the problem]

[0007] According to one feature of the present disclosure, the robot has a trunk, left and right first legs movable relative to the trunk, left and right second legs movable in front of the left and right first legs relative to the trunk, left and right third legs movable in front of the left and right second legs relative to the trunk, and walking control means. The walking control means controls each of the first leg, the second leg, and the third leg relative to the trunk by each actuator.

[0008] The walking control means moves the torso forward relative to the ground by controlling the actuator to alternate between a "contact movement motion" in which the tip of the leg moves backward relative to the torso while in contact with the ground, and a "swing leg movement motion" in which the tip is clear of the ground and moves forward relative to the torso and contacts the ground.

[0009] Furthermore, the walking control means realizes a "first gait" in which a state in which the number of legs performing ground contact movement is three appears. In addition, the walking control means realizes a "second gait" in which a state in which the number of legs performing ground contact movement is four or more is maintained, by reducing the ratio of "front leg swing time," which is the time in which the second and third legs perform swing leg movement, to "rear leg swing time," which is the time in which the first leg performs a swing leg movement, compared to when the first gait is realized.

[0010] The walking control means starts the swing leg movement of the target leg when it is determined that the "swing leg condition" is satisfied for the target leg performing the ground contact movement. In the case where the target leg is the first leg, the swing leg condition is satisfied when the "center of gravity condition" is satisfied and the other first leg on the left or right is performing the ground contact movement. In addition, in the case where the target leg is the second or third leg, the swing leg condition is satisfied when the center of gravity condition is satisfied and the leg adjacent to the rear of the target leg (rear adjacent leg) and the other leg on the left or right of the target leg (left and right opposing legs) are both performing the ground contact movement. The center of gravity condition is satisfied when the center of mass of the trunk is included in a polygon having vertices at the contact points between the tip of the leg performing the ground contact movement other than the target leg and the ground and the ground, as viewed from above.

[0011] The robot has six or more legs, for example, a so-called six-legged walking robot. The six-legged walking robot has a left hind leg and a right hind leg as the first legs, a left middle leg and a right middle leg as the second legs, and a left front leg and a right front leg as the third legs. The center of gravity condition is a condition that is established when there is a high possibility that the robot will not fall over even if the target leg starts a swing leg movement motion (i.e., even if the tip of the target leg leaves the ground).

[0012] As can be understood from the swing leg condition of the first leg, the left and right first legs do not perform swing leg movement at the same time. That is, the left and right first legs perform swing leg movement alternately. For the swing leg condition of each of the left and right second legs to be established, it is necessary that the rear adjacent leg (i.e., the corresponding one of the left and right first legs) has started the ground contact movement (i.e., the swing leg movement has ended). Similarly, for the swing leg condition of each of the left and right third legs to be established, it is necessary that the swing leg movement of the rear adjacent leg (i.e., the corresponding one of the left and right second legs) has ended.

[0013] Therefore, for example, the swing leg movement of the left third leg (left hind leg) ends before the swing leg movement of the left second leg (left middle leg) adjacent to the front starts. Next, the swing leg movement of the left first leg (left front leg) adjacent to the front starts after the swing leg movement of the left second leg (left middle leg) ends.

[0014] If the ratio of the front leg swing time to the rear leg swing time is relatively small, the period (swing time) during which the left middle leg and the left front leg perform a swing leg movement motion is shortened. As a result, the swing leg time of the left front leg and the swing leg time of the left hind leg do not overlap each other. In other words, the swing leg time of the left front leg ends before the swing leg time of the left hind leg starts again. Therefore, the number of the left legs performing a swing leg movement motion does not become two or more. The same is true for the right legs (i.e., the right hind leg, the right middle leg, and the right front leg), so the number (total) of the left and right legs performing a swing leg movement motion is maintained at two or less. Therefore, in this case, the second gait is realized in which the state in which the number of legs performing a ground contact movement motion is maintained at four or more is realized.

[0015] On the other hand, when the ratio of the front leg swing time to the rear leg swing time increases, the swing time of the left middle leg and the left front leg becomes longer, and as a result, the swing time of the left front leg and the swing time of the left rear leg overlap each other. In other words, the swing time of the left rear leg starts again before the swing time of the left front leg ends. Therefore, a state occurs in which the number of the left legs performing the swing leg movement becomes two. When the number of the left legs performing the swing leg movement is two (i.e., when both the left rear leg and the left front leg are performing the swing leg movement), the right rear leg is not performing the swing leg movement, so the number of the right legs performing the swing leg movement does not become two (i.e., it is 0 or 1). Therefore, in this case, the first gait is realized in which the number of the legs performing the swing leg movement is three (i.e., the number of the legs performing the ground contact movement is three).

[0016] According to the second gait, the number of legs performing ground contact movement is greater than that of the first gait, so stable walking can be expected even when walking on sloped ground and uneven ground with an uneven surface. On the other hand, according to the first gait, the number of legs performing ground contact movement is smaller, so the control load on the walking control means is reduced compared to the second gait, and as a result, there is a possibility that the walking speed (forward speed) can be improved. Therefore, according to the robot (multi-legged walking robot) according to the present disclosure, it becomes possible to adaptively switch the gait between the first gait and the second gait.

[0017] According to another feature of the present disclosure, the walking control means stores one “proportionality coefficient.” In addition, the walking control means pre-stores a “reference rear leg swing time” for a rear leg swing time, a “reference first swing time” for a first gait, and a “reference second swing time” for a second gait that is shorter than the reference first swing time.

[0018] Furthermore, when realizing a first gait, the walking control means controls the actuators so that (1) the rear leg swing time is equal to the "target rear leg swing time" determined by multiplying the reference rear leg swing time by a proportionality coefficient, and (2) the front leg swing time is equal to the "target first swing time" determined by multiplying the reference first swing time by a proportionality coefficient. On the other hand, when realizing a second gait, the walking control means controls the actuators so that (1) the rear leg swing time is equal to the target rear leg swing time, and (2) the front leg swing time is equal to the "target second swing time" determined by multiplying the reference second swing time by a proportionality coefficient.

[0019] The desired rear leg swing time, the desired first swing leg time, and the desired second swing leg time are each determined in advance so that the first gait and the second gait are each realized, for example. In this case, the walking control means stores values ​​obtained by dividing these values ​​by the initial value of the proportionality coefficient as the reference rear leg swing time, the reference first swing leg time, and the reference second swing leg time. Therefore, the robot according to this feature can realize each of the first gait and the second gait by controlling the actuators based on the stored reference rear leg swing time, the reference first swing leg time, and the reference second swing leg time, as well as the proportionality coefficient.

[0020] In this feature, the walking control means may increase the proportionality coefficient when the rear leg swing time is longer than the target rear leg swing time or the front leg swing time is longer than the target first swing time during the first gait. On the other hand, the walking control means may increase the proportionality coefficient when the rear leg swing time is longer than the target rear leg swing time or the front leg swing time is longer than the target second swing time during the second gait.

[0021] In this case, if the swing leg time of any of the first to third legs becomes longer, the proportionality coefficient becomes larger, and thus the values ​​of the target rear leg swing time, the target first swing leg time, and the target second swing leg time become larger. That is, the swing leg times of all the first to third legs become similarly longer. An event in which the swing leg time becomes longer may be caused, for example, by deterioration of one of the actuators. Or, it may be caused by a slowing down of the movement speed of the leg performing the ground contact movement movement when the robot walks uphill or on an uneven ground. In such a case, the swing leg times of all the first to third legs become longer in accordance with an increase in the value of the proportionality coefficient, and as a result, it is possible to avoid the movement of only a specific leg becoming slow. Therefore, according to the robot according to this feature, there is a high possibility that stable walking can be maintained even if the swing leg time becomes longer for some reason.

[0022] According to another feature of the present disclosure, the walking control means determines that the center of gravity condition is established when the target leg is the left second leg and all of the "first leg group" consisting of the left first leg, the right second leg, and the left third leg are moving on the ground when the target leg is the right second leg and all of the "second leg group" consisting of the right first leg, the left second leg, and the right third leg are moving on the ground when the target leg is the right second leg and the center of gravity condition is established when the target leg is the right second leg and all of the "second leg group" consisting of the right first leg, the left second leg, and the right third leg are moving on the ground when the target leg is the right second leg.

[0023] Furthermore, the walking control means determines that the center of gravity condition is established when the target leg is to the left of the first leg, all of the second leg group are performing ground contact movement, and when the right of the first leg, the left and right second legs, and the left of the third leg are all performing ground contact movement. In addition, the walking control means determines that the center of gravity condition is established when the target leg is to the right of the first leg, all of the first leg group are performing ground contact movement, and when the left of the first leg, the left and right second legs, and the right of the third leg are all performing ground contact movement.

[0024] Furthermore, the walking control means determines that the center of gravity condition is established when the target leg is to the left of the third leg, all of the second leg group are performing ground contact movement, and when the left of the first leg, the left and right second legs, and the right of the third leg are all performing ground contact movement. In addition, the walking control means determines that the center of gravity condition is established when the target leg is to the right of the third leg, all of the first leg group are performing ground contact movement, and when the right of the first leg, the left and right second legs, and the left of the third leg are all performing ground contact movement.

[0025] In other words, if all of the "opposing three legs" that do not include the target leg among the first leg group and the second leg group are moving on the ground, it is determined that the center of gravity condition is met. In addition, if all of the legs (opposing four legs) other than the "target leg that is one of the first leg and the third leg" and the "diagonal leg of the target leg" among the first leg to the third leg are moving on the ground, it is determined that the center of gravity condition is met. The diagonal leg is the leg "on the left-right opposite side of the trunk of the target leg" and the leg "on the front-rear opposite side of the trunk of the target leg" (i.e., the other one of the first leg and the third leg). For example, if the target leg is the above-mentioned left front leg, the diagonal leg of the target leg is the right rear leg.

[0026] For example, if the base of the first leg (i.e., the end opposite to the tip) is fixed near the rear end of the body and the base of the third leg is fixed near the front end of the body, when one of these conditions is met, there is a high possibility that the center of gravity condition is actually met. Therefore, with the robot according to this feature, it is possible to determine whether the center of gravity condition is met by simple processing.

[0027] According to another feature of the present disclosure, the walking control means executes at least one of: (1) control to switch from a state in which a first gait is realized to a state in which a second gait is realized when at least one of the legs is performing a ground contact movement; and (2) control to switch from a state in which the second gait is realized to a state in which the first gait is realized when at least one of the legs is performing a ground contact movement.

[0028] If at least one of the first to third legs is performing a ground contact movement motion, there is a high possibility that the torso is moving forward. In addition, as described above, the gait can be switched between the first and second gaits by changing the front leg swing time. Therefore, according to the robot of this feature, the gait can be switched without interrupting walking.

[0029] According to another feature of the present disclosure, the walking control means obtains, as a target position of a contact point between the tip of the second leg and the ground, a position that is the same as or close to a contact point between the tip of the third leg adjacent to the front and the ground. In addition, the walking control means obtains, as a target position of a contact point between the tip of the first leg and the ground, a position that is the same as or close to a contact point between the tip of the second leg adjacent to the front and the ground.

[0030] In other words, when the contact points (target ground contact positions) between the tips of the left and right third legs and the ground are determined, the contact points between the tips of the second and first legs and the ground are also determined. Therefore, with the robot according to this feature, it is possible to determine the contact points between the tips of the first to third legs and the ground by simple processing. [Brief description of the drawings]

[0031] [Figure 1] FIG. 1 is a perspective view of a robot according to an embodiment of the present disclosure. [Diagram 2] FIG. 2 is a block diagram of a robot and a walking operation terminal for the robot. [Diagram 3] FIG. 2 is a schematic plan view showing the reachable area of ​​the robot's legs. [Figure 4] FIG. 2 is a simplified diagram showing the links and rotary joints of the legs of the robot. [Diagram 5] FIG. 1 is a schematic plan view showing a robot walking. [Figure 6] 11 is a time chart showing the transition of the control mode of each leg when the robot moves forward in a first gait. [Figure 7] 13 is a time chart showing the transition of the control mode of each leg when the robot moves forward in the second gait. [Figure 8] 13 is a time chart showing the transition of the control mode of each leg when the robot moves forward with another gait. [Figure 9] 1 is a time chart showing the transition of the control mode of each leg when the robot's gait changes from a first gait to a second gait. [Figure 10] 11 is a time chart showing the transition of the control mode of each leg when the robot's gait changes from the second gait to the first gait. [Figure 11] 13 is a time chart showing the transition of the control mode of each leg when a swing leg time exceedance occurs. [Figure 12] 13 is a flowchart showing a walking control processing routine executed by a control device of the robot. [Figure 13] 13 is a flowchart showing a leg state management processing routine executed by a control device of the robot. [Figure 14] 11 is a time chart showing an example of transition of the control mode of each leg when the robot is an eight-legged walking robot. [Figure 15] 13 is a time chart showing another example of the transition of the control mode of each leg when the robot is an eight-legged walking robot. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0032] A preferred embodiment of the present invention will be described in detail below with reference to Figs. 1 to 13. The same reference numbers in the description refer to the same elements having the same functions, without duplicate description. In this embodiment, the robot 1 shown in Fig. 1 includes a body 2, and a left front leg 3, a left middle leg 4, a left hind leg 5, a right front leg 6, a right middle leg 7, and a right hind leg 8, which are movable members movable relative to the body 2. That is, the robot 1 is a six-legged walking robot.

[0033] The left hind leg 5 and the right hind leg 8 are also referred to as the "first leg" for convenience. The left middle leg 4 and the right middle leg 7 are also referred to as the "second leg" for convenience. The left front leg 3 and the right front leg 6 are also referred to as the "third leg" for convenience. In the following explanation, a trunk coordinate system (trunk coordinate values) is used. The trunk coordinate system is an orthogonal coordinate system in which the geometric center of the trunk 2 is the origin O, the front-to-back direction with respect to the trunk 2 is the x-axis, the left-to-right direction is the y-axis, and the up-down direction is the z-axis.

[0034] As can be seen from FIG. 2, which is a block diagram of the robot 1, the robot 1 includes a control device 9, an optical camera 10, an IMU 11, and a wireless communication device 12 (12a).

[0035] The control device 9 mainly includes a microcomputer equipped with a CPU, RAM, and non-volatile memory. The CPU sequentially executes a predetermined program (routine) to read data, perform numerical calculations, and output the calculation results. The RAM temporarily stores data referenced by the CPU. The non-volatile memory is composed of a ROM and a rewritable flash memory, and stores the programs executed by the CPU and look-up tables (maps) referenced when the programs are executed.

[0036] The optical camera 10 is disposed at the front end protrusion of the fuselage 2 (see FIG. 3). The optical camera 10 acquires (takes) an image (forward image) of the area in front of the fuselage 2, and transmits a signal representing the forward image to the control device 9. The IMU 11 detects x-axis acceleration Ax, y-axis acceleration Ay, and z-axis acceleration Az, which are accelerations in the directions of the x-axis, y-axis, and z-axis, respectively, and x-axis angular velocity ωx, y-axis angular velocity ωy, and z-axis angular velocity ωz, which are rotational velocities (angular velocities) in the directions of the x-axis, y-axis, and z-axis, respectively. The combination of these values ​​detected by the IMU 11 is also referred to as a fuselage movement value M (Ax, Ay, Az, ωx, ωy, ωz). The IMU 11 is configured by a known inertial measurement device.

[0037] The wireless communication device 12 (12a) performs data communication with a wireless communication device 12 (12b) included in the walking operation terminal 30 described later. Specifically, the wireless communication device 12 (12a) transmits information (data) input from the control device 9 to the wireless communication device 12 (12b), and outputs information received from the wireless communication device 12 (12b) to the control device 9.

[0038] As can be seen from Fig. 4, which is a simplified diagram of the left front leg 3, the left front leg 3 includes a first link 3a, a second link 3b, a third link 3c, and a base 3d. One end of the base 3d is fixed to the body 2. The other end of the base 3d is connected to one end of the first link 3a via a rotary joint (first joint). The other end of the first link 3a is connected to one end of the second link 3b via a rotary joint (second joint). The other end of the second link 3b is connected to one end of the third link 3c via a rotary joint (third joint). The other end of the third link 3c (i.e., the tip of the left front leg 3) is adapted to come into contact with the ground (ground surface).

[0039] The angle between the rotation axis of the first joint and a plane perpendicular to the z-axis in the trunk coordinate system (for example, a plane represented by z=0, hereinafter also referred to as the "horizontal trunk plane") is 45°. In top view, the rotation axis of the first joint extends diagonally forward from the trunk 2, and the angle between the rotation axis of the first joint and a plane perpendicular to the x-axis in the trunk coordinate system (for example, a plane represented by x=0, hereinafter also referred to as the "vertical trunk plane") is 30°. The rotation axis of the second joint and the rotation axis of the third joint are each parallel to the horizontal trunk plane. The rotation axis of the first joint and the rotation axis of the second joint are perpendicular to each other. The rotation axis of the second joint and the rotation axis of the third joint are parallel to each other.

[0040] The rotation angle α, which is the rotation angle of the first joint (i.e., the rotation angle of the first link 3a relative to the base 3d), is controlled by the torque generated by the motor 20 (motor 20a) (see FIG. 2). The motor 20 is configured by a well-known DC servo motor. The rotation angle β, which is the rotation angle of the second joint (i.e., the rotation angle of the second link 3b relative to the first link 3a), is controlled by the torque generated by the motor 20 (motor 20b). The rotation angle γ, which is the rotation angle of the third joint (i.e., the rotation angle of the third link 3c relative to the second link 3b), is controlled by the torque generated by the motor 20 (motor 20c). For convenience, each of the motors 20 is also referred to as an "actuator."

[0041] The angle sensor 21 (angle sensor 21a) detects a rotation angle α of a first joint of the left front leg 3, and outputs a signal representing the rotation angle α to the control device 9. The angle sensor 21 is configured by a well-known resolver device. The angle sensor 21 (angle sensor 21b) detects a rotation angle β of a second joint of the left front leg 3, and outputs a signal representing the rotation angle β to the control device 9. The angle sensor 21 (angle sensor 21c) detects a rotation angle γ of a third joint of the left front leg 3, and outputs a signal representing the rotation angle γ to the control device 9.

[0042] Hereinafter, the combination of rotation angle α, rotation angle β, and rotation angle γ is also referred to as rotation angle L(α,β,γ). Based on the rotation angle L, the lengths of the first link 3a to the third link 3c, and the angle between the rotation axis of the first joint and the vertical plane of the torso, the control device 9 acquires (calculates) the leg tip position F(xf,yf,zf), which is the torso coordinate of the leg tip of the left front leg 3, by a well-known method.

[0043] The area on the ground that can be reached by the tip of the left front leg 3 when the robot 1 is in the "reference state" (i.e., the range to which the tip of the left front leg 3 can move) is shown in a simplified form by area R3 in Fig. 3. The reference state is a state in which the robot 1 is walking on a flat and level ground, the torso 2 is horizontal to the ground, and the distance between the ground and the center of gravity Gb of the torso is a predetermined reference distance Ds. The center of gravity Gb of the torso is the center of mass (center of mass) of the torso 2, and in this embodiment, is approximately equal to the geometric center of the torso 2 (i.e., the origin O in the torso coordinate system).

[0044] Each of the left middle leg 4, the left hind leg 5, the right front leg 6, the right middle leg 7, and the right hind leg 8 shown in FIG. 1 has a configuration similar to that of the left front leg 3. The left middle leg 4 includes a first link 4a, a second link 4b, and a third link 4c. The rotation axis of the first joint of the left middle leg 4 extends directly to the side from the torso 2. That is, the rotation axis of the first joint of the left middle leg 4 is parallel to the vertical plane of the torso. The control device 9 controls the motor 20 (motor 20d, motor 20e, and motor 20f) provided in the left middle leg 4 to control the rotation angle L (α, β, γ) of the left middle leg 4. In addition, the control device 9 acquires the foot tip position F (xf, yf, zf) of the left middle leg 4 based on the actual rotation angle L (α, β, γ) detected by the angle sensor 21 (angle sensor 21d, angle sensor 21e, and angle sensor 21f). The reachable area on the ground of the tip of the left middle leg 4 when the robot 1 is in the reference state is shown in simplified form by area R4 in FIG.

[0045] The left hind leg 5 includes a first link 5a, a second link 5b, and a third link 5c. The rotation axis of the first joint of the left hind leg 5 extends diagonally backward in a top view, and the angle between the rotation axis of the first joint of the left hind leg 5 and the vertical plane of the torso is 30°. The control device 9 controls the motor 20 (motor 20g, motor 20h, and motor 20i) equipped in the left hind leg 5 to control the rotation angle L (α, β, γ) of the left hind leg 5. In addition, the control device 9 obtains the foot tip position F (xf, yf, zf) of the left hind leg 5 based on the actual rotation angle L (α, β, γ) detected by the angle sensor 21 (angle sensor 21g, angle sensor 21h, and angle sensor 21i). The reachable area on the ground of the foot tip of the left hind leg 5 when the robot 1 is in the reference state is shown in a simplified form by an area R5.

[0046] The right front leg 6 includes a first link 6a, a second link 6b, and a third link 6c. The rotation axis of the first joint of the right front leg 6 extends diagonally forward in a top view, and the angle between the rotation axis of the first joint of the right front leg 6 and the vertical plane of the torso is 30°. The control device 9 controls the motor 20 (motor 20j, motor 20k, and motor 20m) equipped in the right front leg 6 to control the rotation angle L (α, β, γ) of the right front leg 6. In addition, the control device 9 obtains the foot tip position F (xf, yf, zf) of the right front leg 6 based on the actual rotation angle L (α, β, γ) detected by the angle sensor 21 (angle sensor 21j, angle sensor 21k, and angle sensor 21m). The reachable area on the ground of the foot tip of the right front leg 6 when the robot 1 is in the reference state is shown in a simplified form by an area R6.

[0047] The right middle leg 7 includes a first link 7a, a second link 7b, and a third link 7c. The rotation axis of the first joint of the right middle leg 7 extends directly to the side from the torso 2. That is, the rotation axis of the first joint of the right middle leg 7 is parallel to the vertical plane of the torso. The control device 9 controls the motor 20 (motor 20n, motor 20p, and motor 20q) of the right middle leg 7 to control the rotation angle L (α, β, γ) of the right middle leg 7. In addition, the control device 9 obtains the foot tip position F (xf, yf, zf) of the right middle leg 7 based on the actual rotation angle L (α, β, γ) detected by the angle sensor 21 (angle sensor 21n, angle sensor 21p, and angle sensor 21q). The reachable area on the ground of the foot tip of the right middle leg 7 when the robot 1 is in the reference state is shown in a simplified form by an area R7.

[0048] The right hind leg 8 includes a first link 8a, a second link 8b, and a third link 8c. The rotation axis of the first joint of the right hind leg 8 extends diagonally backward in a top view, and the angle between the rotation axis of the first joint of the right hind leg 8 and the vertical plane of the torso is 30°. The control device 9 controls the motor 20 (motor 20r, motor 20s, and motor 20t) equipped in the right hind leg 8 to control the rotation angle L (α, β, γ) of the right hind leg 8. In addition, the control device 9 obtains the foot tip position F (xf, yf, zf) of the right hind leg 8 based on the actual rotation angle L (α, β, γ) detected by the angle sensor 21 (angle sensor 21r, angle sensor 21s, and angle sensor 21t). The reachable area on the ground of the foot tip of the right hind leg 8 when the robot 1 is in the reference state is shown in a simplified form by an area R8.

[0049] The walking operation terminal 30 includes a wireless communication device 12 (12b), an operation control device 31, a display 32, and an input device 33. The operation control device 31 is configured by a general-purpose computer including a CPU, RAM, and non-volatile memory. Images displayed on the display 32 are controlled by the operation control device 31. The input device 33 includes a keyboard and a mouse. The input device 33 outputs a signal representing an operation performed by the operator on the input device 33 to the operation control device 31.

[0050] When the robot 1 is walking, the operation control device 31 displays an image showing a forward image captured by the optical camera 10 on the display 32. While viewing the image displayed on the display 32, the operator determines the direction of travel of the robot 1 (either going straight or turning left or right) and, if turning the robot 1, a value correlating to the turning radius (turning radius correlation value), and inputs the information into the operation control device 31 using the input device 33. The operation control device 31 transmits the input information to the control device 9 via the wireless communication devices 12 (12a and 12b).

[0051] In addition, the operator can switch the gait of the robot 1 between a "first gait" and a "second gait" described below by operating the input device 33. When an operation to switch the gait is performed by the operator, the operation control device 31 transmits a message to that effect to the control device 9. Hereinafter, information transmitted to the control device 9 to switch the gait is also referred to as a "gait change request."

[0052] The control device 9 determines a target ground contact position of the "next ground contact leg" so that the robot walks in the traveling direction input (requested) by the operator. The next ground contact leg is one of the left front leg 3 and the right front leg 6, and is the leg whose tip will next touch (contact) the ground. The control device 9 controls the rotation angles L (α, β, γ) of the left front leg 3 to the right hind leg 8 so that the tip (tip) of the next ground contact leg touches (reaches) the target ground contact position, while determining a target ground contact position for the new next ground contact leg (i.e., the other of the left front leg 3 and the right front leg 6).

[0053] In addition, the control device 9 determines the target ground contact positions of the legs other than the left front leg 3 and the right front leg 6 (i.e., the left middle leg 4, the left hind leg 5, the right middle leg 7, and the right hind leg 8, hereinafter collectively referred to as "following legs") by a "ground contact point following method". The ground contact point following method is a method for acquiring (determining) a position near the ground contact position of the following leg and the leg adjacent to the front of the torso 2 as the target ground contact position of the following leg.

[0054] The touchdown point tracking method will be described with reference to FIG. 5. In FIG. 5, the robot 1 moves forward in a first gait. In FIG. 5, the black circles indicate the touchdown positions immediately after the leg tips touch the ground, and the white circles indicate the touchdown positions immediately before the leg tips leave the ground. In other words, the leg tip positions indicated by the black circles at a certain point in time then move backward with respect to the torso 2 and change to the leg tip positions indicated by the white circles. In the following description, the left front leg 3, the right middle leg 7, and the left hind leg 5 are also collectively referred to as the "first leg group." In addition, the right front leg 6, the left middle leg 4, and the right hind leg 8 are also collectively referred to as the "second leg group."

[0055] When the trunk 2 is at trunk position 2a1, the respective leg tips of the first leg group are indicated by white circles, points PL11, PR21, and PL31, while the respective leg tips of the second leg group are indicated by black circles, points PR11, PL21, and PR31.

[0056] Point PL21, which is the ground contact point of the left middle leg 4, is located adjacent to the rear side of point PL11, which is the ground contact point of the left front leg 3. In addition, point PR31, which is the ground contact point of the right hind leg 8, is located adjacent to the rear side of point PR21, which is the ground contact point of the right middle leg 7. In other words, the ground contact points of the trailing legs, the left middle leg 4 and the right hind leg 8, are determined by the ground contact point following method.

[0057] In other words, after the tip of the left front leg 3 moves relative to the torso 2 and reaches the reachable area of ​​the left middle leg 4 (for example, the area where the tip of the left front leg 3 overlaps with area R3 and area R4 in FIG. 3), the left middle leg 4 touches the ground. Similarly, after the tip of the right middle leg 7 moves relative to the torso 2 and reaches the reachable area of ​​the right hind leg 8 (for example, the area where area R7 and area R8 overlap), the right hind leg 8 touches the ground.

[0058] Thereafter, the tips of the second set of legs move backward relative to the torso 2, and further, the position of the torso 2 in which the tips of the first set of legs have newly touched the ground is indicated by torso position 2a2. As a result of the tips of the second set of legs moving backward relative to the torso 2, the torso 2 moves (forward) from the position indicated by torso position 2a1 to the position indicated by torso position 2a2.

[0059] At this time, the tips of the second group of legs are indicated by white circles, points PR11, PL21, and PR31, while the tips of the first group of legs are indicated by black circles, points PL12, PR22, and PL32, which are new ground contact points.

[0060] Point PR22, which is the ground contact point of the right middle leg 7, is located adjacent to the rear side of point PR11, which is the ground contact point of the right front leg 6. In addition, point PL32, which is the ground contact point of the left hind leg 5, is located adjacent to the rear side of point PL21, which is the ground contact point of the left middle leg 4. In other words, the ground contact points of the right middle leg 7 and the left hind leg 5, which are the trailing legs, are determined by the ground contact point following method.

[0061] Furthermore, the leg tips of the first leg group move backward relative to the torso 2, and the position of the torso 2 in which the leg tips of the second leg group have newly touched the ground is indicated by torso position 2a3. As a result of the leg tips of the first leg group moving backward relative to the torso 2, the torso 2 moves from the position indicated by torso position 2a2 to the position indicated by torso position 2a3.

[0062] At this time, the tips of the first group of legs are indicated by white circles at points PL12, PR22, and PL32, while the tips of the second group of legs are indicated by black circles at points PR12, PL22, and PR32, which are new ground contact points.

[0063] Point PL22, which is the ground contact point of the left middle leg 4, is located adjacent to the rear side of point PL12, which is the ground contact point of the left front leg 3. In addition, point PR32, which is the ground contact point of the right hind leg 8, is located adjacent to the rear side of point PR22, which is the ground contact point of the right middle leg 7. In other words, the ground contact points of the left middle leg 4 and right hind leg 8, which are the trailing legs, are determined by the ground contact point following method.

[0064] The walking control of the robot 1 by the control device 9 will be described in more detail. When the robot 1 is walking, the control device 9 manages the control mode (control state) of each of the left front leg 3 to the right hind leg 8. The control modes are classified into a first control mode to a fourth control mode.

[0065] The first control mode is a control state in which the leg tip that is on the ground is lifted off the ground and moved to a predetermined leg tip standby position Ls. The combination of rotation angles L(α, β, γ) when the leg tip is at the leg tip standby position Ls is represented by the leg tip standby position Ls(αs, βs, γs). The second control mode is a control state in which the leg tip waits in the leg tip standby position Ls.

[0066] The third control mode is a control state in which the leg tip is moved from the leg tip standby position Ls to the target ground contact position. Hereinafter, the leg in the first to third control modes is also referred to as the "swing leg." For convenience, the action of lifting the leg tip of the swing leg off the ground, moving it forward relative to the torso 2, and then touching it to the ground is also referred to as the "swing leg movement action."

[0067] The fourth control mode is a control state in which the leg tips are in contact with the ground and are moved from front to rear relative to the torso 2. In other words, the torso 2 moves forward by moving the leg tips of the legs in the fourth control mode among the left front leg 3 to the right hind leg 8 (hereinafter also referred to as "grounded legs"). For convenience, the action of moving the leg tips of the grounded legs (which are in contact with the ground) backward relative to the torso 2 is also referred to as the "grounded movement action."

[0068] The control device 9 executes a "leg state management process" at every predetermined control period Δt in order to manage the control mode of each of the left front leg 3 to the right hind leg 8. When executing the leg state management process, the control device 9 determines whether or not a transition condition, which will be described later, is satisfied for each of the left front leg 3 to the right hind leg 8, and changes the control mode of the leg for which the transition condition is satisfied.

[0069] The condition for switching the control mode from the first control mode to the second control mode (hereinafter also referred to as the "second transition condition") is met when the leg tip reaches the leg tip waiting position Ls (αs, βs, γs). The condition for switching the control mode of the left front leg 3 and the right front leg 6 from the second control mode to the third control mode (hereinafter also referred to as the "third transition condition") is met when the target ground contact position has been determined and is included in the area (area on the ground) that the leg tips can reach. On the other hand, the third control transition condition for the following leg is met when the leg tip of the leg adjacent to the front of the torso 2 is included in the area that the leg can reach (in this case, the three-dimensional area above the ground).

[0070] The condition for switching the control mode from the third control mode to the fourth control mode (hereinafter also referred to as the "fourth transition condition") is met when the leg tip reaches the target ground contact position. The condition for switching the control mode from the fourth control mode to the first control mode is hereinafter also referred to as the "first transition condition" or the "swing leg condition."

[0071] If the leg for which it is determined whether the swing leg condition is satisfied is referred to as the "target leg," then the swing leg condition when one of the left hind leg 5 and the right hind leg 8 is the target leg is satisfied when the other of the left hind leg 5 and the right hind leg 8 (i.e., the leg on the opposite side to the left or right of the torso 2) is a grounded leg and the "center of gravity condition" described below is satisfied. When any of the left front leg 3, left middle leg 4, right front leg 6, and right middle leg 7 is the target leg, the swing leg condition is satisfied when the leg on the opposite side to the left or right of the torso 2 and the leg adjacent to the rear of the torso 2 are both grounded legs and the center of gravity condition is satisfied.

[0072] Hereinafter, the legs on the left and right opposite sides of the torso 2 are also referred to as "left and right opposing legs." In addition, the leg adjacent to the rear of the torso 2 is also referred to as the "rear adjacent leg."

[0073] The center of gravity condition is a condition that is met when the trunk center of gravity Gb is included in a "support polygon" in a top view. The support polygon is a polygon whose vertices are the contact points between the grounded legs other than the target leg and the ground. The control device 9 determines that the center of gravity condition is met when any of the following conditions (a1) and (a2) and conditions (b1) to (b4) is met.

[0074] Condition (a1): When the target leg is included in the first leg group, all of the legs in the second leg group are grounded legs. Condition (a2): When the target leg is included in the second leg group, all of the first leg group are grounded legs.

[0075] Condition (b1): When the target leg is the left front leg 3, the left middle leg 4, the left hind leg 5, the right front leg 6, and the right middle leg 7 are all grounded legs. Condition (b2): When the target leg is the right front leg 6, the left front leg 3, the left middle leg 4, the right middle leg 7 and the right hind leg 8 are all grounded legs. Condition (b3): ​​When the target leg is the left hind leg 5, the left front leg 3, the left middle leg 4, the right middle leg 7 and the right hind leg 8 are all grounded legs. Condition (b4): When the target leg is the right hind leg 8, the left middle leg 4, the left hind leg 5, the right front leg 6 and the right middle leg 7 are all grounded legs.

[0076] The swing leg condition will be described more specifically with reference to Fig. 6. Fig. 6 shows an example in which the control modes of the left front leg 3 to the right hind leg 8 of the robot 1 change over time while the robot 1 is walking.

[0077] First, it is examined whether or not the swing leg condition of the "target leg, the right front leg 6" is satisfied at time tb1. The trunk 2 at time tb1 is indicated by a trunk position 2b1. In Fig. 6 (and Fig. 7 described later), the tip of the grounded leg (foot tip) is indicated by a black circle, and the tip of the swing leg is indicated by a white circle.

[0078] At this point in time, the legs on the ground other than the target leg are the left front leg 3, the left middle leg 4, the right middle leg 7, and the right hind leg 8. Since the "left front leg 3, which is the left-right opposing leg" and the "right middle leg 7, which is the adjacent rear leg" of the target leg (right front leg 6) are both on the ground, if the center of gravity condition is also satisfied, the swing leg condition of the target leg (right front leg 6) is satisfied.

[0079] The target leg (right front leg 6) is included in the second leg group, but the left hind leg 5, which is included in the first leg group, is not a grounded leg, so condition (a2) is not satisfied. On the other hand, condition (b2) is satisfied in this case. That is, the center of gravity condition is satisfied, so the swing leg condition of the target leg (right front leg 6) is satisfied. Since the swing leg condition of the right front leg 6 is satisfied, the control device 9 transitions the right front leg 6 from the fourth control mode to the first control mode. That is, the right front leg 6 changes from a grounded leg to a swing leg.

[0080] The support polygon in this example is indicated by a rectangle Sa. As can be seen from the trunk 2 (and the state of each leg) at the trunk position 2b1, the trunk center of gravity Gb is included in the support polygon (rectangle Sa) in the top view. In other words, it can be seen that the center of gravity condition is actually satisfied.

[0081] Next, it is considered whether or not the swing leg condition of the "left middle leg 4 and right hind leg 8, which are the target legs," is satisfied at time tb2. The torso 2 at time tb2 is indicated by the torso position 2b2. At this time, the grounded legs other than the target leg are the left front leg 3, the left hind leg 5, and the right middle leg 7. The "right middle leg 7, which is the left-right opposing leg" and the "left hind leg 5, which is the rear adjacent leg" of the "left middle leg 4, which is the target leg," are both grounded legs. On the other hand, the "left hind leg 5, which is the left-right opposing leg" of the "right hind leg 8, which is the target leg," is a grounded leg. Therefore, if the center of gravity condition is also satisfied, the swing leg condition of the target legs (left middle leg 4 and right hind leg 8) is satisfied.

[0082] The target legs (left middle leg 4 and right hind leg 8) are included in the second leg group, while all legs included in the first leg group are grounded legs, so condition (a2) is met. That is, the center of gravity condition is met. Therefore, the swing leg condition of the target legs (left middle leg 4 and right hind leg 8) is met, so the control device 9 transitions the left middle leg 4 and right hind leg 8 from the fourth control mode to the first control mode.

[0083] The support polygon in this example is indicated by triangle Sb. As can be seen from the trunk 2 (and the state of each leg) at trunk position 2b2, the trunk center of gravity Gb is included in the support polygon (triangle Sb) in top view. In other words, it can be seen that the center of gravity condition is actually satisfied. As can be seen from these examples of the support polygon (i.e., quadrangle Sa and triangle Sb), if any of conditions (a1) and (a2) and conditions (b1) to (b4) are satisfied, the center of gravity condition is satisfied in many cases.

[0084] 6, the robot 1 is moving forward (walking) in the first gait. When the robot 1 walks in the first gait, a state in which the number of legs on the ground is "3" (specifically, one of a state in which only the first leg group has legs on the ground and a state in which only the second leg group has legs on the ground) occurs.

[0085] For example, as can be seen from the trunk 2 (and the state of each leg) at time tb3 indicated by trunk position 2b3, the first leg group is the grounded leg and the second leg group is the free leg at this time. On the other hand, as can be seen from the trunk 2 (and the state of each leg) at time tb4 indicated by trunk position 2b4, the second leg group is the grounded leg and the first leg group is the free leg at this time. For this reason, the first gait is also called a "three-legged gait".

[0086] 7 shows an example of transition of the control modes of the left front leg 3 to the right hind leg 8 when the robot 1 moves forward in the second gait. When the robot 1 moves forward in the second gait, the number of legs on the ground is maintained at "4" or more.

[0087] For example, as understood from the body 2 (and the states of each leg) at the time tc1 indicated by the body position 2c1, at this point, the left front leg 3, the left rear leg 5, the right front leg 6, and the right middle leg 7 are the grounded legs. On the other hand, as understood from the body 2 (and the states of each leg) at the time tc2 indicated by the body position 2c2, at this point, the left front leg 3, the left middle leg 4, the right middle leg 7, and the right rear leg 8 are the grounded legs.

[0088] The control of the walking posture and switching of the robot 1 by the control device 9 will be described. In the following description, the length of the period during which each of the left rear leg 5 and the right rear leg 8 performs the floating leg movement operation (that is, the time from the transition from the fourth control mode to the first control mode until the transition from the third control mode to the fourth control mode) is, for convenience, also referred to as the "rear leg floating leg time". The length of the period during which each of the left front leg 3, the left middle leg 4, the right front leg 6, and the right middle leg 7 performs the floating leg movement operation is, for convenience, also referred to as the "front leg floating leg time".

[0089] When the control device 9 realizes the first walking posture, if "floating leg time exceedance" described later does not occur, the control device 9 controls the tip position F of the floating leg (that is, the rotation angle L) so that the rear leg floating leg time becomes equal to the predetermined reference rear leg floating leg time Tr and the front leg floating leg time becomes equal to the predetermined reference first floating leg time Tf1 (see FIG. 6). On the other hand, when the control device 9 realizes the second walking posture, if floating leg time exceedance does not occur, the control device 9 controls the rotation angle L of the floating leg so that the rear leg floating leg time becomes equal to the reference rear leg floating leg time Tr and the front leg floating leg time becomes equal to a predetermined reference second floating leg time Tf2 (that is, 0 < Tf2 < Tf1) that is smaller than the reference first floating leg time Tf1 (see FIG. 7).

[0090] An example of the transition of the control modes of each of the left front leg 3 to the right rear leg 8 when the front leg floating leg time becomes equal to a time Tf3 (that is, Tf2 < Tf3 < Tf1) that is smaller than the reference first floating leg time Tf1 and larger than the reference second floating leg time Tf2 is shown in FIG. 8. In the example of FIG. 8, the rear leg floating leg time is equal to the reference rear leg floating leg time Tr.

[0091] As can be seen from FIG. 8, a portion of the period (swing time) when the left front leg 3 is in swing overlaps with the swing time of the left hind leg 5. Similarly, a portion of the swing time of the right front leg 6 overlaps with the swing time of the right hind leg 8. Meanwhile, in the example of FIG. 6, much of the swing time of the left front leg 3 overlaps with the swing time of the left hind leg 5. Similarly, much of the swing time of the right front leg 6 overlaps with the swing time of the right hind leg 8. In the example of FIG. 7, the swing time of the left front leg 3 does not overlap with the swing time of the left hind leg 5. Similarly, the swing time of the right front leg 6 does not overlap with the swing time of the right hind leg 8.

[0092] As can be seen from these examples, the greater the ratio of the front leg swing time to the rear leg swing time, the greater the overlap rate of the "swing time of one of the left front leg 3 and the right front leg 6" with the "swing time of one of the left rear leg 5 and the right rear leg 8." Therefore, the reference rear leg swing time Tr, the reference first swing leg time Tf1, and the reference second swing leg time Tf2 are adapted (determined) in advance so that the first gait and the second gait are respectively realized and the vehicle walks at a target walking speed St, which will be described later, and are stored in the non-volatile memory of the control device 9. In this embodiment, the reference rear leg swing time Tr, the reference first swing leg time Tf1, and the reference second swing leg time Tf2 are 1.4 seconds, 1.3 seconds, and 0.6 seconds, respectively.

[0093] 9 and 10 show examples in which the operator switches the gait between a first gait and a second gait while the robot 1 is walking. In the example of Fig. 9, when the robot 1 was moving forward at the first gait, the control device 9 received a gait change request requesting switching of the gait to the second gait at time td1. Therefore, the control device 9 switches the front leg swing time from the reference first swing leg time Tf1 to the reference second swing leg time Tf2 in the period after time td1.

[0094] On the other hand, in the example of Fig. 10, when the robot 1 was moving forward in the second gait, the control device 9 received a gait change request requesting switching of the gait to the first gait at time td2. Therefore, the control device 9 switches the front leg swing time from the reference second swing leg time Tf2 to the reference first swing leg time Tf1 in the period after time td2. As can be seen from the examples of Fig. 9 and Fig. 10, it is possible to switch the gait of the robot 1 without interrupting walking (specifically, while at least one leg is performing a ground contact movement motion).

[0095] A method for controlling the swing legs will be described together with the swing leg time excess. As described above, when the next grounded leg (i.e., one of the swing legs) touches the ground at the target grounded position, the control device 9 acquires (determines) the target grounded position of the new next grounded leg based on the turning radius correlation value requested (input) by the operator. In addition, when any of the following legs transitions from a grounded leg to a swing leg, the control device 9 acquires the target grounded position of that leg based on the grounding point tracking method. The body coordinate of the target grounded position is also called the target leg tip position Ft(xt, yt, zt).

[0096] The control device 9 executes a process of updating each of the acquired target leg tip positions Ft of the free legs every control period Δt based on one or more torso movement values ​​M detected by the IMU 13. That is, the actual movement amount of the torso 2 during the control period Δt is reflected in the target leg tip position Ft. Specifically, as the torso 2 moves forward, the target leg tip position Ft moves from the front to the rear with respect to the torso 2.

[0097] The swing time exceedance is an event in which the rear leg swing time becomes longer than the reference rear leg swing time Tr, or the front leg swing time becomes longer than either the reference first swing time Tf1 or the reference second swing time Tf2 (corresponding to the gait at that time). The swing time exceedance occurs due to, for example, the operating speed of the motor 20. As one example, the swing time exceedance may occur due to a decrease in the upper limit of the torque that can be generated due to deterioration of any of the motors 20.

[0098] Alternatively, when the robot 1 walks uphill or on uneven ground, the walking speed of the robot 1 may decrease (i.e., the moving speed of the tip of the grounded leg decreases), and an overrun of the floating leg time may occur. More specifically, as understood from the above-described floating leg conditions, the left rear leg 5 and the right rear leg 8 do not float simultaneously, and legs other than the left rear leg 5 and the right rear leg 8 transition from the grounded leg to the floating leg when the adjacent rear leg transitions from the floating leg to the grounded leg. In other words, the time (walking cycle time) from when a certain leg becomes a floating leg until it becomes a grounded leg and then becomes a floating leg again is substantially equal to the walking cycle time of other legs. Therefore, when the moving speed of the tip of the grounded leg decreases due to the condition of the ground on which the robot 1 is walking, the floating leg time of the rear legs and the floating leg time of the front legs become longer.

[0099] An example in which an overrun of the floating leg time occurs due to the robot 1 walking uphill is shown in FIG. 11. In FIG. 11, the robot 1 that was walking on a flat and even ground in the first step entry entered the uphill section at time td3, and as a result, an overrun of the floating leg time has occurred. Specifically, the floating leg time of the rear legs has become a time Trc that is greater than the reference floating leg time Tr of the rear legs, and the floating leg time of the front legs has become a time Tf1c that is greater than the reference first floating leg time Tf1 (i.e., Tr < Trc and Tf1 < Tf1c).

[0100] When an overrun of the floating leg time occurs, the control device 9 sets the value of the proportional coefficient k to a value greater than the initial value ko. In the present embodiment, the initial value ko is "1". If no overrun of the floating leg time has occurred, the control device 9 sets the proportional coefficient k to a value equal to the initial value ko.

[0101] The control device 9 obtains (calculates) the difference between the "target floating leg time" determined based on the proportional coefficient k and the "current floating leg time" as the remaining time Tu (i.e., remaining time Tu = target floating leg time - current floating leg time), and obtains the target tip speed Rt of the floating leg based on the obtained remaining time Tu.

[0102] The target swing time of the left hind leg 5 and the right hind leg 8 (i.e., the target value of the hind leg swing time) is a value (time) obtained by multiplying the reference hind leg swing time Tr by a proportionality coefficient k, and for convenience, is also referred to as the "target hind leg swing time" (i.e., target hind leg swing time = reference hind leg swing time Tr × proportionality coefficient k).

[0103] The target swing leg time of "legs other than the left hind leg 5 and the right hind leg 8" when the first gait is realized (i.e., the target value of the front leg swing time) is a value obtained by multiplying the reference first swing leg time Tf1 by the proportionality coefficient k. For convenience, this time is also referred to as the "target first swing leg time" (i.e., the target first swing leg time = the reference first swing leg time Tf1 × the proportionality coefficient k).

[0104] The target swing leg time of "legs other than the left hind leg 5 and the right hind leg 8" when the second gait is realized (i.e., the target value of the front leg swing time) is a value obtained by multiplying the reference second swing leg time Tf2 by the proportionality coefficient k. For convenience, this time is also referred to as the "target second swing leg time" (i.e., the target second swing leg time = the reference second swing leg time Tf2 × the proportionality coefficient k).

[0105] The current swing leg time is the time from the transition from the grounded leg to the swing leg to the current time. If the remaining time Tu of a certain swing leg becomes a negative value due to the occurrence of an excess of the swing leg time, the control device 9 sets (changes) the proportionality coefficient k to a larger value so that the remaining time Tu of that leg becomes "0".

[0106] If the remaining time Tu of a certain free leg (more specifically, a leg in the first control mode or the third control mode) is "0", the control device 9 sets the target leg tip velocity Rt to a value equal to a predetermined maximum leg tip velocity Rmax. If the remaining time Tu is a positive value, the control device 9 obtains the value (velocity) obtained by dividing the distance (target distance) from the leg tip of the free leg to the target ground contact position by the remaining time Tu as the target leg tip velocity Rt. However, if the velocity obtained by dividing the target distance by the remaining time Tu is greater than the maximum leg tip velocity Rmax, the control device 9 sets the target leg tip velocity Rt to a value equal to the maximum leg tip velocity Rmax.

[0107] When the target leg speed Rt is acquired, the control device 9 controls the rotation angle L of the free leg so that the actual moving speed of the leg tip becomes equal to the target leg speed Rt and the leg tip approaches the leg tip waiting position Ls (in the first control mode) or the target ground contact position (in the third control mode).

[0108] Next, a method for controlling the grounded leg will be described. The control device 9 obtains a target center-of-gravity position Gt (xg, yg, zg) which is a target position of the trunk center-of-gravity Gb every time a control period Δt elapses. The target center-of-gravity position Gt is the trunk coordinate of the trunk center-of-gravity Gb at the time when the tip of the next grounded leg reaches the target grounded position.

[0109] When the target center-of-gravity position Gt is acquired, the control device 9 acquires a target trunk movement value ΔC (xc, yc, zc, Δφ, Δψ, Δθ) which is a target value of the movement amount of the trunk 2 during the period from the current time until the control period Δt has elapsed. The target trunk movement value ΔC is expressed by a combination of the trunk coordinates (xc, yc, zc) of the trunk center of gravity Gb after the control period Δt has elapsed, and the rotation angle Δφ about the x-axis, the rotation angle Δψ about the y-axis, and the rotation angle Δθ about the z-axis of the trunk 2 after the control period Δt has elapsed.

[0110] The control device 9 acquires a target trunk movement value ΔC so that the trunk center of gravity Gb approaches the target center of gravity position Gt and the state of the robot 1 approaches a reference state. At this time, the control device 9 acquires the target trunk movement value ΔC so that the forward speed (walking speed) of the robot 1 (specifically, the trunk 2) becomes equal to a predetermined target walking speed St and the magnitudes of Δφ, Δψ, and Δθ do not exceed predetermined upper rotation limits.

[0111] When the target trunk movement value ΔC is obtained, the control device 9 obtains the leg tip movement value Fd (xd, yd, zd) of each of the grounded legs. The leg tip movement value Fd is the trunk coordinate value of the leg tip of the grounded leg at the time when the trunk 2 moves and rotates as represented by the target trunk movement value ΔC. When the leg tip movement value Fd of each of the grounded legs is obtained, the control device 9 controls the rotation angle L of the grounded leg so that the leg tip position F after the control period Δt has elapsed is equal to the leg tip movement value Fd.

[0112] (Specific operation) Next, specific operations of the control device 9 related to leg condition management processing will be described with reference to Fig. 12 and Fig. 13. The CPU of the control device 9 (hereinafter, simply referred to as "CPU") executes a "walking control processing routine" shown in the flowchart of Fig. 12 every time a control period Δt elapses.

[0113] At an appropriate timing, the CPU starts the process from step 1200 in Fig. 12 and proceeds to step 1205 to determine whether or not a gait change request has been received during the period from the previous execution of this routine to the present time (hereinafter also referred to as the "interval period"). That is, the CPU determines whether or not an operation to change the gait of the robot 1 has been performed by the operator on the input device 33 during the interval period.

[0114] If a gait change request has been received, the CPU determines "Yes" in step 1205 and proceeds to step 1210 to change (the target value of) the front leg swing time. Specifically, the CPU adopts, from among the desired first swing time (first gait) and the desired second swing time (second gait), a value corresponding to the gait requested by the operator as the target value of the front leg swing time. Next, the CPU proceeds to step 1215.

[0115] On the other hand, if a gait change request has been received during the interval period, the CPU determines "No" in step 1205 and proceeds directly to step 1215. Therefore, in this case, the front leg swing time is not changed.

[0116] In step 1215, the CPU executes leg state management processing. Specifically, the CPU executes a "leg state management processing routine" which is shown in the flowchart of Fig. 13 and will be described later. When the processing in Fig. 13 ends, the CPU proceeds to step 1220 and determines whether or not the swing leg time has exceeded.

[0117] If a swing leg time exceedance has occurred, the CPU determines "Yes" in step 1220 and proceeds to step 1225, where it determines the proportionality coefficient k based on the swing leg that has exceeded the target swing leg time the most. For example, if there is only one swing leg that has exceeded the target swing leg time, the CPU sets the value of the proportionality coefficient k so that the remaining time Tu of that swing leg becomes "0". On the other hand, if there are multiple swing legs that have exceeded the target swing leg time, the CPU sets the value of the proportionality coefficient k so that the remaining time Tu of the swing leg with the largest (negative) remaining time Tu among those swing legs becomes "0". Next, the CPU proceeds to step 1235.

[0118] On the other hand, if the swing time has not exceeded, the CPU determines "No" in step 1220 and proceeds to step 1230 to set the proportionality coefficient k to a value equal to the initial value ko. Next, the CPU proceeds to step 1235. The CPU stores the value of the proportionality coefficient k set by the process of step 1225 or step 1230 in the RAM.

[0119] In step 1235, the CPU acquires the target ground contact position of the free leg and the target trunk movement value ΔC. Specifically, the CPU updates the target ground contact position (i.e., the target leg tip position Ft) and the target center of gravity position Gt of the free leg based on one or more trunk movement values ​​M detected by the IMU 13 during the interval period.

[0120] However, if the tip of the next grounded leg reaches the target grounded position during the interval period (i.e., if the next grounded leg switches from one of the left front leg 3 and the right front leg 6 to the other), the CPU newly acquires the target grounded position and the target center of gravity position Gt of the next grounded leg. In addition, if there is a following leg that has transitioned from a grounded leg to a free leg during the interval period, the CPU acquires the target grounded position of that following leg based on the ground point tracking method.

[0121] Furthermore, the CPU acquires a target trunk movement value ΔC based on the target center-of-gravity position Gt. If any of the left front leg 3 to the right hind leg 8 transitions from a grounded leg to a swing leg during the interval period, the CPU stores that time in the RAM as the "swing leg time" of that leg.

[0122] Next, the CPU proceeds to step 1240 and selects one of the left front leg 3 to the right hind leg 8 as the target leg to be controlled. Furthermore, the CPU proceeds to step 1245 and determines whether the control mode of the target leg is the second control mode. If the control mode of the target leg is the second control mode, the CPU determines "Yes" in step 1245 and proceeds directly to step 1275 and determines whether all of the left front leg 3 to the right hind leg 8 have been selected as the target leg by the processing of step 1240.

[0123] If there is a leg that has not yet been selected as the target leg, the CPU judges "No" in step 1275 and proceeds to step 1240, where it selects one of the legs that has not yet been selected as the target leg from among the left front leg 3 to the right hind leg 8.

[0124] If the judgment condition of step 1245 is not satisfied (i.e., if the control mode of the target leg is not the second control mode), the CPU judges "No" at step 1245 and proceeds to step 1250, where it judges whether the control mode of the target leg is the fourth control mode.

[0125] If the control mode of the target leg is the fourth control mode, the CPU judges "Yes" in step 1250 and proceeds to step 1265 to obtain the leg tip movement value Fd of the target leg (ie, the moving leg) based on the desired torso movement value ΔC.

[0126] Next, the CPU proceeds to step 1270 and controls the target leg based on the leg tip movement value Fd. Specifically, the CPU controls the motor 20 associated with the target leg while referring to the detection value of the angle sensor 21 associated with the target leg (i.e., the actual rotation angle L) so that the leg tip position F after the control period Δt has elapsed becomes equal to the leg tip movement value Fd. That is, the control device 9 executes so-called feedback control. Next, the CPU proceeds to step 1275.

[0127] On the other hand, if the judgment condition of step 1250 is not satisfied (i.e., if the control mode of the target leg is the first control mode or the third control mode), the CPU judges “No” in step 1250 and proceeds to step 1255, where it obtains the target leg tip velocity Rt based on the remaining time Tu of the target leg.

[0128] Specifically, the CPU calculates the target swing time of the target leg (i.e., any one of the target rear leg swing time, the target first swing leg time, or the target second swing leg time) based on the proportionality coefficient k set by the processing of step 1225 or step 1230. In addition, the CPU acquires the current swing time of the target leg based on the swing leg time stored in the RAM by the processing of step 1235. Furthermore, the CPU acquires a value obtained by dividing the "distance from the tip of the target leg to the target ground contact position (target distance)" by the "remaining time Tu calculated as the difference between the target swing leg time and the current swing leg time" as the target leg tip speed Rt (i.e., target leg tip speed Rt = target distance / remaining time Tu).

[0129] Next, the CPU proceeds to step 1260 and controls the target leg based on the acquired target leg tip velocity Rt. Specifically, the CPU controls the motor 20 associated with the target leg while referring to the detection value of the angle sensor 21 associated with the target leg so that the moving velocity of the target leg tip becomes equal to the target leg tip velocity Rt and the leg tip approaches one of the leg tip waiting position Ls or the target ground contact position. Furthermore, the CPU proceeds to step 1275.

[0130] If the judgment condition of step 1275 is met (i.e., if all of the left front leg 3 to the right hind leg 8 have already been selected as the target legs), the CPU judges "Yes" in step 1275 and proceeds to step 1295, ending the processing of this routine.

[0131] Next, the leg state management processing routine of Fig. 13 will be described. When the CPU proceeds to step 1215 of Fig. 12, it starts the processing from step 1300 of Fig. 13 and proceeds to step 1305, where it selects one of the left front leg 3 to right hind leg 8 that has not yet been selected as the target leg by the same processing as in step 1240. Next, the CPU proceeds to step 1310, where it determines whether the control mode of the target leg is the first control mode.

[0132] If the control mode of the target leg is the first control mode, the CPU judges "Yes" in step 1310 and proceeds to step 1315 to judge whether or not a second transition condition for the target leg is satisfied. If the second transition condition is satisfied, the CPU judges "Yes" in step 1315 and proceeds to step 1320 to change the control mode of the target leg to the second control mode. Then, the CPU proceeds to step 1375.

[0133] On the other hand, if the second transition condition is not satisfied, the CPU determines “No” in step 1315 and proceeds directly to step 1375 .

[0134] In step 1375, the CPU determines whether or not the processing of this routine has been executed for all of the left front leg 3 to the right hind leg 8. If the processing has not been executed for all of the legs, the CPU determines "No" in step 1375 and proceeds to step 1305, where it selects one of the left front leg 3 to the right hind leg 8 for which processing has not yet been executed as the target leg.

[0135] On the other hand, if the processing of this routine has been executed for all of the left front leg 3 to the right hind leg 8, the CPU makes a "Yes" determination in step 1375 and proceeds to step 1395 to end the processing of this routine. That is, the CPU proceeds to step 1220 in FIG.

[0136] If the judgment condition of step 1310 is not satisfied (i.e., the control mode of the target leg is not the first control mode), the CPU judges "No" at step 1310 and proceeds to step 1325 to judge whether the control mode of the target leg is the second control mode. If the control mode of the target leg is the second control mode, the CPU judges "Yes" at step 1325 and proceeds to step 1330 to judge whether the third transition condition of the target leg is satisfied.

[0137] If the third transition condition is satisfied, the CPU judges "Yes" at step 1330 and proceeds to step 1335 to change the control mode of the target leg to the third control mode. Then, the CPU proceeds to step 1375. On the other hand, if the third transition condition is not satisfied, the CPU judges "No" at step 1330 and proceeds directly to step 1375.

[0138] If the judgment condition of step 1325 is not satisfied (i.e., the control mode of the target leg is not the second control mode), the CPU judges "No" at step 1325 and proceeds to step 1340 to judge whether the control mode of the target leg is the third control mode. If the control mode of the target leg is the third control mode, the CPU judges "Yes" at step 1340 and proceeds to step 1345 to judge whether the fourth transition condition of the target leg is satisfied.

[0139] If the fourth transition condition is satisfied, the CPU determines "Yes" in step 1345 and proceeds to step 1350, where it changes the control mode of the selected leg to the fourth control mode. Then, the CPU proceeds to step 1375. On the other hand, if the fourth transition condition is not satisfied, the CPU determines "No" in step 1345 and proceeds directly to step 1375.

[0140] If the judgment condition in step 1340 is not met (i.e., the control mode of the target leg is the fourth control mode), the CPU judges "No" in step 1340 and proceeds to step 1355 to judge whether all the control modes of the "opposing three legs" of the target leg are the fourth control mode. The opposing three legs are the ones of the first leg group and the second leg group that do not include the target leg. In other words, the CPU judges whether either of the above-mentioned conditions (a1) and (a2) is met.

[0141] If the control modes of all three opposing legs of the target leg are the fourth control mode (i.e., if the swing leg condition (first transition condition) of the target leg is satisfied), the CPU determines "Yes" in step 1355 and proceeds to step 1370 to change the control mode of the target leg to the first control mode. Then, the CPU proceeds to step 1375.

[0142] On the other hand, if the control modes of all the three opposing legs of the target leg are not the fourth control mode (i.e., if neither condition (a1) nor condition (a2) is satisfied), the CPU judges "No" in step 1355 and proceeds to step 1360 to judge whether the target leg is a front leg or a hind leg. That is, the CPU judges whether the target leg is any one of the left front leg 3, left hind leg 5, right front leg 6, and right hind leg 8.

[0143] If the target leg is a front leg or a rear leg, the CPU judges "Yes" in step 1360 and proceeds to step 1365 to judge whether all the control modes of the "four opposing legs" of the target leg are the fourth control mode. The four opposing legs are the four legs among the left front leg 3 to the right rear leg 8 other than the target leg and the "diagonal leg" of the target leg. The diagonal leg is the leg on the left-right opposite side of the body 2 of the target leg and the leg on the front-rear opposite side of the body 2 of the target leg (i.e., any one of the left front leg 3, the left rear leg 5, the right front leg 6, and the right rear leg 8). For example, if the target leg is the left front leg 3, the diagonal leg of the target leg is the right rear leg 8. That is, the CPU judges whether any one of the above-mentioned conditions (b1) to (b4) is satisfied.

[0144] If the control modes of all of the four opposing legs of the target leg are the fourth control mode, the CPU judges "Yes" in step 1365 and proceeds to step 1370. On the other hand, if the control modes of all of the four opposing legs of the target leg are not the fourth control mode (i.e., none of conditions (b1) to (b4) are satisfied), the CPU judges "No" in step 1365 and proceeds directly to step 1375. Note that if the judgment condition in step 1360 is not satisfied (i.e., the target leg is either the left middle leg 4 or the right middle leg 7), the CPU judges "No" in step 1360 and proceeds directly to step 1375.

[0145] As explained above, the first gait is realized by adopting the desired first swing leg time as the target value for the front leg swing time. On the other hand, the second gait is realized by adopting the desired second swing leg time as the target value for the front leg swing time. Therefore, the robot 1 according to this embodiment can adaptively switch gaits between the first and second gaits without interrupting walking.

[0146] In addition, when the swing time of a certain swing leg is exceeded, the value of the proportionality coefficient k becomes larger than the initial value ko, so that the swing times of the other swing legs also become longer, resulting in a high possibility of maintaining stable walking of the robot 1. Furthermore, if any of the conditions (a1) and (a2) and the conditions (b1) to (b4) are satisfied, it is determined that the center of gravity condition is satisfied, so that it is possible to determine whether the center of gravity condition is satisfied by simple processing.

[0147] Although the embodiments of the present invention have been described with reference to the above structure, it will be apparent to those skilled in the art that many alternatives, improvements, and modifications can be made without departing from the scope of the present invention. Therefore, the embodiments of the present invention may include all alternatives, improvements, and modifications that do not depart from the spirit and scope of the appended claims. For example, the embodiments of the present invention are not limited to the specific structure described above, and may be modified as follows:

[0148] The robot 1 in the above embodiment is a six-legged walking robot. Alternatively, the robot 1 may have legs other than the left front leg 3 to the right hind leg 8. That is, the robot 1 may be an eight-legged walking robot or a ten-legged walking robot. For example, when the robot 1 further has a "left front-most leg" adjacent to the front of the body 2 of the left front leg 3 and a "right front-most leg" adjacent to the front of the body 2 of the right front leg 6, examples of transitions of the control modes of the respective legs are shown in Figs. 14 and 15.

[0149] In the example of FIG. 14, a target first swing leg time is used as a target value for the front leg swing time. As a result, a gait is realized in which the number of legs performing ground contact movement is "4" (i.e., half the total number of legs). This gait corresponds to the first gait in the above-mentioned robot 1 (i.e., a six-legged walking robot). On the other hand, in the example of FIG. 15, a target second swing leg time is used as a target value for the front leg swing time. As a result, a gait is realized in which the number of legs performing ground contact movement is maintained at "5" (i.e., half the total number of legs plus 1) or more. This gait corresponds to the second gait in the above-mentioned robot 1.

[0150] In addition, each of the six legs of the robot 1 has three rotational joints. Alternatively, the position of the leg tip may be controlled by a leg having a linear joint (slide joint) and / or a spherical joint (ball joint). In this case, instead of the motor 20 (i.e., an electric motor), a hydraulic cylinder and / or a hydraulic motor may be used as an actuator.

[0151] In addition, the robot 1 is equipped with an optical camera 10 that captures a forward image as an external sensor. Alternatively, an infrared depth sensor, an ultrasonic sonar, or the like may be used as the external sensor. In this case, the control device 9 may acquire the condition (gradient, unevenness, etc.) of the road surface of the planned walking path of the robot 1 acquired by the external sensor, and determine the target ground contact position according to the acquired road surface condition. Furthermore, the control device 9 may automatically switch the gait between a first gait and a second gait according to the condition of the road surface of the planned walking path.

[0152] In addition, the control device 9 switches the target value of the front leg swing time between a target first swing time (first gait) and a target second swing time (second gait) when switching gaits. Alternatively, the control device 9 may gradually change the target value of the front leg swing time from the target first swing time to the target second swing time when switching gaits from the first gait to the second gait. Similarly, the control device 9 may gradually change the target value of the front leg swing time from the target second swing time to the target first swing time when switching gaits from the second gait to the first gait.

[0153] In addition, the control device 9 acquires the center of gravity Gb of the torso at the time when the tip of the next grounded leg reaches the target grounded position as the target center of gravity position Gt. Alternatively, the control device 9 may acquire the target grounded position of the "next grounded leg" that is the leg that will touch the ground after the next grounded leg (in addition to the target grounded position of the next grounded leg), and acquire the center of gravity Gb of the torso at the time when the tip of the next grounded leg reaches the target grounded position as the target center of gravity position Gt.

[0154] In addition, the control device 9 acquires the target ground contact position of the next ground contact leg based on the moving direction input by the operator to the operation control device 31. Alternatively, the control device 9 may acquire a planned walking path for the robot 1 so as to reach a target point input in advance, and acquire the target ground contact position of the next ground contact leg so that the robot 1 walks along the planned walking path. Alternatively, the control device 9 may control each of the left front leg 3 to the right hind leg 8 according to the target ground contact position of the next ground contact leg input each time by the operator.

[0155] In addition, the control device 9 controls each of the left front leg 3 to the right hind leg 8 so that the forward speed of the robot 1 is equal to the target walking speed St regardless of the gait. Alternatively, the control device 9 may switch the forward speed of the robot 1 depending on the gait. For example, the control device 9 may control each of the left front leg 3 to the right hind leg 8 so that the forward speed in the first gait is faster than the forward speed in the second gait. In this case, the reference hind leg swing time adopted in the first gait may be set to a value smaller than the reference hind leg swing time adopted in the second gait.

[0156] In addition, the control device 9 acquires, as the target ground contact position of the following leg, a position adjacent to the rear of the ground contact position of the leg adjacent forward with respect to the torso 2. Alternatively, the control device 9 may acquire the target ground contact position of the following leg such that the tip of the following leg comes into contact (abuts) with the tip of the leg adjacent forward. In other words, the control device 9 may acquire, as the target ground contact position of the following leg, a position (substantially) the same as the ground contact position of the leg adjacent forward.

[0157] Alternatively, the control device 9 may obtain, as the target ground contact position of the following leg, a position away from the ground contact position of the adjacent leg in front of the torso 2. In other words, the control device 9 does not need to adopt the ground contact point tracking method when determining the target ground contact position of the following leg. Even in this case, it is preferable to control the walking cycle times of the left front leg 3 to the right hind leg 8 so that they are approximately equal to each other.

[0158] In addition, the control device 9 obtains a value (speed) obtained by dividing the distance (target distance) from the tip of the free leg to the target ground contact position by the remaining time Tu as the target leg tip speed Rt. Alternatively, the control device 9 may obtain a value obtained by multiplying the value obtained by dividing the target distance by the remaining time Tu by a predetermined coefficient Dk (where Dk>1, for example, Dk=1.2) as the target leg tip speed Rt. The value of the coefficient Dk may be determined based on the response speed of the motor 20 and the detection speed (time lag) of the angle sensor 21 (i.e., taking into consideration the so-called control delay).

[0159] Additionally, in the control device 9, the initial value ko was set to "1." Alternatively, the initial value ko may be a value greater than 1. For example, if the initial value ko is set to "100," each of the reference rear leg swing time Tr, the reference first swing leg time Tf1, and the reference second swing leg time Tf2 may be set to a value equivalent to 1 / 100 of the above-mentioned value.

[0160] 12 (i.e., the processes for controlling the motor 20) every time the control period Δt elapses. Alternatively, the control device 9 may repeatedly execute the process for controlling the motor 20 (so-called feedback control) in a period shorter than the control period Δt. [Explanation of symbols]

[0161] 1. Robot 2. Torso 3...Left front leg (3rd leg) 3a…1st link 3b…Second link 3c…Third link 3d…base 4...Left middle leg (second leg) 5…Left hind leg (first leg) 6...Right front leg (3rd leg) 7...Right middle leg (second leg) 8…Right hind leg (first leg) 9...Control device (walking control means) 10...Optical camera 12...Wireless communication device 20...Motor (actuator) 21...Angle sensor 30…Pedestrian-operated terminal 31...Operation control device 32…Display 33...Input device

Claims

1. The torso and Left and right first legs movable relative to the trunk; left and right second legs movable with respect to the trunk in front of the left and right first legs; left and right third legs movable with respect to the trunk in front of the left and right second legs; a walking control means for controlling each of the first leg, the second leg, and the third leg with respect to the trunk by each actuator; In a robot having The walking control means includes: the torso is moved forward relative to the ground by controlling the actuators so as to alternately perform a ground contact movement action in which the tip of the leg moves backward relative to the torso while in contact with the ground, and a swing leg movement action in which the tip of the leg moves forward relative to the torso while being separated from the ground and the tip of the leg comes into contact with the ground, realizing a first gait in which a state in which the number of legs performing the ground contact movement motion is three appears; a second gait in which a state in which the number of the legs performing the ground contact moving motion is four or more is maintained is realized by making a ratio of a front leg swing time, which is a time during which the second leg and the third leg perform the swing leg moving motion, to a rear leg swing time, which is a time during which the first leg performs the swing leg moving motion, smaller than that in the case in which the first gait is realized; starting the swing leg movement of the target leg when it is determined that a swing leg condition is satisfied for the target leg performing the ground contact movement; The swing condition is: a center of gravity condition is established when the center of mass of the trunk is included in a polygon having vertices each corresponding to a contact point between the tip of a leg other than the target leg and the ground that is performing the ground contact movement, and when the target leg is the first leg, the other of the left and right first legs is performing the ground contact movement; When the target leg is the second leg or the third leg, the leg adjacent to the rear of the target leg and the other leg on the left or right of the target leg are both performing the ground contact movement motion. robot.

2. 2. The robot according to claim 1, The walking control means includes: One proportionality coefficient is stored, a reference rear leg swing time for the rear leg swing time, a reference first swing leg time for the first gait, and a reference second swing leg time for the second gait which is shorter than the reference first swing leg time are stored in advance; when realizing the first gait, (1) the actuators are controlled so that the rear leg swing time becomes equal to a target rear leg swing time determined by multiplying the reference rear leg swing time by the proportionality coefficient, and (2) the front leg swing time becomes equal to a target first swing leg time determined by multiplying the reference first swing leg time by the proportionality coefficient; When the second gait is realized, (1) the actuator is controlled so that the rear leg swing time becomes equal to the desired rear leg swing time, and (2) the front leg swing time becomes equal to a desired second swing time determined by multiplying the reference second swing time by the proportionality coefficient. robot.

3. The robot according to claim 2, The walking control means includes: increasing the proportionality coefficient when the rear leg swing time is longer than the desired rear leg swing time or the front leg swing time is longer than the desired first swing time during the first gait; increasing the proportionality coefficient when the rear leg swing time is longer than the desired rear leg swing time or the front leg swing time is longer than the desired second swing time during the second gait; robot.

4. The robot according to any one of claims 1 to 3, The walking control means includes: when the target leg is the left side of the second leg, it is determined that the center of gravity condition is satisfied when all of a first leg group consisting of the left side of the first leg, the right side of the second leg, and the left side of the third leg are performing the ground contact movement motion; when the target leg is the right of the second leg, it is determined that the center of gravity condition is satisfied when all of a second leg group consisting of the right of the first leg, the left of the second leg, and the right of the third leg are performing the ground contact movement; when the target leg is to the left of the first leg, if all of the second leg group are performing the ground contact movement motion, and if the right of the first leg, the left and right second legs, and the left of the third leg are all performing the ground contact movement motion, it is determined that the center of gravity condition is satisfied; when the target leg is to the right of the first leg, if all of the first leg group are performing the ground contact movement motion, and if the left of the first leg, the left and right second legs, and the right of the third leg are all performing the ground contact movement motion, it is determined that the center of gravity condition is satisfied; when the target leg is to the left of the third leg, if all of the second leg group are performing the ground contact movement motion, and if the left of the first leg, the left and right second legs, and the right of the third leg are all performing the ground contact movement motion, it is determined that the center of gravity condition is satisfied; When the target leg is to the right of the third leg, it is determined that the center of gravity condition is satisfied when all of the first leg group are performing the ground contact movement motion, and when all of the right of the first leg, the left and right second legs, and the left of the third leg are performing the ground contact movement motion. robot.

5. The robot according to any one of claims 1 to 4, The walking control means includes: (1) control of switching from a state in which the first gait is realized to a state in which the second gait is realized while at least one of the legs is performing the ground contact movement motion; and (2) control of switching from a state in which the second gait is realized to a state in which the first gait is realized while at least one of the legs is performing the ground contact movement motion; A robot that performs at least one of the above.

6. The robot according to any one of claims 1 to 5, The walking control means includes: acquire, as a target position of a contact point between the tip of the second leg and the ground, a position that is the same as or near a contact point between the tip of the third leg adjacent to the front of the second leg and the ground; acquiring, as a target position of a contact point between the tip of the first leg and the ground, a position that is the same as or close to a contact point between the tip of the second leg adjacent to the front and the ground; robot.

Images