JP6221539B2 - Wheel type moving body - Google Patents

Description

  The present specification relates to a wheel type moving body that moves on a road surface by driving a wheel.

  Patent Document 1 discloses a technique for stabilizing the posture of the vehicle body during turning by controlling the posture roll angle around the roll axis of the vehicle body. In this technology, the posture roll angle of the vehicle body is controlled so that the position of the intersection of the extension line of the combined centrifugal force and gravity vector acting on the center of gravity of the vehicle body and the ground contact surface of the wheel is maintained in the region between the left and right wheels. doing.

Japanese Patent Laid-Open No. 2006-136902

  In the technique of Patent Document 1 described above, the posture roll angle of the vehicle body is controlled on the assumption that the posture roll angle of the vehicle body does not change suddenly and the posture roll angular acceleration of the vehicle body is zero. However, when the posture roll angle of the vehicle body changes suddenly with the sudden turning operation, the posture roll angular acceleration of the vehicle body becomes a magnitude that cannot be ignored. As a result, the posture stability of the vehicle body is lowered, and the vehicle body may fall depending on the magnitude of the posture roll angular acceleration. The present specification discloses a technique capable of stably controlling the posture of the vehicle body even when performing a sudden turning operation.

  The wheel type moving body disclosed in the present specification moves on the road surface by driving the wheels. The wheel-type moving body is rotatable around the axle, and at least two wheels arranged at intervals in the axle direction, a wheel driving device that drives the wheels, and the wheels are rotatably supported. A vehicle body rotatable around the roll axis, an actuator for adjusting a posture roll angle around the roll axis of the vehicle body, a wheel drive device, and a control device for controlling the actuator. The control device is based on the target posture roll angle calculated in consideration of the posture roll angular acceleration of the vehicle body so that the ZMP (zero moment point) of the wheel-type moving body is located within a preset setting range. Control the actuator.

  In the wheel type moving body, the target posture roll angle is calculated in consideration of the posture roll angular acceleration of the vehicle body so that the ZMP of the wheel type moving body is located within a preset setting range. Then, the actuator is driven based on the calculated target posture roll angle, and the posture roll angle of the vehicle body is controlled. For this reason, even if the posture roll angle of the vehicle body changes suddenly due to a sudden turn of the wheel-type moving body, the change in posture roll angle (posture roll angular acceleration) is taken into consideration and the target posture roll angle Is calculated. As a result, the ZMP of the wheel-type moving body can be stably controlled within the set range, and the posture of the vehicle body can be stably controlled.

The front view which shows typically the structure of the vehicle which concerns on an Example. The side view of the vehicle shown in FIG. The block diagram which shows the structure of the control system which concerns on an Example. The block diagram which shows the function of a control system. The model figure which shows the control model of a vehicle (state in which the vehicle is turning). It is a figure which shows an example of the numerical simulation of the vehicle which concerns on an Example, and is a figure which shows the relationship between time and ZMP. It is a figure which shows an example of the numerical simulation of the vehicle which concerns on an Example, and is a figure which shows the relationship between time and a roll angle. It is a figure which shows an example of the numerical simulation of the vehicle which concerns on an Example, and is a figure which shows the relationship between time and turning angular velocity.

  The main features of the embodiments described below are listed. The technical elements described below are independent technical elements and exhibit technical usefulness alone or in various combinations, and are not limited to the combinations described in the claims at the time of filing. Absent.

(Characteristic 1) In the wheel type moving body disclosed in the present specification, the control device calculates the wheel type in consideration of the posture roll angular acceleration of the vehicle body so that the ZMP of the wheel type moving body is located in the set range. The wheel drive device may be controlled based at least on the target centripetal acceleration in the centripetal direction of the moving body. According to such a configuration, the target acceleration in the centripetal direction of the wheel type moving body is controlled so that the ZMP is within the set range (that is, the turning speed of the wheel type moving body is controlled). For this reason, the posture of the vehicle body can be controlled more stably.

(Feature 2) In the wheel type moving body disclosed in the present specification, the control device includes a target posture roll angle, a target posture roll angular acceleration that is a second-order differential value of the target posture roll angle, and a wheel type moving body. Based on the target posture roll angle and the target centripetal acceleration calculated so that the ZMP calculated from the target speed in the traveling direction and the target angular velocity in the turning direction of the wheel type moving body is within the set range, the actuator and the wheel drive The device may be controlled. According to such a configuration, the posture of the vehicle body can be controlled more stably. The ZMP may be calculated from the target speed in the traveling direction of the wheel type moving body and the target steering angle of the wheel type moving body.

(Feature 3) In the wheel type moving body disclosed in the present specification, the control device is based on a target posture roll angle calculated based on model predictive control using a model of the wheel type moving body and its controller. The actuator may be controlled. According to such a configuration, the posture roll angle of the vehicle body can be accurately predicted, and the posture of the vehicle body can be controlled more stably.

(Characteristic 4) The wheel type moving body disclosed in the present specification may be an inverted pendulum type moving body in which the center of gravity of the vehicle body is located higher than the axle.

  The present embodiment will be described with reference to the drawings. As shown in FIGS. 1 and 2, the vehicle 100 is an inverted two-wheel moving body, and includes a vehicle body 10 to which left and right wheels 12 a and 12 b are attached. The present embodiment is characterized by the roll angle control of the vehicle body 10, and has other configurations (for example, inverted two-wheel control (speed control in the traveling direction of the vehicle body 10 (direction orthogonal to the axles 13a and 13b)), turning angular velocity. The control (angular velocity control in the turning direction of the vehicle body 10) and the like can be configured in the same manner as in the prior art. For this reason, in the following description, the configuration related to the roll angle control of the vehicle body 10 will be mainly described, and description of other configurations will be omitted as appropriate.

  The wheels 12 a and 12 b are rotatably attached to the side surface of the vehicle body 10. Specifically, the wheels 12a and 12b are attached to the tip ends of the axles 13a and 13b. The axles 13 a and 13 b are arranged on the same axis, extend parallel to the vehicle width direction of the vehicle body 10, and are rotatably supported by the vehicle body 10. Since the wheels 12a and 12b are attached to the front ends of the axles 13a and 13b, the wheels 12a and 12b are arranged at intervals in the direction in which the axles 13a and 13b extend (vehicle width direction). The wheels 12a and 12b are independently driven by motors 14a and 14b (an example of a wheel driving device). By driving the wheels 12a and 12b, the vehicle 100 travels on the traveling surface R while maintaining an inverted state. Further, the vehicle 100 can turn left and right by changing the rotational speed of the wheels 12a and 12b. The rotation angles of the wheels 12a and 12b are detected by encoders 15a and 15b (shown in FIG. 3). The rotation angles of the wheels 12a and 12b detected by the encoders 15a and 15b are input to the control device 20 (shown in FIG. 3).

  The vehicle body 10 is rotatable with respect to the axles 13a and 13b as described above. That is, the axles 13a and 13b are pitch axes, and the vehicle body 10 is rotatable around the axles 13a and 13b (pitch axes). The center of gravity G of the vehicle body 10 is higher than the axles 13a and 13b, and the vehicle 100 is an inverted two-wheeled moving body. The vehicle body 10 is rotatable around the roll axis. The roll shaft is orthogonal to the axles 13a and 13b and passes through the approximate center of the vehicle body 10 in the vehicle width direction. The vehicle body 10 includes a gyro sensor (attitude angle sensor) 16 that detects the pitch angular velocity and the roll angular velocity of the vehicle body 10, an actuator (roll control actuator) 18 that controls the roll angle of the vehicle body 10, motors 14a and 14b, A control device 20 (illustrated in FIG. 3) for controlling the actuator 18 is provided.

  The gyro sensor 16 is a biaxial gyro sensor that detects an angular velocity (pitch angular velocity) around the axle 10 (pitch axis) of the vehicle body 10 and an angular velocity (roll angular velocity) around the roll axis of the vehicle body 10. As shown in FIG. 3, the gyro sensor 16 is electrically connected to the control device 20. The pitch angular velocity signal and the roll angular velocity signal output from the gyro sensor 16 are input to the control device 20.

  The actuator 18 is a device for adjusting the roll angle of the vehicle body 10. The actuator 18 applies a torque τ around the roll axis to the vehicle body 10. As shown in FIG. 3, the actuator 18 is electrically connected to the control device 20. Based on the control command value output from the control device 20, the actuator 18 generates a torque τ around the roll axis. For the actuator 18, for example, a geared servo motor can be used. The control device controls the roll angle of the vehicle body 10 by controlling the actuator 18 according to the roll angle of the vehicle body 10 input by the gyro sensor (posture angle sensor) 16. As a result, the running stability during turning is improved.

  The control device 20 is configured by a computer including a CPU, a ROM, a RAM, and the like. The control device 20 is installed in the vehicle body 10. The control device 20 is electrically connected to the gyro sensor 16, the actuator 18, the motors 14a and 14b, and the encoders 15a and 15b. When the control device 20 drives the motors 14a and 14b, the vehicle 100 travels on the traveling surface R while maintaining an inverted state, and when the control device 20 drives the actuator 18, the vehicle 10 rotates around the roll axis. The posture angle (roll angle) is controlled. Hereinafter, the control device 20 will be described in detail.

  As shown in FIG. 4, the control device 20 includes, as its function, a multiplication unit 22, a model prediction control unit 24, a command value calculation unit 26, an inverted two-wheel control unit 100a, a turning angular velocity control unit 100b, a roll A posture angle control unit 100c is provided.

The multiplying unit 22 calculates the centripetal acceleration command value a _f ^ref of the vehicle 100 by multiplying the traveling direction velocity command value v _f ^ref given to the vehicle 100 and the turning direction angular velocity command value ω _f ^ref . That is, in the present embodiment, the target motion trajectory of the vehicle 100 is given by the speed command value v _f ^{ref in} the traveling direction of the vehicle 100 (the direction orthogonal to the axles 13a and 13b) and the turning direction angular velocity command value ω _f ^ref . . The multiplication unit 22 multiplies these values to calculate a centripetal acceleration command value a _f ^ref (= v _f ^ref · ω _f ^ref ). Conversion by the multiplication unit 22 into the centripetal acceleration command value a _f ^ref facilitates handling in the model prediction control unit 24 described later.

The model prediction control unit 24 predicts the ZMP trajectory of the vehicle 100 based on the centripetal acceleration command value a _f ^ref input from the multiplication unit 22, and a set range (to be described later) in which the predicted ZMP is set in advance. The corrected centripetal acceleration command value a ^ref of the vehicle 100 and the roll angle command value θ _roll ^ref of the vehicle body 10 are calculated so as to be located within. The model prediction control unit 24 uses a motion model of the vehicle 100 in order to predict the future trajectory of the ZMP. By using the motion model, the trajectory of the ZMP is calculated in consideration of the roll angular acceleration of the vehicle body 10. The detailed configuration of the model prediction control unit 24 will be described in detail later.

The command calculation unit 26 calculates a corrected traveling direction speed command value v ^ref and a corrected turning direction angular speed command value ω ^ref based on the corrected centripetal acceleration command value a ^ref calculated by the model prediction control unit 24. That is, the centripetal acceleration command value a is the product of the traveling direction velocity command value v and the turning direction angular acceleration command value ω. Therefore, the command calculation unit 26 converts the corrected centripetal acceleration command value a ^ref calculated by the model prediction control unit 24 into a corrected traveling direction speed command value v ^ref and a corrected turning direction angular velocity command value ω ^ref . For example, if the corrected traveling direction speed command value v ^ref = the traveling direction speed command value v _f ^ref , the corrected turning direction angular speed command value ω ^ref = the corrected centripetal acceleration command value a ^ref / the corrected traveling direction speed command value v ^ref Become. The corrected traveling direction speed command value v ^ref may be different from the traveling direction speed command value v _f ^ref according to the magnitude of the corrected centripetal acceleration command value a ^ref . In the above-described example, the traveling direction velocity command value v and the turning direction angular acceleration command value ω are calculated from the centripetal acceleration command value a. However, the present invention is not limited to this example. That is, since the centripetal acceleration command value can be calculated from the traveling direction speed command value and the target steering angle, the traveling direction speed command value v and the target steering angle may be calculated from the centripetal acceleration command value a.

The inverted two-wheel control unit 100 a drives the motors 14 a and 14 b based on the corrected traveling direction speed command value v ^ref calculated by the command calculation unit 26, so that the traveling direction of the vehicle 100 is maintained while maintaining the inverted state of the vehicle body 10. To control the speed. Specifically, the rotation angle of the wheels 12a and 12b is input from the encoders 15a and 15b to the inverted two-wheel control unit 100a, and the angular velocity around the pitch axis of the vehicle body 10 is input from the gyro sensor 16. Therefore, the motors 14 a and 14 b are driven based on the state quantities fed back and the corrected traveling direction speed command value v ^ref calculated by the command calculation unit 26. As a result, the vehicle body 10 maintains the inverted state and moves at a desired speed in the traveling direction. As a specific configuration of the inverted two-wheel control unit 100a, a conventionally known control configuration (state feedback control, H∞ control, PID control, etc.) can be appropriately employed. For this reason, detailed description of the inverted two-wheel controller 100a is omitted.

The turning angular velocity control unit 100b controls the speed of the vehicle body 10 in the turning direction by driving the motors 14a and 14b based on the corrected turning direction angular velocity command value ω ^ref calculated by the command calculation unit 26. Specifically, the rotation angles of the wheels 12a and 12b are input from the encoders 15a and 15b to the turning angular velocity control unit 100b. Therefore, the motors 14 a and 14 b are driven based on the state quantity fed back and the corrected turning direction angular velocity command value ω ^ref calculated by the command calculation unit 26. As a result, the vehicle body 10 turns at a desired turning angular velocity. It should be noted that a conventionally known control configuration (for example, PID control, state feedback control, etc.) can be appropriately employed as the specific configuration of the turning angular velocity control unit 100b. For this reason, detailed description of the turning angular velocity control unit 100b is omitted.

The roll posture angle control unit 100c controls the posture (roll angle) of the vehicle body 10 by driving the actuator 18 based on the roll angle command value θ _roll ^ref calculated by the model prediction control unit. Specifically, an angular velocity around the roll axis of the vehicle body 10 is input from the gyro sensor 16 to the roll attitude angle control unit 100c. Therefore, the actuator 18 is driven based on the state quantity fed back and the roll angle command value θ _roll ^ref calculated by the model prediction control unit. Thereby, the posture (roll angle) of the vehicle body 10 is controlled to a desired roll angle. For this reason, when the vehicle body 10 performs the turning motion, the roll angle of the vehicle body 10 is controlled in conjunction with the turning motion, so that the vehicle 100 can stably perform the turning motion. As a specific configuration of the roll attitude angle control unit 100c, a conventionally known control configuration (for example, state feedback control, PID control, etc.) can be appropriately employed. For this reason, detailed description of the roll attitude angle control unit 100c is omitted.

Next, the model prediction control unit 24 will be described in detail. The model prediction control unit 24 predicts the ZMP (zero moment point) of the vehicle 100 using the motion model around the roll axis of the vehicle 100, and the predicted ZMP is positioned within the set range. A corrected centripetal acceleration command value a ^{ref of the} vehicle 100 and a roll angle command value θ _roll ^ref of the vehicle body 10 are calculated. First, a procedure for deriving a ZMP trajectory from the motion model will be described, then a procedure for designing a prediction model using the derived ZMP trajectory will be described, and finally, a corrected centripetal acceleration command value a ^ref using the prediction model will be described. A procedure for calculating the _roll angle command value θ _roll ^ref will be described.

(1) Derivation of ZMP Equation from Motion Model FIG. 5 shows a motion model around the roll axis of the vehicle 100. In FIG. 5, the intersection of the perpendicular passing through the roll axis and the running surface R is the origin [0, 0], the coordinates of the ZMP are q = [q _y , q _z ], and the coordinates of the center of gravity of the vehicle body 10 are p = [p _y , p _z ]. Further, h is the distance between the center of gravity and the roll center, and r is the radius of the wheels 12a and 12b. Further, the weight of the vehicle body 10 is M and the moment of inertia is J. When the ZMP equation is constructed from the above model, the following equation (1) is obtained.

Here, f _z is a component in the vertical direction (z direction) of the reaction force f from the traveling surface R, and _fy is a component in the horizontal direction (y direction) of the reaction force f. On the other hand, p _y , p _z , q _z , f _y , and f _z when the vehicle 100 is turning are given by the following equation (2). In Expression (2), a is the centripetal acceleration of the vehicle 100.

Substituting these values into equation (1) and solving for the y-coordinate (q _y ) of ZMP yields the following equation (3).

Here, since θ _roll << 1, since it is assumed that sin θ _roll = θ _roll and cos θ _roll = 1, the following equation (4) is obtained.

Furthermore, since h · (d ² θ _roll / dt ² ) · θroll is much smaller than g, when g−h · (d ² θ _roll / dt ² ) · θroll is regarded as g, equation (4) is Equation (5) is obtained.

When the above equation (5) is arranged with respect to θroll, a, and (d ² θ _roll / dt ² ), the following equation (6) is obtained.

As is clear from the equation (6), the y-coordinate (q _y ) of ZMP is expressed by a linear equation. Also, ZMP y coordinate _{(q y),} using a roll angle theta _roll, the roll angle acceleration is second-order differential value of the roll angle ^{_{θ roll (d 2 θ roll /}} dt 2), the centripetal acceleration a calculated can do. In this embodiment, the setting range (the range in which the position of the ZMP is controlled) is, for example, the range between the left and right wheels 12a and 12b (ie, [−w / 2, 0] to [w / 2, 0] (tread width range). In addition, in order to secure a desired stability margin, the width is set appropriately narrow within a range between the left and right wheels 12a and 12b (that is, a range of [−w / 2, 0] to [w / 2, 0]). (For example, [−k · w / 2, 0] to [k · w / 2, 0] (k is a positive number less than 1)).

(2) Prediction Model Design Next, the prediction model used for predicting the ZMP trajectory (coordinate value q _y ) by the model prediction control unit 24 will be described. The prediction model receives a roll angle command value and a centripetal acceleration command value, and outputs a ZMP trajectory (z coordinate value q _y ) and a centripetal acceleration a. That is, as already described, the model prediction control unit 24 of this embodiment receives the centripetal acceleration command value a _f ^{ref and} outputs the corrected centripetal acceleration command value a ^ref and the roll angle command value θ _roll ^ref . Therefore, the centripetal acceleration command value a ^ref and the roll angle command value θ _roll ^{ref ref} are input to the prediction model, and the ZMP trajectory (z coordinate value q _y ) and the centripetal acceleration a must be output accordingly. I must. Therefore, in this embodiment, a prediction model is constructed based on the roll angle control system model.

Here, it is assumed that the transfer function from the roll angle command value θ _roll ^ref to the roll attitude angle θ _roll is designed by the following equation (7). In addition, although Formula (7) was the transfer function of the cubic system, you may use transfer functions other than this.

  When the state equation is obtained from the equation (7), the following equation (8) is obtained.

  When equation (8) is discretized, the following equation (9) is obtained.

Here, assuming that the responsiveness of the centripetal acceleration a is excellent, the relationship between the centripetal acceleration command value a ^ref and the centripetal acceleration a can be regarded as equivalent to a one-step delay. That is, the relationship shown in the following equation (10) is established. By introducing the relationship of Expression (10), it is possible to design the prediction model with a small amount of state. Note that the relationship between the centripetal acceleration command value a ^ref and the centripetal acceleration a may be expressed by a high-order transfer function as shown in Equation (7).

  Substituting equation (10) into equation (9) yields the following equation (11).

  In order to simply show the following mathematical expression expansion, Equation (11) is redefined as shown in the following Equation (12).

  From the above, if the output equation shown in the following equation (13) is given to the state equation of the equation (11), the trajectory (qy) and the centripetal acceleration (a) of the ZMP can be controlled.

(3) Procedure for calculating the command values a ^ref and θ _roll ^ref using the prediction model When the prediction model is designed as described above, the control command values a ^ref and θ _roll ^ref must be calculated using the prediction model. I must. In the present embodiment, the corrected centripetal acceleration command value a ^ref and the roll angle command value θ _roll ^ref are calculated so that the ZMP trajectory predicted by the prediction model is within the set range. Therefore, a method of predicting the state quantity (ZMP trajectory (q _y )) using the prediction model will be described, and then a procedure for calculating the control command values a ^ref and θ _roll ^ref from the predicted state quantity will be described.

(3-1) Prediction of State Quantity Using Prediction Model Now, x ^ _mpc [k + n | k] is predicted at the kth step, k + n step state quantity, and u ^ [k + n | k] is at the kth step. Assuming that the derived control input is the k + n step, the state quantity from k + 1 to k + N _p steps predicted at the k step is expressed by the following equation (14).

However, if it is assumed that the control input is derived only by _Nu steps and thereafter takes a constant value (that is, assuming that the steady state is reached after _Nu steps elapse), the control from _Nu steps to N _p −1 is performed. The input is expressed by the following equation (15).

In model predictive control, in order to suppress discontinuity of control input, control input variation Δu [k] = u [k] −u [k−1] is generally regarded as a control input. Therefore, u ^ [k + i | k] (i = 0, 1, 2,..., N _u −1) is expressed by the following equation (16).

  Substituting the above equations (15) and (16) into equation (14) and rearranging results in the following equation (17).

Therefore, the predicted value of the control amount y for N _p steps is given by the following equation (18).

  Then, when Expression (17) and Expression (18) are integrated, the following Expression (19) is obtained.

Here, Y [k] and ΔU [k] in the above equation (19) are given by the following equation (20). The above N _p and N _u can be appropriately set in consideration of the controllability of the vehicle 100 and the calculation load of the control device 20.

(3-2) Evaluation Function and Inequality Condition In the present embodiment, the following items are considered when calculating the control command values a ^ref and θ _roll ^ref using the state quantities predicted as described above. That is,
(1) The ZMP trajectory converges to 0 in a steady state (converges to the origin [0, 0] in FIG. 5),
(2) The ZMP trajectory must be within the set range in the transient state.
(3) The centripetal acceleration follows the control command value in a steady state,
(4) The roll angle command value and the turning angular velocity command value are not discontinuous. The above (1), (3), and (4) are considered by the evaluation function, and the above (2) is considered by the inequality condition. Expression (21) is an example of an evaluation function. In Equation (21), the weight matrices Q and R are for suppressing the convergence of ZMP to 0, the followability of the centripetal acceleration to the control command value, and the discontinuity of the roll angle command value and the centripetal acceleration command value. Set as appropriate.

  In the above equation (21), R [k] is a command vector and is given by the following equation (22).

In addition, r [k + i | k] (i = 1, 2,..., N _p ) in the equation (22) is a command vector of the i-th step and is represented by the following equation (23). Here, a _f ^ref is a centripetal acceleration command value given at the k-th step, and it can be assumed that the same centripetal acceleration command value is given continuously for N _p steps in the evaluation function.

  On the other hand, the inequality condition is given by the following equation (24).

  Here, G in the equation (24) is a constant matrix and can be derived as follows. That is, the inequality condition for Y [k] can be given by the following equation (25).

  Here, Ymax is an upper limit value (for example, position [w / 2, 0] of wheel 12a in FIG. 5), and Ymin is a lower limit value (for example, position [−w / 2, 0] of wheel 12b in FIG. 5). ). When formula (25) is transformed, the following formula (26) is obtained.

  By comparing the equations (24) and (26), the constant matrix G of the equation (24) can be easily derived. Further, when G = [Γ g] where g is the last column of G and Expression (19) is substituted into Expression (24), the following Expression (27) can be obtained.

Since Equation (27) is a linear inequality constraint on ΔU [k], ΔU [k] that minimizes the evaluation function of Equation (21) while satisfying the inequality condition of Equation (27) is a quadratic programming problem. Can be derived by solving Since ΔU [k] derived in this way is a control input for N _p steps, the process of giving the control input for the first one step to the control target is executed for each sample period.

The model prediction control unit 24 configured as described above outputs the corrected centripetal acceleration command value a ^ref and the roll angle command value θ _roll ^ref . Then, by driving the motors 14a and 14b and the actuator 18 based on these control command values, the ZMP trajectory (z coordinate value q _y ) of the vehicle 100 can be set within the set range.

Next, the simulation result of the roll angle control by the control device 20 described above will be described. In the simulation, while the vehicle is traveling straight at 6.0 [km / h] (= 1.667 [m / s]), at the time t = 4.0 [s], the turning angular velocity command value ω _f ^ref (= 1.176 [rad / s] (corresponding to 0.2 G)). Y _max and Y _min were set to +0.05 [m] and −0.05 [m], respectively.

6 to 8 show simulation results, FIG. 6 shows the trajectory q _y of ZMP, FIG. 7 shows the roll angle command value θ _roll ^ref and the roll angle θ _roll , and FIG. 8 shows the turning angular velocity command value ω ^ref and the turning. The angular velocity ω is shown. As is apparent from FIGS. 6 to 8, the ZMP trajectory quickly converges to 0 while the ZMP of the vehicle 100 is maintained within the set range (that is, −0.05 to 0.05 [m]). In addition, the turning angular velocity quickly converges to the command value. From the above results, it was confirmed that in the vehicle 100 of the present embodiment, a quick turning operation can be performed while maintaining the stability of the vehicle body 10.

  As is apparent from the above description, in the vehicle 100 of the present embodiment, the control command value is calculated by model predictive control based on the motion model around the roll axis of the vehicle body 10, so that the change in roll angle (posture roll) during the transition period is calculated. The control command value is calculated in consideration of (angular acceleration). As a result, the posture of the vehicle 100 can be stably controlled, and an agile turning operation can be realized.

  Although the present embodiment has been described in detail above, these are merely examples and do not limit the scope of the claims. The technology described in the claims includes various modifications and changes of the specific examples illustrated above.

  For example, in the above-described embodiment, an inverted two-wheel type moving body is used. However, the technique disclosed in the present specification is a wheel-type moving body having an actuator that controls the posture (roll angle) in the roll direction. Good. For example, the number of wheels is not limited to two, and may be three or four, or may be a moving body on which a person is boarded.

  In the above-described embodiment, the vehicle 100 is caused to turn by changing the drive amount of the left and right wheels 12a and 12b. However, the technology disclosed in the present specification is not limited to such an example. . For example, a steering wheel may be provided separately from the drive wheels 12a and 12b, and the vehicle 100 may be turned by adjusting the steering angle of the steering wheel. In this case, for example, a steering angle adjustment device that adjusts the steering angle of the steering wheel may be further provided, and the control device may drive the steering angle adjustment device so that the steering angle of the steering wheel becomes the target steering angle.

  Further, in the above-described embodiment, the control device 20 is configured under the condition that the current value flowing through the actuator 18 is not restricted. However, the technique disclosed in this specification is based on the condition that the current value flowing through the actuator 18 is restricted. It can also be used below. An example of the configuration of the control device (specifically, the model prediction control unit) in such a case will be described.

  In the motion model around the roll axis shown in FIG. 5, the equation of motion is expressed by the following equation (28).

Here, J is the moment of inertia of the vehicle body 10 around the center of gravity, M is the weight of the vehicle body 10, h is the distance from the center of the roll to the center of gravity, g is the acceleration of gravity, θ _roll is the roll angle, and v is the vehicle 100 in the traveling direction. The speed, ω is the turning angular speed, and τ is the output torque of the actuator 18. Here, the relationship of the following equation (29) exists between the output torque τ and the current i of the actuator 18.

_{Here, n act} is the gear reduction ratio of the actuator 18, _{k t} is the torque constant. Assuming that the moment of inertia of the rotor of the motor is small and the reduction ratio n _act is not large, the current i is expressed by the following equation (30) by the equations (28) and (29).

  Therefore, when Expression (30) regarding the current i is added to Expression (13) in the above-described embodiment, Expression (31) is derived. Thereafter, the model prediction control unit when the current i of the actuator 18 is limited can be designed by using the equation (30) instead of the equation (13).

  The technical elements described in this specification or the drawings exhibit technical usefulness alone or in various combinations, and are not limited to the combinations described in the claims at the time of filing. In addition, the technology illustrated in the present specification or the drawings achieves a plurality of objects at the same time, and has technical utility by achieving one of the objects.

DESCRIPTION OF SYMBOLS 10 Car body 12a, 12b Wheel 13a, 13b Axle 14a, 14b Motor 15a, 15b Encoder 16 Gyro sensor 18 Actuator 20 Control apparatus

Claims (5)

A wheel type moving body that moves on the road surface by driving wheels,
At least two wheels that are rotatable about an axle and are spaced apart in the direction of the axle;
A wheel drive device for driving the wheel;
A vehicle body that rotatably supports the wheel and that can rotate about the roll axis;
An actuator for adjusting the posture roll angle around the roll axis of the vehicle body;
A wheel drive device and a control device for controlling the actuator,
The control device
A target centripetal acceleration input unit for inputting a target centripetal acceleration in the centripetal direction of the wheeled moving body;
A ZMP prediction unit that predicts a ZMP (zero moment point) of a wheel-type moving body in consideration of the posture roll angular acceleration of the vehicle body based on the input target centripetal acceleration;
A target posture roll angle calculation unit that calculates a target posture roll angle so that the predicted ZMP is located within a preset setting range;
An actuator driving unit that drives an actuator based on the calculated target posture roll angle . The control device
A target centripetal acceleration correction unit for correcting the input target centripetal acceleration so that the predicted ZMP is located within a preset setting range;
The wheel type moving body according to claim 1 , further comprising: a wheel driving device driving unit that drives the wheel driving device based on at least the corrected target centripetal acceleration . The ZMP prediction unit predicts ZMP using a prediction model,
In the prediction model, when the corrected target centripetal acceleration and the target posture roll angle are input, the ZMP and the target centripetal acceleration before correction are output.
The target posture roll angle calculation unit calculates the target posture roll angle so that the ZMP predicted using the prediction model is located within a preset setting range,
The wheel type moving body according to claim 2 , wherein the target centripetal acceleration correction unit corrects the target centripetal acceleration so that the ZMP predicted using the prediction model is located within a preset setting range . A steering wheel and a steering angle adjusting device for adjusting a steering angle of the steering wheel;
The target centripetal acceleration input unit inputs the target centripetal acceleration in the centripetal direction of the wheel type moving body by inputting the target speed in the traveling direction of the wheel type moving body and the target steering angle of the wheel type moving body,
The control device includes a target steering angle correction unit that corrects the target steering angle based on the target speed in the traveling direction of the wheel-type moving body and the corrected target centripetal acceleration;
The wheel type moving body according to claim 2 , further comprising a steering angle adjusting device driving unit that drives the steering angle adjusting device based on the corrected target steering angle . The wheel type moving body according to any one of claims 1 to 4 , wherein the wheel type moving body is an inverted pendulum type moving body in which a center of gravity of the vehicle body is higher than an axle.

Images