Walking assist device, control device, and control program

Abstract

To provide a walking assist device capable of estimating a target walking period by quickly responding to a change in walking motion of a pedestrian.SOLUTION: A walking assist device 1 mounted on a leg part of a pedestrian to assist the walking motion of the pedestrian includes: an actuator 16 for giving an assist force to the walking motion of the pedestrian; a walking motion detection part 18 for detecting the walking motion of the pedestrian; a walking period estimation part 211 for estimating a target walking period of the pedestrian on the basis of the detected walking motion of the pedestrian; a control target value generation part 212 for generating a control target value of the actuator 16 on the basis of the estimated target walking period; and a drive control part 20 for controlling the drive of the actuator 16 on the basis of the generated control target value.SELECTED DRAWING: Figure 8

Description

[Technical field]

[0001] The present invention relates to a walking assistance device, a control device, and a control program. [Background technology]

[0002] In recent years, with the increase in the elderly population, it is important to promote walking training and exercise for the elderly, prevent strokes, and maintain the health of the elderly. However, the walking motion of the elderly tends to show gait disorders because the stride length is shorter than that of healthy adults and the height of the toes is lower when the foot is raised. For this reason, walking assistance devices that assist the walking motion of a pedestrian by an actuator have been used in the past. The walking assistance device detects the walking motion of the pedestrian and controls the actuator according to the walking motion to assist the walking motion of the pedestrian.

[0003] For example, Patent Document 1 discloses a technology in which the interactive force between the walker's own force and the assistive force of the suit at each joint of a robotic suit is detected by a torque sensor, the degree of synchronization between the walker and the suit is adjusted, and then the result is analyzed using a neural oscillator model to determine the target angle for each joint, and actuators are feedback controlled based on the difference between the current angle and the target angle at each joint.

[0004] Patent Document 2 discloses a technology for a walking assist device that can apply torque to a walker around the leg joints such as the hip joint and knee joint, in which when the walker changes their walking rhythm, the walking assist rhythm changes to follow this, and an actuator that applies torque is controlled based on the changed walking assist rhythm.

[0005] Patent Document 3 discloses a technology in which sensors attached to the heel and toe of a lower leg attachment that can swing in plantar flexion and dorsiflexion detect whether the heel and toe are in a grounded state, determine which of three or more phases one cycle of a walker's walking movement is in, and apply resistance using ER fluid or gas to the heel and toe of the sole placement part of the lower leg attachment depending on the determined phase.

[0006] The technologies disclosed in Patent Documents 1 to 3 are all technologies that detect the walking motion of a walker, follow the walking motion of the walker, and assist the walking motion with the assisting force of an actuator, but the walking assist device does not take the initiative to promote the walking motion. In order to promote walking training and exercise for the elderly and the like, there is a demand for a walking assist device that encourages walkers to walk in the correct walking motion of healthy people and guides walkers to walk in the correct walking motion. For this reason, the present inventors have been studying a walking assist device that automatically adjusts the target values ​​of the walking cycle and stride length according to the walker's emotions and muscle fatigue state while walking, as shown in Non-Patent Documents 1 and 2. [Prior art documents] [Patent documents]

[0007] [Patent Document 1] JP 2012-066375 A [Patent Document 2] JP 2004-073649 A [Patent Document 3] JP 2006-326185 A [Non-patent literature]

[0008] [Non-Patent Document 1] Yunfan Li, et. al., Development of automatic controlled walking assistive device based on fatigue and emotion detection, JRM, Vol.34, No.6, pp.1383-1397,2022. [Non-Patent Document 2] Yunfan Li, Yukai Gong, and Eiichiro Tanaka, Wearable emotional fatigue assessment system for automatic control of walking assistance device, MDT2021, 1203, (2021). Summary of the Invention [Problem to be solved by the invention]

[0009] However, the walking assistance devices disclosed in Non-Patent Documents 1 and 2 detect the emotions and muscle fatigue state of the walker and estimate an optimal target walking cycle based on the detection results, so there is a problem in that it takes time to estimate the target walking cycle and it is not possible to quickly respond to changes in the walking movement of the walker.

[0010] An object of the present invention is to provide a walking assist device, a control device, and a control program that can quickly respond to changes in the walking motion of a walker and estimate a target walking cycle. [Means for solving the problem]

[0011] The walking assist device of the present invention is a walking assist device that is attached to the legs of a walker and assists the walking movement of the walker, and includes an actuator that applies an assist force to the walking movement of the walker, a walking movement detection unit that detects the walking movement of the walker, a walking cycle estimation unit that estimates a target walking cycle of the walker based on the walking movement of the walker detected by the walking movement detection unit, a control target value generation unit that generates a control target value of the actuator based on the target walking cycle estimated by the walking cycle estimation unit, and a drive control unit that controls drive of the actuator based on the control target value generated by the control target value generation unit.

[0012] The control device of the present invention is a control device that controls a walking assist device attached to the legs of the walker, and includes an actuator that applies an assist force to the walking movement of a walker, a drive control unit that controls the drive of the actuator, and a walking movement detection unit that detects the walking movement of the walker, and includes a walking cycle estimation unit that estimates a target walking cycle of the walker based on the walking movement of the walker detected by the walking movement detection unit, a control target value generation unit that generates a control target value of the actuator based on the target walking cycle estimated by the walking cycle estimation unit, and a control target value output unit that outputs the control target value generated by the control target value generation unit to the drive control unit.

[0013] The control program of the present invention is a computer-readable control program executed on a computer that controls a walking assist device attached to the legs of a walker, the computer comprising: an actuator that applies an assist force to the walking movement of a walker, a drive control unit that controls the drive of the actuator, and a walking movement detection unit that detects the walking movement of the walker, and causes the computer to function as a walking cycle estimation unit that estimates a target walking cycle of the walker based on the walking movement of the walker detected by the walking movement detection unit, a control target value generation unit that generates a control target value of the actuator based on the target walking cycle estimated by the walking cycle estimation unit, and a control target value output unit that outputs the control target value generated by the control target value generation unit to the drive control unit. Effect of the Invention

[0014] According to the present invention, the walking period estimation unit estimates a target walking period based on an already detected walking movement, and the control target value generation unit generates a control target value based on the estimated target walking period. Therefore, the target walking period can be estimated in rapid response to changes in the walking movement of the walker, such as the walking period and stride length. [Brief description of the drawings]

[0015] [Figure 1] 1 is a perspective view showing a structure of a walking assistance device according to an embodiment of the present invention; [Diagram 2] 1 is a plan view showing a structure of a walking assistance device according to an embodiment. [Diagram 3] 1A and 1B are diagrams showing the structure of a lower leg attachment part of a walking assistance device according to an embodiment, in which the upper right figure is a plan view, the lower left figure is a rear view, and the lower right figure is a side view. [Figure 4] 1 is a side view showing a state in which a walking assistance device according to an embodiment is worn by a walker. [Diagram 5] FIG. 2 is a side view showing the structure of the sole-of-foot placement portion and the lower leg attachment portion in the embodiment. [Figure 6] 4A to 4C are schematic diagrams for explaining detection of a walking motion by a walking motion detection unit in the embodiment. [Figure 7] FIG. 2 is a side view showing the external appearance of the structure of a controller in the embodiment. [Figure 8] FIG. 2 is a block diagram showing a structure of a controller in the embodiment. [Figure 9] 3 is a schematic diagram showing a configuration of a GUI screen of a portable information terminal in the embodiment. FIG. [Figure 10] 13 is a graph showing a walking ratio map in the embodiment. [Figure 11] 11 is a graph showing the relationship between stride length and change in ankle (ankle joint) joint angle when speeds are different in an embodiment. [Figure 12] 1 is a graph showing the relationship between the walking phase in each phase and the joint angle amplitude value of the walker in an embodiment, with the walking cycle of the walker being divided into five phases. [Figure 13] 13 is a graph showing the relationship between the walking phase and the joint angle amplitude value of the walker in FIG. 12 and a simplified graph of this relationship. [Figure 14] 1 is a graph showing the relationship between the angle change of the ankle (ankle joint) and the stride length in five phases of a walking movement in an embodiment. [Figure 15] 11 is a graph showing control target values ​​(target joint angle amplitude values) generated based on angle changes in five phases in an embodiment. [Figure 16] FIG. 2 is a schematic diagram showing a gait estimation method according to the first embodiment. [Figure 17] 1 is a graph showing the relationship between initial stance time and gait cycle in the first embodiment. [Figure 18] FIG. 11 is a schematic diagram showing a gait estimation method according to a second embodiment. [Figure 19] FIG. 13 is a schematic diagram showing a gait estimation method according to a third embodiment. [Figure 20] FIG. 13 is a schematic diagram showing a gait estimation method according to a fourth embodiment. [Figure 21] FIG. 13 is a schematic diagram showing a gait estimation method according to the fifth embodiment. [Figure 22] 4 is a flowchart for explaining the operation of the embodiment. [Figure 23] 5A to 5C are schematic diagrams showing a method for setting an initial operation of a walking assistance device using a portable information terminal in an embodiment. [Figure 24] 5A to 5C are schematic diagrams showing changes in walking period according to the number of steps of a walker in an embodiment. [Diagram 25] 1 is a graph comparing a target walking period according to the walking estimation method of the first embodiment with an actually measured walking period in Example 1. [Figure 26] 13 is a graph comparing a target walking period according to a walking estimation method of a second embodiment with an actually measured walking period in Example 2. [Figure 27] 13 is a graph comparing a target walking period according to a walking estimation method of a third embodiment with an actually measured walking period in Example 3. [Figure 28] 13 is a graph comparing a target walking period according to a walking estimation method of a fourth embodiment with an actually measured walking period in Example 4. [Figure 29] 13 is a graph comparing a target walking period according to a walking estimation method of a fifth embodiment with an actually measured walking period in Example 5. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0016] [1] Basic structure FIG. 1 shows a walking assistance device 1 according to this embodiment. In the following description, the same parts and members as those already described will not be described. The basic structure and the structure of the controller unit 19 are common to the first to fifth embodiments described later. The walking assistance device 1 is a device that is attached to the legs of a walker and supports the walking movement of the walker by applying an auxiliary force to the walking movement of the walker. The walking assistance device 1 includes a sole placement unit 11, a lower leg attachment unit 12, straps 13, 14, 15, a servo motor 16, a torque limiter 17, a walking movement detection unit 18, and a controller unit 19. The walking assistance device 1 is attached to each of the left and right legs of a walker, but the following description will be given of the right side of the walking assistance device 1. The left side has the same configuration as the right side, and differs only in that it is symmetrical.

[0017] The sole placement part 11 is a part on which the sole of the foot of the walker is placed, and is a plate-like body made of synthetic resin that extends from the heel to the toe of the sole of the foot of the walker as shown in FIG. 2, and the lower surface opposite to the upper surface on which the sole of the foot is placed is the contact surface with the walking surface. A heel 111 is provided on the outer periphery of the rear part of the sole placement part 11 so as to protrude upward and cover the outer periphery of the heel of the walker. In addition, a metal plate-like protruding part 112 that protrudes from the upper surface of the sole placement part 11 is provided on the outer side of the rear part of the sole placement part 11. The upper end of the plate-like protruding part 112 protrudes to the joint position of the ankle (ankle joint) of the walker, and a hole 113 is formed at a position corresponding to the joint position of the ankle (ankle joint) (the position of the ankle (ankle joint) joint position PA, which is the intersection of the vertical line VL and the horizontal line HL in FIG. 5). A torque limiter 17 is provided on the outer side of the plate-like protruding part 112.

[0018] As shown in the side view (lower right), plan view (upper view), and rear view (lower left view) of Fig. 3, the lower leg attachment section 12 has a vertical extension section 121 made of a metal rod-shaped body with a U-shaped cross section, and supports the weight of the walker acting on the leg. The base end of the vertical extension section 121 is rotatably connected to a hole 113 formed in the plate-shaped protruding section 112, and a strap 13 is attached to the upper end. The strap 13 is made of a belt-shaped member made of cloth, and is wrapped around the lower leg of the walker. Note that the strap 13 is not shown in the side view of Fig. 3.

[0019] As shown in Fig. 2, the strap 14 is made of a cloth belt-like member provided between the widthwise ends of the heel 111 of the sole-placing portion 11. When the sole of a walker's foot is placed on the upper surface of the sole-placing portion 11, the strap 14 is wrapped around the area between the heel and the instep of the walker's foot. The strap 15 is made of a cloth belt-like member provided between the widthwise ends of the front side of the sole-placing portion 11. When the sole of a walker's foot is placed on the upper surface of the sole-placing portion 11, the strap 15 is wrapped around the front part of the instep of the walker's foot.

[0020] In the walking assistance device 1, straps 13, 14, and 15 are wrapped around the lower leg, between the heel and the instep, and the front part of the instep, as shown in Fig. 4. By fastening the straps 13, 14, and 15, the device is attached to each part of the walker's leg (the walker's lower leg, heel, and front part of the sole). When the ankle joint (foot joint) rotates due to the walker's walking motion, the sole placement part 11 rotates integrally with the walker around the hole 113.

[0021] The servo motor 16 as an actuator is provided at a position corresponding to the ankle joint (ankle joint) position of the walker, and applies a rotational force serving as an assisting force to the sole placement unit 11 according to the movement of the joint during the walker's walking motion. Specifically, the servo motor 16 is composed of a motor having a rotating shaft, and as shown in FIG. 5, the rotating shaft is disposed at the ankle joint (ankle joint) position PA of the walker. In FIG. 5, when the rotating shaft of the servo motor 16 rotates clockwise (θ downward), a rotational force is generated that rotates the tip portion of the sole placement unit 11 downward from the horizontal line HL, thereby assisting the plantar flexion during the walker's walking motion. On the other hand, when the rotating shaft of the servo motor 16 rotates counterclockwise (θ upward), a rotational force is generated that rotates the tip portion of the sole placement unit 11 upward from the horizontal line HL, thereby assisting the dorsiflexion during the walker's walking motion. The rotating shaft of the servo motor 16 is connected to a torque limiter 17.

[0022] The torque limiter 17 is an overload protection device attached to the outer surface of the plate-shaped protrusion 112. Although not shown, the torque limiter 17 includes an input part connected to the rotation shaft of the servo motor 16, an output part attached to the plate-shaped protrusion 112, a damping member sandwiched between the input part and the output part, and a pair of elastic members sandwiched in the rotation direction of the input part. When the servo motor 16 is not overloaded, the input part and the output part operate as one unit due to the frictional force of the damping member and the pair of elastic members. On the other hand, when the servo motor 16 is overloaded, the input part is allowed to rotate relative to the output part against the frictional force of the damping member and the biasing force of the elastic member. In addition, as the torque limiter of this embodiment, the one disclosed in JP 2021-166617 A can be suitably used, but is not limited thereto, and may be a torque limiter of another structure.

[0023] 1 and 2, the walking motion detection unit 18 includes a heel detection sensor 181 and a toe detection sensor 182 provided on the upper surface of the sole placement unit 11. The heel detection sensor 181 is provided immediately below the ankle joint (foot joint) of the walker, and the toe detection sensor 182 is provided immediately below the ball of the foot on the sole of the walker. The heel detection sensor 181 and the toe detection sensor 182 are composed of pressure sensors that detect the walker's own weight acting on the heel and ball of the foot as pressure and output on / off electrical signals.

[0024] Heel detection sensor 181 and toe detection sensor 182 output an electrical signal that is on when pressure is detected and that is off when no pressure is detected. A threshold is set for the electrical signal detected by heel detection sensor 181 and toe detection sensor 182, respectively, and the on / off detection is performed based on whether the threshold is exceeded. The threshold can be changed by operating a threshold icon on mobile information terminal 22, which will be described later, and the sensitivity can be adjusted according to the user (see FIG. 9).

[0025] As shown in FIG. 6, when the walker walks, the heel detection sensor 181 is on and the toe detection sensor 182 is off at the initial contact. Next, when the entire sole of the foot is on the ground (Foot flat), both the heel detection sensor 181 and the toe detection sensor 182 are on. Furthermore, when the walker is plantar flexing (Heel off), the heel detection sensor 181 is off and the toe detection sensor 182 is on. Then, when the sole of the walker is completely floating (Toe off), both the heel detection sensor 181 and the toe detection sensor 182 are off. In this way, the walking motion detection unit 18 can detect the state of the walker's walking motion by a combination of the on and off states of the heel detection sensor 181 and the toe detection sensor 182.

[0026] [2] Structure of controller unit 19 As shown in Fig. 7, the controller unit 19 is housed in a recess of the vertical extension 121 of the lower leg attachment part 12 together with the servo motor 16, and is covered with a resin cover 191. The controller unit 19 is electrically connected to an output signal line 183 from the heel detection sensor 181 of the walking motion detection part 18 and an output signal line 184 from the toe detection sensor 182, and estimates a target walking cycle based on the detection value detected by the walking motion detection part 18, generates a control target value based on the estimated target walking cycle, and performs drive control of the servo motor 16 based on the generated control target value. The controller unit 19 includes a drive control part 20 and a command generation part 21 as shown in Fig. 8.

[0027] The drive control unit 20 outputs an analog voltage or current signal to the servo motor 16 to control the drive of the servo motor 16. The drive control unit 20 includes a PID control unit 202. The PID control unit 202 breaks down the input control command into a proportional term, an integral term, and a derivative term, and adjusts the gain of each term to control the drive of the servo motor 16. By adjusting the gain, the output of the servo motor 16 is prevented from exceeding the control command (overshoot), and robustness of the output of the servo motor 16 is ensured. At this time, an encoder (not shown) provided on the rotating shaft on the output side of the servo motor 16 detects an error signal (Error) of the output of the servo motor 16 relative to the control command, and feeds it back to the input side of the PID control unit 202 to offset the error signal from the next control command, thereby adjusting the gain of the PID control unit 202.

[0028] The command generation unit 21 as a control device generates a control target value to be given to the servo motor 16 based on the walking motion of the walker detected by the walking motion detection unit 18. The command generation unit 21 includes a walking cycle estimation unit 211, a control target value generation unit 212, and a control target value output unit 213. The walking cycle estimation unit 211, the control target value generation unit 212, and the control target value output unit 213 are configured as a computer-readable program executed on a CPU (Central Processing Unit) of the controller unit 19.

[0029] The walking period estimation unit 211 estimates a target walking period based on the walking period of the walker detected by the walking movement detection unit 18. A method of estimating a target walking period will be described in detail later.

[0030] The portable information terminal 22 is wirelessly connected to the walking cycle estimation unit 211, and as shown in FIG. 9, by operating icons 222 and 223 on a GUI screen 221 of the portable information terminal 22, initial setting values ​​when starting walking assistance by the walking assist device 1 (for example, initial values ​​such as the ratio of walking cycle and the target ankle angle amplitude for icon 222) can be input.

[0031] The control target value generating unit 212 generates a control target value to be given to the servo motor 16 based on the target walking period estimated by the walking period estimating unit 211. The control target value is given as a target joint angle amplitude value of the sole placement unit 11. The procedure for calculating the target joint angle amplitude value from the target walking period will be described in detail below.

[0032] The control target value generating unit 212 calculates a target cadence based on the target walking cycle estimated by the walking cycle estimating unit 211. The target cadence is calculated as the reciprocal of the target walking cycle. Once the target cadence is calculated, the control target value generating unit 212 calculates a target step length according to the target cadence.

[0033] The walking ratio is the ratio of cadence to step length, and is known to be constant over a wide range of walking speeds. Walking at a constant walking ratio can optimize the energy efficiency, walking stability, and attention of the pedestrian. The relationship between walking ratio and step length can be calculated, and the walking ratio of healthy people varies from 0.005 to 0.007, as shown in the walking ratio map WMP in Figure 10. In contrast, past research has shown that for elderly people, energy efficiency, walking stability, and attention of the pedestrian can be optimized when the walking ratio is approximately 0.006.

[0034] Next, the control target value generating unit 212 calculates a target joint angle amplitude value (control target value) corresponding to the target step length from the relationship between the step length and the joint angle amplitude value (ankle angle amplitude) of the joint position of the leg of the walker. The relationship between the step length and the joint angle amplitude value of the joint position of the leg is determined by performing an experiment to measure the angle change of the subject using a joint angle measuring device that can be attached to three joints of the lower limbs (hip joint, knee joint, ankle joint), and calculating not only the change in each joint angle but also the step length during walking according to the three measured joint angles.

[0035] Specifically, the stride length is calculated by the following procedure. When the hip joint angle is θhp, the knee joint angle is θknee, and the ankle (ankle) angle is θankle, as shown in FIG. 11, the hip joint position is set as the origin (0,0), and data on the ankle (ankle) position (xi, yi) is obtained from the change in each joint angle. The obtained ankle (ankle) position (xi, yi) data is logged multiple times, and the walking trajectory of the walker is plotted and calculated. Next, the ankle (ankle) joint angle amplitude value and stride length walking data are picked up for 50 completed cycles under each of fast, normal, and slow speed conditions, and the corresponding ankle (ankle) joint angle amplitude value (ankle joint angle) and stride length data are collected, so that the stride length can be calculated. Note that Figure 11 is based on Non-Patent Document 3 “J. Zhuang, “Study on walking assistance considering emotion evaluation with a device of compact design,” Ph.D. dissertation, Waseda University, 2020.”

[0036] The inventors simplified the relationship between the walking phase (Ratio of walking cycle) of a walker and the joint angle amplitude value (Ankle joint angle) of the ankle joint in each of phases 1 to 5 as shown in graph GR1 in Fig. 12, and modeled it with only straight lines as shown in graph GR2 in Fig. 13. Here, there is a large gap between graph GR1 and graph GR2 in phases 4 and 5, but the gap is intentionally made large in consideration of the problem of elderly people walking that the dorsiflexion movement of the ankle joint (ankle joint) in the swing phase (movement from phase 4 to phase 5) that can be performed only by the anterior tibial muscle becomes insufficient due to muscle weakness and fatigue, and there is a high possibility that the toes cannot be lifted even on a small step and get caught and fall.

[0037] According to the aforementioned Non-Patent Document 3, as shown in FIG. 14, walking motion in one walking cycle is divided into five phases, and the relationship between the walking phase (Ratio of walking cycle) in each phase and the ankle joint angle amplitude value (Ankle joint angle) at that time is analyzed. As a result, the heel (ankle joint) joint angle amplitude value of a walker is the ankle (ankle joint) joint angle amplitude value as shown in graph GR1 of FIG. 12 in walking phases 1 to 5. In FIG. 14, the target stride length S T and the joint angle change of the ankle (foot joint) θ varT According to Non-Patent Document 3, the linear equation for is as follows:

[0038] θ varT =7.92×S T -11.47: Phase 1 (0-10% of the gait cycle) -57.16×S T +65.28: Phase 2 (10-40% of the gait cycle) -111.16×S T +37.62: Phase 3 (40-60% of gait cycle) 73.77×S T +-28.33: Phase 4 (60-85% of gait cycle) 79.75×S T -60.64: ​​Phase 5 (85-100% of the gait cycle)

[0039] In this embodiment, as shown in the graph GR3 of FIG. 15, the target joint angle amplitude value (Motor target angle) which is the control target value of the servo motor 16 is generated to precede the walking motion of the walker shown in the graph GR2. Specifically, the angular velocity of the servo motor 16 in each phase is set to be twice the angular velocity of the walking motion of the walker to generate the target joint angle amplitude value of the servo motor 16. Note that the target angle amplitude value of the servo motor 16 differs depending on the angular velocity of the walker wearing it, so it can be set arbitrarily. When the joint angle amplitude value of the servo motor 16 reaches the target joint angle amplitude value, the walker waits until the walking motion of the walker reaches the target joint angle amplitude value, and when the walking motion of the walker catches up with the target joint angle amplitude value, the walker moves to the walking motion of the next phase.

[0040] As shown in the graph GR3 of Fig. 15, by setting the target joint angle amplitude value as the control target value in each phase, the walking motion of the pedestrian can be actively encouraged, and the pedestrian can walk in a correct walking motion similar to that of a healthy person. Also, since the target joint angle amplitude value waits until the walking motion of the pedestrian reaches the target joint angle amplitude value before moving to the next phase, the pedestrian can walk in a correct walking motion without feeling fatigue or psychological burden.

[0041] That is, in general, a walking assistance device attached to an exoskeleton is not connected to the skeleton no matter how closely it is attached to the device, so an error in movement is always generated. Even if the walking assistance device could move exactly the same as the walker wearing it without any error, it cannot be said that it is assisting the walker in any way. In contrast, in the walking assistance device 1 of this embodiment, the straps 13, 14, and 15 that secure the legs of the walker are pulled by the movement of the walking assistance device 1, and the walker can feel the direction, timing, and angular velocity of the next movement by the straps 13, 14, and 15, so that the walker is always encouraged to move the legs a little faster at any walking cycle and any walking speed. Therefore, a walker wearing the walking assistance device 1 can walk with a correct walking movement close to that of a healthy person.

[0042] The target joint angle amplitude value generated by the control target value generation unit 212 is output to the control target value output unit 213, and the control target value output unit 213 outputs the target joint angle amplitude value to the drive control unit 20 as an electrical signal.

[0043] [3] Method for estimating target walking cycle by the walking cycle estimation unit 211 Next, a description will be given of a method for estimating a target walking period by the above-mentioned walking period estimation unit 211. There are a number of possible methods for estimating a target walking period, and since each estimation method is different, the following description will be given as a first embodiment to a fifth embodiment.

[0044] [3-1] First embodiment In the first embodiment, as shown in Fig. 16, the walking cycle estimation unit 211 estimates a target walking cycle according to the time from the heel touching the ground to the toe touching the ground (initial stance phase time) of a walker detected by the walking motion detection unit 18. It has been known that a change in the time from the initial heel touching the ground to the toe touching the ground in a walking motion reflects the walker's intention to change his / her walking style, and in this embodiment, based on this knowledge, the target walking cycle is estimated by the initial stance phase time.

[0045] Specifically, the walking assistance device 1 described above was attached to the legs of two able-bodied subjects, and the subjects were made to walk freely for 100 steps while changing their gait without assistance from the servo motor 16. It was confirmed that there is a relationship between the initial stance phase time and the walking cycle (in the case of the second subject in Example 1 described below) as shown in Fig. 17. There is a linear relationship between the initial stance phase time and the walking cycle, with a correlation coefficient of 0.9 or more, and it was found that the target walking cycle can be estimated by measuring the initial stance phase time.

[0046] According to this embodiment, the target walking period can be estimated by simply measuring the initial stance time among the detected walking periods, so that the target walking period can be estimated in an extremely short time, and there is an effect that it is possible to quickly respond to changes in the walking motion of the walker. In particular, the current walking period starting from the contact of the heel can be instantly estimated and corrected. However, when an elderly person who has a reduced stride length and only performs the minimum necessary plantar flexion motion walks with a flat foot (heel and toe landing simultaneously), it is difficult to estimate the target walking period with the method according to the first embodiment, so it is preferable to apply it when the person is walking with a measurable initial stance time.

[0047] [3-2] Second embodiment In the second embodiment, the walking cycle estimation unit 211 estimates a target walking cycle by averaging multiple walking cycles detected by the walking motion detection unit 18. It is generally known that elderly people have a lower rate of change in gait when walking and a smaller range of gait change compared to younger people. Based on this knowledge, the present inventors have decided to estimate the target walking cycle by the walking cycle estimation unit 211. For example, as shown in FIG. 18, the walking cycle estimation unit 211 averages the previous three walking cycles (2nd to 4th) already detected by the walking motion detection unit 18, and estimates the next fourth (5th) target walking cycle.

[0048] According to this embodiment, the target walking period cannot be estimated very quickly as in the first embodiment, but since the target walking period is estimated based on the detected walking period of the walker, the walking assistance device 1 can be operated according to the wearer of the walking assistance device 1. The estimation of the target walking period according to this embodiment is suitable for walking training of patients with hemiplegia. That is, hemiplegia patients tend to have their toes extended due to joint contracture, and walk while swinging their toes sideways so as not to hit the ground, which is called a swinging walk. At this time, since the toes contact the ground from the toes, there is a possibility that they will get caught on even a slight step and fall. For this reason, it is necessary to input the target movement as an initial movement and train to follow the target movement as much as possible. As the training progresses, the walking speed increases, the walking period becomes shorter, and the stride length increases, so training can be performed by gradually changing the walking period and stride length according to the walking speed while repeating the same movement each time without making any significant changes.

[0049] [3-3] Third embodiment In the third embodiment, the walking cycle estimation unit 211 estimates the target walking cycle by performing linear approximation of the relationship between the number of strides of the walker already detected by the walking motion detection unit 18 and the walking cycle for each number of strides. For example, as shown in FIG. 19, the walking cycle estimation unit 211 plots three numbers of strides of the walker already detected by the walking motion detection unit 18 and the walking cycle for each number of strides on a map MP representing the relationship between the number of strides and the walking cycle, and performs linear approximation (linear regression: y=ax+b) of these plots to estimate the fourth target walking cycle. Note that, in FIG. 19, linear approximation of three walking cycles is performed, but this is not limiting, and the target walking cycle may be estimated by performing linear approximation of plots of four or more walking cycles.

[0050] According to this embodiment, the target walking period is estimated taking into consideration the changes in the three detected walking periods, so that the target walking period can be estimated while taking into consideration the tendency of changes in the walking motion of the walker. In particular, when the walking motion of the walker is smoothly changing, the estimated target walking period can be made to accurately follow the actually measured walking period.

[0051] [3-4] Fourth embodiment In the fourth embodiment, the walking cycle estimation unit 211 estimates the target walking cycle by performing curve approximation of the relationship between the number of strides of the walker already detected by the walking motion detection unit 18 and the walking cycle at each number of strides. For example, as shown in FIG. 20, the walking cycle estimation unit 211 plots three numbers of strides of the walker already detected by the walking motion detection unit 18 and the walking cycle at each number of strides on a map MP that represents the relationship between the number of strides and the walking cycle, and performs curve approximation (Nonlinear regression: y=ax 2 +bx+c) to estimate the fourth target walking cycle. Note that in FIG. 20, curve approximation is performed on three walking cycles, but this is not limiting, and curve approximation may be performed on plots of four or more walking cycles to estimate the target walking cycle. Also, in this embodiment, curve approximation is performed using a quadratic function, but the curve approximation is not limited to this. Curve approximation may be performed using a cubic or higher order function, or an exponential or logarithmic function.

[0052] According to this embodiment, the target walking period is estimated in consideration of the changes in the three detected walking periods, so that the target walking period can be estimated in consideration of the tendency of changes in the walking motion of the walker, as in the third embodiment. When the walking motion of the walker suddenly changes, the estimated target walking period can be made to accurately follow the actually measured walking period.

[0053] [3-5] Fifth embodiment In the fifth embodiment, the walking cycle estimation unit 211 compares a first target walking cycle obtained by performing linear approximation of the relationship between the number of strides of the pedestrian already detected and a plurality of walking cycles at each number of strides with a second target walking cycle obtained by performing curve approximation, and selects a target walking cycle that is close to the walking cycle detected immediately before, to estimate the target walking cycle. For example, as shown in FIG. 21, the walking cycle estimation unit 211 plots three numbers of strides of the pedestrian already detected by the walking motion detection unit 18 and the walking cycles at each number of strides on a map MP that represents the relationship between the number of strides and the walking cycle, and performs linear approximation (linear regression: y=ax+b) of these plots to obtain a first target walking cycle W for the fourth time. TL Estimate.

[0054] Next, the walking cycle estimation unit 211 plots the number of strides of the walker for three times detected by the walking movement detection unit 18 and the walking cycle for each number of strides on a map MP that indicates the relationship between the number of strides and the walking cycle, and performs curve approximation (Nonlinear regression: y=ax 2 +bx+c) to complete the fourth second target walking cycle W TQ The first target walking period W TL and the second target walking period W TQ When the walking cycle estimation unit 211 estimates the walking cycle of the third step (x3) detected immediately before, the walking cycle estimation unit 211 estimates the second target walking cycle W TQ is selected and estimated as the target walking cycle.

[0055] According to this embodiment, the walking period estimation unit 211 estimates the first target walking period W TL and the second target walking period W estimated by curve approximation. TQ Among these, the second target walking cycle WTQ By selecting the above and estimating the target walking cycle, it is possible to accommodate both smooth and sudden changes in the gait of a walker, and it is possible to estimate an optimal target walking cycle in response to various changes in the gait of a walker.

[0056] [4] Function and effect of walking assistance device 1 Next, the operation of the walking assist device 1 will be described based on the flowchart shown in Fig. 22. Using the portable information terminal 22, initial values ​​at the start of walking assistance by the walking assist device 1 are input (step S1). Specifically, as shown in Fig. 23, by operating an icon 222 on a GUI screen 221 of the portable information terminal 22, the ratio of walking cycle and the motor target angle amplitude are adjusted and input according to the walking motion of phase 1 to phase 5 in the walking cycle.

[0057] The walking cycle estimation unit 211 estimates a target walking cycle by any of the methods from the first embodiment to the fifth embodiment based on the input initial value (step S2). After the target walking cycle is estimated, the control target value generation unit 212 calculates a target walking rate which is the reciprocal of the walking cycle (step S3). The control target value generation unit 212 calculates a target stride length according to the target walking rate of the walking motion by referring to the calculated target walking rate and the walking ratio map WMP shown in FIG. 10 (step S4).

[0058] The control target value generating unit 212 generates a target joint angle amplitude value (Motor target angle) that is a control target value based on the calculated target stride length (Step S5). The drive control unit 20 controls the drive of the servo motor 16 based on the generated target joint angle amplitude value (Step S6). The torque limiter 17 determines whether the load torque acting on the servo motor 16 exceeds a predetermined torque limit value (Step S7), and if it is determined that the torque limit value is exceeded (Step S7: Yes), it allows the input side part to which the rotation shaft of the servo motor 16 is connected to rotate relative to the output side part, and stops the operation of the walking assist device 1.

[0059] In step S7, if it is determined that the load torque acting on the servo motor 16 does not exceed the torque limit value (step S7: No), the walking motion is detected by the walking motion detection unit 18 (step S8). The walking period estimation unit 211 determines whether the walking motion of the walker reaches the target joint angle amplitude value (step S9). If it is determined that the walking motion of the walker does not reach the target joint angle amplitude value (step S9: No), the walking period estimation unit 211 waits for the estimation of the next target walking period. If it is determined that the walking motion of the walker reaches the target joint angle amplitude value (step S9: Yes), the walking period estimation unit 211 returns to step S2 based on the detection value detected in step S8 to estimate the target walking period of the next phase, and repeats steps S3 to S5 based on the estimated target walking period to generate the target joint angle amplitude value. Then, the generation of the target joint angle amplitude value that is the control target value is performed from phase 1 to phase 5 shown in FIG. 15.

[0060] According to this embodiment, the walking period estimation unit 211 estimates a target walking period based on the detected walking period, and the control target value generation unit 212 generates a target joint angle amplitude value that serves as a control target value based on the estimated target walking period. Therefore, it is possible to quickly respond to changes in the walking movement of the walker and correct the target values ​​of the walker's walking period and stride length.

[0061] [5] Modifications of the embodiment The present invention is not limited to the above-described embodiments, but includes the following modifications. In the above-described embodiments, the servo motor 16 is used as an actuator that provides an assist force to the walking motion of the walker, but the present invention is not limited to this. For example, a walking assist device may be configured using an actuator that supplies a fluid or the like to the heel and toe of the sole placement portion to provide an assist force to the walking motion of the walker.

[0062] In the above embodiment, the servo motor 16 of the walking assistance device 1 is configured to apply a rotational force to the ankle joint position, but the present invention is not limited to this. For example, the present invention may be applied to a walking assistance device equipped with a servo motor that applies a rotational force to the hip joint or knee joint of the walker's legs. In this case, the angle and angular velocity of the walker's leg joints (hip joint, knee joint, etc.) can be detected by the rotation angle of the servo motor or an IMU (Inertial Measurement Unit) attached to the leg.

[0063] In the above embodiment, the heel detection sensor 181 and the toe detection sensor 182, which are pressure sensors, are used as the walking motion detection unit 18, and the walking motion of the walker is detected by combining the detection values ​​of these sensors 181 and 182, but the present invention is not limited to this. For example, an angle sensor that directly measures the joint angle of the ankle (ankle joint) of the walker may be used as the walking motion detection unit, and the walking motion may be detected from the detection value of the joint angle of the ankle (ankle joint). Furthermore, the detection of the walker's motion may be detected by the rotation angle of the servo motor 16 or an IMU attached to the instep of the walker, as described above.

[0064] In the above-described embodiment, the controller unit 19 is built into the walking assist device 1, but the present invention is not limited to this. For example, only the drive control unit that controls the driving of the actuator may be built into the walking assist device, and the walking cycle estimation unit and the control target value generation unit may be executed on the CPU of a computer that can wirelessly communicate with the drive control unit.

[0065] In the above embodiment, the detected walking cycle is divided into five phases, but the present invention is not limited to this. For example, the detected walking cycle may be divided into six or more phases to generate a control target value. In addition, for example, among the five divided phases, phase 1 and phase 2 may be regarded as one phase and a linear relationship may be obtained as a whole. In addition, the specific structure and shape of the present invention may be other structures, etc. within the scope that can achieve the object of the present invention. EXAMPLES

[0066] [Evaluation method] The inventors had two subjects wear the walking assist device 1 on their legs and evaluated the results of actual walking for the target walking cycle estimation methods of the first to fifth embodiments described above. Each subject walked with the same walking tendency shown in FIG. 24, with the number of strides being 30 (60 for both legs), with the walking cycle gradually decreasing for the first 10 steps, the walking cycle being constant for the next 10 steps, and the walking cycle gradually increasing for the final 10 steps. The walking ratio was set to 0.0067 in the walking ratio map WMP shown in FIG. 10. The experimental results were as follows: Ri For the estimated target walking period W Ti The accuracy rate of r p was evaluated using formula (1), where n is the total number of steps (=30 steps) and i is the number of strides (=number of strides).

[0067]

number

[0068] [Example 1] In Example 1, the target walking period was estimated by the method for estimating the target walking period of the first embodiment. D : Initial stance phase time) and target walking cycle (W T The linear relationship of the first subject was W T =2.1×T D +680.75, the second subject was W T =2.1×T D +724.1 (see Figure 17). For the first subject, the accuracy rate of the estimated target walking cycle against the measured real walking cycle was r p is 95.54%, and for the second subject, the accuracy rate r p was 95.10% (see Figure 25).

[0069] [Example 2] In Example 2, the target walking period was estimated by the method for estimating the target walking period of the second embodiment. Assistance by the walking assist device 1 was started when the first three steps were completed. The accuracy rate r p was 93.34%, and the accuracy rate of the second subject, r p The accuracy was 93.36% (see FIG. 26). However, in Example 2, it was necessary to wait until the first three steps were completed before collecting data, and since the data collection cycle was long, a certain delay was observed.

[0070] [Example 3] In Example 3, the target walking period was estimated by the method for estimating the target walking period of the third embodiment. p is 96.47%, and the accuracy rate of the second subject r p The accuracy was 96.51% (see FIG. 27). It was confirmed that the delay in the case of the second embodiment could be eliminated in the third embodiment. However, as shown in FIG. 27, it was confirmed that the difference between the actually measured walking period and the estimated target walking period became large at the inflection point, where the walking period changes nonlinearly and unevenly.

[0071] [Example 4] In Example 4, the target walking period was estimated by the walking period estimation method of the fourth embodiment. p The accuracy rate of the second subject was 89.71%. p The accuracy was 91.92% (see FIG. 28). In Example 4, it was confirmed that the difference between the actually measured walking period and the estimated walking period was small in the portion where the walking period changes nonlinearly and nonuniformly (inflection point), but the difference between the actually measured walking period and the estimated target walking period was large in the portion where the walking period is constant (Walking cycle constant).

[0072] [Example 5] In Example 5, the target walking period was estimated by the method for estimating the target walking period of the fifth embodiment. p is 96.56%, and the accuracy rate of the second subject, r p The accuracy was 96.09% (see FIG. 29). In FIG. 28, the black dots indicate the first target walking period W TL The plot of the white circle is the second target walking period W TQ In the fifth embodiment, the second target walking cycle W TQ is selected, and in the part where the walking cycle is constant, the first target walking cycle W TL is selected.

[0073] Comparing the third and fifth examples, the accuracy rate r p Since it was difficult to evaluate which of the three subjects had the highest accuracy, we conducted a follow-up test on the third and fifth examples using three subjects with different heights and weights. As a result, the accuracy rate r p The average value of the accuracy rate r p The average value of was 96.14%. From the above, it was confirmed that the fifth embodiment can eliminate the differences between the actually measured walking period and the estimated target walking period that occurred in the third and fourth embodiments, and reduce the difference between the actually measured walking period and the estimated target walking period. [Explanation of symbols]

[0074] 1. Walking aids 16 Servo motor (actuator) 18 Walking motion detection unit 20 Drive control unit 181 Heel detection sensor 182 Toe detection sensor 211 Gait cycle estimation unit 212 Control target value generation unit

Claims

1. A walking assistance device that is attached to a leg of a walker and assists the walker in walking, comprising: an actuator that applies an assist force to the walking motion of the walker; a walking motion detection unit that detects the walking motion of the pedestrian; a walking cycle estimation unit that estimates a target walking cycle of the walker based on the walking motion of the walker detected by the walking motion detection unit; a control target value generating unit that generates a control target value of the actuator based on the target walking period estimated by the walking period estimating unit; A drive control unit that controls driving of the actuator based on the control target value generated by the control target value generation unit. Walking aid.

2. the control target value generation unit generates a control target value such that the operation of the actuator precedes the walking motion of the walker detected by the walking motion detection unit, and waits until the walking motion of the walker reaches the control target value. The walking assistance device according to claim 1 .

3. the control target value generation unit divides the target walking period estimated by the walking period estimation unit into a plurality of phases corresponding to a walking motion of the walker, and generates a control target value for the actuator corresponding to each phase. The walking assistance device according to claim 2 .

4. the actuator is composed of a servo motor that applies a rotational force to a joint position of a leg of the walker, the control target value generation unit calculates a target walking rate based on the target walking period estimated by the walking period estimation unit, calculates a target stride length according to the target walking rate, and calculates a target rotation angle amplitude value of the actuator as a control target value based on a relationship between the stride length of the walker measured in advance and a joint angle amplitude value of a joint position of a leg of the walker. The walking assistance device according to claim 3.

5. The walking cycle estimation unit estimates a target walking cycle according to a time from a contact of a heel of the walker with the ground to a contact of a toe of the walker with the ground, which is detected by the walking motion detection unit. The walking assistance device according to any one of claims 1 to 4.

6. The walking cycle estimation unit estimates a target walking cycle by averaging a plurality of walking cycles detected by the walking motion detection unit. The walking assistance device according to any one of claims 1 to 4.

7. The walking cycle estimation unit estimates a target walking cycle by performing a linear approximation of a relationship between the number of steps of the walker detected by the walking motion detection unit and a plurality of walking cycles at each number of steps. The walking assistance device according to any one of claims 1 to 4.

8. The walking cycle estimation unit estimates a target walking cycle by performing curve approximation of a relationship between the number of steps of the walker detected by the walking motion detection unit and a plurality of walking cycles at each number of steps. The walking assistance device according to any one of claims 1 to 4.

9. the walking cycle estimation unit compares a first target walking cycle obtained by performing linear approximation of a relationship between the number of steps of the walker detected by the walking motion detection unit and a plurality of walking cycles at each number of steps with a second target walking cycle obtained by performing curve approximation, selects a walking cycle that is closest to the walking cycle detected immediately before, and estimates the target walking cycle. The walking assistance device according to any one of claims 1 to 4.

10. An actuator that applies an assist force to the walking motion of a walker; A drive control unit that controls the drive of the actuator; a walking motion detection unit that detects the walking motion of the walker, and controls a walking assist device that is attached to the leg of the walker, a walking cycle estimation unit that estimates a target walking cycle of the walker based on the walking motion of the walker detected by the walking motion detection unit; a control target value generating unit that generates a control target value of the actuator based on the target walking period estimated by the walking period estimating unit; a control target value output unit that outputs the control target value generated by the control target value generation unit to the drive control unit, Control device.

11. An actuator that applies an assist force to the walking motion of a walker; A drive control unit that controls the drive of the actuator; a walking motion detection unit that detects the walking motion of the walker; and a computer-readable control program executed on a computer that controls a walking assist device that is attached to the legs of the walker, The computer, a walking cycle estimation unit that estimates a target walking cycle of the walker based on the walking motion of the walker detected by the walking motion detection unit; a control target value generating unit that generates a control target value of the actuator based on the target walking period estimated by the walking period estimating unit; a control target value output unit that outputs the control target value generated by the control target value generation unit to the drive control unit; Control program.

Images