JP6645658B2 - Mobile training support device - Google Patents

Description

  The present invention has processing means for generating information necessary for guiding a user to a desired walking posture or running posture, and outputs information output by the image, sound, and stimulation to the user. The present invention relates to a mobile training support device provided with means for presenting as a training.

  Furthermore, the present invention, a running body independent of the training subject, and mounted on the running body, enhance the movement recognition of the training subject, measurement, evaluation, presentation, guidance, to efficiently guide the walking motion The present invention relates to a mobile training support device including a reproducing unit.

  In Patent Document 1, the load torque and the walking speed of an annular belt serving as a walking surface are measured in a normal walking state, and the rotational speed and the load torque of the annular belt are changed based on the measured torque and the walking torque according to the physical condition of the trainee. A walking training device for performing walking training and fall prevention training is shown.

  Patent Document 2 discloses that a walking pattern generation unit performs training based on the trainee's joint angle and the trainee's lower limb length at any point in a basic walking pattern stored in advance in a walking pattern data storage unit. The position of the lower limb of the trainee is calculated, the force sensor data measured by the force sensor is changed according to the position of the lower limb of the trainee, and the training is performed based on the position of the lower limb of the trainee and the changed force sensor data. The controller determines the command value to the lower limb driving unit attached to the lower limb, and in the controller, drives the lower limb driving unit according to the command value to measure the floor reaction force at the time of standing walking. A walking training device for training walking by a joint operation of joints according to the walking phase of the walking is shown.

  Patent Literature 3 discloses a walking training support device having a supporting function of supporting a patient and a function of displaying a foot position of a patient, and the like. Running a walking assist device at a constant speed according to the patient, illuminating the target position of the next step of the patient with light or generating a sound that takes a walking rhythm or displaying an image at a fixed time interval in front of the patient Having a photographing foot position display device for photographing the patient's foot position and displaying it in front of the patient together with the target foot position, and detecting and detecting the horizontal position of the patient's foot It is shown that a detection foot position display device for displaying the foot position together with the target foot position in front of the patient is provided.

  Patent Literature 4 has a notification unit that provides sound or light guidance to a trainee in accordance with the walking distance of the trainee, and motivates and trains the trainee by encouraging voice or shouting. There is disclosed a walking training device that makes the sound rhythmic, excites the trainee, and prevents the trainee from getting tired of the training.

  Patent Literature 5 discloses a plurality of walking surfaces including an endless track belt on which a trainee can walk, a driving unit that drives these walking surfaces, a measurement unit that measures the trainee's walking function or biological information, Setting means capable of setting operating conditions of the walking training machine for setting training conditions, information processing means for calculating information of the measuring means, information measured by the measuring means and the set walking training set There is shown a walking training device that includes a display unit that displays the operating conditions of the machine and the information calculated by the information processing unit in association with each other, and that can improve the results of training.

  Patent Literature 6 discloses an image capturing device that captures a walking image of a trainee walking on a walking training machine, an image arithmetic device that performs arithmetic processing on an image of the image capturing device, and a display that displays the arithmetic result of the image arithmetic device. Device, and a storage device for storing the calculation result. The image calculation device calculates the floor contact position of each foot and the temporal change of the floor contact position for each foot from the image captured by the image calculation device. Walking with a footprint analyzer that can grasp the training situation quantitatively by displaying the time-dependent change of the surface contact position for each foot and compare it with past data to apply the training data to a diagnostic device A training device is shown.

  Patent Document 7 discloses a footprint analyzer similar to that of Patent Document 6, in which the image calculation device determines a floor contact position and a floor contact position of each foot from a captured moving image and / or a static image that does not change. It is shown that the position that does not touch the surface is calculated, and the display device displays the floor contact position of each foot and / or the position that does not touch the floor.

  Patent Literature 8 discloses a hand load measuring device that measures a load applied from a pedestrian's hand in a walker that assists walking by supporting a pedestrian's body, and each foot mounted on a pedestrian's foot. Provide a foot load measurement device that measures the weight of the pedestrian and a foot behavior measurement device that measures the behavior of the pedestrian's foot, and perform walking training while recognizing the current state of walking and checking the degree of recovery of the legs and hips. It is shown.

  Patent Literature 9 discloses an actual step length deriving unit that derives an actual step length, which is an actual step length of a user who walks on a pseudo walking path, in a walking training apparatus having a belt-type pseudo walking path, and that the user performs walking training. Evaluation of walking ability in walking training, comprising target step deriving means for deriving a target step as a recommended value of the step in performing the step, and step stride comparing means for comparing the derived actual step and the target step to the user. Is shown to do.

  Patent Literature 10 discloses a walking support device that integrates with a training instructor, instructs walking by using video projection, and guides walking, in a self-propelled bogie that functions as a movable walking stick that supports during walking. It is shown.

JP 2008-036054 JP 2005-210186 A JP-A-2003-164544 JP 2003-024469 JP-A-2002-345994 JP-A-2002-345785 JP-A-2002-345784 JP-A-2001-276155 JP-A-2001-238982 JP 2011-229838A

  In the walking training support device proposed in the prior art, an installation-type walking training machine using a parallel bar or a belt driving device as disclosed in Patent Documents 1, 4, 5, 6, 7, 9 and the like has a large facility. And there are many restrictions on the installation conditions.

  Patent Documents 2, 3, and 8, which use a walking support device that individually supports a trainee, walk in a normal walking environment where ordinary persons other than the trainees pass indoors and outdoors. Technical support for training (shape of walkway, ensuring safety, opening and closing doors, avoiding objects on walkways, etc.) has not been fully realized. In addition, the walking support machine type walking training machine has many restrictions on the installation position of information transmission equipment using sound and light signals to the walking trainee and the installation position of sensors such as measurement of the walking state of the trainee, and is functionally limited. There is a drawback that this is often done. In addition, since a support device is required for each trainee, it is difficult to perform walking training together by a plurality of people.

  Furthermore, in the walking training support devices proposed in Patent Literature 3, Patent Literature 4, and Patent Literatures 5 to 7, the trainee indicates the target position of the next step with light, or sounds encouraging voice or rhythm sound. It is shown that the training effect can be enhanced by displaying the training effect from the individual information of the walking trainer or the actual measurement situation. It cannot be said that a mechanism for enhancing the training effect by enhancing the trainee's own motor cognition or psychologically motivating to exercise has been sufficiently studied.

  Specifically, it is difficult for a gait training device that presents an image on a monitor to present the trajectory of the trainee in the walking environment, the past and present walking trajectory, foot movement, and future stepping position on the walking surface in actual size. It is. In addition, a walking training device that presents sound using earphones or speakers can present the landing timing and walking pace of the foot in a walking environment, but a device without a walking environment recognition mechanism is suitable for a walking environment. It is difficult to give a turn stop timing instruction or to warn of the presence of an obstacle.

  In any case, the walking training program of the related art has a poor ability to adapt to the actual training state of the trainee, and thus has little effect of increasing the motivation to exercise by dynamically recognizing the movement of the trainee.

  In the prior art of Patent Document 3, gait training especially for Parkinson's disease patients is also proposed. In Parkinson's disease, freezing phenomena, difficulty in starting walking, and forward lunging phenomena appear. In response to this, a mechanism has been proposed that reproduces the stored movements to run a walking indicator, instructs walking by light irradiation and sound generation, and measures and displays foot position and sole pressure. ing. However, the foot position measurement device disclosed herein does not consider enhancing the recognition of landing on the sensory organs of the foot in order to enhance the training effect. In addition, since motion recognition is limited to the feet, consideration should be given to assisting and instructing motion recognition in coordination with walking posture and balance of the whole body, upper limb movement such as opening and closing doors and grasping objects. It has not been.

  In the walking support device proposed in Patent Literature 10, a walking stick is integrated with a training instructor, and footing is instructed using video projection to guide walking. In this technique, it is difficult to use the training target with another supporting device such as a parallel bar due to the integration of the training target and the device, and the behavior control is subordinate to the training target. It is difficult to measure whole-body motion and present and instruct a wide range of motion state using a walking environment.

  To summarize these, the prior art has the following problems. (1) No consideration is given to guiding to an arbitrary walking route or ensuring the safety of the route. (2) It is difficult to present the timing of turning stop suitable for the walking environment and to warn of an obstacle. (3) Training (4) Training according to the trainee's situation, because it is difficult for the trainee to adapt to walking, and to improve the walking style based on the intention of the trainee to move and the current walking form without difficulty. (5) It is difficult to support or instruct the movement recognition of the whole body in walking posture and balance, coordination with the upper limb movement, because the movement recognition is limited to the feet, (6) ) The use of a walking training device and other supporting devices is not considered.

In order to solve the above-mentioned problem, the present invention is configured as follows. The invention according to claim 1 is a depth image acquisition unit that acquires a depth image of a whole body of a trainee who performs walking or running training from an arbitrary acquisition position,
A three-dimensional position acquisition unit that calculates a distance from the acquisition position in the plurality of body parts of the trainee based on the depth image to obtain a three-dimensional position of the body part, and a three-dimensional position of the body part. Processing means for measuring the kinematic whole body movement of the trainee based on the change and calculating a guidance position of a body part necessary for guiding the walker or running of the trainee based on the measurement result And video presentation means for visually providing the guidance position of the body part calculated by the processing means.

  The invention of claim 2 is the mobile training support device according to claim 1, further comprising an acceleration sensor attached to a body part of the trainee, wherein the processing unit performs the training based on information of the acceleration sensor. Measuring the kinematic whole body movement of the trainee in the body part of the trainee, and calculating the guidance position of the body part together with the three-dimensional position.

  According to a third aspect of the present invention, there is provided the movement training support device according to the first or second aspect, further comprising an angular velocity sensor and a geomagnetic sensor attached to a body part of the trainee, and wherein the processing unit includes information of the acceleration sensor. In addition, based on the information from the angular velocity sensor and the geomagnetic sensor, the whole body movement amount of the trainee, the movement distance of the body part, and the stride are calculated.

  According to a fourth aspect of the present invention, in the mobile training support device according to any one of the first to third aspects, the depth image acquiring unit, the three-dimensional position acquiring unit, the processing unit, and the video presenting unit include It is mounted on a runnable traveling body, the traveling body further includes a driving unit configured by driving wheels, a control unit that controls a driving state of the driving unit, a corridor in a facility or an outdoor. Detecting means for detecting the surrounding environment with a laser sensor, such as obstacle information in a pedestrian passage, the control means, based on the distance information with the trainee detected by the detecting means, The driving unit is controlled so that the appropriate interval is maintained.

  According to a fifth aspect of the present invention, in the mobile training support device according to the second or third aspect, the image presentation means is mounted on a self-propelled traveling body, and the depth image acquisition means, The three-dimensional position acquisition unit and the processing unit are fixedly installed in the training facility independently of the traveling body, and the traveling body further includes a driving unit configured by driving wheels, Control means for controlling the driving state of the drive unit, detection means for detecting the surrounding environment, second processing means for processing information in each means mounted on the traveling body, the processing means, Transmission means for transmitting and receiving various information to and from the second processing means, wherein the control means maintains an appropriate interval based on distance information from the trainee detected by the detection means Control the drive so that A shall, said processing means is characterized in that it is intended to process the information of the second processing unit information and various sensors mounted on the body part of the trainee.

  According to a sixth aspect of the present invention, in the mobile training support device according to the fourth or fifth aspect, the traveling body includes storage means for storing environmental map information, and the control means is configured to execute the control based on the environmental map information. Controlling the drive unit to follow and drive the trainee, wherein the processing means calculates the position of the trainee in the environmental map information, and the video presenting means Is characterized in that the calculation result is presented as an image.

  According to a seventh aspect of the present invention, in the mobile training support apparatus according to the sixth aspect, the position of the trainee in the environmental map information calculated by the processing means is an individual position of a specific part of the trainee. A set of information, the individual position information of the specific part is environmental map information, the position information of the mobile in the environmental map information, and the relative positional relationship from the mobile to the specific part. It is characterized by being calculated based on

  According to an eighth aspect of the present invention, in the mobile training support device according to the sixth or seventh aspect, the storage means of the moving object stores a route for guiding the trainee, and the guidance route includes: It is characterized in that the route information is artificially created planned route information or route information created by the processing means based on position information in the environmental map information on a route in which the traveling body has been moved in advance.

  According to a ninth aspect of the present invention, in the mobile training support device according to any one of the first to eighth aspects, the depth image obtaining means further obtains a depth image of a walking support device used by the trainee. Wherein the three-dimensional position obtaining means obtains a three-dimensional position of the walking support device, and the processing means determines a distance to a feature point for the walking support device based on the change in the three-dimensional position. The information is calculated, and the video presenting means visually presents a guidance position where the walking support device should be moved together with the guidance position of the body part.

  According to a tenth aspect of the present invention, in the mobile training support device according to any one of the first to ninth aspects, the depth image obtaining means further obtains a depth image of the whole body of the training instructor using an infrared depth sensor. The three-dimensional position acquiring means calculates the distance from the acquired positions of the plurality of body parts of the training instructor to acquire the three-dimensional positions of the body parts.

  According to an eleventh aspect of the present invention, in the mobile training support device according to the tenth aspect, the evaluation means for evaluating the foot exercise and / or the body balance of the trainee based on the three-dimensional position of the body part is further provided. It is characterized by having.

  According to a twelfth aspect of the present invention, in the mobile training support device according to any one of the first to eleventh aspects, the processing means includes: The present invention is characterized in that a guidance position of a body part necessary for guiding the travel is calculated.

  According to a thirteenth aspect of the present invention, in the movement training support device according to the twelfth aspect, the guidance position calculated by the processing means is an ideal movement state of the body part during walking or running of the trainee. And an intermediate position between the actual state of movement of the body part at the time of walking or running of the trainee, and depending on the degree of difference between the ideal movement of the body part and the movement of the real body part. One or more guidance positions are calculated.

  According to a fourteenth aspect of the present invention, in the mobile training support device according to the thirteenth aspect, the video presenting unit is configured to provide the image presenting unit together with or separately from the guidance position when the trainee actually walks or runs. The moving state of the body part at the time is presented as an image.

  According to a fifteenth aspect of the present invention, in the mobile training support device according to any one of the first to fourteenth aspects, the mobile training support apparatus further includes a sound presentation unit that presents an auditory stimulus to the trainee. The means is for creating acoustic information to be presented in accordance with the temporal timing of providing a guidance position to the trainee, and the sound presenting means is configured to present the voice or sound selected by the trainee to the video presentation. It is characterized by being presented instead of or together with the video presentation by means.

  The invention according to claim 16 is the movement training support device according to claim 15, further comprising an acceleration sensor attached to the trainee's foot, wherein the processing unit is configured to perform processing based on information from the acceleration sensor. Outputting a signal to operate the sound presenting means when it is determined that the trainee's foot has landed, and recognizing that the trainee is in a landing state by operating the sound presenting means. Characterized in that:

  Further, the invention of claim 17 is the movement training support device according to claim 16, further comprising a stimulus presentation means attached to a part of the body of the trainee to provide a tactile stimulus, The processing means outputs a signal for activating the stimulus presentation means when the landing is confirmed from the state of the foot by the acceleration sensor.

  According to an eighteenth aspect of the present invention, in the movement training support device according to the seventeenth aspect, the stimulus presenting means is mounted on a plurality of body parts of the trainee and presents a haptic stimulus by vibration. Preferably, the processing means outputs a signal for operating the body part to be induced among the body parts of the trainee so as to present a haptic stimulus by vibration.

  According to a nineteenth aspect of the present invention, in the mobile training support device according to any one of the first to eighteenth aspects, an optical camera for photographing the walking or running state of the trainee is further provided, and the image presentation is performed. The means is for presenting a still image or a moving image taken by the optical camera instead of or together with the guidance position.

  According to a twentieth aspect of the present invention, in the movement training support device according to the nineteenth aspect, the optical camera captures an image of the face of the trainee, and the processing means includes a muscle on the face of the trainee. The change of the muscle is measured and compared with the change information pre-classified from the change in the muscle to determine the continuation, suspension, promotion or suppression of the training, and the video presenting means determines the content of the determination. It is characterized by presenting a visually converted video.

  The following effects can be obtained by the present invention. (1) Walking training (walking training described below includes running training) can be supported at any place where the trainee can walk. (2) By mounting equipment necessary for walking training on a self-propelled traveling body, the physical load on the trainee is reduced. (3) By guiding the walking route independently of the trainee, it is possible to ensure the safety of the walking route and to select a walking environment suitable for training. (5) By presenting a signal to a walking trainee based on recognition of the walking environment, it is possible to present a timing of turning stop suitable for the walking environment and to warn of an obstacle. (6) With the adaptive function of the walking trainer to walking, It is possible to implement a walking training based on the intention of the trainee's movement and a training method for sequentially improving the current walking form. (7) Direct stimulation of the trainee's sensory organs by stimulation according to the trainee's situation (8) By recognizing whole-body movement, it is possible to support and instruct movement recognition for the walking posture and balance of the whole body, and coordination of walking and upper limb movement. (9) Since the self-propelled running body functions independently of the trainee, walking training can be performed irrespective of a walking support device such as a brace, a walking stick, or a support device used by the trainee. (10) By recognizing the walking state of the training instructor and the walking trainer together, it is possible to guide the walking posture based on the training intent of the training instructor and to evaluate the movement of the walking trainer in accordance with the training intention.

It is a mobile training support device of the present invention (when taking the form of a running body). It is a use state of this invention. 1 is an electrical component configuration of the present invention. 3 is a flow of a process of the present invention. It is detection of a moving state by the depth image of the present invention. 3 is a diagram illustrating a state recognition and a path plan of a mobile environment according to the present invention. It is a situation of stimulus presentation of the present invention. 5 is an image and sound effect of the present invention. 9 is an experimental example of moving state detection by acceleration measurement according to the present invention. This is a procedure for realizing the extended mirror image therapy. It is an image of an extended measuring means. It is three-dimensional measurement of walking in the embodiment. It is an image of an extended evaluation means. It is an image of an expanding presentation means. It is an image of an extended guidance means, and an environment map, a teaching route, and a reproduction route learned in the embodiment. It is an image of an expansive reproduction means and an extended body image presented on a wall surface or a floor surface in the embodiment. It is a presentation of an evaluation amount of a walking state at the time of reproduction. It is a spatial arrangement of the traveling body in the embodiment. This is the presentation of a video pattern at the time of guidance.

  As an embodiment of the mobile training support device according to the present invention described below, the trainee (104) walks and runs while moving independently and independently of the trainee (104) and the training instructor (105). Using a traveling body (hereinafter referred to as a traveling body (100)) equipped with a device for supporting the training of the moving operation of a vehicle in an environment (hereinafter referred to as an environment system (300)) having an equivalent training support function. It presents what is to be trained, and is further attached to a trainee (104) or a trainer (105) or a mobile support device (106) to extend the motion measurement function and stimulus presentation function of the running body and the environmental system. The device (hereinafter referred to as the mounting unit (200)) is presented.

  Hereinafter, the best mode for carrying out the invention will be described with reference to the drawings. Note that the structure and shape of the illustrated device are merely examples, and the present invention is not limited to the illustrated device.

  FIG. 1 is a view schematically showing a traveling body (100) according to one embodiment of the present invention. The traveling body (100) is realized as an autonomously movable robot equipped with components indicated by reference numerals in FIG. In other words, the vehicle can travel with two drive wheels (125, 126) and one driven wheel (127), and can move forward and backward using the drive wheels (125, 126) as front wheels. The drive wheels (125, 126) can be steered, and the direction can be freely changed by steering the drive wheels (125, 126). In the main body of the traveling body (100), an infrared sensor (111), an optical camera (112), a laser sensor (113), two microphones (115, 116), an attitude sensor (118), a projector ( 121) and a speaker (122) are selectively mounted. A rear laser sensor (114) is selectively mounted on the rear side of the main body, and a contact sensor (117) is also selectively mounted on the main body in such a manner as to completely surround the main body in the horizontal direction. is there. An image projected from the projector (121) is projected on a floor or the like to form a projection surface 101.

  FIG. 2A shows a state of the movement training when the traveling body (100) is used, and FIG. 2B shows a state of the movement training when the environment system (300) is used. In these mobile trainings, a state where the mounting unit (200) is mounted on the trainee (104) is illustrated. The running body (100) recognizes a person or an environment as shown in FIG. 2 (a), and moves ahead of the trainee while dynamically presenting the walking state and instruction contents by video and sound. Thus, the trainee has the ability to appropriately guide the trainee to walk while ensuring the safety of the moving route. As shown in FIG. 2B, the environment system (300) is a form in which a function equivalent to that of the mobile object (100) is realized using a sensor or a communication network embedded in a pedestrian training environment. The mobile unit (100) and the environmental system (300) can be used in combination, or one of the two can be used alone.

  FIG. 3 shows the configuration of the electrical unit of the mobile training support device (302) of the present invention. The traveling body (100) has a detection unit (detection means) (135) for detecting a state of a person or an environment, and includes an infrared sensor (111), an optical camera (112), a laser sensor (113, 114), a left and right microphone. (115, 116), a contact sensor (117), a posture sensor (118), and the like. As shown in FIG. 1, the traveling body (100) includes a control unit (control means) (134) for generating a drive command, left and right front wheel motor units (123, 124) having an articulated link mechanism, and drive wheels (125). , 126) and a driving unit (driving device) (133) composed of passive wheels (127). The traveling body (100) further includes a projector (121) and a speaker (122) as a presentation unit (sound and video presentation means) (136) for generating a video and audio effect as shown in FIG. The traveling unit (100) includes a calculation unit (processing unit) (133) that integrates recognition of the state of a person or an environment and an action plan, and a communication unit (132) that transmits information to the inside or outside of the traveling unit (100). (Including a communication network and a wireless device). Further, the traveling body (100) has a power supply unit (131) for supplying electric power to these electrical units, and includes a DC converter for an external electric AC source and an internal DC power supply.

  The mounting unit (200) has a motion measuring unit (201) that measures the motion of a trainee, a training instructor, or a walking support device, and includes an acceleration sensor, an angular velocity sensor, and a geomagnetic sensor. The mounting unit (200) has a stimulus presentation section (202) for presenting a stimulus to a trainee or a training instructor, and includes a vibration motor and an electric stimulator.

  The external computing device (301) is provided as necessary, and is provided in a computing unit (133) installed in an environment for mobile training or in a remote location (such as a control room of a nursing / training facility) or an external computing device. It has a power supply unit (131) for supplying power and a communication unit (132) for transmitting information to and from an external computing device.

  The components of the above-mentioned traveling body are not limited to the form of the traveling body, and can be laid in the environment of the movement training. The environmental system (300) is a system in which the running body (100) and the external computing device (301) are laid in the environment, and the combination of the environmental system (300) and the mounting unit (200) is the movement training of the present invention. A support device (302).

  The main body of the traveling body (100) is provided with a detection unit (135) for measuring a person or an environment. The detection unit includes an infrared sensor (111), an optical camera (112), front and rear laser sensors (113, 114), a contact sensor (117), a posture sensor (118), and left and right microphones (115, 116). The infrared sensor (111) measures the whole body movement of a person such as a trainee or a training instructor. The optical camera (112) measures images such as trainees and walking environments. The laser sensors (113, 114) measure a spatial structure of a walking environment including a leg of a person such as a trainee or a training instructor or an obstacle on a walking path. The contact sensor (117) measures contact with an obstacle or a person. The posture sensor (118) measures the posture of the running body or the end of an internal joint link. The microphones (115, 116) listen to the instruction content and the environmental sound by the voice of the trainee or the training instructor.

  The traveling body (100) is provided with a driving mechanism, and includes left and right motor units (123, 124) for controlling the rotation of the left and right driving wheels (125, 126). The motor units (123, 124) include an encoder for measuring the amount of rotation of the motor and a motor driver for controlling driving of the motor. The control unit (134) transmits a driving command to the motor driver, and the motor driver controls driving of the motor unit. The rotation amount of the motor measured by the motor unit is transmitted to the control unit via the motor driver. The main body of the traveling body (100) includes driving wheels (125, 126) to which power is transmitted and a passive wheel (127) that supports the main body and rotates passively. The optical camera (112) and the projector (121) of the presentation unit are provided with a joint link mechanism for dynamically changing the posture. The posture of the joint link mechanism is controlled by the motor units (123, 124), similarly to the drive wheels (125, 126).

  The traveling body (100) is provided with a power supply device (power supply unit) (131), and includes a power supply switch, an AC / DC converter, and a rechargeable battery. The traveling body (100) can be supplied with power from a rechargeable battery mounted on the traveling body (100) and with power from an external power supply (not shown) via a power cable and a DC / AC converter. Can be switched by a power switch. A small DC power supply such as a lithium polymer rechargeable battery is used for the rechargeable battery.

  The traveling body (100) and the external computing device (301) include a computing device (calculating unit: processing means) (133). The calculation function of the traveling body (100) is realized by a small device such as an on-board board PC or a small microcomputer for weight reduction. The calculation device (133) of the traveling body (100) performs low-order information processing (such as input / output data processing of sensors and motors) and high-order information processing (recognition of a walking state, an environmental state, an appliance state, and a walking route). Planning, generation of audiovisual effects, etc.). The higher-level information processing is performed by transferring data to the outside via a communication device (communication unit) (132) and a work, instead of using the computing device (133) mounted on the traveling body (100). , Can be performed by the external computing device (301).

  The calculation processing of the running body (100) is modularized by independent processes communicating with each other for each functional unit. These calculation processes are executed not only by the traveling body (100) but also by the mounting unit (200) and the external computing device (301) described later, and are linked to each other, so that the movement training realized by the traveling body alone is performed. The computing power of support can be expanded.

  FIG. 4 shows each flow for operating this embodiment, and a flowchart (401) of FIG. 4A shows a flow of processing of the traveling body (100) or the entire environmental system (300). , (B) shows a flowchart (410) for detecting a measurement subject (trainer or the like), and (c) shows a flowchart (420, 420) when the use form of the traveling body (100) is classified into three types. 421, 422), and (d) is a flowchart (423) for driving the drive wheels (125, 126) and the joint link mechanism by the control unit (control means), and a flowchart for displaying a video / audio pattern ( 424). Therefore, each configuration will be described in detail with reference to these flowcharts.

  As shown in FIG. 4 (a), the running object (100) or the environmental system (300) includes an operation mode in which the trainer or the training instructor teaches the movement route in the environment of the driving training, and an operation mode of the trainer. It has an operation mode to guide the trainee while re-planning the movement route taught based on the walking state and the current state of the environment, and is selectively set by the user (trainer or training instructor) at the start of the operation. . In the following description, the function of the traveling body (100) will be mainly described.

  The trainee or the training instructor selects teaching or guidance of the mobile training for the traveling body (100), and makes the traveling body (100) function. In any of the teaching and guidance operation modes, the traveling body (100) detects the state of a person, an environment, equipment, and the inside of the traveling body by various sensors (402). Based on the detection result, the traveling body (100) recognizes the state of the person, the environment, and the equipment (recognition of the running state of the trainer or the training instructor, recognition of the state of the running environment, recognition of the state of the running support equipment, and the like) Then, calculation processing (processing means) such as an action plan of the traveling body (a movement route plan and a posture change plan of the joint link) is performed (403). Based on the recognition and the plan, the traveling body performs drive control such as an action command (movement control, joint link posture control) (404) (execution of control means), and gives a trainer or training instructor. The audiovisual pattern is generated and the audiovisual pattern is presented (405) (execution of presentation means). The traveling body repeatedly performs a series of these processes until the end of the operation is selected.

  The above-described overall processing (401) is performed not only when the traveling body (100) is used but also in the environment system (300) (FIG. 2B) in which these elements are embedded. However, in that case, the movement plan of the traveling body itself and the execution of the drive command are omitted.

  The moving state of the trainer or training instructor (hereinafter referred to as a measurement target) is determined by the detection unit (including the human measurement unit, the environment measurement unit, the image measurement unit, the sound measurement unit, and the contact / posture measurement unit). Measured. In the detection of the moving state, a moving speed of the person to be measured, a time interval between landings of the legs, and a moving distance per one step (step length) are detected as the characteristic amounts representing the moving state. The procedure (410) shown in FIG. 4B represents the flow of the detection of the moving state.

  The detection of the movement state of the measurement target includes a processing procedure (411) of a method using depth image measurement of the infrared sensor (111) and a processing procedure (412) of a method using acceleration measurement of the mounting unit. When (106) is used, the processing procedure (413) of the method of detecting the moving state of the instrument and correcting the depth image is used.

  The detection of the moving state can be performed in any one of the processing procedure (411) using the depth image and the processing procedure (412) using the acceleration measurement in three dimensions of the body part (mainly the body) of the measurement target. The position and speed amplitudes are measured (S-11, 21), the moving speed is calculated (S-12.22), the landing is detected (S-13-23), and the landing time interval is measured (S- 14.24). The landing time interval is the time interval between the landing time and the previous landing time. In addition, at the start of movement (first step) or when the vehicle has been stopped for a certain period of time or more, the last landing time is set to the time of the immediately preceding movement, and the landing time interval is calculated. The distance moved (the amount of movement) during the landing time interval is calculated (S-15, 25), and is set as the stride. Finally, the step length is accumulated to calculate the movement distance from the start of the movement (S-16, 26).

  The processing procedure (411) of the method based on depth image measurement has the advantage that it is not necessary to attach the device unit (200) to the person to be measured, and the movement of the person to be measured is limited to the field of view of the infrared sensor (111). Is a disadvantage. On the other hand, the method using acceleration measurement (412) is advantageous in that the movement of the measurement target is free within the transmission area of the mounting unit (200) of several tens of meters. The disadvantage is that it must be worn. Further, these methods can be used in combination. For example, the landing can be detected by the method of the procedure (412), and the movement amount during that can be measured from the position information of the method of the procedure (411).

  In the processing procedure (411) of the method using the depth image measurement, the position of the body part of the trainee is detected from the measured depth image. FIG. 5 shows the situation of the person to be measured on the depth image. By using existing infrared sensors (such as Microsoft's Kinect (TM) and ASUS's Xtion PRO LIVE (TM)), body images (for example, head, neck, shoulder x2, elbow x2, hand x2) can be obtained from depth images. , Torso, waist, thigh x2, knee x2, foot x2). These three-dimensional positions are corrected for the influence of a change in the position or posture of the infrared sensor (111) due to the movement of the running body (100) or the like, so that an absolute coordinate system with respect to the moving environment (for example, a position with a walking environment as an origin) , A coordinate system having one perpendicular axis and two perpendicular axes on the floor surface). The reference point of the trainee is the position of the torso extracted from the depth image. Assuming that the movement vector in the horizontal plane of the trunk is r and the velocity vector is v, v is estimated by Equation 1.

Here, t is the current time, and δt is an appropriate short time interval. The detection of the landing in the processing procedure (410) of FIG. 4B is performed when the vertical component of the three-dimensional position of one foot takes the minimum value at an appropriate time interval before and after the current time t. Let it be t. The measurement noise of the vertical component x is reduced by noise processing such as a low-pass filter. Time interval t _c -t _p of the detected last landing time t _p and t of the current landing is the time interval of the landing, the moving distance δr between them is estimated by the number 2, defined as the stride this I do.

The movement state of the trainee can be measured in more detail by integrating information of the infrared sensor (111) and the optical camera (112). The optical image of the optical camera (112) and the depth image of the infrared sensor (111) are integrated by performing positioning of the measurement region in advance, and the depth image and the optical image are used for region extraction and texture extraction, respectively. The region (505) in FIG. 5 is a region determined based on the feature points of the skeleton model suitable for the subject in the depth image. By extracting a local image corresponding to this region (505) from the optical image, Still images and moving images of body areas such as faces, hands, and feet can be acquired. When the trainee uses the moving support device, an image of the moving support device (106) can be extracted by the device state recognition described below, and the depth image in which the area of the moving support device (106) has been corrected is used as the moving state. By using the detection procedure (411), a person can be extracted without the influence of the moving support device (106).

  The kinematic state of the whole body and the appearance of the image obtained here can be used for recognizing exercise and instructing exercise to a trainee during walking training. It is useful for kinematic analysis such as posture of the whole body during walking and coordination between walking and upper limb movement. By performing facial expression recognition on the acquired facial expression, the mental state at the time of walking training can be estimated. The recognition of the walking state may be performed by a plurality of persons, and the walking state of both the trainee and the training instructor such as a physiotherapist can be recognized as described later.

  In the procedure (412) of the method using the acceleration measurement illustrated in FIG. 4B, the position of the body part of the measurement target is detected from the acceleration measured by the mounting unit (200). The mounting unit (200) measures the motion of the body part using the acceleration sensor, and transfers the data to the traveling body (100) or the external computing device (301) via the communication device. Although the motion measurement of the measurement subject can be performed by an infrared sensor (111) or an optical camera (112) mounted on the traveling body, it is difficult to directly detect the landing on the floor with these visual sensors. . On the other hand, when the mounting unit (200) is mounted on the person to be measured, the landing can be directly detected by detecting the impact generated by the landing. The acceleration sensor is not limited to landing, and can also detect an impact on the arm, the moving support device (106), and the like due to physical contact.

  By providing the mounting unit (200) with an angular velocity sensor and a geomagnetic sensor in addition to the acceleration sensor, the acceleration of the mounting part can be measured without depending on the attitude of the acceleration sensor. Based on the measured value of the acceleration, the moving amount, the stride, and the moving distance of the person to be measured in the procedure (412) can be calculated.

First, the acceleration of the body part wearing the sensor is measured from the acceleration value measured by the acceleration sensor. Similarly to the above, the reference position of the measurement target is set as the torso of the person to be measured, and the mounting unit (200) is mounted on the torso. Since the measurement value of the acceleration sensor of the mounting unit (200) is a value of a coordinate system unique to the acceleration sensor, the posture is estimated by using the acceleration sensor, the angular velocity sensor, and the geomagnetic sensor in combination, and the movement is performed. Correction to the measurement value in the coordinate system for the environment. To estimate the posture, a Kalman filter is used for these sensor measurement values. As described above, the acceleration a excluding the gravitational acceleration in the coordinate system for the moving environment is obtained. Next, speed estimation is performed. The speed and the distance are obtained by repeating the integration of the acceleration a, but have a value including an error. Therefore, the error is reduced by utilizing the fact that the magnitude of the speed variation correlates with the magnitude of the speed in the traveling direction. The velocity amplitude v _i ′ (indicating i = 1, 2, 3 represents the traveling direction in the horizontal plane, the direction perpendicular to the traveling direction in the horizontal plane, and the vertical velocity component) is calculated by Expression 3. You.

Here, a _i (i = 0, 1, 2) represents an acceleration vector corresponding to v _i (i = 0, 1, 2). v _i ^a represents the short-term average of acceleration. ΔT represents a time interval at which short-time averaging is performed.

A method of estimating the moving speed from the component of the moving direction of the speed amplitude (hereinafter referred to as a conventional method) has been proposed. In the present invention, components of all directions of the moving (moving direction, moving in the horizontal plane) are proposed. Focusing on the fact that there is also a correlation between the velocity amplitude and the traveling direction velocity for each component (direction perpendicular to the traveling direction, vertical direction), the present invention estimates the moving speed using v ′ having these three components. I do. The estimated value v ^e of the moving speed is obtained by Expression 4.

Here, f is a nonlinear function or a linear function having the parameter set W. The nonlinear function can be implemented by a higher-order polynomial, a multi-layer perceptron, or the like. The function f can be obtained by optimizing W so as to satisfy the input / output data collected in advance. The moving distance δr (t) between the time 0 and the time t is obtained by Expression 5.

However, v 'is the moving direction of travel of the v _i. Assuming that a time having a peak value with respect to the time change of the posture-corrected acceleration a is a landing time, a landing time interval is calculated by the calculation of the moving state recognition, and the above-mentioned δr Step length is obtained by performing integration. Furthermore, the moving distance r during the moving training is obtained by adding the stride from the start of the training according to Equation 6.

Here, δr _i represents the movement distance (step length) from t _i−1 to t _i , t ₀ is the movement start time, t _N is the movement end time, and N is the number of steps from the movement start to the movement end.

  Since the vehicle (100) and the environmental system (300) of the present invention function physically independently of the mobile support device (106) used by the trainee, the trainee can use the existing mobile support suitable for mobile training. The device (106) can be used with the present invention. As the mobile support device (106), various types of mobile support devices such as a device directly attached to the leg, a wand-shaped mobile support device, and a mobile support device for grasping and standing can be used. It is also possible to change the moving support used. The traveling body (100) and the environmental system (300) recognize the movement state of the movement support device (106) in accordance with the movement state of the trainee, and thereby, the movement state including the operation of the movement support device (106) is recognized. You can recognize and exercise instructions. That is, it is possible to support movement training in which the movement support device (106) is regarded as a part of the foot or the third foot.

  Since the moving support device (106) has various shapes and functions, a mechanism for learning and recognizing the shape and behavior of the moving support device (106) to be used is required. The movement support device (106) can be detected by a method using an acceleration sensor mounted on the mounting unit (200) or a method using an optical image of an optical camera (112) and an infrared sensor (111). is there. In the method using the mounting unit (200), by mounting the mounting unit (200) on the mobile support device (106), the acceleration and speed of the mobile support device (106) can be directly obtained by the procedure (412). .

  In the method using the optical camera (112) and the infrared sensor (111), the area of the mobile support device is specified, and the movement state of the device is detected while performing tracking. The flow of the detection is shown in a procedure (413) in FIG. In the detection means based on the measurement of the depth image in the above-mentioned procedure (411), extraction of the body part of the measurement target person may fail due to the influence of the presence of the mobile support device (106). This occurs because the human skeleton model used for extracting the body part does not assume the use of tools such as the mobile support device (106).

  In order to solve this problem, in the procedure (413) of the method using the optical image, the area of the moving support device (106) is corrected from the depth image to infinity, and the corrected depth image is used as the depth image. It is used for detecting the state of the person to be measured in the procedure (411) of the method.

  In the procedure (413) of the method using the optical image and the depth image, an area of the mobile support device is specified at the start of the mobile training, and the area is detected during the mobile training by detecting the area. First, the movement support device (106) is presented at the start of the movement training, and an image is taken with the infrared sensor (111) and the optical camera (112). Next, the optical marker (label of a specific color) (504) attached to the moving support device (106) is detected by color identification on the optical image (S-31). Note that the optical marker (label of a specific color) (504) is attached to the moving support device (106) in advance. Next, a region that is relatively close on the depth image is regarded as a region of the moving support device (106) (S-32), and the region of the moving support device (106) on the depth image and the positional relationship between the optical marker are determined. Then, a skeletal model of the moving support device (106) having the position of the optical marker as a feature point is created (S-33). During the movement training, first, an optical marker is extracted on the optical image, and a skeletal model is applied on the depth image by using them as feature points, thereby collating the region of the movement support device (106) (S-34). The matched feature points are tracked (S-35). The motion state is calculated based on the three-dimensional position, velocity, and acceleration of the feature point of the moving support device (106) being tracked, and is regarded as the motion state of the mobile support device (106). However, the three-dimensional position is a value converted into an absolute coordinate system for the moving environment. Finally, the depth of the region of the moving support device (106) on the depth image is corrected to be infinity (S-36). During the movement training, the steps (S-31 to S-33) up to the generation of the skeleton model are omitted.

  The traveling body (100) or the environmental system (300) detects the moving state of the trainee (104), the training instructor (105), and the moving support device (106) by the procedure (402) of the detecting means described above. Then, the processing procedure (403) of recognizing the moving environment and planning the movement is executed. FIG. 6 shows the state recognition of the moving environment and the route planning. The running body (100) automatically creates a map of the moving environment while moving autonomously, and identifies the position and orientation of the running body (100). The laser sensors (113, 114) measure depth data in a horizontal plane of the moving environment starting from the current position. As the laser sensor used in the present invention, an existing laser sensor (URG (trademark) manufactured by Hokuyo Co., Ltd.) can be used. It is possible to simultaneously generate a map of the moving environment and identify its own position / posture from the measured local depth data. Existing methods (such as ROS-gmapping (wiki.ros.org)) are used for simultaneous map generation and self-position and orientation estimation.

  The traveling body (100) appropriately recognizes a situation and appropriately executes an action plan on the assumption that the operation mode is teaching or guidance. As shown in a procedure (401) in FIG. 4A, the teaching and guidance of the moving route are selected by a person at the start of the traveling body.

  The procedure (420) in FIG. 4C represents recognition / planning at the time of teaching by the power assist operation. In the power assist operation (420), the load instructor applied to the running body (100) by the training instructor (105) is measured by a force sensor, and the translation body and the rotating component of the force are used as the teaching intention to control the running body (100). A movement target position for driving is calculated.

  The procedure (421) in FIG. 4C represents recognition / planning at the time of teaching by following operation. In the following operation (421), the position of the training instructor is measured, and a movement target position for driving the traveling body (100) is calculated with the predicted movement position of the training instructor (105) as the teaching intention. As the predicted movement position of the training instructor (105), a movement speed vector in a horizontal plane is measured, and a movement position after a short time has elapsed is set as a movement target position. The predicted movement position is measured by the infrared sensor (111) and the laser sensors (113, 114). In the follow-up operation procedure (421), an action plan for tracking the running object is performed based on the change in the three-dimensional position of the measurement target acquired by the detection means. The tracking of the measurement target by the laser sensors (113, 114) is performed by recognizing an arc-shaped object (corresponding to a foot of the tracking target) extracted in a short-distance region as a tracking target, and in a horizontal plane of the tracking target. This is realized by setting the movement target position on an extension of the speed vector.

  The procedure (422) in FIG. 4C represents the recognition / planning of the trainee (104) at the time of guidance. When the trainee (104) is guided, the distance between the trainee (104) and the running body (100) is recognized by the method using the infrared sensor (111) in the above-mentioned procedure (421), and the movement target position is set. If the trainee moves too much, the presentation of the later-described audiovisual pattern will be affected. Therefore, if the trainee (104) approaches more than necessary, the trainee will follow the movement route, and if not, will wait on the spot.

  Reference numeral (600) in FIG. 6 represents a route plane for explaining a situation at the time of teaching the travel route and re-planning the travel route at the time of guidance. Reference numeral (601) in FIG. 6 represents a path plane for explaining a situation at the time of teaching the movement path. At the time of teaching the movement route, the above-mentioned power assist operation (420) by the training instructor (105) and the follow-up operation (421) by the lead are performed. The traveling body (100) creates a map of the moving environment at the same time as moving. As for the teaching of the movement route, in addition to the training instructor (105) demonstrating and teaching the movement, the training instructor (105) gives the waypoint (613) on the map generated by the traveling body, and A moving route that connects the waypoints to the system may be generated. In that case, first, the traveling body (100) needs to autonomously search in advance for the moving environment to generate a map. Alternatively, the system of the traveling body (100) may generate the waypoint (613) on the map to construct the movement route.

  Reference numeral (602) in FIG. 6 represents a route plane for explaining the guidance of the moving route. At the time of the movement training, the trainee is guided by dynamically re-planning the movement path based on the movement path planned in advance. In general, the environment at the time of teaching the moving path and the environment at the time of execution may be different, and dynamic adjustment according to the state of the moving path is necessary. For example, there may be a case where an obstacle (615) such as a person (612) or an object enters the moving route, a case where the trainee (104) intends to change the moving route, or a case where it is desired to stop the moving training. When an obstacle (615) is detected on the planned travel route, guidance is executed while avoiding the obstacle (615) and dynamically changing the route so as to restore the previously planned travel route. . When the contact sensor detects a collision with an obstacle including the trainee (104), the trainee temporarily retracts to avoid the obstacle, or stops and stops and instructs the trainer (104) or the training instructor (105). To wait.

  Although it has been difficult to enhance the recognition of the exercise state during exercise and to dynamically instruct the exercise according to the situation in the conventional exercise, in the present invention, the traveling body (100) and the environment The system (300) recognizes the movement state of the trainee (104), and returns the trainee (104) 's movement information by a video / audio signal or a stimulus to a body part. The cognition is strengthened and the exercise plan becomes clear. In addition, by recognizing the exercise state of the trainer (104) and the exercise state of the mobile support device (106) used by the trainee (104) as well as the exercise state of the trainee (104), more effective exercise information can be obtained. Can be presented. Exercise information is presented using stimuli such as video, sound, vibration, and electricity.

  FIG. 7 shows the stimulus presentation of the present invention. The visual stimuli (701, 702) and the auditory stimuli (703, 704) supplement somatic sensations obtained from the sense of balance, the kinetic sensation of the muscle spindle, and the tactile sensation of the sole skin. The visual stimuli (701, 702) and the auditory stimuli (703, 704) can be presented without the running object (100) or the environmental system (300) contacting the trainee (104). Further, by using the mounting unit (200), it is possible to apply a tactile stimulus (705, 706) such as direct vibration or electricity to the body part (for example, an ankle or a wrist) of the trainee (104). it can.

  From the trainee (104), the training instructor (105), and the moving state detecting means (106), the moving state detection means detects a moving image pattern (701) representing the next stepping position when the trainee (104) steps on the foot. ) And an acoustic pattern (703) representing the landing timing. When the training instructor (105) steps on the foot, the same video pattern and sound pattern can be presented. Therefore, the present invention can be used for exercise training of a trainer as well as voluntary training of the trainee himself. In addition, by presenting a video pattern (702) representing the movement target position and an acoustic pattern (704) representing the timing of landing of the device, the movement supporting device (106) also includes the movement support device (106). Can give cognitive and exercise instructions.

  The place where the mounting unit (200) is mounted is determined in consideration of the condition of the trainee (104), a body part requiring rehabilitation, a training policy, and the like. For example, it is attached to the torso, left and right wrists, ankles, and moving support equipment to be used. The stimulus command is managed by the traveling body (100) or the external computing device (301), and the stimulus command is transmitted from these devices to the mounting unit (200), and the mounting unit (200) executes the stimulus command. The calculation processing of the unit (200) can be reduced.

  A stimulus presentation unit (202) provided in the wearing unit (200) gives a direct haptic stimulus (705, 706) to a specific part of the body by vibrating a motor before swinging a foot or at the time of landing. be able to. In addition, electrical stimulation or the like may be used to present the stimulus in addition to the tactile force stimulus caused by vibration. In any case, by using these stimuli, the trainee (104) can improve the recognition of the landing by himself, and can enhance the training effect. The stimulus presentation by the wearing unit (200) can also be executed based on the exercise state performed by the training instructor (105) by attaching the wearing unit (200) to the foot of the training instructor (105). Thereby, the trainee (104) can intuitively perceive the landing timing and the location of the body part where the exercise is instructed (for example, the foot to be stepped on).

  The vehicle (100) or the environmental system (300) performs the movement training while providing the trainee (104) with a visual stimulus and an acoustic stimulus. The presented video pattern is not limited to the shape indicated by the reference numerals (701) and (702), and any video pattern can be projected. FIG. 8 shows an audio-visual effect that is useful in performing the movement training. The shape of the video pattern (701) indicating the stepping position of the foot is arbitrary, and may be a simple footprint pattern or an optical image of the foot of the trainee (104). The appearance of the trainee's (104) foot may vary daily due to the effects of slippers, shoes, socks, etc. worn during training, but the trainee (104) photographed with the optical camera (112) of the running body at the start of training. By using the foot image of 104), it is easy to imagine stepping on the foot, and the realism of the movement training is enhanced. The footprint image pattern (701) may be presented for several steps at the same time as shown in FIG. 8, or may be presented one step at a time for each step. A reference line (807) serving as a measure of the stride length can be presented together with these footprint image patterns. Similarly, an image pattern (804) representing the stride, step speed, and walking distance detected by the moving state detecting means, an environment map generated by the moving environment recognition / planning processing means, and a moving path. (805) A communication screen (802) for displaying video / audio information of a training instructor who is in a remote place or a video of the exerciser's own exercise state can also be presented. Information such as a start position, a current position, and a destination position can be added to the environment map (805), and the direction (803) of a future movement route can be presented. As a result, the movement training becomes clear to the trainee (104), and it is useful for the persons (612) in the vicinity to grasp the future movement of the running body (100) and the trainee (104). Training is more secure.

  It is important to correctly recognize one's own walking rhythm in walking training in order to recover the walking rhythm. In particular, recognition of the timing of landing is important in walking, and presenting a walking rhythm during or after movement by auditory stimulation is highly effective in movement training. By using acoustic patterns of different pitches and timbres for landing on the left and right feet, it is also possible to independently enhance recognition of landing on the left and right feet. By generating a pace sound such as a metronome, a standard walking cycle can be presented. Further, the acoustic stimulus is not limited to the acoustic pattern (703) for presenting the landing timing, and the training parameters are set, the training is started, the training is continued, and the training is performed for the trainee (104) and the training instructor (105). It can also be used as an audio pattern or audio guide for notifying the end of the game. In conjunction with the projection of the video pattern, the training can be performed while talking with a person in a remote place.

  The presentation of the video and audio patterns while moving is effective in giving a sense of accomplishment and worthwhile to the walking training, and in improving the motivation of the walking training. In addition, the function to guide the walking training to the place where the trainee wants to walk is a function that could not be realized by the walking training support device performed in the conventional equipment, such as going to the destination or going around the favorite place. Through this act, the pleasure of walking training and the motivation to continue walking training are improved.

  Next, an embodiment (an extended function related to walking training support) according to a walking training support device in which functions are extended using a traveling body will be described.

  As disclosed in the above-mentioned patent document, the conventional walking training support devices do not sufficiently consider the whole body movement in the walking evaluation, and continuously link the current walking state of the trainee and the target walking state. Walking guidance and effective presentation of a walking state utilizing a walking environment have not been realized.

  The extended function for walking training support using the running object of the present embodiment is an extensible function for enhancing the movement recognition of the training target person using the running object independent of the training target person and efficiently guiding the walking movement. It has various means for measurement, evaluation, presentation, guidance, and reproduction, and solves the problem described in the preceding section. Exercise recognition refers to the training target person himself / herself recognizing the state of exercise, and the exercise state covers all three-dimensional postures of the body (tilt and movement of each part described later).

  Regarding this point, for example, in the above-mentioned Patent Document 10 (Japanese Patent Application Laid-Open No. 2011-229838 “Independent driving support device”), a moving stick is integrated with the training instructor, and footing is instructed using video projection. To induce walking. In this technique, it is difficult to use the training target with another supporting device such as a parallel bar due to the integration of the training target and the device, and the behavior control is subordinate to the training target. It has been difficult to measure whole body movements and present and instruct a wide range of exercise state using a walking environment.

  In the present embodiment, a training method based on three-dimensional measurement and reconstruction of walking motion is realized. The present embodiment autonomously controls a running body independent of a trainee and measures and integrates three-dimensional body motion information of the trainee in a free space with various sensors to integrate the trainee and the trainer. Is reconstructed as a movement of the body image that is easy to understand intuitively. By adjusting the motion of the body image according to the training intention and presenting it to the trainee before and after walking, the walking of the trainee is guided or the walking motion recognition is enhanced.

  The realization of the training method based on the three-dimensional measurement and the reconstruction makes it possible to evaluate the whole body motion during walking in free space, and the training subject and the running body are independent, and the training subject Can use the most desirable support device for the condition. In addition, since the traveling body travels in the walking environment ahead of the trainee, the safety of the walking floor can be confirmed in advance, and a space for walking can be secured in advance even when there are people around. In addition, the presence and physicality of the preceding running object become the psychological walking target of the trainee, along with the physical walking target that the running body presents to the trainee using images, sounds, and vibrations. It also contributes to motivating the trainees to be trained as being with them.

  The walking training method realized in the present embodiment is referred to as Augmented Mirror Therapy (AMT). In extended mirror image therapy, human whole body movements are measured three-dimensionally and recorded in spatial correspondence with the walking environment (referred to as “extended measurement means”), and the measured movements are evaluated based on training intentions. Make adjustments and reconstruct (represented as augmented evaluation means and reconstructed body movement information as augmented body image) and present the augmented body image in a form that humans can intuitively recognize To perform a walking process (referred to as an extended presentation means) and present an extended body image during walking training to induce walking (referred to as an extended guidance means). This is a training method to restore the motor function by presenting and promoting understanding of the current state and ideal walking movement of the trainee (indicated as an extended reproduction means). It should be noted that the above-described expanded body image includes not only the spatial characteristics of the body but also its motion information. Each of these units is processed by a processing unit mounted on the traveling body, but a part of the processing may be processed by a processing unit fixed independently of the traveling body. When processing is performed by different processing means, the transmission and reception are performed between the two by the transmission means.

  In addition, in the present embodiment, since the recovery training can be supported along with the trainee in the free space according to the form of the running body, the highest effect is expected for the walking exercise in which the trainee moves in the space, By using the function while the vehicle is stationary, the present invention can be applied to a function recovery exercise for an exercise with no spatial movement or a small amount of movement, for example, an upper limb exercise such as a hand movement.

  FIG. 10 shows a procedure for realizing the extended image therapy. The device that realizes the walking training method has a “setting mode”, a “guidance mode”, and a “reproduction mode” as modes for executing functions. In the "setting mode", input and browsing of each set value in using the apparatus, teaching of a walking route, generation of a map, and the like are performed. The set value is, for example, a value for setting a reference position of the traveling body, or a basic numerical value such as a moving distance or a moving speed. In the “guidance mode”, the positional relationship between the running object and the trainee at the time of guidance, for example, a guidance method such as forward guidance or side guidance is selected, and the walking state is evaluated while measuring walking and the environment. Instruct training while strengthening motor recognition while indicating (presenting) goals and the like. In the "reproduction mode", the user selects a method of reproducing the extended body image, for example, reproduction on the floor or on the wall, and reproduces the current walking motion and the training target walking motion recorded during the guidance to perform the training. Increase the understanding of the target audience.

  FIG. 11 shows an image of the extended measuring means. As shown in the figure, the traveling body three-dimensionally measures the whole body motion during walking while moving in free space together with the trainee. The three-dimensional measurement of walking is realized by integrating the measurement of the position and orientation of the traveling body expressed in the coordinate system of the environment with the measurement of the whole body movement of the trainee in the coordinate system of the traveling body. Thereby, the walking motion of the trainee can be recorded in association with the walking environment.

  FIG. 12 shows the results of the three-dimensional measurement of walking in the present embodiment. The position and orientation measurement of the traveling body is realized by a laser sensor mounted on the traveling body and a known processing unit (Simultaneous Localization and Mapping: SLAM). With SLAM, it is possible to create an environmental map and at the same time estimate the position and attitude of the vehicle.

FIG. 18 shows a spatial arrangement of the traveling body in the embodiment of the present invention. The traveling body has a cylindrical shape, and this drawing shows an arrangement in which the traveling body is viewed from the side and the top. In this figure, point O is the reference position of the traveling body, point A is the light source position of the projector, point G is the measurement position of the laser sensor, line segment AD is the line segment on the optical axis of the projector, and line segment DB is the image space. The (X ^R , Z ^R ) plane corresponds to the vertical axis and is the measurement plane of the laser sensor. The reference position of the line member in the coordinate system of the environment ^{(x R, y R, z} R) When an angle from the laser sensor theta, a distance r points (X ^R, Z ^R) perpendicular foot onto the plane (i.e. The coordinates (x, y, z) of the point on the floor surface are expressed by Expression 7.

As described above, the position of the perpendicular foot lowered from the position measured by the laser sensor to the floor by the traveling body is expressed in the environment coordinate system.

  The measurement of the whole body motion of the trainee is realized using various sensors mounted on the running body and sensors mounted on the trainee. The position of the foot below the knee is measured mainly by a laser sensor. The whole body movement on the knee is mainly measured by an infrared sensor. The detection of landing on a foot and the local movement of a body part that is difficult to measure by the above-described sensor are mainly measured by an acceleration sensor, an angular velocity sensor, or the like, which is worn on a trainee. The appearance of the whole body movement is mainly measured with an optical camera. Although the measurement information obtained from the various sensors has overlapping portions, information on the parts and types that can be measured with the highest accuracy by each sensor is extracted and integrated.

  In the measurement of the foot position in the preceding paragraph, the detection accuracy can be improved by utilizing the shape characteristics of the foot. Assuming that the cross section of the foot in the horizontal plane is an ellipse, the foot can be detected by extracting an elliptic arc pattern in consideration of the occlusion of laser measurement. In addition, if the standing position of the trainee is instructed using the video pattern and the pattern is extracted from the vicinity thereof, the detection becomes easy. Further, by tracking the pattern once detected while predicting the movement, the foot position can be continuously detected even if the trainee moves.

  By integrating the above-mentioned measurement of the position and orientation of the running body with the measurement of the whole body movement of the trainee, the whole body movement of the trainee is modeled as a movement formed by a three-dimensional body image described in the coordinate system of the environment. be able to. In addition, the surrounding environment of the walking target obtained from the generated environment map and the photographing information of the optical camera, for example, the situation of the walking floor surface, the presence of surrounding people and obstacles, etc. Record along with. As a result, it is possible to evaluate the influence of the walking environment on the walking state, such as the presence or absence of interpersonal communication at the time of guidance, the cause of departure from the guidance route, and the like.

  FIG. 13 shows an image of the extended evaluation means. The traveling body evaluates the position of the center of gravity and the inclination of the body at the time of guidance using the position information and the mass distribution information of each part of the whole body. For example, body balance such as forward leaning of the whole body or backward leaning of the pelvis is an evaluation amount that is difficult to estimate only by measuring the foot motion, but in the present embodiment, the stride, the stride, the step speed obtained from the measurement of the foot motion. By associating the number of steps, the toe angle, the walking phase, the walking time, the walking distance, and the like with the above-described whole-body exercise, more advanced walking evaluation can be performed. For example, it is known that the toe's movable angle at the time of walking has a correlation with the fall risk, and in particular, the risk of walking can be evaluated from the toe's movable angle at the time of stepping. In addition, the quality of walking evaluation can be enhanced by combining the whole body movement measurement with general walking evaluation, for example, Timed up and go test (Podsiadlo & Richardson 1991).

  FIG. 14 shows an image of the extended presentation means. The running body presents the above-described expanded body image and the evaluation amount of the walking state to the trainee as stimulation by video, sound, and vibration. The running body is equipped with projection means such as a projector, and can present an image at a position such as a wall surface or a floor surface that can be recognized by a trainee. Thus, for example, when projecting onto a floor, the following processing is performed. The coordinate system (ξ, η) of the image space of the projector corresponding to the coordinates (x, y, z) of the foot of the perpendicular that the traveling body has lowered from the position measured by the laser sensor to the floor is expressed by Expression 8. .

However, a bar is set up in advance at the point where the point in the image space is projected, and (ξ, η, d) = (0,0, d ''), (ξ, η, d) = (0, n ' , d '), d'',n', and d 'are measured. Thus, the traveling body can display an image pattern immediately below an arbitrary position measured by the laser sensor within a range where the projector can project.

  Information that the traveling body presents to the trainee when guiding walking is described. In the video, the target positions of the left and right foot steps are presented. The presentation position of the video pattern is adjusted by associating the real space with the image space. Further, the color, size, and shape of the video pattern are adjusted according to the walking state and the standing position of the trainee. For example, as shown in FIG. 19, when the evaluated position of the center of gravity enters the polygon formed by the target position of the next foot and the support device, the target position of the next foot and the support device is displayed or the color scheme is displayed. May be set to be emphasized. When presenting an image on a wall surface, the walking posture of the trainee may be projected. The current walking target sound indicates the target timing of landing. By using different pitches and timbres corresponding to each part such as the right foot, the left foot, and the cane, the acoustic pattern can be presented with the target timing of each part being distinguished. In the vibration, the landing timing of the trainee detected by the traveling body is presented. By feeding back the landing detection timing, the training target person's recognition of the landing is strengthened, and it becomes easy to synchronize the target timing with the actual landing.

For example, using the presentation means described above, the following walking exercise can be practiced.
A Practice of adjusting the landing location 1 Teach the target location of the landing with a video pattern.
2 The landing position and the landing timing are detected, and the next target position is presented.
3 Practice so that the landing position matches the target position.
4. After practice, move the running body to reproduce and observe the target position and the detected position.
5 Adjust the distance between the landing target positions and repeat the practice.
B Practice for adjusting the landing rhythm 1 Teach the target timing of the landing using acoustic patterns.
2 Detect the landing timing and give a vibration stimulus.
3 Practice so that the timing of sound and vibration coincide.
4. Reproduce the sound and vibration after practice and feel the stimulus.
5. Adjust the interval of the landing target timing and repeat the practice.

  FIG. 15 shows an image of the extended guidance means, and an environment map, a teaching route, and a reproduction route learned in the embodiment. The walking route is taught by the training manager moving the running body with power assist in the setting mode. Note that the walking route may be taught in a form in which the traveling body follows the training manager. The traveling body moves while generating a map by the above-described SLAM, and at the same time, records a trajectory formed by the estimated own position as a walking route. Thereby, the walking route is expressed in a coordinate system based on the environment, and is associated with the environment map. In the guidance mode or the reproduction mode, the trainee moves on the taught walking route while keeping the distance to the trainee, and guides the trainee. The target position of the foot and the walking speed of the target at the time of the guidance are adjusted based on the judgment of the training manager, or the moving body is automatically adjusted based on the current measurement information. In addition, the guidance means can provide a rough target position to the trainee by leading the traveling body, and can provide that the passage of the traveling body is safe. When the presentation means provides video or other perception information as information to be guided, in addition to guiding to a state that should be an ideal behavioral posture, a plurality of differences from the actual behavioral posture are also considered. And the steps may be appropriately guided in accordance with the training proficiency of the walking trainee.

  FIG. 16 shows an image of the expansion reproducing means and an expanded body image presented on the wall surface or the floor surface in the embodiment. The running body spatially and temporally reconstructs and presents the walking motion of the trainee as an expanded body image. The movement of the traveling body is controlled so as to reproduce the extended body image three-dimensionally in the real space. The running body encourages the trainee to understand the current state and the ideal walking state by presenting information on the position and exercise of the extended body image to the real space and the body of the trainee while moving.

  By the extended reproduction means, the trainee can observe a walking state, for example, a walking posture, from an objective viewpoint in a stationary state, so that the understanding of the current walking state is enhanced. In addition, by adjusting the parameters for generating the extended body image, it is possible to reproduce a motion state intermediate between the current motion state and the ideal motion state. For example, by reproducing the walking motion by gradually increasing the current step length, step speed, and the like, it is possible to imagine a process from the current exercise state to reaching the training target.

  FIG. 17 shows an image in which the evaluation amount of the walking state is presented by the expansion reproducing means. For example, evaluation amounts such as a landing position, a center of gravity position, and a body inclination at the time of guidance can be presented on the floor surface while moving in the real space. By observing these biases, it is possible to evaluate, for example, the asymmetry of walking with respect to the symptoms of hemiplegia. In addition, by converting the measurement values of the acceleration sensor and the angular velocity sensor worn by the trainee into sound and presenting the sound, it is possible to intuitively indicate the movement of the body part such as the swing of the feet and the swing of the hands.

  In the above-described embodiment, the correlation between the velocity amplitude and the traveling direction velocity has been described for each component in all directions (movement traveling direction, direction perpendicular to the traveling direction in the horizontal plane, vertical direction) of the movement by the acceleration sensor. Therefore, based on these correlations, an experiment was performed to detect the stride length during walking. In the experiment, a general person's ankle (both ankles) acceleration sensor, which looks like a trainee, was installed, and the movement direction (X) measured from the acceleration sensor, the direction perpendicular to the movement direction in the horizontal plane (Y), and the vertical direction The acceleration component is analyzed for all the directions (Z) to detect the stride. FIG. 9 shows the result.

FIG. 9A is a graph showing acceleration time-series data and detection of landing. The continuous line at the top of the graph represents the absolute value of the magnitude of the acceleration, and the point at the bottom of the graph represents the time at which the landing was detected. In both graphs, the horizontal axis represents time (seconds). The vertical axis at the top of the graph represents the magnitude of the acceleration (m / s ² ).

  FIG. 9B is a graph showing the correlation between the speed amplitude and the moving speed. The horizontal axis represents the speed amplitude (m / s) of each of the X, Y, and Z coordinate components, and the vertical axis represents the moving speed (short-time average moving speed) (m / s). Each straight line in the graph is a regression line that approximates the distribution of data points in each component. FIG. 9B shows that three components of the speed amplitude in the X, Y, and Z directions are correlated with the moving speed.

  FIG. 9C shows a comparison of the stride length with the conventional method. The conditions of the experiment are shown with the number of steps when the subject walks on the horizontal axis, and the stride length (m) on the vertical axis. For example, the value of the walking number 4 in FIG. 9C means the stride of the fourth step during walking. The polygonal line of series 1 is a stride measured using an optical measurement method (Vicon (trademark)) with an accuracy of 1 mm or less, and this step is regarded as a true value in this experiment. Series 2 is a graph showing the result of the conventional method, and series 3 is a graph showing the result of the detection means of the present invention. FIG. 9 (c) shows that the result of the detecting means of the present invention is closer to the true value than the conventional method, indicating the effectiveness of the detecting means of the present invention.

  In the present invention, a movement training support device that guides the movement by recognizing the movement state of the trainee, the movement environment, and the state of the movement support device used for the movement is realized. According to the present invention, the exercise instruction of the training instructor is transmitted to the trainee via the running body and the mounting unit, and the trainer repeatedly exercises the movement recognition while understanding his / her own movement because the exercise recognition is repeatedly enhanced. be able to. Since the present invention can be used in combination with an existing mobile support device, users with various symptoms can be assumed, and commercialization and commercialization of this device are promising. In addition to the support of the movement training, the present invention can also be applied to attending walking of a dementia patient and cognitive support of a whole body movement not limited to movement. Further, since the present invention is expected to be applied to an amusement game as a mobile audiovisual recognition generation device and to be used for training of assembly work at a production site, the present invention is considered to have high industrial applicability.

100 Running body (body)
101 Projection plane 102 Field of view
103 Projection area
104 Trainer 105 Training instructor 106 Moving support device 111 Infrared sensor
112 Optical Camera
113 Laser sensor (front)
114 Laser sensor (rear)
115 microphone (right)
116 microphone (left)
117 Contact Sensor
118 Attitude sensor 121 Projector
122 speaker
123 Motor unit (for right front wheel)
124 motor unit (for front left wheel)
125 drive wheel (front left wheel)
126 drive wheel (front right wheel)
127 passive wheel (rear wheel)
131 Power supply
132 communication unit 133 calculation unit (processing means)
134 control unit (control means)
135 detection unit (detection means)
136 Presentation unit (presentation means)
200 mounting unit
201 Motion measurement unit (acceleration, angular velocity, geomagnetic sensor)
202 Stimulation presentation unit (vibration motor / electric stimulator)
300 Environment system 301 External computing device 302 Mobile training support device 401 Overall processing 402 State detection (detection means)
403 State recognition and action plan (processing means)
404 Drive control (control means)
405 Audiovisual presentation (presentation means)
410 Details of Detecting Means 411 Detecting Moving State Based on Depth Image Measurement 412 Detecting Moving State Based on Acceleration Measurement 413 Detecting Moving State of Moving Supporting Device and Depth Image Correction 420 Recognition / Planning 421 during Teaching in Power Assist Operation Recognition / planning at the time of instruction by trainee 422 Recognition / planning at the time of guidance of trainee 423 Details of drive control 424 Details of video / audio presentation 501 Depth image 502 Skeletal model 503 Feature points of skeletal model 504 Extraction of optical markers 505 Area 600 Teaching and guidance of moving route 601 Teaching of moving route 602 Guiding of moving route 611 Person 613 near wall 612 Via point 614 Teach route 615 Obstacle 616 Execution route 701 Image pattern 702 showing target movement position of foot Moving device Image pattern 703 indicating target position Indicates landing timing of foot Sound pattern 704 Sound pattern 705 representing the landing timing of the equipment Stimulation pattern 706 representing the landing timing of the foot Stimulation pattern 801 representing the landing timing of the equipment Video projection area 802 Communication screen 803 Video pattern 804 representing the moving direction Video pattern representing the moving state 805 Map of moving environment 806 Current foot position of trainee 807 Reference line of moving target

Claims (19)

Depth image acquisition means for acquiring a depth image of the whole body of a trainee who performs walking or running training from an arbitrary acquisition position,
A three-dimensional position acquisition unit that calculates a distance from the acquisition position in the plurality of body parts of the trainee based on the depth image and acquires a three-dimensional position of the body part,
Measuring the kinematic whole body movement of the trainee based on the change in the three-dimensional position of the body part, and based on the result of the measurement , the body required to guide the trainee to walk or run. Processing means for calculating the guidance position of the part,
Image presentation means for visually providing the guidance position of the body part calculated by the processing means ,
A traveling body that guides the trainee and runs on the floor by itself, and a traveling body that functions physically independently of a walking support device used by the trainee,
The image presentation means is mounted on the traveling body,
A mobile training support device, characterized in that:   The apparatus further includes an acceleration sensor attached to a body part of the trainee, wherein the processing unit measures a kinematic whole body movement of the trainee in the body part of the trainee based on information of the acceleration sensor. The mobile training support device according to claim 1, wherein the guidance position of the body part is calculated together with the three-dimensional position. Further, an angular velocity sensor and a geomagnetic sensor to be attached to a body part of the trainee are provided, and the processing unit is configured to detect the trainee's information based on the information of the angular velocity sensor and the geomagnetic sensor in addition to the information of the acceleration sensor. 3. The mobile training support apparatus according to claim 2 , wherein the mobile training support apparatus calculates a whole body moving amount, a moving distance of the body part , and a stride. The depth image acquisition unit, the three-dimensional position acquisition unit, and the processing hand stage is intended to be mounted on the traveling body,
The traveling body further includes a driving unit configured by driving wheels, a control unit that controls a driving state of the driving unit, and a detection unit that detects a surrounding environment. The mobile training support device according to any one of claims 1 to 3, wherein the driving unit is controlled based on the detected distance information to the trainee so as to maintain an appropriate distance from the trainee. Before Symbol depth image obtaining means, the three-dimensional position acquiring means and the processing means, wherein the traveling body is intended to be fixedly installed in the training site independently,
The traveling body further includes a driving unit including driving wheels, a control unit that controls a driving state of the driving unit, a detection unit that detects a surrounding environment, and information on each unit mounted on the traveling body. And a transmission unit for transmitting and receiving various kinds of information between the processing unit and the second processing unit, wherein the control unit detects the detection by the detection unit. Based on the information on the distance to the trainee, the driving unit is controlled so that the interval is appropriately maintained, and the processing unit stores the information of the second processing unit and the body part of the trainee. The mobile training support device according to claim 2 or 3, which processes information of various mounted sensors. The running body includes a storage unit that stores environmental map information, and the control unit controls a driving unit based on the environmental map information so that the trainee follows and runs while leading the trainee. 6. The movement according to claim 4, wherein the processing means calculates a position of the trainee in the environmental map information, and the video presenting means presents a result of the calculation as a video. Training support device. The position of the trainee in the environmental map information calculated by said processing means, a collection of discrete positional information of the specific portion of the trainee, the individual position information of the specific portion, the environmental map The mobile training support according to claim 6, wherein the mobile training support is calculated based on the information, position information of the traveling body in the environmental map information, and a relative positional relationship from the traveling body to the specific part. apparatus. It said storage means of said running body is for storing the route information indicating a route for guiding the trainee moves the route information, artificially created planned route information, or the running body in advance 8. The mobile training support device according to claim 6 , wherein the route information is a route information representing the route performed , and is route information created by the processing unit based on the position information in the environmental map information. The depth image acquisition unit further is intended to obtain the depth image before Kifu line support device, the three-dimensional position acquiring means is adapted to acquire the three-dimensional position of the walking support device, said processing The means is for calculating distance information on a characteristic point for the walking support device based on the change in the three-dimensional position, and the image presentation means should move the walking support device together with the guidance position of the body part. The mobile training support device according to any one of claims 1 to 8, wherein the guidance position is visually presented. The depth image acquisition unit further acquires a depth image of the whole body of the training instructor, and the three-dimensional position acquiring unit calculates a distance from the acquisition position in a plurality of body parts of the training instructor. The mobile training support apparatus according to any one of claims 1 to 9 , wherein the three-dimensional position of the body part is obtained by using the method.   The mobile training support device according to claim 10, further comprising an evaluation unit configured to evaluate a foot exercise and / or a body balance of the trainee based on a three-dimensional position of the body part. The processing means, wherein said three-dimensional position in the trainer to claim 10 or 11 is to compute the inductive position of the body parts required to induce the walking or running of the trainee as teacher data Mobile training support equipment.   The guidance position calculated by the processing means is the movement state of the ideal body part at the time of walking or running of the trainee, and the moving state of the body part at the time of actual walking or running of the trainee. The movement training according to claim 12, which is an intermediate position, wherein one or a plurality of guidance positions are calculated according to a degree of a difference between an ideal body part movement and a real body part movement. Support equipment.   14. The image presentation means according to claim 13, wherein the image presentation means presents, as an image, the movement state of the body part during the actual walking or running of the trainee, together with the guidance position or separately from the guidance position. Mobile training support device.   Further, the apparatus further includes sound presenting means for presenting an auditory stimulus to the trainee, and the processing means creates sound information to be presented in accordance with a temporal timing for providing a guidance position to the trainee. 15. The sound presenting means according to claim 1, wherein the sound presenting means presents voice or sound selected by the trainee instead of or together with the video presenting by the video presenting means. The mobile training support device according to the above.   Further, an acceleration sensor attached to the trainee's foot is provided, and the processing unit is configured to, based on information of the acceleration sensor, determine that the trainee's foot has landed, 16. The mobile training support device according to claim 15, wherein the device outputs a signal for causing the trainee to recognize that the trainee is in a landing state by operating the sound presenting means.   Further, there is provided stimulus presenting means which is attached to a part of the trainee's body to provide a haptic stimulus, and wherein the processing means displays the stimulus when the landing is confirmed from the foot state by the acceleration sensor. 17. The mobile training support device according to claim 16, which outputs a signal for operating the means.   The stimulus presenting means is attached to a plurality of body parts of the trainee and presents a haptic stimulus by vibration, and the processing means is the body part to be induced among the body parts of the trainee. 18. The mobile training support device according to claim 17, wherein the device outputs a signal for operating the device to present a tactile stimulus by vibration.   Furthermore, an optical camera for photographing the state of walking or running of the trainee is provided, and the image presenting means replaces the guide position or together with the guide position, a still image taken by the optical camera or The mobile training support device according to any one of claims 1 to 18, which presents a moving image.