Cognitive function training robot and use method thereof

Abstract

The invention discloses a cognitive function training robot and a use method thereof, and relates to the technical field of cognitive function training, the cognitive function training robot comprises an autonomous following mode and an active training mode, the autonomous following mode is used for replacing a person to carry out voice communication exercise with a patient, real-time tracking and voice reminding are carried out on the human body in the walking process of the patient, and the active training mode is used for carrying out active training. When a user starts an autonomous following mode, a high-definition camera on a touch interaction screen shoots a front scene in real time, a target user is locked through image recognition and the movement direction of the target user is tracked, and a depth camera measures and calculates the real-time distance between the user and the robot based on the binocular vision principle. And distance data are fed back to a control system of the mobile chassis. By installing the high-definition camera and the depth camera, a user following function is realized, the problem of training scene solidification of traditional equipment is solved, and the use time of a user is prolonged.

Description

Technical Field

[0001] The present invention relates to the technical field of cognitive function training, in particular to a cognitive function training robot and a use method thereof. Background Art

[0002] With the aging of the population and the increase in the number of patients with neurological diseases, the demand for rehabilitation training for cognitive dysfunction is growing. Currently, most cognitive training devices on the market are fixed tablets or computer software. These traditional devices have problems such as single interaction form, fixed training scenarios, and lack of environmental adaptability, resulting in fragmented user experience and limited rehabilitation effects.

[0003] Patent CN104407375B discloses a cognitive function training device for clinical psychiatric use. The above patent realizes the automatic replacement of cognitive pictures at intervals, avoiding the labor loss of manual replacement, and after the training is completed, the cognitive pictures can be automatically stored and organized. At the same time, the device can clean and remove dust from the cognitive pictures, avoiding the tedious manual cleaning.

[0004] The above-mentioned patent storage box is fixedly connected to a fixed block, and the fixed block is rotatably connected to a threaded sleeve, the threaded sleeve is threaded with a screw, and the storage box is provided with a driving mechanism that cooperates with the threaded sleeve, the upper end of the screw is fixedly connected to a cover plate that cooperates with the storage box, and the cover plate is fixedly connected to a protective box, and there are two support shafts rotatably connected between the protective box and the cover plate, which can automatically replace the cognitive pictures at intervals, avoiding the labor loss of manual replacement, and after the training is completed, the cognitive pictures can be automatically stored and organized. At the same time, the device can clean and dust the cognitive pictures, avoiding the tedious manual cleaning, but the above-mentioned patent still has room for improvement in improving user experience through multiple modes.

[0005] To this end, the present application proposes a cognitive function training robot and a method of using the robot, which can improve user experience through multiple modes. Summary of the Invention

[0006] The purpose of the present invention is to provide a cognitive function training robot and a method of using the same to solve the technical problems raised in the above background technology.

[0007] To achieve the above objectives, the present invention provides the following technical solutions: a method for using a cognitive function training robot, including an autonomous following mode and an active training mode. The autonomous following mode is used to replace a person in voice communication training with a patient, and performs real-time tracking of the human body, voice reminders, and accompanying training during the patient's walking process;

[0008] The autonomous following mode includes the following steps:

[0009] When the user starts the autonomous following mode, the high-definition camera on the touch interactive screen captures the scene in front in real time, locks the target user through image recognition and tracks their movement direction, and the first motor drives the hinge to rotate, so that the mounting plate drives the touch interactive screen to pitch and adjust, ensuring that the high-definition camera is always facing the user's upper body to capture dynamic posture. At the same time, the depth camera measures the real-time distance between the user and the robot based on the principle of binocular vision, and feeds the distance data back to the control system of the mobile chassis. The two second motors at the bottom of the mobile chassis drive the symmetrically distributed tires to rotate differentially according to the direction command, controlling the robot to move forward, turn or stop.

[0010] Preferably, the active training mode is used to carry out structured cognitive training through multimodal interaction, specifically including the following tasks:

[0011] Task A: Graphic matching training based on a touch screen interactive screen, where users drag screen elements to complete logical associations;

[0012] Task B: Human posture training based on a high-definition display. The display displays the training movements, and the user imitates them. The posture is evaluated using image data, and corrections and encouragement are provided through a loudspeaker.

[0013] Task C: Use the HD camera and rear camera to capture environmental objects, display them on the HD display and mark them with red frames, and generate a physical object recognition question-answering task.

[0014] Preferably, the autonomous following mode further includes an environment adaptation step:

[0015] The laser radar scans obstacles in the travel path at a frequency of 10Hz and combines the close-range detection data of the ultrasonic sensor to generate a dynamic obstacle avoidance trajectory. When an obstacle is detected within 1 meter in the direction of travel, the mobile chassis controls the two second motors to differentially turn on the spot. During the steering process, the shock-absorbing bracket ensures that the pitch angle fluctuation of the touch interactive screen is less than ±2°.

[0016] Preferably, the usage method further includes a user identity binding process:

[0017] During first use, the high-definition camera collects the user's facial features to build a three-dimensional model, which is combined with the physiological parameters input through the touch interactive screen to generate an encrypted user profile. During subsequent startups, historical training data is automatically loaded through facial recognition. When the recognition confidence level is lower than 85%, the touch interactive screen pops up a password verification interface for secondary authentication.

[0018] Preferably, the task C specifically includes:

[0019] The rear camera captures environmental objects at a 120° wide angle. After extracting at least three feature objects through the image segmentation algorithm, the high-definition display screen displays the object images according to their spatial orientation. Each time a question is asked, one of them is marked with a red frame. The speaker asks the question, and the touch interactive screen generates options based on the marked objects. The user needs to select the option that matches the voice question through the touch interactive screen within 10 seconds.

[0020] Preferably, the method of use further includes an exception handling process:

[0021] When the emergency stop button is triggered, the second motor cuts off power output within 0.2 seconds, and the hydraulic damper of the shock absorber bracket is automatically locked. The lidar immediately starts a 360° panoramic scan and marks a safe docking point. During the docking process, the speaker continues to play soothing voice until the system is completely shut down.

[0022] Preferably, the autonomous following mode includes a human posture recognition and anti-fall mechanism, specifically including:

[0023] The three-dimensional coordinates of the user's skeleton points are collected in real time through depth cameras and high-definition cameras to build a dynamic posture model;

[0024] The fall prevention mechanism includes three levels of response:

[0025] Level 1 response: When the robot detects an obstacle on the ground in front of the user through the lidar, the speaker will issue a voice reminder;

[0026] Level 2 response: When the user falls, the speaker plays an 85dB warning sound and activates the red strobe light of the emergency stop button on the top of the housing;

[0027] Level 3 response: When the user does not stand up for a period of time, the touch interactive screen automatically pops up the emergency contact interface and makes a call.

[0028] A cognitive function training robot comprises a housing, a mobile chassis and a touch interactive screen, wherein the mobile chassis is mounted on the bottom of the housing outer wall, and the touch interactive screen is mounted on the top of the housing outer wall via a connecting assembly;

[0029] The connecting assembly includes: a fixing frame, a hinge and a mounting plate;

[0030] A fixing bracket is fixedly installed on the top of the outer wall of the shell, and a mounting plate is connected to the top of the outer wall of the fixing bracket through a hinge. A first motor is installed at both ends of the outer wall of the hinge. A touch interactive screen is fixedly connected to the front of the outer wall of the mounting plate. A high-definition camera is installed inside the front of the touch interactive screen, a rear camera is installed inside the back of the touch interactive screen, and a speaker is installed inside the back of the touch interactive screen.

[0031] Preferably, the mobile chassis is circular, with three grooves distributed along the circumference of 110° on the bottom, the second motor-driven tires are installed in the two symmetrical grooves, and the middle groove is equipped with a universal wheel that can deflect by ±45°. The top of the outer wall of the mobile chassis is connected to the shell through a shock-absorbing bracket, and the shock-absorbing bracket has a built-in adjustable damping hydraulic cylinder, and its damping coefficient is adjusted in conjunction with the ground roughness data detected by the lidar.

[0032] Preferably, the front of the shell is provided with a high-definition display and a depth camera, the top is integrated with a switch button and an emergency stop button, the side plane of the mobile chassis is provided with a charging plate, 6 ultrasonic sensors are evenly distributed on the cylindrical surface, and the laser radar is installed in the connection gap between the mobile chassis and the shell, and the scanning plane is parallel to the horizontal plane.

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] 1. The present invention realizes the user-following function by installing a high-definition camera and a depth camera, solves the problem of fixed training scenes in traditional equipment, and increases the user's usage time;

[0035] 2. The present invention achieves multimodal human-computer interaction by installing a touch control screen and a speaker, solving the problem of single interaction form of traditional devices, improving the user experience immersion and the diversity of training tasks;

[0036] 3. This invention achieves three-dimensional environmental perception and obstacle avoidance capabilities by installing laser radar and ultrasonic sensors, solving the problem of insufficient spatial adaptability of traditional equipment and improving the robot's movement safety in autonomous following mode;

[0037] 4. The present invention realizes the functions of motion shock absorption and emergency braking by installing a shock-absorbing bracket and an emergency stop button, solves the problem of insufficient equipment operation stability, and improves the data collection quality and the safety of use in emergency situations. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a front structural schematic diagram of the present invention;

[0039] Figure 2 It is a schematic diagram of the back structure of the present invention;

[0040] Figure 3 This is a schematic diagram of the top-view structure of the touch interactive screen of the present invention;

[0041] Figure 4 This is a schematic diagram of the side structure of the touch interactive screen of the present invention;

[0042] Figure 5 Schematic diagram of the housing structure of the present invention;

[0043] Figure 6 This is a schematic structural diagram of the mobile chassis of the present invention;

[0044] Figure 7 This is a schematic diagram of the autonomous following mode workflow of the present invention;

[0045] Figure 8 Schematic diagram of the active training mode workflow of the present invention.

[0046] In the figure: 1. Shell; 2. Mobile chassis; 3. Touch interactive screen; 4. Fixing bracket; 5. Hinge; 6. Mounting plate; 7. First motor; 8. HD camera; 9. Rear camera; 10. Speaker; 11. Tire; 12. Second motor; 13. Universal wheel; 14. Groove; 15. HD display; 16. Depth camera; 17. Switch button; 18. Emergency stop button; 19. Shock-absorbing bracket; 20. LiDAR; 21. Charging plate; 22. Ultrasonic sensor. DETAILED DESCRIPTION

[0047] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0048] In the description of the present invention, it should be noted that the terms "upper," "lower," "inner," "outer," "front end," "rear end," "both ends," "one end," "the other end," and the like, indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings and are intended solely to facilitate and simplify the description of the present invention. They are not intended to indicate or imply that the devices or components referred to must have, be constructed, or operate in a specific orientation, and therefore should not be construed as limiting the present invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0049] In the description of the present invention, it should be noted that, unless otherwise expressly specified or limited, the terms "installed," "provided with," "connected," etc., should be understood in a broad sense. For example, "connected" may refer to a fixed connection, a detachable connection, or an integral connection; it may refer to a mechanical connection or an electrical connection; it may refer to a direct connection or an indirect connection through an intermediate medium; it may refer to internal communication between two components. Those skilled in the art will be able to understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0050] See also Figure 1 and Figure 7The present invention provides an embodiment of a method for using a cognitive function training robot, including an autonomous following mode and an active training mode. The autonomous following mode is used to replace a person in voice communication training with a patient, and to perform real-time tracking of the human body, voice reminders, and accompanying training during the patient's walking process.

[0051] The autonomous following mode includes the following steps:

[0052] When the user activates the autonomous following mode, the high-definition camera 8 on the touch interactive screen 3 captures the scene in front in real time, locks the target user through image recognition and tracks their movement direction, and the first motor 7 drives the hinge 5 to rotate, so that the mounting plate 6 drives the touch interactive screen 3 to pitch and adjust, ensuring that the high-definition camera 8 is always facing the user's upper body to capture dynamic postures. At the same time, the depth camera 16 measures the real-time distance between the user and the robot based on the principle of binocular vision, and feeds the distance data back to the control system of the mobile chassis 2. The two second motors 12 at the bottom of the mobile chassis 2 drive the symmetrically distributed tires 11 to rotate at a differential speed according to the direction command, controlling the robot to move forward, turn or stop;

[0053] The autonomous following mode also includes an environment adaptation step:

[0054] The laser radar 20 scans obstacles in the travel path at a frequency of 10Hz and combines the close-range detection data of the ultrasonic sensor 22 to generate a dynamic obstacle avoidance trajectory. When an obstacle is detected within 1 meter in the direction of travel, the mobile chassis 2 controls the two second motors 12 to steer in place at a differential speed. During the steering process, the shock-absorbing bracket 19 ensures that the pitch angle fluctuation of the touch interactive screen 3 is less than ±2°.

[0055] The autonomous following mode includes a human posture recognition and anti-fall mechanism, specifically including:

[0056] The depth camera 16 and the high-definition camera 8 are used to collect the three-dimensional coordinates of the user's skeleton points in real time to build a dynamic posture model;

[0057] The fall prevention mechanism includes three levels of response:

[0058] Level 1 response: When the robot detects an obstacle on the ground in front of the user through the laser radar 20, the speaker 10 issues a voice reminder;

[0059] Secondary response: When the user falls, the speaker 10 plays an 85dB warning sound and activates the red strobe light of the emergency stop button 18 on the top of the housing 1;

[0060] Level 3 response: When the user does not stand up for a period of time, the touch interactive screen 3 automatically pops up the emergency contact interface and makes a call;

[0061] Furthermore, when the user activates autonomous following mode, the high-definition camera 8 locks onto the user's upper body through image recognition. The first motor 7 drives the hinge 5 to adjust the pitch angle of the touch interactive screen 3, ensuring that the camera continuously collects user posture data. At the same time, the mobile chassis 2 uses binocular vision data from the depth camera 16 to adjust the distance to the user in real time. The second motor 12 drives the tires 11 to rotate differentially to achieve following. The lidar 20 scans the path for obstacles at a frequency of 10Hz. When an obstacle is detected within 1 meter of the direction of travel, the mobile chassis 2 controls the tires 11 to perform differential steering. The shock absorber bracket 19 suppresses the pitch fluctuation of the touch interactive screen 3 to within ±2°. The depth camera 16 simultaneously builds a three-dimensional model of the user's skeletal points. When the robot detects an obstacle on the ground in front of the user through the lidar 20, the speaker 10 issues a voice reminder. If the user is detected to have fallen, the speaker 10 immediately plays an 85dB warning sound and activates the red strobe light of the emergency stop button 18. If the user does not get up after a preset time, the touch interactive screen 3 automatically pops up the emergency contact interface and makes a call.

[0062] See also Figure 1 and Figure 8 The present invention provides an embodiment of a method for using a cognitive function training robot, wherein the active training mode is used to carry out structured cognitive training through multimodal interaction, specifically including the following tasks:

[0063] Task A: Graphic matching training based on touch screen 3, where users drag screen elements to complete logical associations;

[0064] Task B: Human posture training based on the high-definition display 15. The high-definition display 15 displays the training movements. After the user imitates them, the human posture is evaluated using image data, and corrections and encouragement are given using the speaker 10.

[0065] Task C: Use the HD camera 8 and the rear camera 9 to capture environmental objects, display them on the HD display 15 and mark them with red frames, and generate a physical object recognition question-answering task;

[0066] The housing 1 is provided with a high-definition display screen 15 and a depth camera 16 on the front, with a switch button 17 and an emergency stop button 18 integrated on the top. A charging plate 21 is provided on the side plane of the mobile chassis 2, and six ultrasonic sensors 22 are evenly distributed on the cylindrical surface. A laser radar 20 is installed in the gap between the mobile chassis 2 and the housing 1, with the scanning plane parallel to the horizontal plane.

[0067] Furthermore, when the user selects Task B in the active training mode, the high-definition display screen 15 plays preset standardized movements, such as raising hands and squatting. When the user imitates the movements, the depth camera 16 collects three-dimensional bone point data of the user's entire body through binocular vision, and calculates the deviation value of the joint angle and the standard movement in real time. The touch interactive screen 3 adjusts the pitch angle so that the high-definition camera 8 faces the user's face to capture expressions and line of sight, and assists in evaluating concentration. If the user's limb deviation exceeds the threshold or the movement lags behind, the speaker 10 outputs a voice prompt, such as "Raise your left arm 9 cm". At the same time, the high-definition display screen 15 superimposes a red arrow mark on the corresponding limb position to dynamically guide correction. If the user has a high degree of completion, encouragement will be given. After the training, the joint motion trajectory data recorded by the depth camera 16 is fused and analyzed with the expression data of the high-definition camera 8 to generate a posture score and attention report. The results are displayed through the touch interactive screen 3 and stored in the user profile.

[0068] See also Figure 1 and Figure 6 , an embodiment provided by the present invention: a cognitive function training robot, the use method also includes an exception handling process:

[0069] When the emergency stop button 18 is triggered, the second motor 12 cuts off power output within 0.2 seconds, and the hydraulic damper of the shock-absorbing bracket 19 automatically locks. The lidar 20 immediately initiates a 360-degree panoramic scan and marks a safe docking point. During the docking process, the speaker 10 continuously plays a soothing voice until the system is completely shut down.

[0070] The mobile chassis 2 is circular, with three grooves 14 distributed along the circumference at 110° on the bottom. The second motor 12 is installed in the two symmetrical grooves 14 to drive the tire 11. The middle groove 14 is installed with a universal wheel 13 that can deflect by ±45°. The top of the outer wall of the mobile chassis 2 is connected to the shell 1 through a shock-absorbing bracket 19. The shock-absorbing bracket 19 has an adjustable damping hydraulic cylinder built in. Its damping coefficient is linked to the ground roughness data detected by the laser radar 20.

[0071] Furthermore, when the robot is operating, the LiDAR 20 continuously scans ground features and obstacle distances to build a dynamic spatial map. If an unexpected situation occurs, the user triggers the emergency stop button 18, immediately disconnecting the power supply circuit of the second motor 12 and causing the two symmetrical tires 11 to stop rotating within 0.2 seconds. At the moment of emergency stop, the shock-absorbing bracket 19 absorbs the longitudinal impact force generated by the sudden stop of the tires 11 through internal hydraulic damping, suppressing the inertial forward tilt of the shell 1, ensuring the stability of the robot and avoiding image blur. When there is gravel or unevenness on the ground, the spring assembly of the shock-absorbing bracket 19 laterally buffers the bumps and vibrations, allowing the LiDAR 20 to maintain horizontal scanning accuracy within the chassis gap. The LiDAR 20 initiates a 360-degree panoramic scan and, combined with the data from the ultrasonic sensor 22, marks the nearest safe stopping point, such as a corner or flat area. The mobile chassis 2 controls the differential speed of the tires 11 to adjust the direction and move to the target point at a low speed of 0.3 m / s. During the docking process, the speaker 10 plays the voice prompt "Entering safety mode" in a loop until the system is completely shut down.

[0072] See also Figure 1 and Figure 6 The present invention provides an embodiment of a cognitive function training robot, wherein the mobile chassis 2 is circular and has three grooves 14 distributed along a circumference of 110° on the bottom. A second motor 12 is installed in the two symmetrical grooves 14 to drive the tire 11. The middle groove 14 is installed with a universal wheel 13 that can deflect by ±45°. The top of the outer wall of the mobile chassis 2 is connected to the shell 1 through a shock-absorbing bracket 19. The shock-absorbing bracket 19 has a built-in adjustable damping hydraulic cylinder, and its damping coefficient is adjusted in conjunction with the ground roughness data detected by the laser radar 20.

[0073] The housing 1 is provided with a high-definition display screen 15 and a depth camera 16 on the front, with a switch button 17 and an emergency stop button 18 integrated on the top. A charging plate 21 is provided on the side plane of the mobile chassis 2, and six ultrasonic sensors 22 are evenly distributed on the cylindrical surface. A laser radar 20 is installed in the gap between the mobile chassis 2 and the housing 1, with the scanning plane parallel to the horizontal plane.

[0074] Furthermore, when the robot's battery level is lower than the threshold, the LiDAR 20 locates the position of the charging pile through the SLAM algorithm, and combines the short-range detection of the ultrasonic sensor 22, covering a range of 0.2-3 meters, to move toward the charging pile along the preset path. During the movement, the two second motors 12 drive the symmetrical tires 11 for differential steering to avoid dynamic obstacles detected by the LiDAR 20, such as pedestrians or furniture. At the same time, the universal wheel 13 in the middle deflects synchronously with the steering angle of the tire 11 to offset the centrifugal force of the chassis. When the robot approaches the charging pile to within 0.5 meters, the ultrasonic sensor 22 switches to high-precision ranging mode. The charging piece 21 is guided to gradually align with the charging pile contacts. The shock-absorbing bracket 19 absorbs the slight vibrations caused by the friction between the tire 11 and the ground through the compression spring to ensure the stability of the charging piece 21 when in contact. If there is an obstacle in the blind spot of the laser radar 20 in the charging path, such as a glass door or a low step, the ultrasonic sensor 22 triggers an emergency obstacle avoidance signal within the 30cm distance threshold, the second motor 12 performs reverse pulse braking, the tire 11 stops instantly and fine-tunes in the reverse direction, and the path is replanned. After charging is completed, the charging piece 21 is out of the charging state, the laser radar 20 reloads the environmental map, and returns to the training area.

[0075] See also Figure 1 and Figure 8 The present invention provides an embodiment of a cognitive function training robot, wherein the active training mode is used to carry out structured cognitive training through multimodal interaction, specifically including the following tasks:

[0076] Task A: Graphic matching training based on touch screen 3, where users drag screen elements to complete logical associations;

[0077] Task B: Human posture training based on the high-definition display 15. The high-definition display 15 displays the training movements. After the user imitates them, the human posture is evaluated using image data, and corrections and encouragement are given using the speaker 10.

[0078] Task C: Use the HD camera 8 and the rear camera 9 to capture environmental objects, display them on the HD display 15 and mark them with red frames, and generate a physical object recognition question-answering task;

[0079] The task C specifically includes:

[0080] The rear camera 9 captures environmental objects at a 120° wide angle. After extracting at least three characteristic objects using an image segmentation algorithm, the high-definition display screen 15 displays the object images according to their spatial orientation. Each question is marked with a red frame. The speaker 10 speaks the question, and the touch screen 3 generates options based on the marked objects. The user must select the option that matches the voice question through the touch screen 3 within 10 seconds.

[0081] Furthermore, when executing Task A, graphic matching training is started, the first motor 7 drives the hinge 5 to adjust the touch interactive screen 3 to a suitable angle, and the speaker 10 plays the task instructions, such as "Please connect the animals with their habitats." The touch interactive screen 3 generates a split-screen interface, with dynamic biomes, such as swimming fish, displayed on the left, and environmental thumbnails, such as oceans and forests, displayed on the right. When the user drags an element, the built-in pressure sensor of the touch interactive screen 3 detects the touch point coordinate offset at intervals of 5ms, and calculates the graphic outline overlap in real time through an algorithm. When the user drags the shark icon to the ocean area, if the outline matching degree is ≥75%, the high-definition display 15 plays a 3D wave special effect, and the speaker 10 emits a positive sound effect. If a penguin is mistakenly dragged to the desert, the touch interactive screen 3 uses local pixel expansion technology to flash the wrong area in red. At the same time, the high-definition camera 8 captures the user's facial expression. If a frown is detected three times in a row, the graphic complexity is automatically reduced.

[0082] In addition, when executing Task C, during the environmental scanning phase, the rear camera 9 captures the indoor scene at a 110° wide-angle, and the high-definition camera 8 synchronously focuses to identify the user's clothing. The images captured by the two cameras are displayed on the high-definition display screen 15. Objects with an identification confidence level greater than 80% are marked with red boxes, and a red box is displayed. Based on the content of the red box, the touch interactive screen 3 generates options and asks questions to the user. The questions are progressive prompts. When the user touches to answer the questions, if the answer is incorrect, the touch interactive screen 3 highlights the correct option and marks the incorrect option in red. At the same time, the speaker 10 plays the answer analysis, and all interaction data will be uploaded to the training and evaluation system.

[0083] Working Principle: First, when the robot starts, the switch button 17 on the top of the shell 1 activates the system. The laser radar 20 on the mobile chassis 2 performs a 360-degree scan of the surrounding environment. Combined with the short-range detection data of the ultrasonic sensor 22, it constructs a three-dimensional indoor map and marks the boundaries of obstacles. The touch interactive screen 3 adjusts the pitch angle through the first motor 7 at both ends of the hinge 5, so that the high-definition camera 8 faces the user's face and simultaneously collects height data and facial feature information. The shock-absorbing bracket 19 absorbs ground vibrations to ensure the stability of the upper components of the shell 1 during the collection process, thus establishing a spatial benchmark and user profile for subsequent training.

[0084] Next, in autonomous following mode, the depth camera 16 measures the distance between the user and the robot in real time. When the distance exceeds a threshold of 1.5 meters, the two second motors 12 of the mobile chassis 2 drive the symmetrical tires 11 to rotate at a differential speed. The central universal wheel 13 deflects synchronously with the steering angle, and the robot follows the user along the path planned by the lidar 20. If the user selects active training mode, the touch screen 3 interacts with the user, the high-definition display 15 displays images, and the pressure sensor of the touch screen 3 captures the touch point trajectory in real time. When the user drags the icon to the wrong area, the speaker 10 plays a correction prompt.

[0085] Finally, at the end of the training, the charging piece 21 on the side of the mobile chassis 2 is docked with the charging pile through the precise positioning of the ultrasonic sensor 22, the shock-absorbing bracket 19 buffers the mechanical impact during docking, and the touch interactive screen 3 integrates the facial expression data collected by the high-definition camera 8, the joint movement trajectory and training task completion data recorded by the depth camera 16 to generate a multi-dimensional evaluation chart, which is encrypted and uploaded to the medical database through the wireless module inside the shell 1. The emergency stop button 18 is in a triggerable state throughout the process. When the emergency stop button 18 is triggered, the system automatically cuts off the power supply of the second motor 12 and starts the protection mechanism.

[0086] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. A method for using a cognitive function training robot, characterized in that: It includes autonomous following mode and active training mode. The autonomous following mode is used to replace the voice communication training between the person and the patient, and performs real-time tracking of the human body, voice reminders, and accompanying training during the patient's walking process. The autonomous following mode includes the following steps: When the user starts the autonomous following mode, the high-definition camera (8) on the touch interactive screen (3) captures the scene in front in real time, locks the target user through image recognition and tracks the direction of movement, and the first motor (7) drives the hinge (5) to rotate, so that the mounting plate (6) drives the touch interactive screen (3) to adjust the pitch, ensuring that the high-definition camera (8) is always facing the user's upper body to collect dynamic postures. At the same time, the depth camera (16) measures the real-time distance between the user and the robot based on the binocular vision principle, and feeds the distance data back to the control system of the mobile chassis (2). The two second motors (12) at the bottom of the mobile chassis (2) drive the symmetrically distributed tires (11) to rotate at a differential speed according to the direction command, and control the robot to move forward, turn or stop.

2. The method for using a cognitive function training robot according to claim 1, characterized in that: The active training mode is used to carry out structured cognitive training through multimodal interaction, specifically including the following tasks: Task A: Graphic matching training based on the touch interactive screen (3), where the user drags screen elements to complete logical associations; Task B: Human posture training based on a high-definition display (15). The high-definition display (15) displays the training movements. After the user imitates, the human posture is evaluated through image data, and the speaker (10) is used to correct and encourage; Task C: Use the high-definition camera (8) and the rear camera (9) to shoot environmental objects, display them on the high-definition display (15) and mark them with red frames, and generate a physical object recognition question-answering task.

3. The method for using a cognitive function training robot according to claim 1, characterized in that: The autonomous following mode also includes an environment adaptation step: The laser radar (20) scans obstacles in the travel path at a frequency of 10 Hz and generates a dynamic obstacle avoidance trajectory in combination with the close-range detection data of the ultrasonic sensor (22). When an obstacle is detected within 1 meter in the direction of travel, the mobile chassis (2) controls the two second motors (12) to perform differential in-situ steering. During the steering process, the shock-absorbing bracket (19) ensures that the pitch angle fluctuation of the touch interactive screen (3) is less than ±2°.

4. The method for using a cognitive function training robot according to claim 1, characterized in that: The usage method also includes the user identity binding process: During the first use, the high-definition camera (8) collects the user's facial features to establish a three-dimensional model, and generates an encrypted user profile together with the physiological parameters input by the touch interactive screen (3). During subsequent startups, historical training data is automatically loaded through face recognition. When the recognition confidence is lower than 85%, the touch interactive screen (3) pops up a password verification interface for secondary authentication.

5. The method for using a cognitive function training robot according to claim 2, characterized in that: The task C specifically includes: The rear camera (9) captures environmental objects at a 120° wide angle. After extracting at least three characteristic objects through an image segmentation algorithm, the high-definition display screen (15) displays the object images according to their spatial orientation. Each time a question is asked, one of the objects is marked with a red frame. The speaker (10) asks the question, and the touch screen (3) generates options based on the marked objects. The user needs to select the option that matches the voice question through the touch screen (3) within 10 seconds.

6. The method for using a cognitive function training robot according to claim 1, characterized in that: The usage method also includes an exception handling process: When the emergency stop button (18) is triggered, the second motor (12) cuts off the power output within 0.2 seconds, and the hydraulic damper of the shock-absorbing bracket (19) is automatically locked. The laser radar (20) immediately starts a 360° panoramic scan and marks a safe docking point. During the docking process, the speaker (10) continues to play a soothing voice until the system is completely shut down.

7. The method for using a cognitive function training robot according to claim 1, characterized in that: The autonomous following mode includes a human posture recognition and anti-fall mechanism, specifically including: The three-dimensional coordinates of the user's skeleton points are collected in real time by a depth camera (16) and a high-definition camera (8) to construct a dynamic posture model; The fall prevention mechanism includes three levels of response: Level 1 response: When the robot detects an obstacle on the ground in front of the user through the laser radar (20), the speaker (10) issues a voice reminder; Secondary response: When the user falls, the speaker (10) plays an 85dB warning sound and activates the red strobe light of the emergency stop button (18) on the top of the housing (1); Level 3 response: When the user does not stand up for a period of time, the touch screen (3) automatically pops up the emergency contact interface and makes a call.

8. A cognitive function training robot, suitable for the method for using a cognitive function training robot according to any one of claims 1 to 7, characterized in that: It comprises a housing (1), a mobile chassis (2) and a touch interactive screen (3), wherein the mobile chassis (2) is mounted on the bottom of the outer wall of the housing (1), and the touch interactive screen (3) is mounted on the top of the outer wall of the housing (1) via a connecting component; The connecting assembly comprises: a fixing frame (4), a hinge (5) and a mounting plate (6); A fixing frame (4) is fixedly installed on the top of the outer wall of the shell (1), and the top of the outer wall of the fixing frame (4) is connected to a mounting plate (6) via a hinge (5). A first motor (7) is installed at both ends of the outer wall of the hinge (5). A touch interactive screen (3) is fixedly connected to the front of the outer wall of the mounting plate (6), a high-definition camera (8) is installed inside the front of the touch interactive screen (3), a rear camera (9) is installed inside the back of the touch interactive screen (3), and a speaker (10) is installed inside the back of the touch interactive screen (3).

9. The cognitive function training robot according to claim 8, characterized in that: The mobile chassis (2) is circular, and has three grooves (14) distributed along the circumference at 110 degrees at the bottom. A second motor (12) is installed in two symmetrical grooves (14) to drive the tire (11). A universal wheel (13) that can be deflected by ±45 degrees is installed in the middle groove (14). The top of the outer wall of the mobile chassis (2) is connected to the shell (1) through a shock-absorbing bracket (19). The shock-absorbing bracket (19) has an adjustable damping hydraulic cylinder built in, and its damping coefficient is linked to the ground roughness data detected by the laser radar (20) and adjusted.

10. The cognitive function training robot according to claim 8, characterized in that: The housing (1) is provided with a high-definition display screen (15) and a depth camera (16) on the front, and a switch button (17) and an emergency stop button (18) are integrated on the top. A charging plate (21) is provided on the side plane of the mobile chassis (2), and six ultrasonic sensors (22) are evenly distributed on the cylindrical surface. A laser radar (20) is installed at the connection gap between the mobile chassis (2) and the housing (1), and the scanning plane is parallel to the horizontal plane.

Images