Upper limb rotation training device and active intention based scene interaction system

Abstract

The application discloses a scene interaction system based on upper limb rotation training equipment and active intention, which comprises upper limb rotation training equipment and a controller. The equipment comprises an operating rod, a rotation driving module and a multi-modal sensing unit, and can collect upper limb horizontal turning angular velocity and active force value. The controller receives multi-modal signals to obtain user upper limb safe angle range, maximum acceleration, maximum active force value and state coefficient; real-time signal acquisition is performed to determine current acceleration intensity, to judge whether the movement is in a safe range, to calculate active intention index based on normalized force value, angular velocity and current acceleration intensity, and to determine whether it is an effective intention; if yes, the user state coefficient, curve curvature coefficient and angular velocity are fused, a virtual control object is taken as a gamified scene carrier, and a mapping relationship between upper limb rotation action and virtual curve passing scene is established. The system establishes a fine and individualized mapping relationship between user upper limb rotation action and virtual control object model curve passing scene.

Description

Technical Field

[0001] This invention relates to the field of scene interaction technology, specifically to a scene interaction system based on upper limb rotation training equipment and active intention. Background Technology

[0002] Effective exercise training is crucial for improving users' motor function, especially active training modes combined with interactive scenarios, which can significantly enhance user engagement and training effectiveness. However, current applications of upper limb rotation training equipment, exercise training and assessment technologies, and interactive scenarios face several pressing issues that need to be addressed:

[0003] Passive training is the primary method, making it difficult to quantify active participation. Furthermore, the training difficulty cannot be dynamically adjusted based on individual user abilities, resulting in problems such as strong passive mechanicality, low active participation, and a mismatch between difficulty and user condition. Traditional sports training often relies on passive modes (such as equipment-driven limb movements), but the improvement of exercise effects highly depends on the user's active intention and level of participation. Due to the lack of effective means to quantify the user's active intention, it is impossible to accurately determine the user's level of active participation during training. Moreover, the training difficulty uses fixed parameters that cannot be adapted to the user's condition in real time.

[0004] More importantly, existing upper limb rotation training equipment lacks a deep integration design with scene interaction. It has not built an integrated system of equipment, active intent recognition, and scene interaction, which makes it impossible for users' upper limb rotation movements to achieve precise adaptation and linkage feedback with virtual scenes (such as virtual turning, scene control, etc.). At the same time, the lack of scene-based interactive guidance and feedback mechanisms during training further reduces users' motivation to participate actively. It is also impossible to achieve visualization and fun evaluation of training effects through scene interaction, making it difficult to form a closed loop of "equipment training - intent recognition - scene interaction - effect feedback" and failing to meet users' personalized and scene-based training needs. Summary of the Invention

[0005] The technical objective of this invention is to address the technical problems of existing sports training methods being passive, lacking scene interaction integration, and not incorporating subjective intentions, by providing a scene interaction system based on upper limb rotation training equipment and active intentions.

[0006] To achieve the above-mentioned technical objectives, the present invention adopts the following technical solution.

[0007] This application provides a scenario interaction system based on an upper limb rotation training device and active intention, the system comprising:

[0008] An upper limb rotation training device includes an operating lever, a rotation drive module that assists the user in horizontal upper limb turning movements, and a multimodal sensing unit for collecting multimodal signals when the user performs horizontal upper limb turning movements using the operating lever. The multimodal signals include the angular velocity of the horizontal upper limb turning movement, the motion acceleration, and the active force value.

[0009] The controller is configured to perform the following steps:

[0010] S1: Pre-training calibration data process: Receive the multimodal signals collected by the multimodal sensing unit during the user's horizontal turning motion, and obtain the safe angle range, maximum motion acceleration value, maximum active force value, and user state coefficient of the user's upper limb rotation joint activity;

[0011] S2: Training process: Receive multimodal signals from the user's upper limb horizontal turning motion collected in real time by the multimodal sensing unit; determine the current acceleration intensity value based on the user's current active force value, current motion acceleration, maximum active force value, and maximum motion acceleration value; determine whether the horizontal turning motion is within a safe angle range based on the upper limb horizontal turning motion angular velocity value; calculate the active intention index based on the normalized active force value, upper limb horizontal turning motion angular velocity value, and current acceleration intensity value; and determine whether the active intention is valid based on the active intention index.

[0012] S3: If the intent is valid, the user state coefficient, curve curvature coefficient and upper limb horizontal turning motion angular velocity are integrated, and a simplified virtual control object model is used as the gamified scene carrier to establish a mapping relationship between the user's upper limb rotation action and the virtual control object model's cornering scene.

[0013] This application constructs an integrated interactive system based on upper limb rotation training equipment and active intentions. By combining the joint safety angle range, maximum active force value, maximum motion acceleration value, user state coefficient, and effective active intentions, it achieves coordinated linkage between equipment operation, active intention recognition, scene interaction, and training difficulty adjustment. The specific beneficial effects are as follows:

[0014] 1. Safe and controllable, adapted to individual physiological characteristics: This system relies on the data collection capabilities of the upper limb rotation training equipment, and sets the action boundaries by combining the user's exclusive safety angle range and force exertion ability to avoid joint damage caused by excessive exercise; at the same time, it dynamically and personally adapts to the physiological limits of users at different training stages through user state coefficients to ensure the safety and pertinence of the system operation.

[0015] 2. Precisely respond to active intentions and improve training efficiency: This system adjusts the training difficulty and scene interaction feedback in real time based on the intensity of effective active intentions, ensuring that the training intensity, scene interaction rhythm and user's active exercise intention are precisely matched. This avoids frustration caused by excessive difficulty and prevents training efficiency from being reduced by insufficient difficulty, thus fully mobilizing the user's active participation.

[0016] 3. Achieve deep linkage between equipment, intent, and scene to build an integrated interactive system: This system organically combines the motion acquisition, active intent recognition, motion amplitude constraint, and scene interaction difficulty adjustment of the upper limb rotation training equipment, so that the entire training process not only conforms to the individual physiological characteristics of the user, but also accurately responds to their active movement intent, thereby establishing a refined and individualized mapping relationship between the user's upper limb rotation movements and the virtual control object model and scene interaction (such as cornering, scene control, etc.).

[0017] 4. Enrich training scenarios and improve feedback loop: This system provides users with visual and engaging training scenarios through the scenario interaction module, enhancing the fun and sustainability of training. At the same time, the scenario interaction process can provide real-time feedback on the user's training effect, providing scenario response data and test routes for subsequent cognitive tasks and training program optimization, further improving the closed-loop system of "training-recognition-interaction-feedback-optimization", and enhancing the practicality and application value of the entire system.

[0018] It should be understood that the summary section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0019] The accompanying drawings described herein are for illustrative purposes only and are not intended to limit the scope of this application in any way. Furthermore, the shapes and scales of the components in the drawings are merely illustrative to aid in understanding this application and do not specifically limit the shapes and scales of the components. Those skilled in the art, guided by the teachings of this application, can select various possible shapes and scales to implement this application according to specific circumstances. In the drawings:

[0020] Figure 1 A schematic diagram illustrating the user interaction process of the scene interaction system based on upper limb rotation training equipment and active intention provided in this embodiment;

[0021] Figure 2 This is a schematic diagram of the overall structure of the upper limb rotation training device used in the embodiment;

[0022] Figure 3 It is a car gear table showing the cornering speed at curve 1 at a certain upper limb rotation speed when the user controls the vehicle model using the upper limb rotation training device.

[0023] Figure 4 Is with Figure 3 When the same user operates the same upper limb rotation training device to control the same vehicle model, the cornering speed at curve 2 is represented by the car gear table at the same upper limb rotation speed.

[0024] Figure 5 Is with Figure 3 When the same user operates the same upper limb rotation training device to control the same vehicle model, the cornering speed at curve 3 is represented by the car gear table at the same upper limb rotation speed.

[0025] Figure 6 This is a schematic diagram illustrating the process of detecting motion time and total reaction time during training in the embodiment.

[0026] Figure 7 This is a schematic diagram of the force feedback calculation and adaptive adjustment process in the embodiment;

[0027] Figure 8 A schematic diagram of a training device structure that can provide force feedback is provided for an embodiment.

[0028] Figure 9 This is a schematic diagram illustrating the adjustment of training difficulty and safety monitoring in the embodiment;

[0029] Reference numerals: 1. Operating lever; 2. Bearing housing; 3. Pressure sensor; 4. Interface module; 41. Power interface; 42. Encoder interface; 5. Rotary drive module; 6. Base; 7. Rocker arm bracket; 8. Drive shaft; 9. Transmission unit; 91. Driving bevel gear; 92. Driven bevel gear. Detailed Implementation

[0030] To enable those skilled in the art to better understand the technical solutions in this application, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this application.

[0031] The following embodiments are provided, taking users with limb training needs as the application target, and detailing the specific implementation process of a scene interaction system based on upper limb rotation training equipment and active intention. It should be noted that the core of the system control logic in any of the above embodiments is to combine user physiological characteristics with effective active intention to achieve safe and personalized scene interaction training control. The above technical logic is universal; for ordinary users without training needs, parameters such as the joint safety angle range, training difficulty threshold, and user state coefficient in this embodiment can be adaptively adjusted according to their daily exercise and physical training needs and can be directly applied without changing the core control steps of this system. The scene interaction system based on upper limb rotation training equipment and active intention in any of the above embodiments can also be directly applied to the exercise training needs of ordinary users.

[0032] This application provides a scenario interaction system based on an upper limb rotation training device and active intent. The system includes an upper limb rotation training device and a controller. The upper limb rotation training device includes a joystick, a rotation drive module to assist the user in horizontal upper limb turning movements, and a multimodal sensing unit for collecting multimodal signals when the user uses the joystick to perform horizontal upper limb turning movements. The multimodal signals include the angular velocity θ of the horizontal upper limb turning movement, the motion acceleration E, and the active force value F.

[0033] The controller is configured to perform the following steps:

[0034] S1: Pre-training calibration data process: Receive multimodal signals collected by the multimodal sensing unit during the user's horizontal turning motion, and obtain the safe angle range, maximum motion acceleration value, maximum active force value, and user state coefficient of the user's upper limb rotation joint activity;

[0035] S2: Training Process: The multimodal sensing unit collects multimodal signals from the user's upper limb horizontal turning motion in real time; the current acceleration intensity value is determined based on the user's current active force value, current motion acceleration, maximum motion acceleration value, maximum active force value, and maximum motion acceleration value; the horizontal turning motion angular velocity value is used to determine whether the horizontal turning motion is within a safe angle range; then, based on the normalized active force value, upper limb horizontal turning motion angular velocity value, and current acceleration intensity value, the active intention index is calculated; and the active intention index is used to determine whether it is a valid active intention.

[0036] S3: If the intent is valid, the user state coefficient, curve curvature coefficient and upper limb horizontal turning motion angular velocity θ are integrated, and a simplified virtual control object model is used as the gamified scene carrier to establish a mapping relationship between the user's upper limb rotation action and the virtual control object model's cornering scene.

[0037] In this embodiment, a mapping relationship is established between the actual limb horizontal turning motion angular velocity value and the training scenario difficulty parameter value. At the same time, the training scenario difficulty is dynamically adjusted based on the changes in the user state coefficient and active intention index to achieve personalized adaptation of the movement range and training difficulty.

[0038] The entire upper limb rotation training equipment, such as Figure 2 As shown, it resembles a hand-cranked wheelchair-assist vehicle.

[0039] like Figure 8 As shown, the upper limb rotation training device in this embodiment includes two symmetrically arranged operating rods 1, bearing seats 2, rocker arm brackets 7, transmission shafts 8 and transmission units 9, and a rotation drive module 5.

[0040] The two control levers 1 are rigidly connected to the transmission shaft 8 through two rocker arm brackets 7 respectively. The operator controls the longitudinal rotation of the control levers 1, which is converted into the rotational motion of the transmission shaft 8 through the rocker arm brackets 7. The operator controls the swinging and turning action of the control levers 1, which drives the rotation drive module 5 to rotate horizontally.

[0041] In some embodiments, the multimodal sensing unit includes a pressure sensor 3. The operating force exerted by the operator on the control lever 1 is collected by the pressure sensor 3. The interface module 4 includes a power interface 41 and an encoder interface 42, with the power interface 41 providing power input. The drive shaft 8 is supported by a bearing housing 2 to ensure rotational accuracy. The transmission unit 9 stably transmits the torque generated by the operation of the control lever 1 through gear meshing, ensuring effective transmission of steering power. The transmission unit 9 includes a driving bevel gear 91 and a driven bevel gear 92, wherein the driving bevel gear 91 is fixed to the drive shaft 8 and changes the transmission direction by meshing with the driven bevel gear 92, transmitting motion to the multimodal sensing unit.

[0042] The multimodal sensing unit also includes an inertial measurement unit (IMU) (not shown in the figure) and an angle sensor (not shown in the figure). The IMU can be integrated into two joysticks 1 to realize motion acceleration detection, steering angle detection, and limb motion data output. The angle sensor can be coaxially mounted with the driven bevel gear 92 to provide feedback on the forward and backward angles of the virtual control object model. The angle sensor, such as a steering wheel magnetic encoder (1024 lines / revolution), outputs angle sensing signals through the encoder interface 42.

[0043] In some embodiments, the rotary drive module 5 includes a motor and a drive unit, which can be used to output lateral horizontal steering power. Optionally, the rotary drive module 5 also includes an electromagnetic damper (not shown in the figure), which is coaxially mounted with the motor and connected to the control module to output dynamic damping force to achieve force feedback.

[0044] In some embodiments, the rotary drive module 5 may not include a motor, but only a steering transmission bushing assembly with meshing teeth and an electromagnetic damper for output force feedback. The bushing is mounted on the base 6 via bearings to ensure smooth rotation.

[0045] In some embodiments, the angle sensor may be mounted coaxially with the motor power output shaft to measure the steering angle.

[0046] In this embodiment, the base 6 provides support to support components such as the rocker arm bracket 7 and the rotation drive module 5, providing a stable foundation.

[0047] In some embodiments, the action intent index in step S2 is calculated using a multi-signal weighted average, as shown in the formula:

[0048] I=α×E normal +β×F normal +γ×θ normal；

[0049] Where I represents the proactive intent index, and E... normal F is the normalized value of the acceleration intensity. normal θ is the normalized active force value. normal γ represents the normalized angular velocity value of the upper limb during horizontal turning motion, where α is the acceleration value weight, β is the active force value weight, and γ is the horizontal turning motion angular velocity weight.

[0050] In some embodiments, α=0.3, β=0.3, γ=0.3.

[0051] Active Intent Index ,like Figure 1 As shown, in some embodiments, when It was determined to be a valid proactive intent.

[0052] In this embodiment, the motion acceleration value E and the current active force value F are fused together to calculate the motion acceleration intensity value E. new The formula is as follows: ;

[0053] Where E max F represents the maximum acceleration value of the user's upper limbs during horizontal turning movements using the joystick. max E represents the maximum active force exerted by the user when performing horizontal upper limb turning movements using the joystick (personalized calibration), E represents the current motion acceleration value of the user when performing horizontal upper limb turning movements using the joystick, and F represents the current active force exerted by the user when performing horizontal upper limb turning movements using the joystick. new This is the current motion acceleration intensity value that combines this acceleration with the active force exertion.

[0054] In this embodiment, the maximum motion acceleration value E can be obtained through three pre-training sessions. max Maximum active force value F max .

[0055] The current acceleration intensity value E is respectively new After normalizing the active force value F and the angular velocity θ, we obtain E. normal F normal and θ normal E normal F normal and θ normal The value range is 0 to 1.

[0056] In some embodiments, step S2, after calculating the active intent index, further includes an intent reliability determination step: using the formula: ;

[0057] Where R is the intent reliability parameter, I(t) is the active intent index at time t, t0 is the start time of intent monitoring, t1 is the end time of monitoring, and the monitoring duration can be set, such as t1-t0=0.5s; in some embodiments, when R>intent reliability threshold R th When R is 0.5, it is determined to be a reliable active intent, triggering S3 to achieve dynamic response in the training scenario; when R ≤ the intent reliability threshold R th If the value is 0.5, it is determined to be an unreliable intent, and S3 will not be triggered to respond to the scene. Instead, multimodal signals will be continuously collected to recalculate R.

[0058] In some embodiments, to address the issue of asynchrony between intent and action, step S2 further includes correcting the timeliness parameter I(t) of the intent at time t, as shown in the following formula: ;

[0059] Where I is the active intention index, t delay Let λ be the action response delay (in seconds), and λ be the attenuation coefficient. In some embodiments, λ may be 0.5.

[0060] In some embodiments, the virtual control object model (which may be a simplified physical model of a vehicle, an animal, a spaceship, etc. in some embodiments) provides a gamified training scenario carrier for the upper limb rotation training device, which then achieves "gamification of training" by mapping the user's upper limb rotation movements to the control logic of the virtual control object model.

[0061] If a virtual vehicle model is used, parameters for the cornering scenario (such as virtual vehicle speed v, corner radius r, and vehicle steering angle δ) can be provided as the basis for training difficulty and interaction.

[0062] In this embodiment, the upper limb rotation training device collects the angular velocity θ of the user's upper limb during horizontal turning motion, and the controller converts it into control commands for the vehicle model (such as mapping the angular velocity θ during horizontal turning motion to the vehicle steering angle δ).

[0063] As an example, step S3 uses a gamified training scenario. Taking a vehicle as the virtual control object model, the specific steps are: the gamified training scenario is a vehicle cornering scenario, establishing a mapping relationship between the user's upper limb rotational movements and the vehicle cornering scenario. As an example, the mapping relationship includes converting the angular velocity θ of the user's upper limb horizontal steering movement into the steering angle δ of the virtual control object model, using the following formula:

[0064] δ=k p ×k c ×θ;

[0065] Where δ is the steering angle (°) of the virtual control object in the game (such as a vehicle), θ is the angular velocity (°) of the user's upper limbs during horizontal steering motion, and k p k represents the user state coefficient. c This is the curvature coefficient of the curve.

[0066] As an example, gentle curve k c =0.8, sharp bend k c =1.2.

[0067] In some embodiments, the user state coefficient k can be dynamically and adaptively adjusted. p For example, in the early stages of training, k p Take 0.6, later k p Take 1.

[0068] In some embodiments, the user state coefficient k can also be dynamically updated. p If recalculated every 3 training sessions Where ΔS represents the improvement in the overall training effect score.

[0069] Force feedback during training is a key means to help users perceive movement and promote functional recovery. However, the force feedback mechanisms of existing devices are mostly based on preset parameters and cannot be dynamically adjusted by combining the intensity of the user's active intention (such as the degree to which the user attempts to exert force) and training scenario parameters (such as different types of movements, training difficulty levels, etc.). This makes it difficult to achieve precise training that provides "as much force feedback as the user needs" and fails to meet the personalized needs of different users at different training stages.

[0070] To solve the above problems, in some embodiments, S4: If it is a valid active intention, generate an adaptive force feedback value based on the user status coefficient and the active intention index, and input it to the rotation drive module of the upper limb rotation training device to apply a dynamic force to the user's upper limb to achieve the collaborative output of the user's active force and the device's auxiliary force. The generation of the adaptive force feedback based on the user status coefficient and the active intention index includes: determining the basic resistance according to the user status coefficient, the mass of the virtual control object model (such as a vehicle), the virtual speed in the training scenario, the reference distance in the training scenario, and the scenario operation angle mapped based on the angular velocity θ during the horizontal turning movement of the upper limb; determining the adaptive force feedback value according to the basic resistance and the user's active intention index; referring to the maximum active force value, setting the force feedback safety threshold to ensure that the adaptive force feedback value does not exceed the force feedback safety threshold and avoid the adaptive force feedback value causing the movement to exceed the safe angle range.

[0071] As an example, the formula for determining the basic resistance (simulating the centrifugal force during cornering) is as follows:

[0072] ; <​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​​Referencing the maximum safe force feedback value, ensure that the adaptive force feedback value does not exceed the force feedback safety threshold. The expression is: F feedback For the maximum safety force feedback value, such as =15N. If the adaptive force feedback value exceeds the force feedback safety threshold ( If this occurs, the safety protection will be triggered, and training will be suspended.

[0079] In some embodiments, the horizontal turning angle of the upper limb can be calculated based on the angular velocity value θ of the horizontal turning motion of the upper limb. The safe range of motion for the user's upper limb rotation joints can be set, such as... , where Φ max For maximum safety angle, .

[0080] In some embodiments, a rotation direction correction factor is included. Simulates the lateral force components during cornering, including: ;in, To simulate the lateral force components during cornering, For the optimal steering angle, F basic The lateral force input by the user generates lateral resistance when the actual angle deviates from the target, thus enhancing lateral resistance training.

[0081] Improving overall limb motor function is a complex process involving the coordinated recovery of motor and cognitive functions (for example, stroke patients often have both motor control and cognitive impairments). However, existing assessment methods often focus only on motor dimensions (such as joint range of motion and muscle strength) or assess cognitive function separately, lacking technical means to simultaneously assess the cognitive-motor coupling effect. This fails to comprehensively and accurately reflect the user's training progress and is not conducive to providing a comprehensive basis for optimizing limb training programs. This application uses a cognitive-motor coupling difficulty formula ( Simultaneous evaluation of dual-dimensional training effects can comprehensively and accurately reflect the user's training progress.

[0082] In some embodiments, such as Figure 9 As shown, step S3 also includes a training effect detection process. The detection process takes the user's vehicle cornering scenario in the gamified training scenario as the object, and specifically includes: S31: Start timing from when the user-controlled virtual vehicle enters the starting position of the curve until the user-controlled virtual vehicle drives out of the test position. This time period is recorded as the total reaction time. S32: After the timing ends, a detection cycle begins, which monitors the angular velocity of the user's upper limbs as they operate the handlebars in real time. It records all time segments where the angular velocity of the horizontal turning motion exceeds a preset threshold (the threshold can be calibrated, such as 30% of the user's pre-trained maximum rotational angular velocity). All time segments are then summed to obtain the motion execution time. S33: The detection process begins when the virtual vehicle enters the curve and ends when the virtual vehicle exits the curve. and Calculate decision reaction time :

[0083] This is used to evaluate the effectiveness of user cognitive-motor coordination training. The process of detecting movement time and total reaction time is as follows: Figure 6 As shown. The ideal reaction time requires a decision reaction time. 1.5s (later training <1.0s).

[0084] S34: Evaluate the effectiveness of user cognitive-motor coordination training based on decision-making reaction time, and determine a comprehensive training effectiveness score, including: Where S represents the overall training performance score. This score combines accuracy, precision, and reaction speed, with a maximum score of 100. For cognitive-motor coupling scoring.

[0085] In the embodiments, ;

[0086] The normalized value of the motion error is given by the following formula: , where t c It is the detection start time, t end It is the end of training, θ t θ is the angular velocity of the upper limb horizontal turning motion at time t during training. target D represents the target angular velocity of the upper limb's horizontal turning motion; D represents the cornering difficulty level of the virtual control object model; D cognize Error rate for cognitive tasks.

[0087] In the embodiment, the cognitive task error rate D cognize The formula for determining it is as follows:

[0088]

[0089] in, For cognitive task error rate, This represents the number of times a cognitive task test result is correct. This represents the total number of tests conducted on the cognitive task.

[0090] As an example, the cognitive task in this embodiment can take two forms: First, a normal track is designated, and the reaction time during cornering is used to determine the outcome, such as measuring the overall reaction time at a corner. Second, a distraction track is designated. "Attention anchors" (such as colored road signs or temporary speed limit signs) are set on the track, and players must keep a close eye on them and perform corresponding actions (such as slowing down or using turn signals). Missing or misreading these points results in point deductions, thus training sustained attention and selective attention. In this mode, the number of times the user continuously misses or misreads a point is recorded, and the cognitive error rate is calculated. The two forms of cognitive tasks have different complexities, C.

[0091] In some embodiments, the cornering difficulty level D of the virtual control object model is dynamically adjusted based on the virtual speed v, the cornering radius r, and the cognitive task complexity C of the virtual control object model. The adjustment method for the cornering difficulty level D of the virtual control object model is as follows:

[0092] ;

[0093] Among them, v, , k c Here, v represents the scene classification parameter, and v represents the speed (m / s) of the virtual control object model (virtual vehicle). The radius of the curve is (m). The complexity of the cognitive task is 0~1.

[0094] In this embodiment, the cornering difficulty of the virtual control object model can be adjusted according to the cornering difficulty level D of the virtual control object model.

[0095] In some embodiments, the user state coefficient k is used. p The cognitive complexity C can achieve coordinated regulation of "motor-cognition-feedback".

[0096] For specific details, please refer to the parameter table in Table 1 below for dynamic adjustment.

[0097] Table 1 Scene Classification Parameter Table

[0098]

[0099] Figures 3-4 This represents the cornering speed of the same vehicle, the same user, and the same upper limb rotation speed at different bends. The speed readings on the gear selector (left) and tachometer (right) at different bends can be seen. These values ​​will gradually increase as training progresses. Figure 3 The speed limit is shown as 3rd gear at 70km / h on the first curve. Figure 4 The speed limit is shown as 70 km / h in third gear on the curve. Figure 5 The speed is shown as 103 km / h in 4th gear on the curve.

[0100] The physical feedback (such as centrifugal force when cornering) of the virtual control object model (such as the vehicle model) is transformed into resistance that the user can perceive through the force feedback system of the upper limb rotation training device. At the same time, the scene dynamics of the vehicle model (such as changes in cornering difficulty) are linked with the training difficulty adjustment of the training device to jointly achieve the goal of driving active training with gamified scenarios.

[0101] In some embodiments, step S3 further includes monitoring fatigue level, with the fatigue level detection value calculated using the following formula: ;

[0102] F tired E represents the fatigue test value. current E represents the current acceleration intensity value. start F represents the initial maximum acceleration value. current F represents the current active force value. start This represents the maximum active force value in the initial period.

[0103] F tired Fatigue threshold (e.g., 0.3, representing the fatigue detection value) If the user's fatigue level drops by more than 30% from the initial value, then the user is considered fatigued, and a rest prompt is triggered.

[0104] The controller described in the above embodiments can be implemented by a computer chip or physical entity, or by a product with certain functions. A typical implementation device is a computer. Specifically, the computer can be, for example, a personal computer, a laptop computer, a tablet computer, or any combination of these devices.

[0105] The above provides a detailed description of the upper limb rotation training device and active intention-based scene interaction system provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the concept of this application and should not be construed as a limitation on the scope of protection of this application.

Claims

1. A scene interaction system based on upper limb rotation training equipment and active intention, characterized in that, The system includes: An upper limb rotation training device includes an operating lever, a rotation drive module to assist the user in horizontal upper limb turning movements, and a multimodal sensing unit for collecting multimodal signals when the user performs horizontal upper limb turning movements using the operating lever. The multimodal signals include the angular velocity of the horizontal upper limb turning movements, the acceleration of the movements, and the value of the active force exerted. The controller is configured to perform the following steps: S1: Pre-training calibration data process: Receive the multimodal signals collected by the multimodal sensing unit during the user's horizontal turning motion, and obtain the safe angle range, maximum motion acceleration value, maximum active force value, and user state coefficient of the user's upper limb rotation joint activity; S2: Training process: Receive multimodal signals from the user's upper limb horizontal turning motion collected in real time by the multimodal sensing unit; determine the current acceleration intensity value based on the user's current active force value, current motion acceleration, maximum active force value, and maximum motion acceleration value; determine whether the horizontal turning motion is within a safe angle range based on the upper limb horizontal turning motion angular velocity value; calculate the active intention index based on the normalized active force value, upper limb horizontal turning motion angular velocity value, and current acceleration intensity value; and determine whether the active intention is valid based on the active intention index. S3: If the intent is valid, the user state coefficient, curve curvature coefficient and upper limb horizontal turning motion angular velocity are integrated, and a simplified virtual control object model is used as the gamified scene carrier to establish a mapping relationship between the user's upper limb rotation action and the virtual control object model's cornering scene.

2. The scene interaction system according to claim 1, characterized in that, The mapping relationship mentioned in step S3 includes converting the angular velocity θ of the user's upper limb during horizontal turning motion into the turning angle δ of the game's virtual control object model, as expressed below: δ=k p ×k c ×θ; Where k c k is the curvature coefficient of the curve. p For user status coefficients.

3. The scene interaction system according to claim 1, characterized in that, The controller is also configured to execute step S4: S4: If it is a valid active intention, an adaptive force feedback value is generated based on the user state coefficient and the active intention index, and input to the rotation drive module of the upper limb rotation training device, so as to achieve the coordinated output of the user's active force and the device's auxiliary force by applying dynamic force to the user's upper limb. The adaptive force feedback value is generated based on the user state coefficient and the active intent index, including: The basic resistance is determined based on the steering angle δ of the virtual control object model, the mass of the virtual control object model, the virtual speed of the virtual control object model in the training scenario, the cornering radius, and the user state coefficient. The adaptive force feedback value is determined based on the basic resistance and the user's active intention index.

4. The scene interaction system according to claim 1, characterized in that, The active intent index is calculated using a multi-signal weighting method, and the formula is as follows: I=α×E normal +β×F normal +γ×θ normal , Where I represents the proactive intent index, and E... normal F is the normalized value of the acceleration intensity. normal θ is the normalized active force value. normal γ represents the normalized angular velocity of the upper limb during horizontal turning motion, where α is the acceleration value weight, β is the active force value weight, and γ is the upper limb horizontal turning motion angular velocity weight.

5. The scene interaction system according to claim 1, characterized in that, In step S2, after calculating the proactive intent index, the step of determining the reliability of the intent is also included: using the formula: ; Where R is the intention reliability parameter, I(t) is the active intention index at time t, t0 is the intention monitoring start time, and t1 is the monitoring end time; When R > intentional reliability threshold R th When R is determined to be a reliable active intention, S3 is triggered; when R ≤ the intention reliability threshold R th If the intention is deemed unreliable, S3 will not be triggered temporarily, and multimodal signals will continue to be collected to recalculate R.

6. The scene interaction system according to claim 5, characterized in that, The controller is also configured to correct the active intention index I(t) at time t, as shown in the following formula: ； Where I is the active intention index, t delay λ represents the action response delay, and λ is the attenuation coefficient.

7. The scene interaction system according to claim 1, characterized in that, Step S3 also includes a training effect detection process, which takes the scenario of a user navigating a curve using a virtual control object model in a gamified training scenario as the object, and specifically includes: S31: Timing starts when the user-controlled virtual control object model enters the starting position of the curve and stops when the user-controlled virtual control object model leaves the detection position on the track. This time period is recorded as the total reaction time. ; S32: Total reaction time After the timing ends, a detection cycle begins, which monitors the angular velocity of the user's upper limb horizontal turning motion as they operate the joystick in real time. All time segments with a horizontal turning motion angular velocity greater than a preset threshold are recorded, and all these time segments are summed to obtain the pure motion execution time. ; S33: The detection process begins when the virtual control object model enters the starting position of the curve and ends when the virtual control object model travels off the track. Based on... and Calculate decision reaction time ; S34: Evaluate the user's cognitive-motor coordination training effect based on the decision-making reaction time, and determine the comprehensive training effect score, using the following formula: ; Where S is the overall score for training effectiveness; The formula for cognitive-motor coupling scoring is as follows: ； The normalized value of the motion error is given by the following formula: , where t c It is the detection start time, t end It is the end of training, θ t θ is the angular velocity of the upper limb horizontal turning motion at time t during training. target D represents the target angular velocity of the upper limb's horizontal turning motion; D represents the cornering difficulty level of the virtual control object model; D cognize Error rate for cognitive tasks.

8. The scene interaction system according to claim 7, characterized in that, Cognitive task error rate D cognize The formula for determining it is as follows: ； in, For cognitive task error rate, This represents the number of times a cognitive task test result is correct. This represents the total number of tests conducted on the cognitive task.

9. The scene interaction system according to claim 7, characterized in that, The controller is also configured to dynamically adjust the cornering difficulty level D of the virtual manipulator model based on the virtual speed v, cornering radius r, and cognitive task complexity C of the virtual manipulator model, as shown in the following formula: ； D represents the difficulty level of cornering in the virtual control object model.

10. The scene interaction system according to claim 1, characterized in that, Step S3 also includes calculating the fatigue test value, using the following formula: ； Among them, F tired E represents the fatigue test value. current E represents the current acceleration intensity value. start F represents the initial maximum acceleration value. current F represents the current active force value. start This represents the maximum active force value in the initial stage; F tired If the value exceeds the fatigue threshold, the user is considered fatigued.

Images