Virtual tactile simulation method of tactile glove

Abstract

The invention relates to a virtual tactile simulation method of a tactile glove, which comprises the following steps of: establishing a perception response pedigree model by analyzing the surface attribute of a virtual object; projecting touch interaction of the user in the virtual scene into a multi-level force feedback target value; sensing spectrum parameterization is introduced, and a feedback sensitive interval is set according to the minimum distinguishable difference JND principle of sensing psychology; dividing the finger tip region into a plurality of sub-sensing regions by adopting a region specific response function modeling mechanism; a flexible film strain sensor is used for detecting the finger pulp contact position and distribution in real time; judging whether the contact is concentrated, diffused or slipped or not, and dynamically adjusting a feedback strategy of a corresponding area; introducing a feed-forward predictive modulation mechanism: predicting a contact state in the future 5-10 ms by using motion capture; generating a response signal in advance, and sending the response signal to a driving layer to offset control-sensing lag; a bistable cooperative driving structure is adopted; a channel distribution mechanism of contact importance is adopted; and adjusting resource occupation according to the interaction task semantics.

Description

Technical Field

[0001] The invention relates to a virtual tactile simulation method for a tactile glove. Background Art

[0002] Simulation methods currently used in tactile gloves have made significant progress in recent years, especially in the integration of physical-level methods such as force feedback, electrical stimulation, vibration feedback, and pneumatic feedback. However, in practical interactive applications, these traditional tactile simulation methods still have significant shortcomings and structural drawbacks, which restrict their further development in high-precision, complex interactive and immersive environments and their degree of matching with users' real perception. First, existing technologies have inherent limitations in the regional resolution of tactile feedback. Many tactile gloves still use low-density linear or circular arrangements, which are unable to independently model feedback and drive output based on the physiological differences of different areas of the fingertips. This leads to a homogenized tactile experience. That is, whether the user touches the edge of an object with the center of the fingertip or swipes the outside of the fingertip across a rough surface, the glove system often outputs feedback with a uniform force or frequency, lacking regional differentiation and positioning accuracy. This not only makes it difficult for users to distinguish the structural details of the touched object, but also makes the virtual object surface lack realistic texture.

[0003] Secondly, existing tactile simulation methods lack the ability to dynamically identify interactive behaviors. Most gloves rely on a static threshold trigger mechanism, that is, when the pressure of a sensor exceeds the set value, the feedback module is activated. However, this type of method cannot accurately determine whether the user is performing complex actions such as tapping, clicking, sliding, or pressing. Especially when there is high-frequency and rapid switching of tasks (such as virtual drawing, gesture handwriting, and 3D modeling), the system cannot extract the semantics of contact behavior through perception signals, resulting in delayed and inaccurate feedback strategy execution or abrupt mode switching, which seriously affects the user experience; thirdly, traditional tactile simulation methods are ineffective in driving mode input. Most commercial products only support one form of physical feedback, such as providing single-frequency stimulation through a linear vibrator or creating pressure through pneumatic expansion. This ignores the fact that human touch is a multi-frequency, multi-level perception system. The feedback of the human sensory organs to different materials and actions requires the coordinated simulation of different frequencies and different energy forms to form a complete tactile experience. However, traditional systems cannot simultaneously superimpose force and texture signals at the same contact point, lack modal fusion strategies, and lack a mechanism to handle user perception transitions during modal switching. The final feedback results are either too strong and rigid or monotonously distorted.

[0004] Fourth, in terms of feedback control, most gloves currently use a fixed feedback parameter driving mechanism, that is, no matter how the size, direction, or task scenario of the user's movements changes, the system uses a preset frequency, amplitude, and activation time for tactile output. There is a lack of a feedback scheduling system based on real-time analysis of user behavior, so adaptive and efficient energy management cannot be achieved. Especially when multiple fingers interact at the same time or system resources are limited, feedback conflicts, delays, and even feedback failures are prone to occur; Fifth, the existing technology is also obviously insufficient in delay control and pre-response mechanisms. Touch is one of the most sensitive senses of humans. When the feedback delay exceeds 20 milliseconds, the user will obviously feel unnatural or even have an illusion, and most systems are due to It needs to go through multiple links such as sensor acquisition, signal judgment, and drive control, and there is generally a system response delay of more than 40 milliseconds. This delay cannot be corrected by simple physical compensation. Therefore, traditional methods have almost no processing logic for action prediction or feedforward modulation, and cannot make feedback preparations in advance for fast, continuous or impending interactions; finally, from the perspective of user experience, current tactile simulation gloves also have obvious shortcomings in individual adaptability and scene generalization capabilities. Most systems do not take into account the differences in finger size, skin sensitivity, and physiological structure between different users. They cannot establish an adaptive feedback model based on individual users, and lack scene perception capabilities, resulting in the inability to intelligently switch feedback logic during task changes. Summary of the Invention

[0005] The purpose of the present invention is to provide a virtual tactile simulation method for a tactile glove, thereby solving some of the drawbacks and deficiencies pointed out in the background art.

[0006] The present invention solves the above-mentioned technical problems by adopting the following technical solutions: a virtual tactile simulation method for a tactile glove, comprising: establishing a perceptual response spectrum model (PFSM) by analyzing the surface properties of a virtual object, including hardness, elasticity, roughness, and friction coefficient; projecting the user's tactile interaction in a virtual scene into multi-level force feedback target values in different frequency bands and contact forms;

[0007] Introducing perceptual spectrum parameterization: 0–20Hz represents macroscopic force perception, 30–200Hz represents surface texture, and 200–400Hz represents slip and vibration. Based on the Just Noticeable Difference (JND) principle in perceptual psychology, feedback sensitivity intervals are set. Physical driving frequency domain requirements are inferred from sensory results, establishing a signal coding structure centered on human perception.

[0008] A region-specific response function modeling mechanism is adopted to divide the fingertip area into multiple sub-sensing areas, each of which is assigned different feedback gain weights; a flexible thin film strain sensor is used to detect the contact position and distribution of the fingertips in real time; it is determined whether the contact is concentrated, diffuse or slipping, and the feedback strategy of the corresponding area is dynamically adjusted.

[0009] Furthermore, the corresponding area-based feedback strategy introduces a feedforward predictive modulation mechanism: using motion capture to predict the contact state in the next 5–10 ms; using a lightweight neural network model to estimate the future tactile feedback trend; and generating a response signal in advance and sending it to the drive layer to offset the control-perception lag.

[0010] Furthermore, the predictive modulation mechanism adopts a bistable collaborative drive structure: the main drive unit, including the airbag, produces stable macroscopic force perception; the auxiliary drive unit, including piezoelectric / electromagnetic, produces high-frequency details; when the contact environment changes drastically, the two units switch to the main control mode or activate together to ensure a stable feel.

[0011] Furthermore, the bistable collaborative driving structure adopts a channel allocation mechanism based on contact importance: each finger or fingertip segment is assigned a current interaction priority value; dynamic decisions are made to allocate limited driving resources including voltage, air pressure, and frequency band to key perception points; and resource occupancy is adjusted according to the semantics of the interaction task.

[0012] Furthermore, the region-specific response function modeling mechanism includes:

[0013] A feedback gain function is established for each sub-sensing area, forming a region-specific response function, which is used to adjust the intensity and frequency response of the feedback output. A flexible thin film strain sensor array is embedded in the glove fingertip to obtain the spatial distribution characteristics of the contact between the finger pad and the external environment in real time, including pressure intensity, distribution range, and dynamic changes.

[0014] In the above scheme, based on the physiological differences of the fingertip skin area, the contact surface inside the glove is divided into multiple sub-sensing areas R1, R2, ..., R n , such as the center of the fingertip, the middle of the fingertip, the outer edge, the area near the nail, etc.; each area R i Corresponds to a set of tactile feedback characteristics, including:

[0015] Force sensitivity (e.g., magnitude of response to pressure changes);

[0016] Texture frequency sensitivity (ability to sense high-frequency vibrations);

[0017] Perceived delay tolerance (subjective tolerance for response time);

[0018] For each sub-region R i Assign a regional gain parameter group G i ={α i ,β i ,γ i}, which is used to adjust the intensity and mode bias of the feedback output of the region; combined with the gain groups of each sub-region, a regional response function Ψ(x,t) with spatial normalization and dynamic response adaptability is constructed, which is defined as follows:

[0019]

[0020] in:

[0021] Ψ(x,t): the comprehensive feedback intensity that the system should output at the spatial coordinate x at a certain time t; i: the sub-sensory area number, a total of n areas; α i : static feedback strength factor of region i (determines the maximum output amplitude); λ i : spatial attenuation parameter of the center of the region, reflecting the tactile concentration; μ i : The center coordinate of region i (calibrate the spatial position of the region); β i : feedback frequency factor, corresponding to vibration / texture mode frequency; φ i : Feedback signal initial phase, controlling multi-region synchronization or misalignment; γ i (t): a time-variable gain function, dynamically adjusted based on real-time sensor data, including the contact behavior classification results (e.g., slip → increase dynamic weight);

[0022] In order to realize the dynamic assignment of the above Ψ(x, t) function, a two-dimensional flexible strain sensor array is embedded in the finger cuff, and the array covers each R i The system senses the current pressure value in different areas of the array; the spatial distribution of pressure; the trend of contact changes over time (static pressing vs. dynamic sliding), and adjusts each γ in real time. i The value of (t) makes the overall feedback function Ψ(x,t) output style differentiation, energy optimization, and natural experience in different contact situations.

[0023] Furthermore, based on the data obtained by the flexible thin film strain sensor, the current contact behavior is identified as one of concentrated contact, diffuse contact or sliding contact; based on the identified contact behavior, the corresponding feedback strategy is dynamically selected and applied to perform specific tactile feedback control on the target area, including feedback intensity adjustment, feedback mode switching or drive sequence scheduling.

[0024] Furthermore, the feedback strategy output is executed by multiple tactile drive units, including low-frequency force drive and high-frequency texture drive, forming a multimodal response system based on regional feedback weight regulation;

[0025] In this solution, an embedded flexible thin-film sensor array collects real-time characteristics of each sub-region of the fingertip, such as contact pressure, area, and movement direction. Combining this contact information with a predefined weight matrix of feedback gain functions for each region, the system determines the response requirements of each sub-region in the current interaction (e.g., pressure-based or texture-based).

[0026] At this point, the controller calculates the tactile drive combination mode corresponding to the current sub-area, including the main drive mode, the secondary drive mode, and their relative energy weights. To coordinate the output signals of the two physical drivers so that their response process conforms to the tactile perception curve and maintains physical response continuity, a modal hybrid drive response function is constructed as follows:

[0027]

[0028] in:

[0029] F(t) is the total tactile feedback signal strength (mixed modal result) that the system should output at any time point t; x is the spatial position coordinate of the fingertip (traversing sub-region); T is the total length of the spatial region of the currently activated feedback; ω l (x, t) is the feedback weight function of the low-frequency force mode at position x and time t; ω h (x, t) is the feedback weight function of the high-frequency texture mode; σ l (x, t) is the physical feedback value (such as air pressure) that the low-frequency drive unit should apply under the current task; δ h (x, t) is the texture feedback value (such as voltage or amplitude) that should be applied by the high-frequency drive unit; η(x, t) is the modal difference smoothing control factor, which is used to coordinate the output smoothness of the two drives in the modal switching region; is the weight change rate between the two modes, indicating the main and auxiliary mode switching trend; ∫dx represents the aggregation and integration of sub-perception units in the entire spatial area;

[0030] Design logic and engineering implications:

[0031] The first term ω l ·σ l : Indicates the main contribution of low-frequency drive in the current area;

[0032] The second term ω h ·δ h : Indicates the contribution of high-frequency feedback in the current area;

[0033] Item 3 Control the smooth transition of the output when switching between main modes to prevent sudden changes or misperceptions;

[0034] Integral expression: The system will perform spatial summation over the entire contact area and output a unified, smooth, multi-source fusion feedback response curve.

[0035] Furthermore, the feedback gain function is set according to the tactile perception sensitivity of the human skin area, and the gain value of the central area of the fingertip is higher than that of the edge area of the fingertip; the flexible thin film strain sensor array covers each sub-sensing area to form a continuous two-dimensional pressure sensing network.

[0036] Furthermore, the classification of the contact behavior is determined based on the pressure distribution change trajectory, specifically:

[0037] Concentrated contact: pressure is concentrated in one or two sub-areas;

[0038] Diffusion contact: multiple areas are under pressure simultaneously;

[0039] Sliding contact: The pressure center moves in a specific direction over time.

[0040] Furthermore, the dynamic feedback strategy includes:

[0041] In the case of concentrated contact, high-frequency texture actuation in the target subregion is enhanced;

[0042] In the case of diffusion contact, feedback energy is evenly distributed to simulate surface contact;

[0043] In the case of sliding contact, the feedback units are activated sequentially according to the contact path to simulate the sliding friction feeling.

[0044] Beneficial effects of the present invention:

[0045] By dividing the fingertips into multiple sub-sensing areas and assigning independent feedback gain functions to each, customized drive output is provided based on regional sensitivity differences, achieving a realistic tactile experience from point pressure to sliding. The tactile feedback is significantly more refined than the single-area drive methods of traditional gloves. A flexible sensor array collects contact pressure, area, and displacement trajectory data in real time. Based on the pressure variation pattern, the system automatically determines the user's current interaction behavior (concentrated, diffuse, or sliding contact) and dynamically switches the corresponding feedback strategy, ensuring that the feedback behavior aligns with the interaction semantics, enhancing the realism and immersion of the interaction.

[0046] By constructing a modal hybrid drive response function, low-frequency force drive (such as micro airbags) and high-frequency texture drive (such as piezoelectric devices) are dynamically coordinated and controlled to achieve a natural transition from macro pressure to micro texture, avoiding abrupt tactile switching and effectively improving the fingers' ability to recognize details such as material, resistance, and edges. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1This is a flow chart of the virtual tactile simulation method of the tactile glove of the present invention.

[0048] Figure 2 This is a flow chart of the region-specific response function modeling mechanism of the present invention.

[0049] Figure 3 This is a flow chart of the dynamic feedback strategy of the present invention. DETAILED DESCRIPTION

[0050] The following is a detailed description of the specific embodiments of the present invention with reference to the accompanying drawings.

[0051] Combined with attachment Figure 1 The virtual tactile simulation method for the tactile glove of the present invention is an intelligent feedback mechanism based on the mapping relationship between human perception patterns and physical properties. By analyzing the surface properties of virtual objects, key tactile parameters such as hardness, elasticity, roughness, and friction coefficient are extracted. These physical properties determine the tactile perception characteristics of the human hand when interacting with the object. Based on these properties, the system establishes a perceptual response spectrum model called PFSM (Perceptual Force Spectrum Mapping). Based on the response patterns of the human tactile system to different stimulus frequency bands, this model maps abstract physical parameters to specific perceptual levels, converting subjective tactile experiences such as hardness, slipperiness, roughness, and elasticity into quantifiable force frequency band data. In this model, each tactile interaction generated by the user in a virtual scene is no longer viewed as a unified pressure or vibration, but rather as a set of composite tactile signals with temporal, spatial, and frequency characteristics. These signals are projected onto a perceptual spectrum, with different frequency bands representing different tactile dimensions. For example, low frequencies (0-20Hz) are used to express macroscopic positive force and material elastic response, mid-frequency bands can be used to reproduce friction and viscosity, and high frequencies (above 100Hz) are more suitable for expressing subtle textures and vibrations. Based on this, the system calculates the position and amplitude of each contact behavior within this spectrum model to generate a set of multimodal force feedback target values with hierarchical, distributed, and perceptual alignment. These target values are then used to drive different types of feedback units (such as micro-airbags, piezoelectric plates, or vibrators) in the haptic glove, thereby recreating the user's desired tactile sensation on the surface of real skin with high fidelity.

[0052] By introducing a perceptual spectrum parameterization approach, the system quantifies and structurally models the human skin's response to tactile stimuli of varying frequencies. This system expands traditional feedback strategies based on force or vibration intensity into a multi-band, controllable feedback system that is highly coupled to the human perceptual system. The system divides tactile stimuli into multiple functional regions based on frequency: The 0–20 Hz frequency band corresponds to the macroscopic force perception band, simulating low-frequency, large-displacement physical contact sensations such as pressure, resistance, and hardness. The 30–200 Hz mid-frequency band recreates the texture, graininess, and surface roughness experienced by a finger on various surfaces, activating rapidly adapting receptors in the skin, such as Meissner corpuscles. Finally, the 200–400 Hz high-frequency band focuses on simulating more subtle dynamic tactile experiences, such as slip, microvibration, and high-speed friction, activating deeper receptors, such as Pacinian corpuscles. On the basis of the perceptual spectrum division, the system further sets a feedback sensitivity range for each frequency band based on the principle of least discernible difference in perceptual psychology, ensuring that slight changes in tactile output can still be recognized by the user and produce obvious perceptual differences, thereby avoiding wasted driving resources or redundant feedback signals. At the same time, this method no longer designs tactile logic based on how to drive the device, but reversely deduces the frequency range and output intensity that should be presented at the physical level from the perceptual end point of what humans want to feel, that is, reversely deducing the physical implementation process from the perceptual target. This design logic emphasizes the modeling thinking centered on human perception. Finally, the system constructs a set of signal coding structures based on the above-mentioned reverse deduction path. During the coding process, each virtual interactive behavior is located through the perceptual spectrum, and then translated into a multi-band driving instruction combination to achieve accurate mapping of the driving signal from the perceptual space to the execution space.

[0053] By adopting a region-specific response function modeling mechanism, and taking into account the physiological basis that different parts of the human finger respond differently to tactile stimulation, the entire fingertip area is subdivided into multiple sub-sensory regions, such as the center of the fingertip, the middle of the fingertip, and the edge of the fingertip. Each region has different suitable feedback methods and perception characteristics due to differences in skin thickness, receptor density, and force sensitivity. To address this physiological structural heterogeneity, the system assigns independent feedback gain weights to each sub-region, forming a set of region-specific response functions. This function determines the output drive strength, response frequency, and feedback style of each region during tactile feedback, achieving adaptive human perception based on skin regional characteristics. The glove embeds a high-resolution flexible thin film strain sensor array within the fingertip. The sensors generate an electrical response to tiny deformations of the fingertip, thereby detecting the specific location of the current contact point on the fingertip, the intensity of the applied pressure, the distribution range, and its changing trend. Based on this continuous sensor data, the system further analyzes the contact behavior pattern, determining whether the contact is concentrated (such as clicking), diffuse (such as pressing), or sliding (such as sliding or dragging). It then dynamically adjusts the corresponding feedback strategy based on the sub-region to which the behavior belongs. For example, in the case of concentrated contact, the system will increase the feedback frequency in the central area of the fingertip to highlight texture details. In the case of diffuse contact, it will enhance the low-frequency feedback in multiple areas to simulate surface contact. In the sliding state, it will continuously activate multiple areas according to the sliding path to simulate dynamic friction.

[0054] A feedforward predictive modulation mechanism is introduced to address the control-perception lag problem in traditional haptic feedback systems. This is due to the typical delay of several milliseconds or even tens of milliseconds between a user action and the system's recognition, calculation, and feedback output. This delay can lead to asynchronous, distorted, or unrealistic haptic feedback in high-precision, highly immersive virtual interactions. Leveraging motion capture technology, the system collects data such as the user's current hand motion trajectory, acceleration, and angle change at the very beginning of the change. Combined with historical interaction behavior, a microsecond-response prediction algorithm predicts the contact state between the finger and the virtual object within the next 5 to 10 milliseconds, including contact presence, contact location, contact strength, and contact type (e.g., click, swipe, press, etc.). Based on this, the system integrates a lightweight neural network model, which has been extensively trained to rapidly estimate the haptic feedback trends experienced by the user during upcoming interactions. This model predictively determines key parameters such as force, frequency, feedback area, and feedback type that should be simulated in the future. Once the prediction result is formed, the system immediately generates the corresponding driving instructions in advance and sends them to the low-level driving module, so that the tactile driver (such as airbag, piezoelectric sheet, vibrator) is in a pre-activated state before the actual contact occurs, or at least shortens the response preparation time during the startup process, so that the actual feedback is almost synchronized or slightly ahead of the user's actual tactile perception needs.

[0055] A bistable collaborative drive structure is adopted to solve the difficult problems of limited response range, insufficient detail restoration or prominent feedback delay of a single type of driver in complex tactile scenarios. The structure consists of a main drive unit and a sub-drive unit, each responsible for tactile simulation in different dimensions. The main drive unit usually adopts low-frequency and high-power actuators such as airbags and linear motors, which are dedicated to providing stable macro-force feedback, such as pressure, support force, deformation resistance, etc., to restore the large-scale, low-frequency force signals required by users when holding, touching, and pushing virtual objects. Its advantages are large output force, wide displacement range, and high stability, but low response frequency; the sub-drive unit uses high-frequency and fast-response actuators such as piezoelectric sheets, electromagnetic vibrators or micro-vibration motors, which are dedicated to simulating micro-tactile features such as surface texture, granularity, sliding friction or vibration prompts. It is characterized by high feedback resolution, wide bandwidth, and fast response speed, and can provide delicate local perception supplements. This bistable system uses a logic controller to monitor changes in the user's tactile environment state in real time during operation, such as whether there are sliding interruptions, sudden material switching, rapid clicks, or complex path movements. When the system detects drastic changes in the contact environment, or the current tactile task presents high complexity characteristics in the spatial or temporal dimensions, the system will dynamically switch to the master control mode. That is, it determines which type of drive unit will take the lead in responding based on the dominant perception target of the current feedback task, while retaining another unit as an auxiliary execution or entering a synchronous collaboration mode. Under a higher-priority control strategy, the main and secondary coordinated activations are achieved, so that macro force perception and micro details are output simultaneously during the key interaction stage, thereby ensuring that the continuity, stability, and realism of the feel can still be maintained in contact mutations or high-dynamic scenarios, and no tactile interruption or inconsistent experience will occur due to drive frequency band misalignment or feedback lag.

[0056] By employing a contact importance-based channel allocation mechanism, the system optimizes management of precise tactile feedback in environments with multi-point interaction, high computational loads, or resource constraints. This system treats tactile resources as finite resources, rather than fixed allocations that are evenly distributed or statically scheduled. Therefore, during operation, the system sets a current interaction priority value for each finger or fingertip segment in real time. This priority value is dynamically calculated based on multiple factors, including whether the segment is in contact with a virtual object, whether the contact is stable or fluctuating, whether the user is performing delicate manipulation, whether the region is responsible for the dominant perception task, and the semantic weight of the task context. This priority not only reflects the importance of contact strength and location but also incorporates an understanding of interaction intent. Based on this dynamic priority system, the haptic engine intelligently schedules available drive resources within each control cycle, including key resource parameters such as output voltage intensity, airbag actuation pressure utilization, vibration unit bandwidth utilization, and signal update frequency. This ensures that the sensory points most in need of feedback receive optimal physical drive support, while those not currently involved in critical interactions enter a low-power, low-frequency, or even suspended drive state. This prevents resource fragmentation and reduces energy consumption and processing latency. Furthermore, the system will adjust its resource usage strategy according to the semantic level of the interactive task. For example, when the user is performing text input, fine object manipulation, or tasks with high attention requirements, the system will concentrate more feedback channels and energy distribution on the thumb, index finger, or the knuckle area being operated. In extensive interactions such as dragging objects and sliding to switch interfaces, the feedback range is expanded but the frequency density is reduced. This mechanism gives the tactile feedback system a dynamic regulation capability based on cognitive meaning, enabling it to understand what the interaction is doing rather than just where the contact is made, realizing a truly task-oriented feedback-driven logic, and greatly improving the system's adaptability to complex scenarios and the subjective simulation perception of user experience.

[0057] Example 1:

[0058] In this example, a user is using a virtual fabric tactile application within a VR system. The task is to touch a piece of virtual fabric. The user is wearing the tactile glove of the present invention. During this process, the user must control force and determine surface texture differences. During this process, the user primarily uses their right index finger. Specific contact actions include pressing the fabric with the center of the fingertip, smoothing the surface with the middle of the fingertip, and quickly sliding with the outer fingertip to confirm.

[0059] The system divides the index finger area into four sub-sensing areas R1 to R4, corresponding to:

[0060] R1: Center of fingertips (fabric touch action area)

[0061] R2: middle part of fingertip (smoothing action area)

[0062] R3: Outer edge (sliding action area)

[0063] R4: Near nail area (non-primary contact)

[0064] Each region is assigned the following gain parameter set G i ={α i ,β i ,γ i}, the units are calibrated according to the system as follows:

[0065] α i ∈[0.2,1.0]: Maximum feedback strength coefficient (low is auxiliary feedback, high is main feedback)

[0066] β i ∈[20,400]: feedback frequency (Hz)

[0067] λ i ∈[1,10]: spatial attenuation control (the larger the value, the more concentrated)

[0068] φ i ∈[0,π]: feedback phase adjustment

[0069] γ i (t)∈[0.0,2.0]: Time dynamic adjustment factor (corrected in real time according to sensor response)

[0070] Next, assume that the user is touching the rough surface with his fingertips. The system sensor detects that the maximum pressure occurs at the position x = μ1 = 3.5 cm (i.e., the center area of the fingertip). The sensor feedback shows that the pressure at this point is 0.65 N, which belongs to the concentrated contact mode. The current time is t = 1.2 seconds.

[0071] At this point the system calls the region response function:

[0072]

[0073] Substitute typical parameters (partial median values):

[0074] R1: α1=0.9, λ1=7, μ1=3.5, β1=320, φ1=0.2π, γ1(1.2)=1.8

[0075] R2:α2=0.5,λ2=4,μ2=5.0,β2=160,φ2=0.4π,γ2(1.2)=0.7

[0076] R3:α3=0.3,λ3=6,μ3=6.5,β3=260,φ3=0.6π,γ3(1.2)=0.4

[0077] R4:α4=0.1,λ4=2,μ4=8.0,β4=80,φ4=0.8π,γ4(1.2)=0.2

[0078] At this point, the main feedback is clearly concentrated in R1, the fingertip area, and the feedback amplitude is much higher than other areas. The calculation results will be directly translated into the following driving strategy:

[0079] Activate the piezoelectric patch in the fingertip area to output high-frequency texture vibration at a frequency of 320Hz and a modulation voltage of 350Vpp;

[0080] The micro-airbag charging unit is activated simultaneously, increasing the pressure according to the 0.9N standard to provide stable fabric tactile resistance feedback;

[0081] Reduce the vibration output of the R2 and R3 areas to the minimum threshold of the system to reduce invalid power consumption;

[0082] Marking γ1(t) as a high-weight contact allows for a higher feedback refresh rate and improves texture continuity.

[0083] Subsequently, when the user's finger moves horizontally to wipe the surface, the system detects that the contact center is offset to μ2=5.0 (i.e. R2), and the pressure distribution is wider, which is consistent with the diffuse contact characteristics. The system will increase γ2(t) to 1.5 and reduce γ1(t) to 0.6, transfer the main feedback control from the fingertip to the middle of the fingertip, and reconfigure the vibration frequency to 160Hz to simulate the dumb vibration feeling of smooth brushing on the surface.

[0084] Finally, when the user quickly confirms the boundary texture with the outside of their fingertip, the slip path and time difference are quickly identified as high-dynamic slip. The system increases γ3(t) to 1.6, triggers phase compensation, adjusts φ3 to synchronize the texture update to form a slip-sensing misalignment jitter, outputs a 260Hz texture simulation signal, and briefly reduces the pressure at the airbag port to release the sliding resistance, ensuring that the user's perception is consistent with the visual trajectory.

[0085] After the user has finished marking the fabric's tactile details, he or she begins to evenly stroke the fabric's tactile surface with the pads of their fingers. This process is dominated by the middle area R2 of the fingertips, and the flexible film strain sensor array records that the pressure distribution area activated in this area has increased from the previous 0.5cm 2 Expanded to 1.6cm 2, the number of covered sensor units increased from 8 to 19, while the average pressure changed from 0.65N to 0.48N, and maintained a uniform state within ±0.02N for more than 250ms. The system determines that this state is diffuse contact, that is, the contact range is wide, the pressure distribution is uniform and the duration is long. The system adjusts γ2(t) to 1.5 (from the previous 1.0) in real time, and the corresponding α2 is reduced to 0.6 to control the overall feedback output to be smooth and stable, and prevent short-term fluctuations from causing sudden changes in touch. At the same time, the controller reduces β2 from 160Hz to 120Hz to reduce the sense of vibration and simulate the soft resistance brushing feeling produced by the large-area contact of the palm under the sand surface material; in terms of driving sequence, the system first uses the micro-airbag to generate a slowly rising low-frequency support force, and then superimposes the low-amplitude texture vibration to achieve the simulation effect.

[0086] Next, the user quickly swipes the outside of their fingertip across the fabric's tactile edge to check accuracy. Within 0.12 seconds, the system detects the pressure center rapidly moving from μ = 6.5 cm to μ = 3.9 cm, with an average displacement velocity of 22 cm / s. This is accompanied by a jump in the local pressure pulse, and the contact path forms a clear oblique sliding trajectory. At this point, the system determines the contact type as sliding based on the pressure center displacement rate (dx / dt > 15 cm / s) and pressure pulse duration (< 100 ms) in the sensor output. γ3(t) is immediately increased to 1.7, β3 is set to 280 Hz, and the piezoelectric plate in the secondary drive unit is activated for rapid pulse feedback. The φ3 phase delay is also adjusted to 0.3π to create a slight sense of displacement. Regarding the actuation sequence, the system first deactivates the airbag support to release the positive force burden, then quickly activates local high-frequency vibration, allowing the user to quickly perceive the detailed effect of the contact flying over the granular edge, enhancing the sensitivity and feedback of edge judgment.

[0087] Finally, the user's fingertip returns to the fabric touch area and tries to click repeatedly three times in a narrow area to adjust the fabric touch depth. The sensor detects that the three pressure concentration events are all concentrated within 0.3cm, with each pressure peak exceeding 0.8N and lasting less than 50ms. The system determines that it is a concentrated contact with high-frequency repetition characteristics. At this time, the system strengthens the responsiveness of R1, sets γ1(t) = 2.0, α1 = 1.0, β1 = 360Hz, and uses a feedforward prediction strategy to pre-activate the feedback channel, so that each click produces a clear but non-delayed rebound feeling. 5ms before each click, the piezoelectric driver preloads the voltage to about 90% of the peak value to ensure that the user's finger gets quick feedback as soon as it touches the surface, avoiding the appearance of a delayed empty feeling. This operation lasts for 20ms, and is supported by a 0.2N low-pressure short-pulse airbag to create a real material rebound texture.

[0088] In summary, each identified contact behavior triggers the system to reconstruct and schedule the parameter group, which is specifically reflected in:

[0089] Concentrated contact: small area, high intensity, short-term pressure → fast response, high-frequency main mode + local micro-vibration;

[0090] Diffusion contact: large area, uniform, slowly changing pressure → reduced frequency, main aerodynamic support + additional low-frequency texture;

[0091] Sliding contact: fast displacement and large changes in contact points → cut off the support force, increase the frequency, and adjust the phase to create a sliding feeling.

[0092] These strategy combinations are dynamically adjusted through the parameters in the formula Ψ(x, t) and mapped into the amplitude, frequency, sequence, and response delay of the actual drive output. This allows each interactive behavior to be expressed not only at the visual and action levels, but also to obtain real feedback at the tactile level, thereby establishing a tactile control logic that dynamically adapts in real time based on the changes in the user's actual interactive behavior.

[0093] The user slides the index finger back and forth on the texture edge of the fabric with good touch to verify the material's finish and feel. In this high-demand fine touch recognition scenario, the system begins to call the multimodal drive output strategy to integrate force perception and high-frequency texture output, while enhancing the realism of the virtual material and ensuring that the feedback response is natural, smooth, and the transition is coherent. At this time, the user's fingertip contact path spans two sub-areas, namely R2 (the middle of the fingertip) and R3 (the outer edge of the fingertip). The flexible film strain sensor array records that the path length is approximately T = 2.8 cm, and a total of 18 sensor units are activated, of which the pressure is concentrated in R2 at 0.52 N and the area is about 1.3 cm 2 , the feedback is uniform, indicating that force sensation is dominant; while in the R3 area, although the pressure is only 0.31N, the contact surface produces rapid directional changes, indicating that high-frequency sliding sensation dominates the feedback demand.

[0094] The system dynamically calculates the feedback weight function for each region within this time period based on historical gain settings and real-time data processing:

[0095] In R2:ω l (x, t) = 0.75, ω h (x,t)=0.25 (mainly low frequency)

[0096] In R3:ω l (x, t) = 0.30, ω h (x, t) = 0.70 (mainly high frequency)

[0097] At the same time, judging from the sensor and prediction model that the slip behavior continues to occur, the controller increases the modal adjustment factor η(x,t) to the maximum value of 0.6 in the range to enhance the response flexibility of the modal transition process. Substitute the actual physical signal value into the feedback, where:

[0098] σl (x, t): The average pressure adjustment value in the R2 area is 1.2 bar (simulating continuous pressure), and drops to 0.6 bar in the R3 area (releasing support)

[0099] δ h (x, t): In R2, the piezoelectric frequency is adjusted to 140Hz and the voltage amplitude is 110Vpp, which is a weak texture; in the R3 area, it is increased to 300Hz and 220Vpp to create a sliding grainy feeling.

[0100] At this point, the system integrates and solves the formula:

[0101]

[0102] Select the typical interval value of each term in the integral and substitute it into the calculated approximate estimate. The result is as follows:

[0103] In R2: the main feedback is 0.75×1.2=0.90, the secondary feedback is 0.25×110=27.5, and the modal change smoothing term

[0104] In R3: the main feedback is 0.30×0.6=0.18, the secondary feedback is 0.70×220=154, and the modal change smoothing term

[0105] After calculating the spatial average, the composite tactile response intensity output by the system is:

[0106] R2: F2≈0.90+27.5-0.3=28.1

[0107] R3: F3≈0.18+154+0.24=154.42

[0108] Total feedback strength (Units are system normalized feedback units)

[0109] This feedback value translates directly into:

[0110] In R2: The main low-frequency airbag provides a uniform sense of support, supplemented by mild vibration to maintain the consistency of material touch;

[0111] In R3: The main piezoelectric vibration array is activated to simulate the sliding feeling of particles with high-frequency output, while reducing the airbag support to create a realistic feeling of friction between skin and surface;

[0112] In the boundary area between the two, the modal conversion rate increases rapidly, and the controller fine-tunes the signal output according to the third modal change rate to prevent feedback interruption, lag or illusion.

[0113] The system refreshes the F(t) response curve every 10ms to ensure smooth modal transitions. As a result, users can perceive not only differences in edge hardness during a quick finger slide, but also the smooth transition from support to release and from blurred sliding to enhanced graininess. Subjective user experience ratings exceed 95% for stability and realism.

[0114] Example 2:

[0115] Based on Example 1, after the user completes operations such as fabric touch, stroking, and edge inspection using the tactile glove, they enter the highlighting stage, performing final fingertip pressure and touch modification on a simulated metal surface. Based on anatomical and psychological tactile knowledge, the system divides the user's right index finger pad into five sub-sensory areas, R1 to R5, with the following positions and roles:

[0116] R1: Fingertip center (main sensing area)

[0117] R2: Inside of the fingertip (highly sensitive texture auxiliary area)

[0118] R3: middle part of the fingertip (support and sliding area)

[0119] R4: outer edge of fingertip (low-frequency dynamic control area)

[0120] R5: Near the back of the nail (non-primary feedback area)

[0121] According to the density and physiological sensitivity of tactile receptors in different areas, the system sets a set of feedback gain function parameters α for each area. i , used to adjust the priority of feedback intensity. The value range is:

[0122] α1∈[0.85,1.0],

[0123] α2∈[0.65,0.8],

[0124] α3∈[0.45,0.6],

[0125] α4∈[0.25,0.4],

[0126] α5∈[0.0,0.2],

[0127] During execution, the user slides their finger from the fingertip contact point x = 3.0 cm outward to the area x = 6.2 cm, forming a complete press-slide-release feedback path. The system refreshes the 2D flexible strain sensor array data at a frame rate of one every 5 ms and identifies the following physical contact features:

[0128] At x = 3.0-3.4 cm, corresponding to R1: peak pressure 0.72 N, area 0.35 cm 2 , contact lasts 200ms;

[0129] At x = 3.5-4.4 cm, corresponding to R2 and R3: pressure range 0.48–0.51 N, area slightly expanded to 0.9 cm 2 , sliding speed 16cm / s;

[0130] At x = 5.0-6.2 cm, corresponding to R4: the pressure continues to drop to 0.32 N, and the slip speed further increases to 23 cm / s, which is maintained for less than 50 ms;

[0131] R5 is not activated, and the system is set to default feedback disabled. At this time, the control system performs regional feedback gain multiplication calculation in the feedback scheduling logic. The feedback output of each sub-region is defined as:

[0132] f i (t) = α i ·P i (t)·S i (t)

[0133] in:

[0134] α i is the feedback gain coefficient of the region;

[0135] P i (t) is the pressure value of the region at time t (unit N, 0–1);

[0136] S i (t) is the contact area of the region (in cm 2 , 0–1.5).

[0137] Substitute some typical values to calculate the feedback amplitude:

[0138] f1(t) = 0.95 × 0.72 × 0.35 = 0.2394 (high-precision touch main response of the fingertip)

[0139] f3(t) = 0.55 × 0.50 × 0.90 = 0.2475 (balanced response of the middle section of the fingertip)

[0140] f4(t)=0.30×0.32×0.85=0.0816 (weak response in the slip zone)

[0141] Feedback output value f i(t) is directly used to control the airbag inflation or piezoelectric excitation intensity. After normalization to the system control voltage (maximum 350 Vpp) and air pressure (maximum 2.0 bar), the actual output command for each area is obtained. For example, the fingertip area f1 is amplified to the piezoelectric drive of the high-frequency texture signal: the output frequency is 320 Hz, the amplitude is adjusted to 0.2394 × 350 V ≈ 83.8 Vpp, and the positive force is maintained at 0.8 bar by the micro-airbag. The middle section of the fingertip f3 activates the low-frequency dynamic control module, with an inflation pressure of 1.3 bar to establish the basis for the slip sensation. The outer slip zone f4 is only activated by a slight vibration, the voltage is adjusted to 26 Vpp, and a 15ms delay is used to achieve a phase delay to simulate the sliding delay.

[0142] The two-dimensional flexible strain gauge array covers each sub-area unit with a 16×16 grid and uses a spatial interpolation algorithm to smoothly restore the pressure of the continuous contact path in space. This ensures that the feedback process from fine static pressure of the fingertip to the sliding of the fingertip is free of feedback interruptions or sudden changes, achieving true regional gradient feedback that aligns with the human body's natural perceptual transition habits. User subjective ratings particularly praised the continuity of the tactile sensation from fabric touch to caress. The average experience delay was kept within 14ms, far below the human hand's perceptible threshold for feedback sudden changes (~25ms), verifying that this solution strikes a good balance between engineering implementation and user experience.

[0143] The user performs a virtual signature operation, using his right index finger to write his personal signature on the virtual fabric. This operation process covers complex mixed actions such as short clicks (concentrated contact), light pressure writing (diffuse contact), and rapid strokes and continuous movements (sliding contact).

[0144] The system sets the following judgment rules to classify the three types of contact behaviors in real time:

[0145] Concentrated contact: Pressure is concentrated on one or two sub-sensing areas, with an activation area less than 0.5 cm 2 , pressure concentration index C p >0.8;

[0146] Diffusion contact: The number of active sub-regions is greater than 3 and the area is greater than 1.0 cm 2 , standard deviation of mean pressure σ P <0.15;

[0147] Sliding contact: The pressure center shifts continuously over time, with a displacement rate v = dx / dt > 12 cm / s, and the directionality is consistent. Taking the beginning of the first stroke as an example, the user quickly places their finger down, and between t = 0.0s and t = 0.06s, the system records activation of areas R2 and R3 in the fingertip, with a maximum pressure value of P. max =0.71N, the total activation area is only 0.42cm2 , the regional pressure concentration index is calculated as follows:

[0148]

[0149] Because C p If the touchscreen is >0.8 and the number of zones is less than 3, it is considered concentrated contact. The system immediately triggers a rapid pulse feedback strategy, activating a high-frequency vibration unit (320Hz), low-amplitude air pressure support (0.6bar), and setting the activation time to 40ms, allowing the user to experience the point pressure of the virtual pen touching the fabric for the first time.

[0150] Then, during the main signature stroke phase, the user's fingertips began to write steadily, and the system detected that the activation area expanded to R1 to R4, with a contact area of 1.2 cm 2 , mean pressure P avg =0.52N, standard deviation σ P = 0.09, satisfying the diffuse contact condition. At this point, the system lowers the high-frequency vibration amplitude (180Hz, 0.3 amplitude) and increases the airbag support to 1.2 bar, creating a wide, evenly pressured, and smooth writing experience. In the feedback gain function, the weights of regions α2 and α3 are increased to match the primary sensing area.

[0151] When the user writes the last stroke of the signature, they quickly draw the arc at the end. The system continuously collects the pressure center x(t) with a resolution of 5ms. Its position trajectory is:

[0152] x(t) = [3.2, 3.8, 4.5, 5.3, 6.2] cm, each step is 5 ms

[0153] The average displacement rate is calculated as:

[0154]

[0155] At the same time, the directional continuity is high (increasing in the same direction), and the system immediately identifies this section as a sliding contact, triggering the following response strategy:

[0156] Stop low-frequency airbag support to avoid the positive resistance of fingers on the surface interfering with the sliding feeling;

[0157] Enhanced high-frequency vibration (frequency increased to 360Hz, amplitude 0.5) to simulate the feeling of the last stroke leaving the pen;

[0158] Adjust the feedback phase φ = 0.3π to create a slight hysteresis to simulate the handwriting pause.

[0159] The judgment result is dynamically input into the multimodal drive response function F(t) for fusion control signal output. The system switches between three feedback strategies throughout the signing action based on spatial integration and time weighting, keeping the feedback seamless and free of illusion, and controlling the response delay within 13ms, which is far lower than the JND time threshold (approximately 20-25ms) that the human hand can perceive for tactile lag, thus achieving a real and natural perceptual transition.

[0160] The user entered the final interactive test task: verifying the material and functionality of a virtual nameplate made of simulated glass. This task, which combines short clicks (concentrated contact), large-area presses (diffuse contact), and smooth sweeps (sliding contact), can be used to verify the responsiveness of the dynamic feedback strategy mechanism, including the ability and engineering feasibility of enhancing high-frequency texture feedback during concentrated contact, uniform energy distribution during diffuse contact, and sequential activation of paths during sliding contact.

[0161] First, the user uses the index finger to perform a click test on the corner of the nameplate. The system's flexible thin film strain sensor array collects pressure data in the fingertip R1 (center of the fingertip) and R2 (inside of the fingertip) within 5ms. The maximum pressure is P max =0.78N, area S=0.42cm 2 , and meet the requirements of short contact time (<60ms), concentrated pressure center (C p >0.85). The system immediately identifies it as concentrated contact behavior and calls the strategy module of the present invention to drive the gain factor β of the high-frequency texture in this area. i Adjusted to 350–380Hz, the voltage amplitude V pp The voltage is increased to 120–140V (the specific value is adjusted based on the user sensitivity matching function). In this example, the final values assigned to the fingertip area R1 are β1 = 360Hz, V1 = 135V, and α1 = 0.95. The output texture simulation signal peak is significantly enhanced, while the tactile response time is kept below 12ms, ensuring that the glass point bounce is delivered promptly when clicked. The user subjectively describes the experience as crisp edges and a point-like feeling on tempered glass.

[0162] Next, the user places the entire fingertip flat on the middle of the nameplate to test the recognition function. The sensor array now covers R2 to R4 (the middle of the fingertip to the outer edge), with a total area of 1.5cm 2 , standard deviation of pressure distribution σ P <0.12N, it is judged as diffuse contact, and the system switches to the surface contact strategy. Under this strategy, the controller limits the total feedback energy to the upper power limit P = 100mW (to ensure that the heat load is controllable), and then evenly distributes it within the activation area. Specifically, the output value of each area is:

[0163]

[0164] Set the area parameters as: α2 = 0.75, α3 = 0.6, α4 = 0.4, and the areas are 0.6, 0.5, and 0.4 cm respectively 2 , substituting into:

[0165]

[0166] E3=35.21mW,E4=25.26mW

[0167] The system allocates the airbag driving pressure accordingly: R2 = 1.0 bar, R3 = 0.85 bar, R4 = 0.65 bar, and uniformly sets the high-frequency texture frequency to 180 Hz and low amplitude (<80 Vpp) to achieve flexible and continuous pressure surface feedback, simulating the touch feeling of the user's finger touching the pressure-sensitive surface of a real material. At this stage, the user describes the feeling of real glass temperature and damping when the hand is placed on it.

[0168] Finally, the user quickly swept the nameplate border with the fingertips to test the consistency of the material edge. The system recorded that the pressure center moved continuously from x = 2.1cm to x = 6.8cm within 30ms, with an average speed of v = 157cm / s. The movement path in the straight line was clear and identified as a sliding contact. At this time, the system calls the sliding strategy module, sorts the 16 sensor pixel numbers in the activation path in sequence, builds a sequence table according to the sliding direction, and activates the high-frequency texture output along the path in sequence with a timing rhythm. The feedback vibration frequency is set to 300Hz, and the voltage is pushed with a delay of 5ms from zone to zone, so that the user can obtain a continuous granular path tactile simulation. In addition, the system reduces the positive support force of the airbag at the tail end of the sliding path (from 0.9bar to 0.5bar), and delays the piezoelectric plate closing signal by 10ms, so that the sliding feeling has tail sliding inertia, enhancing the residual feeling after the real sliding.

[0169] Through these three typical actions, the system dynamically responds to contact behavior and uses the strategy module of the present invention to distinguish and adjust:

[0170] Focused contact: High-frequency gain peak is increased to the 350–400Hz range;

[0171] Diffusion contact: driving energy is evenly distributed according to weight, and the gas pressure transitions continuously;

[0172] Sliding contact: Activate feedback nodes sequentially along the path sequence to match the physical trajectory of the fingertip with the vibration rhythm.

[0173] In experimental tests, the dynamic strategy switching control cycle averaged 7ms, with no obvious interruption or delay, and the vibration signal waveform had no frequency jumps in the boundary area, verifying the effectiveness, computational feasibility and real-time perception stability of the dynamic feedback strategy in the scheme.

Claims

1. A virtual tactile simulation method for tactile gloves, characterized in that The following steps are involved: By analyzing the surface properties of virtual objects, including hardness, elasticity, roughness, and friction coefficient, a Perceptual Response Spectrum Model (PFSM) is established. The user's tactile interaction in the virtual scene is projected into multi-level force feedback target values under different frequency bands and contact forms. Introducing perceptual spectrum parameterization: 0–20Hz represents macroscopic force perception, 30–200Hz represents surface texture, and 200–400Hz represents slip and vibration. Based on the Just Noticeable Difference (JND) principle in perceptual psychology, feedback sensitivity ranges are set. Inferring the physical driving frequency domain requirements from the sensory results and establishing a signal coding structure centered on human perception; Adopting the region-specific response function modeling mechanism, the fingertip area is divided into multiple sub-sensory areas, each of which is assigned different feedback gain weights. Use flexible thin film strain sensors to detect the contact position and distribution of the fingertips in real time; determine whether the contact is concentrated, diffuse or slipping, and dynamically adjust the feedback strategy of the corresponding area.

2. The virtual tactile simulation method of the tactile glove according to claim 1, characterized in that The feedback strategy based on the corresponding area introduces a feedforward predictive modulation mechanism: using motion capture to predict the contact state 5–10 ms in the future; using a lightweight neural network model to estimate the future tactile feedback trend; Generate response signals in advance and send them to the driver layer to offset control-perception lag.

3. The virtual tactile simulation method of the tactile glove according to claim 2, characterized in that The predictive modulation mechanism adopts a bistable collaborative drive structure: the main drive unit, including the airbag, produces stable macroscopic force perception; the secondary drive unit, including piezoelectric / electromagnetic, produces high-frequency details; when the contact environment changes drastically, the two units switch to the main control mode or activate together to ensure a stable feel.

4. The virtual tactile simulation method of the tactile glove according to claim 3, characterized in that The bistable collaborative drive structure adopts a channel allocation mechanism based on contact importance: each finger or fingertip segment is assigned a current interaction priority value; limited drive resources including voltage, air pressure, and frequency band are dynamically allocated to key perception points; and resource occupancy is adjusted according to the semantics of the interaction task.

5. The virtual tactile simulation method of the tactile glove according to claim 1, characterized in that The region-specific response function modeling mechanism includes: A feedback gain function is established for each sub-sensing area to form a region-specific response function, which is used to adjust the intensity and frequency response of the feedback output; a flexible thin film strain sensor array is embedded in the glove fingertip to obtain the spatial distribution characteristics of the contact between the finger pad and the outside world in real time, including pressure intensity, distribution range and dynamic changes.

6. The virtual tactile simulation method of the tactile glove according to claim 5, characterized in that According to the data obtained by the flexible thin film strain sensor, the current contact behavior is identified as concentrated contact, diffuse contact or sliding contact; based on the identified contact behavior, the corresponding feedback strategy is dynamically selected and applied to perform specific tactile feedback control on the target area, including feedback intensity adjustment, feedback mode switching or drive sequence scheduling.

7. The virtual tactile simulation method of the tactile glove according to claim 6, characterized in that The feedback strategy output is executed by multiple tactile drive units, including low-frequency force drive and high-frequency texture drive, forming a multimodal response system based on regional feedback weight regulation.

8. The virtual tactile simulation method of the tactile glove according to claim 7, characterized in that The feedback gain function is set according to the tactile perception sensitivity of the human skin area, and the gain value of the central area of the fingertip is higher than that of the edge area of the fingertip; the flexible thin film strain sensor array covers each sub-sensing area to form a continuous two-dimensional pressure sensing network.

9. The virtual tactile simulation method of the tactile glove according to claim 8, characterized in that The classification of the contact behavior is determined based on the pressure distribution change trajectory, specifically: Concentrated contact: pressure is concentrated in one or two sub-areas; Diffusion contact: multiple areas are under pressure simultaneously; Sliding contact: The pressure center moves in a specific direction over time.

10. The virtual tactile simulation method of the tactile glove according to claim 9, characterized in that The dynamic feedback strategy includes: In the case of concentrated contact, high-frequency texture actuation in the target subregion is enhanced; In the case of diffusion contact, feedback energy is evenly distributed to simulate surface contact; In the case of sliding contact, the feedback units are activated sequentially according to the contact path to simulate the sliding friction feeling.

Images