Multi-parameter dynamic coding method for tactile feedback

Abstract

The invention discloses a multi-parameter dynamic coding method for tactile feedback. The method comprises the following steps: arranging a multi-channel electrode array in a tactile sensing area, and establishing a corresponding relation between each channel of a multi-channel array electrode and a spatial position in the tactile sensing area through spatial coding; calibrating the minimum value and the maximum value of various types of electrode stimulation parameters reflecting the touch perception degree to obtain a dual-threshold calibration result; performing intensity coding according to the double-threshold calibration result; constructing a type code, wherein the type code reflects a corresponding relation between the frequency interval and the touch induction type; and generating a parameter dynamic mapping model based on the intensity code and the type code, wherein the parameter dynamic mapping model reflects a corresponding relation between an interaction characteristic between the touch sensing area and a target and an electrode stimulation parameter combination sensed by the touch sensing area. The method can be widely applied to man-machine interaction scenes needing tactile feedback.

Description

Technical Field

[0001] The present invention relates to the field of rehabilitation engineering technology, and more particularly to a multi-parameter dynamic encoding method for tactile feedback. Background Art

[0002] Sensory feedback plays a crucial role in achieving closed-loop human-robot interaction in prosthetic limbs and biomimetic robots. Some amputees have difficulty transmitting and processing sensory feedback. The lack of tactile feedback severely impacts their ability to interact naturally with their surroundings, as tactile information from the hand is crucial for dexterous manipulation and identifying the physical properties of objects. For example, for forearm amputees, restoring lost sensory feedback, particularly tactile feedback, could significantly enhance the manipulation capabilities of prosthetic limbs.

[0003] Existing tactile feedback methods are mainly divided into two categories: invasive and non-invasive. Invasive methods directly stimulate nerves by implanting electrodes or neural interfaces. Although they can provide relatively accurate tactile feedback, they have defects such as surgical risks, biocompatibility issues, and low patient acceptance. Non-invasive methods mainly transmit stimulation signals to the skin surface to activate relevant nerve endings. For example, TENS (transcutaneous electrical nerve stimulation) as the current mainstream non-invasive method provides an effective solution for restoring tactile feedback. This method is highly safe, easy to operate, does not require surgery, and has relatively low equipment costs. Compared with invasive methods, TENS methods have higher user acceptance. However, existing TENS tactile feedback technology still faces bottlenecks such as parameter adjustment relying on subjective experience and insufficient dynamic adaptability. To optimize the tactile feedback effect, researchers have proposed a variety of encoding methods, but none of them systematically solves problems such as multi-parameter coordination and dynamic adaptation. Its practical application is still limited by the shortcomings of the encoding methods.

[0004] Analysis reveals that existing encoding methods for tactile feedback primarily include subjective description methods, static encoding methods based on a single parameter, and bio-inspired encoding methods. Subjective description methods adjust stimulation parameters based on the user's subjective perception and rely on questionnaires or verbal feedback to guide parameter optimization. For example, some researchers have users describe tactile intensity using a rating scale and gradually adjust the pulse amplitude to a target perception threshold. This approach has the disadvantage of being highly dependent on subjective perception and lacking objective evidence and consistency. Static encoding methods based on a single parameter adjust only a single stimulation parameter and its mapping to tactile perception, while keeping other parameters constant. For example, some researchers have varied only the stimulation amplitude to induce varying sensations in amputees. This approach has the disadvantage of being limited in dimension and unable to accommodate multidimensional tactile information, and its static parameters are unable to adapt to dynamic scenarios. Bio-inspired encoding methods mimic the neural encoding mechanisms of natural tactile sensation in vivo to adjust stimulation parameters. For example, some researchers have simulated the high-frequency response of rapidly adapting mechanoreceptors and designed electrical stimulation pulse waveforms that resemble their discharge patterns. However, these methods suffer from limited real-time performance and complex parameter adjustment.

[0005] In summary, existing tactile feedback technologies still have limitations in their encoding methods, severely restricting their practical application in fields such as prosthetic control. For example, static encoding methods rely solely on the mapping of a single parameter (such as pulse amplitude) to tactile perception (e.g., simulating intensity changes solely through amplitude), making it difficult to coordinate the control of multiple parameter combinations such as amplitude, pulse width, and frequency. This results in a limited dimensionality in the expression of tactile information (e.g., the inability to simultaneously express tactile type and intensity information), making it difficult to reproduce the complexity of natural touch. Although bioinspired encoding methods attempt to simulate neural mechanisms, their high computational complexity (e.g., real-time simulation of dynamic responses) makes it difficult to achieve low-latency parameter adjustment. Their dynamic adaptation capabilities are insufficient, making them unable to adapt to real-time changes in interactive scenarios (e.g., adapting feedback intensity when grip strength suddenly changes). Subjective description methods rely on manual experience or user subjective feedback to adjust parameters, lack the support of quantitative models based on mathematical theory, and have poor individual compatibility. Due to significant individual differences, the adjustable parameter range for different users requires a significant amount of time to manually adapt. Summary of the Invention

[0006] The purpose of the present invention is to overcome the above-mentioned defects of the prior art and provide a multi-parameter dynamic encoding method for tactile feedback. The method comprises the following steps:

[0007] Disposing a multi-channel electrode array in the tactile sensing area, and establishing a correspondence between each channel of the multi-channel array electrode and a spatial position in the tactile sensing area through spatial coding;

[0008] For individual users, the minimum and maximum values of multiple types of electrode stimulation parameters reflecting the degree of tactile perception are calibrated to obtain a dual-threshold calibration result, where the minimum value is the evoked threshold representing the user's perception of tactile sensation, and the maximum value is the pain threshold representing the user's perception of pain. The dual-threshold calibration result includes an amplitude safety range, a pulse width safety range, and a frequency safety range defined by the minimum and maximum values;

[0009] Performing intensity coding according to the dual-threshold calibration result, wherein the intensity coding reflects the relationship between the tactile evoked intensity and the average power of the stimulation pulse, wherein the average power of the stimulation pulse is defined by the amplitude, pulse width and frequency;

[0010] constructing a type code for the individual user, wherein the type code reflects a corresponding relationship between a frequency interval and a tactile induced type;

[0011] Based on multiple sensor data and the intensity coding and the type coding, a parameter dynamic mapping model is generated, wherein the parameter dynamic mapping model reflects the correspondence between the interaction characteristics between the tactile sensing area and the target and the electrode stimulation parameter combination.

[0012] Compared with existing technologies, the present invention offers the advantage of a multi-parameter dynamic encoding method for tactile feedback. By constructing a quantitative mapping model between parameter space and tactile perception, combined with automated calibration of parameter threshold intervals and a dynamic adaptive adjustment mechanism, this method systematically addresses issues such as a single parameter adjustment dimension, insufficient individual adaptability, and poor real-time scenario adaptability. This method can be applied to a variety of scenarios requiring high-precision tactile feedback, such as intelligent prosthetic control, virtual reality interaction, and remote surgical robots, providing multi-dimensional perceptual support for human-machine collaborative operations.

[0013] Further features and advantages of the present invention will become apparent from the following detailed description of exemplary embodiments of the present invention with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and, together with the description, serve to explain the principles of the invention.

[0015] Figure 1 is a flow chart of a multi-parameter dynamic encoding method for tactile feedback according to one embodiment of the present invention;

[0016] Figure 2 is a schematic diagram of the overall process of a multi-parameter dynamic encoding method for tactile feedback according to an embodiment of the present invention;

[0017] Figure 3 is an architectural diagram of a multimodal TENS tactile feedback platform according to one embodiment of the present invention;

[0018] Figure 4 is a flowchart of a multi-parameter dynamic encoding method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0019] Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that unless otherwise specifically stated, the relative arrangement of components and steps, numerical expressions and numerical values set forth in these embodiments do not limit the scope of the present invention.

[0020] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the invention, its application, or uses.

[0021] Technologies, methods, and equipment known to ordinary technicians in the relevant art may not be discussed in detail, but where appropriate, the technologies, methods, and equipment should be considered part of the specification.

[0022] In all examples shown and discussed herein, any specific values should be interpreted as merely exemplary and not limiting. Therefore, other examples of the exemplary embodiments may have different values.

[0023] It should be noted that like reference numerals and letters refer to like items in the following figures, and therefore, once an item is defined in one figure, it need not be further discussed in subsequent figures.

[0024] The multi-parameter dynamic coding method for tactile feedback provided by the present invention systematically solves the coding bottleneck of tactile feedback technology, and provides high-precision, personalized tactile feedback functions for scenarios such as intelligent prostheses and bionic robots. In the field of bionic robots, multi-parameter dynamic coding can enhance the anthropomorphic tactile feedback of the manipulator, enabling it to more accurately perceive the material of objects and the force of operation in medical care or industrial grasping. The present invention is conducive to promoting the transformation of tactile feedback technology from the laboratory to the clinical and consumer markets. This technological breakthrough marks the leap of tactile feedback from experience-driven to mechanism-driven, and provides a reusable technical paradigm for closed-loop control of human-computer interaction.

[0025] In general, the multi-parameter dynamic coding method for tactile feedback provided by the present invention includes processes such as spatial coding, threshold calibration, intensity coding, type coding and dynamic coding. For example, in the field of prosthetic control, a spatial coding mapping relationship is first established based on a 5-channel electrode array, and the tactile feedback from the thumb to the little finger is mapped one-to-one to the independent electrode channel, providing a spatial basis for multi-dimensional tactile expression; then, the individualized tactile evoked threshold and pain threshold boundary are calibrated by a progressive parameter adjustment method (such as amplitude step 0.1mA, pulse width step 20μs, frequency step 1Hz), and the safe adjustment boundary of amplitude, pulse width and frequency is determined; subsequently, a three-parameter collaborative control mechanism is adopted to preferentially adjust the amplitude or pulse width to achieve tactile intensity quantization coding, while the frequency is independently used for tactile type coding. Finally, combined with the real-time data of the prosthetic sensor (such as pressure, acceleration, etc.), the parameter dynamic mapping model constructed based on the present invention is used to dynamically adjust the stimulation parameter combination, and continuously learn the user's historical preference data through an adaptive learning algorithm to improve the matching accuracy of the scene requirements and tactile feedback stimulation parameters. This invention can be applied to a variety of human-machine interaction scenarios. For example, in the field of prosthetic control, it has been proven to achieve slip vibration warning, real-time grip pressure matching, and sliding rate feedback. In the field of industrial robots, it can achieve material recognition and abnormal working condition warning.

[0026] Specifically, see Figure 1 and Figure 2 As shown, the provided multi-parameter dynamic encoding method for tactile feedback includes the following steps:

[0027] Step S1: building a multimodal tactile feedback platform.

[0028] To facilitate understanding and verification, we first build a multimodal tactile feedback platform, which can be developed by ourselves or based on an existing platform. Figure 3 As shown in the figure, the TENS tactile feedback platform is used as an example of a multimodal closed-loop control system. It integrates sensing detection, main control calculation, electrical stimulation output, and synchronous data acquisition. The platform architecture mainly includes core modules such as pressure detection sensor, inertial measurement unit (IMU), vibration detection unit, photoelectric sensor, voice input unit, electrophysiological signal acquisition unit, main controller, data acquisition and transmission module, waveform generator, PC, and electrode array.

[0029] For example, the pressure detection sensor uses a flexible piezoresistive array sensor (range 0-50N, resolution 0.1N) embedded in the prosthetic grip contact surface. 2 The C bus transmits pressure data to the main controller in real time with a sampling rate of ≥200Hz, which can detect dynamic changes in grip force and pass the data to the main controller.

[0030] The inertial measurement unit (IMU) integrates a three-axis accelerometer (±16g) and a gyroscope (±2000° / s) and is deployed at the prosthetic fingertip. The sliding window variance algorithm is used to calculate the object's sliding acceleration in real time (with an accuracy of 0.1m / s 2 ) and passes the data to the main controller.

[0031] The vibration detection unit detects vibration signals, such as ultrasonic signals, and environmental vibration interference through a vibration sensor, and transmits the detected sensor signal data to the main controller to enhance tactile feedback or reduce external interference.

[0032] The photoelectric sensor is used to detect the light signal required for tactile feedback, convert the light signal into an electrical signal, and transmit the detected sensor signal data to the main controller, such as detecting light signals of different light intensities and light signals of different wavelengths to enhance tactile feedback.

[0033] The voice input unit uses a high-sensitivity microphone array to support voice command recognition. Through a voice recognition algorithm (such as a deep learning-based speech-to-text model), it converts user commands into control signals and transmits the commands to the main controller.

[0034] The electrophysiological signal acquisition unit is used to collect electrophysiological signals such as myoelectric and electroencephalographic (EEG) signals to enhance tactile feedback. For example, the wireless EMG acquisition unit uses surface EMG sensors deployed on the surface of the residual limb muscles to collect EMG signals in real time (bandwidth 20–500 Hz, sampling rate 1 kHz) to monitor muscle activity and transmit the data to the main controller. The wireless EEG acquisition unit uses a 64-lead wireless EEG signal acquisition system (sampling rate 1000 Hz, bandwidth 0.1–100 Hz) to synchronously record EEG signals and transmit the data to the main controller.

[0035] The main controller, an STM32F407 microcontroller integrated into the prosthesis, dynamically adjusts parameter combinations based on the interaction scenario and encoding method (e.g., changes in grip force cause parameter changes) and controls the electrical stimulation output. Multi-threaded processing is achieved through DMA channels, simultaneously parsing sensor data, receiving host computer commands, calculating encoding methods, and adjusting stimulation parameters.

[0036] The waveform generator is used to generate electrical stimulation pulses with adjustable parameters, including amplitude (0.1-10mA), pulse width (20-500μs) and frequency (1-500Hz).

[0037] The data acquisition and transmission module is used to synchronously record the user's stimulation parameter information, sensor raw data and task completion indicators, and wirelessly transmit the data to the PC for subsequent data processing and analysis.

[0038] In one embodiment, the electrode array uses a 5-channel electrode output for electrical stimulation, and the stimulation sites can be switched programmably. For example, the 5 channels can be independently controlled using an analog switch matrix (ADG5412).

[0039] The PC stores stimulation parameter information, raw sensor data, and task completion metrics, providing data support for further evaluation and optimization of encoding methods. Furthermore, the PC can also handle the recognition and decoding of speech, EMG, and EEG signals. The relevant speech recognition algorithms and EMG / EEG decoding algorithms run on the PC. The decoded instructions are sent to the main controller via the wireless communication module for real-time adjustment of stimulation parameters.

[0040] Step S2: constructing a ternary mapping relationship between intensity, type and scene through multi-parameter dynamic coding, wherein the multi-parameter dynamic coding includes spatial coding, threshold calibration, intensity coding, type coding and dynamic coding.

[0041] In one embodiment, see Figure 4 As shown, the multi-parameter dynamic coding method includes spatial coding, threshold calibration, intensity coding, type coding and dynamic coding.

[0042] For example, in the field of prosthetic control, for spatial coding, a five-channel electrode array is used to establish a topological mapping relationship between finger tactile sensations. The tactile areas corresponding to the thumb, pinky, and thumb are each bound to an independent electrode channel. The main controller switches the tactile sensations of each finger. For example, channel 1 corresponds to the thumb tactile sensation, channel 2 corresponds to the index finger tactile sensation, and so on.

[0043] Threshold calibration can be achieved through a progressive parameter adjustment method to achieve personalized dual-threshold calibration. Calibrated parameters include amplitude (0.1-10mA), pulse width (20-500μs), and frequency (1-500Hz). Step sizes are 0.1mA, 20μs, and 1Hz, respectively. This threshold calibration can determine the tactile evoked threshold and pain threshold, i.e., the minimum and maximum thresholds, and further determine safe parameter adjustment ranges, such as the amplitude, pulse width, and frequency safe adjustment ranges.

[0044] Intensity coding reflects the relationship between the tactile evoked intensity and the average power of the stimulation pulse. The quantitative expression of the tactile evoked intensity can be achieved through a three-parameter collaborative regulation model, and its intensity relationship satisfies I∝PA 2 *PW*f, where PA is amplitude, PW is pulse width, and f is frequency. In one embodiment, intensity is varied by adjusting only amplitude or pulse width (parameter adjustment priority: amplitude > pulse width). Frequency independently regulates tactile type to avoid intensity-type coupling interference. Amplitude and pulse width adjustments must be based on the user's safe parameter adjustment range.

[0045] The type code reflects the correspondence between the frequency range and the tactile trigger type, which can be achieved based on frequency changes. Tactile trigger types include vibration, slip, pressure, and instantaneous pressure. For example, a 1-50Hz code for vibration can be used for object slip warning. A 100-500Hz code for pressure can be used for grip force feedback. Frequency adjustment should refer to the user's safety parameter adjustment range.

[0046] Dynamic coding is used to generate a parameter dynamic mapping model based on the interaction characteristics between the prosthesis and the object, as well as intensity coding and type coding, to adjust the electrode stimulation parameters. For example, based on the real-time detection data of the prosthetic sensor, the main controller extracts the interaction characteristics (such as slip acceleration) through a sliding window Kalman filter, and adjusts the stimulation parameters based on the parameter dynamic mapping model generated by intensity coding and type coding. At the same time, the user's historical preferences can be continuously optimized through an adaptive learning algorithm (such as one based on a reinforcement learning framework) to improve the matching accuracy between scene requirements and tactile feedback stimulation parameters and enhance personalized adaptation capabilities. Through the above-mentioned multi-parameter coding, the intensity-type-scene ternary mapping relationship can be obtained.

[0047] Step S3: Apply the constructed intensity-type-scene ternary mapping relationship to the tactile feedback scenario of human-computer interaction, and adjust the corresponding electrode stimulation parameters according to the feedback results.

[0048] Taking the prosthetic control scenario as an example, constructing a strength-type-scene ternary mapping relationship for tactile feedback mainly includes the following steps.

[0049] Step S31: spatial coding.

[0050] First, the stimulation site is calibrated. The amputee determines the phantom finger sensory area through mechanical pressure stimulation (1cm spherical stylus). A square grid matrix of 8cm×8cm with a side length of 1cm is drawn in this area. Highly sensitive sites are screened and electrically stimulated. Finally, the electrode placement positions of the five phantom fingers are determined. Healthy users can place them on the pads of their five fingers. Subsequently, 5-channel electrodes (10mm in diameter) are placed, and the main controller independently controls the channels according to the preset mapping (A1 = thumb, A5 = little finger), thereby establishing a mapping relationship between finger tactile sensation and electrode channels.

[0051] Step S32: threshold calibration.

[0052] After spatial encoding establishes the mapping relationship between finger tactile sensation and the five-channel electrodes, the evoked threshold (minimum threshold) is first determined. The main controller starts with the baseline parameters (0.1mA, 20μs, 1Hz) and increases them in steps (amplitude 0.1mA / step, pulse width 20μs / step, frequency 1Hz / step) until the user perceives tactile sensation. The main controller records these critical parameters as the evoked threshold. Next, the pain threshold (maximum threshold) is determined, continuing to increase the parameters from the minimum threshold until the user perceives pain. The main controller records these critical parameters as the pain threshold. The range between the two thresholds represents the user's safe parameter adjustment range, and this threshold calibration process can be completed automatically by the main controller.

[0053] Step S33: intensity encoding.

[0054] First, the initialization parameters and relationship calculations are performed. According to the user's dual threshold calibration results (tactile evoked threshold and pain threshold), the initial amplitude, pulse width and frequency and the adjustable parameter range are set. avg ), so the relationship between the pulse average power and the three parameters is calculated and derived.

[0055] For example, the single pulse energy E onepulse It is mainly determined by the amplitude (PA), pulse width (PW), and human body impedance (R), and is expressed as:

[0056] E onepulse =PA 2 ×R×PW(1)

[0057] Pulse average power P avg (Energy delivered per unit time) is the product of the single pulse energy and the pulse frequency (f, pulses per second), expressed as:

[0058] P avg =E onepulse ×f=PA 2 ×R×PW×f(2)

[0059] Then, the intensity is adjusted by changing the average power of the pulse. In one embodiment, the average power is changed by coordinating the three parameters (PA, PW, f) to adjust the intensity. Among them, the amplitude (PA) dominates P due to its square relationship. avg Change is the core parameter for adjusting the intensity; the pulse width (PW) linearly affects P avg Changes are the main parameters for regulating intensity; frequency (f) linearly affects P avg change.

[0060] In short, adjusting the tactile evoked intensity requires adjusting the average pulse power P avg , and adjust P avgThe PA or PW needs to be adjusted. The specific relationship is as follows:

[0061]

[0062] PW∝P avg ∝I(4)

[0063] Step S34: intensity adjustment.

[0064] Based on the above relationship and the safe parameter adjustment range, the main controller adjusts parameters in order of priority (amplitude > pulse width) to achieve intensity changes. It prioritizes PA adjustment (Equation 3). If the amplitude parameter limits are reached, it switches to PW adjustment (Equation 4). f is independently used for type coding to avoid interfering with intensity stability.

[0065] Step S35: type coding.

[0066] After adjusting the tactile induced intensity through intensity coding, it is necessary to further construct an individualized type coding table for the user, such as 1-50Hz for vibration / slip sensation, which can be used for warning of unstable grasping of objects; 100-300Hz for uniform pressing sensation, which can be used for static grip force feedback; 300-500Hz for instantaneous pressing, which can be used for surface texture recognition (such as bumps). Based on this type coding table, the tactile induced type is changed by adjusting the frequency. Subsequently, power balancing adjustment is performed. When switching the tactile type (f), PA / PW is automatically compensated to maintain P avg , keeping the intensity unchanged.

[0067] Step S36, dynamic encoding.

[0068] Dynamic coding includes acquiring interaction features, constructing a dynamic parameter mapping model, adjusting the stimulus parameter combination based on the interaction features and the constructed parameter mapping model, and personalizing and optimizing the parameter mapping model.

[0069] Specifically, prosthetic sensors first detect and extract key features in real time. For example, when a prosthetic limb performs a grasping action, the main controller collects data from pressure, IMU, vibration, light, myoelectricity, and electroencephalogram (EEG) sensors. It then uses a sliding window Kalman filter to extract interactive features such as slip acceleration, grasping force, vibration amplitude, light intensity, voice commands, myoelectricity power spectral density, and EEG amplitude.

[0070] Then, based on the extracted key features, combined with the intensity-coded regulation relationship and type coding table, the following parameter dynamic mapping model is constructed:

[0071] 1) Grip force feedback

[0072] The average pulse power is adjusted according to the detection data of the prosthetic grip pressure sensor (grip force F), and the frequency is fixed at 100Hz to induce a pressing sensation:

[0073]

[0074] Among them, V in The pressure sensor converts the detected pressure F into an analog voltage output; V in and V max Is the minimum and maximum range of the pressure sensor; P min and P max is the user’s evoked threshold and pain threshold. Then, the main controller calculates the user’s pain threshold based on PA and PW. avg The system uses the relationship between the two (Formulas 3 and 4) and adjusts PA or PW according to priority. Furthermore, the system supports dual-modal closed-loop correction: By analyzing muscle power spectral density using real-time EMG sensor data or receiving natural language commands from the user (such as "lack of grip strength"), it automatically triggers parameter compensation (amplitude or pulse width) for grip feedback intensity, achieving stimulation optimization through the coordinated regulation of multiple sensory signals.

[0075] 2) Sliding feedback

[0076] The frequency is dynamically increased based on the sliding acceleration data (sliding acceleration a) from the IMU. The frequency range is 1-50Hz, and the sliding speed information of the object is provided by changing the vibration frequency:

[0077] f=1+a(f min ≥1Hz,f max ≤50Hz)(6)

[0078] In addition, the increasing step size of the stimulation frequency is dynamically adjusted according to the voice signal instruction.

[0079] 3) Slip warning

[0080] The IMU and pressure sensor jointly detect that the object is about to slip. The frequency is adjusted to 10Hz, and the amplitude and pulse width are adjusted to maximize the parameter range to reach the upper limit of the pain threshold, inducing a painful vibration to warn of the slip.

[0081] 4) Emotionally adaptive prosthetic tactile sensation

[0082] When the voice unit detects a shift in pitch, the system triggers a dual-mode pain-safety adjustment strategy: PA is dynamically attenuated to 60% of the baseline value (to prevent emotional overload), while PW is shortened by 50% and the stimulation frequency is increased by 15% to simulate an "electric shock retraction reflex." Myoelectric sensors detect muscle fatigue and achieve emotionally driven tactile intensity adaptation by reducing PA and PW or switching the electrical stimulation site.

[0083] 5) Brain-computer collaborative precision operation

[0084] When the system detects that the P300 component of the EEG signal, representing "precision manipulation intention," exceeds 6 microvolts in amplitude and the power spectral density of the hand's surface electromyography (EMG) signal exceeds 50 μV² / Hz, it automatically triggers an intelligent adjustment mechanism. First, the system compresses the electrical stimulation PW in real time based on EEG signal strength and EMG accuracy, adjusting the PW to 80 μs to match the precise conduction requirements of neural signals. Second, based on the contact force change rate of the hand pressure sensor (increasing by more than 5 N per second), the stimulation current amplitude is strictly limited to a safe range of 1 to 3 mA to avoid muscle overload caused by sudden load changes. If the IMU detects tool slippage, the system immediately activates emergency mode: the electrical stimulation frequency is adjusted to 300 Hz and a reverse phase pulse (delay of 0.5 ms, amplitude of 4 mA) is superimposed. This "emergency brake" is achieved through the high-frequency inhibition of neural synapses, reminding the user to maintain attention and precisely manipulate the tool.

[0085] The parameter dynamic mapping model reflects the correspondence between the interaction characteristics with the target and the electrode stimulation parameter combination, such as the correspondence between the above-mentioned grasping force feedback, sliding feedback, slip warning, emotion-adaptive prosthetic tactile sensation, brain-computer collaborative fine operation and the electrode stimulation parameter combination.

[0086] After obtaining the parameter dynamic mapping model, the parameters can be dynamically adjusted based on the interaction features obtained in real time. For example, the main controller adjusts the stimulation parameter combination based on the interaction features and the constructed parameter mapping model to achieve specific interaction tasks, such as pressing, moving, or gripping.

[0087] Furthermore, the parameter dynamic mapping model can be personalized and optimized. For example, the PC side collects relevant data and continuously optimizes the user's historical preferences through an adaptive learning algorithm (based on the reinforcement learning framework).

[0088] The multi-parameter dynamic encoding method provided by the present invention can achieve breakthroughs in high-precision tactile feedback technology in the following fields through parameter collaborative control and dynamic adaptation mechanism.

[0089] Specifically, in the field of prosthetic control, the system can be applied to prosthetic control systems, achieving precise tactile feedback for amputees' phantom fingers. Specifically, the system first uses a five-channel electrode array for spatial encoding. Based on the distribution of the amputee's phantom finger sensor, the electrode placement from the thumb to the pinky is sequentially determined, establishing a mapping relationship corresponding to the finger tactile sensation. Subsequently, automated threshold calibration is used to calibrate the evoked threshold and pain threshold for each channel, determining a safe parameter adjustment range to ensure that the feedback stimulation effectively transmits tactile information without exceeding the user's tolerance limit. Next, based on the grip force data collected by the prosthetic's built-in flexible piezoresistive sensor, intensity encoding is achieved through a three-parameter coordinated control model of amplitude, pulse width, and frequency. While prioritizing amplitude and switching to pulse width control after reaching the parameter boundary, this ensures a continuous and smooth change in tactile intensity. Furthermore, frequency control is used to construct a personalized type coding table, mapping different tactile types, such as vibration and pressure, to specific frequency ranges, thereby providing feedback on tactile information during operations such as grasping and sliding. More importantly, the system combines the IMU and pressure sensor integrated on the prosthesis to collect the operating status (such as slip acceleration) in real time, and uses the sliding window Kalman filter to extract key features, and provides slip speed feedback based on the constructed mapping model. Especially when an object slip warning is detected, when the multi-mode sensor data jointly detects an abnormal mutation (such as a sudden drop in pressure and a slip acceleration greater than 0.5m / s 2 ), the system triggers emergency coordinated control mode, generating a warning signal using 10Hz high-frequency vibration superimposed on the amplitude / pulse width parameters of the upper pain threshold. Experimental verification has shown that when the grip force increases from 2N to 20N, the stimulation amplitude can be gradually increased from 2.5mA to 4.2mA, significantly enhancing the amputee's perception of pressure changes and tactile intensity. This solution has increased the grasping success rate of amputees by 30%, and the slip warning response delay is ≤ 200ms, significantly improving the real-time and safety of prosthetic operation.

[0090] In the field of industrial robots, the present invention can be applied to the tactile feedback system of the robot's end effector to achieve real-time feedback on the contact status and grasping force of the object. The system arranges a multi-channel electrode array on the robot's hand or grasping area, realizes spatial encoding through a preset mapping relationship, and forms a localized tactile sensing network; for different grasping tasks and working conditions, it uses automated threshold calibration to calculate the induced and pain thresholds of each feedback area respectively, ensuring that the feedback stability and safety are maintained in a high-dynamic industrial environment. The robot's end effector integrates a distributed pressure sensor and a six-axis force / torque sensor to monitor the contact force distribution and posture offset of the grasped object in real time, and constructs a grasping force feedback parameter mapping model based on dynamic coding rules: when the contact force is insufficient, the electrical stimulation intensity is increased according to the amplitude priority principle (Formula 3); when the posture offset exceeds the limit, the 50Hz vibration mode is activated to prompt posture correction; at the same time, the system sets different frequency intervals for personalized type coding databases of different materials such as metal, glass, and flexible materials to achieve automatic material recognition. Finally, the system combines multimodal sensor data like pressure and acceleration collected in real time by the robot, and uses a dynamic coding method to adjust stimulation parameters. When an object slips or experiences abnormal contact, a vibration sensation is immediately triggered, generating an emergency braking signal. A priority arbitration mechanism ensures that safety warning information is prioritized for feedback. Furthermore, the system supports the integration of signals such as vibration, light intensity, EEG, EMG, and voice input, enabling intelligent manipulation of electrical stimulation parameters. By collecting EEG or EMG signals in real time, the system can automatically adjust stimulation amplitude and frequency based on the operator's neural or muscular state. Alternatively, parameter models can be switched via voice commands, further enhancing the system's interactivity and adaptability. In practical applications, this solution has reduced the breakage rate in precision grasping by 45% and increased the response speed to abnormal working conditions by 2.3 times, significantly improving the overall safety and reliability of industrial automation systems.

[0091] To further validate the effectiveness of this invention, mathematical derivation, demonstration, and preliminary experiments were conducted on more than 10 subjects. When the subjects performed a prosthetic control task, the main controller modified the stimulation parameter combination based on the environmental conditions detected by the prosthetic sensors and provided feedback to the subjects to modify their grip posture and grip force, thereby improving task performance.

[0092] Experimental data confirms that the present invention demonstrates improved tactile induction efficiency, enhanced control effects, and adaptability to multiple scenarios. Furthermore, the encoding method requires minimal computation, is easy to integrate, and requires minimal additional cost, demonstrating feasibility for clinical translation and industrialization. In summary, the present invention is able to construct a mapping model between multidimensional parameters and tactile perception, with the ability to dynamically adapt to different usage scenarios and individual differences, thereby providing a more accurate and personalized tactile feedback experience.

[0093] It should be noted that, without violating the spirit and scope of the present invention, those skilled in the art may make appropriate changes or modifications to the above embodiments. For example, the microcontroller STM32F4107 can be replaced with a low-power model (such as the STM32L4 series) or a high-performance model (such as the STM32H7 series) to adapt to the computing power and power consumption requirements of different scenarios. For spatial coding, the stimulation electric stimulation array can use stretchable conductive materials (such as liquid metal-silicone composite electrodes) to adapt to the dynamic changes in the morphology of the amputee's residual limb. For type coding, different frequency pulses (such as 10Hz+100Hz) can be superimposed in a single stimulation cycle to simulate composite touch, thereby constructing more complex sensory modalities and further expanding scene applications. For another example, the step size, parameter range, etc. involved in various coding processes can be appropriately adjusted according to the actual scenario or individual user. In addition, the present invention can also be applied to fields such as medical rehabilitation and neural reconstruction, such as sensory function reconstruction and neural remodeling for stroke patients.

[0094] In summary, the multi-parameter dynamic encoding method for tactile feedback provided by the present invention achieves multi-dimensional expression, dynamic adaptation, and individual compatibility optimization of tactile feedback by integrating the collaborative control mechanism of TENS parameters (pulse amplitude, pulse width, and frequency) with the scene adaptive parameter mapping model. It can be widely used in human-computer interaction scenarios that require tactile feedback. For example, in the field of prosthetic control, the effectiveness of slip warning, real-time matching of grip pressure, and sliding rate feedback has been verified. At the same time, it shows broad application prospects in the field of industrial robots. Compared with the existing technology, the advantages of the present invention are mainly reflected in the following aspects:

[0095] 1) The present invention proposes a five-step progressive method framework consisting of spatial coding, threshold calibration, intensity coding, type coding, and dynamic coding, which realizes the full-dimensional mapping of parameter space and perceptual features. In addition, the spatial coding relationship of tactile feedback is established based on a multi-channel electrode array, and the amplitude-pulse width priority adjustment rule and frequency-independent coding mechanism are pioneered for intensity coding and type coding, respectively, eliminating multi-parameter coupling interference.

[0096] 2) The present invention uses sensor data (pressure, acceleration, vibration, light, voice, EEG, EMG, etc.) to drive the parameter dynamic mapping model in real time to adjust the stimulation parameters, and combines it with an adaptive learning algorithm (such as one based on a reinforcement learning framework) to optimize the scene adaptation accuracy, thereby realizing a dynamic closed-loop adjustment mechanism.

[0097] 3) The present invention can adapt to the needs of multiple scenarios such as prosthetic limb control and industrial robots, and realizes the reusability of tactile feedback technology through standardized coding rules (such as 10Hz coding slip warning and 100Hz coding grip force feedback), that is, it realizes the universality of cross-domain applications.

[0098] 4) The multi-parameter dynamic encoding method proposed in this invention takes into account the advantages of multi-parameter collaboration, dynamic self-adaptation, strong individual adaptability, and automated parameter calibration. Through mathematical theoretical derivation and experimental quantification of tactile perception, it realizes "intensity-type-scene" collaborative mapping, breaking through the limitations of a single parameter adjustment dimension (such as encoding vibration intensity), thereby improving the naturalness, environmental adaptability, and accuracy of tactile feedback.

[0099] 5) The encoding method proposed in the present invention provides feedback on gripping conditions such as gripping force, slippage, and falling, allowing users to continuously correct their gripping gestures and gripping force to achieve more precise prosthetic control. The method is easily integrated into a prosthetic controller, has low computational complexity, fast response, and complete theory, which facilitates programming for automated parameter adjustment and application adaptation in most scenarios, pushing prosthetic control from "mechanical execution" to a humanoid interactive paradigm of "perception-decision-execution."

[0100] 6) The present invention combines sensor data to establish a dynamic parameter mapping model for environmental perception and tactile feedback, realizing interactive scenario-driven parameter adaptive adjustment; introduces dual-threshold automatic calibration technology to quickly determine the individual perception threshold range and construct personalized parameter adjustment boundaries, significantly reducing the time consumption of manual adaptation and enhancing dynamic interaction and individual adaptation capabilities.

[0101] The present invention may be a system, a method and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for causing a processor to implement various aspects of the present invention.

[0102] Computer-readable storage medium can be a tangible device that can keep and store the instructions used by the instruction execution device.Computer-readable storage medium can be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device or any suitable combination thereof.More specific examples (non-exhaustive list) of computer-readable storage medium include: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanical encoding device, for example, a punch card or a convex structure in a groove having instructions stored thereon, and any suitable combination thereof.Computer-readable storage medium used herein is not interpreted as a transient signal itself, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagated by waveguides or other transmission media (for example, light pulses by fiber optic cables), or electrical signals transmitted by wires.

[0103] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to each computing / processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions to be stored in the computer-readable storage medium in each computing / processing device.

[0104] The computer program instructions for performing the operation of the present invention can be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages such as Smalltalk, C++, Python, and conventional procedural programming languages such as "C" language or similar programming languages. The computer readable program instructions can be executed entirely on the user's computer, partially on the user's computer, as an independent software package, partially on the user's computer, partially on a remote computer, or completely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (e.g., using an Internet service provider to connect via the Internet). In some embodiments, an electronic circuit, such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), is personalized by utilizing the state information of the computer readable program instructions, and the electronic circuit can execute the computer readable program instructions, thereby realizing various aspects of the present invention.

[0105] Various aspects of the present invention are described herein with reference to flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present invention. It should be understood that each block of the flowcharts and / or block diagrams, and combinations of blocks in the flowcharts and / or block diagrams, can be implemented by computer-readable program instructions.

[0106] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine, so that when these instructions are executed by the processor of the computer or other programmable data processing device, a device is generated that implements the functions / actions specified in one or more blocks in the flowchart and / or block diagram. These computer-readable program instructions can also be stored in a computer-readable storage medium, where these instructions cause the computer, programmable data processing device, and / or other device to operate in a specific manner. Thus, the computer-readable medium storing the instructions comprises an article of manufacture that includes instructions for implementing various aspects of the functions / actions specified in one or more blocks in the flowchart and / or block diagram.

[0107] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device so that a series of operational steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to implement the functions / actions specified in one or more blocks in the flowchart and / or block diagram.

[0108] The flowcharts and block diagrams in the accompanying drawings show the possible implementation architecture, functions and operations of the systems, methods and computer program products according to multiple embodiments of the present invention. In this regard, each box in the flowchart or block diagram can represent a module, program segment or part of an instruction, and the module, program segment or part of the instruction contains one or more executable instructions for implementing the specified logical function. In some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system that performs the specified function or action, or can be implemented by a combination of dedicated hardware and computer instructions. It is well known to those skilled in the art that implementation by hardware, implementation by software, and implementation by a combination of software and hardware are all equivalent.

[0109] While various embodiments of the present invention have been described above, the foregoing description is intended to be illustrative, non-exhaustive, and not limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is selected to best explain the principles of the embodiments, their practical applications, or technological improvements in the marketplace, or to enable others skilled in the art to understand the embodiments disclosed herein. The scope of the present invention is defined by the appended claims.

Claims

1. A multi-parameter dynamic encoding method for tactile feedback, comprising the following steps: Disposing a multi-channel electrode array in the tactile sensing area, and establishing a correspondence between each channel of the multi-channel array electrode and a spatial position in the tactile sensing area through spatial coding; For individual users, the minimum and maximum values of multiple types of electrode stimulation parameters reflecting the degree of tactile perception are calibrated to obtain a dual-threshold calibration result, where the minimum value is the evoked threshold representing the user's perception of tactile sensation, and the maximum value is the pain threshold representing the user's perception of pain. The dual-threshold calibration result includes an amplitude safety range, a pulse width safety range, and a frequency safety range defined by the minimum and maximum values; Performing intensity coding according to the dual-threshold calibration result, wherein the intensity coding reflects the relationship between the tactile evoked intensity and the average power of the stimulation pulse, wherein the average power of the stimulation pulse is defined by the amplitude, pulse width and frequency; constructing a type code for the individual user, wherein the type code reflects a corresponding relationship between a frequency interval and a tactile induced type; Based on multiple sensor data and the intensity code and the type code, a parameter dynamic mapping model is generated, wherein the parameter dynamic mapping model reflects the correspondence between the interaction characteristics between the tactile sensing area and the target and the electrode stimulation parameter combination.

2. The method according to claim 1, characterized in that The tactile sensing area is the tactile sensing area of the phantom finger of an amputee, the tactile sensing area provided on the end effector of a robot, or the tactile sensing area provided on a wearable tactile feedback device.

3. The method according to claim 1, characterized in that The stimulation pulse average power P avg Expressed as: P avg =E onepulse ×f=PA 2 ×R×PW×f in: It is onepulse =PA 2 ×R×PW Among them, E onepulse is the single pulse energy, PA is the amplitude, PW is the pulse width, R is the human body impedance, and f is the pulse frequency.

4. The method according to claim 3, characterized in that The intensity code is constructed according to the following steps: The amplitude is adjusted within the amplitude safety range to achieve tactile induced enhancement changes based on the following formula until the upper and lower limits of the amplitude parameters are reached: Within the pulse width safety range, the pulse width is adjusted based on the following formula to achieve changes in the tactile induced intensity: PW∝P avg ∝I Where PA is the amplitude, PW is the pulse width, and I represents the tactile evoked intensity.

5. The method according to claim 3, characterized in that The type code is constructed according to the following steps: Within the frequency safety range, the tactile induced type is changed by adjusting the frequency; When switching the tactile induction type, the average power of the stimulation pulse P is maintained by compensating the amplitude and frequency. avg , keeping the intensity constant.

6. The method according to claim 1, characterized in that The parameter dynamic mapping model is constructed according to the following steps: detecting sensing data of the tactile sensing area in real time; Extracting interaction features with the target from the sensor data through a sliding window Kalman filter, and dynamically mapping a parameter model based on the intensity coding and the type coding; For the generated parameter dynamic mapping model, the user's historical preferences are optimized through an adaptive learning algorithm.

7. The method according to claim 3, characterized in that The parameter dynamic mapping model reflects the correspondence between grip force feedback, sliding feedback, slip warning, emotion-adaptive prosthetic tactile sensation, brain-computer collaborative fine manipulation and the electrode stimulation parameter combination; The corresponding relationship between the grip force feedback and the electrode stimulation parameter combination is determined according to the following steps: The average power P of the stimulation pulse is adjusted according to the grip force F detected by the pressure sensor. avg , the frequency is fixed to induce a sense of pressure: Among them, V in The pressure sensor converts the detected pressure F into an analog voltage, V min and V max is the minimum and maximum range of the pressure sensor, P min and P max is the individual user's evoked threshold and pain threshold; Based on the amplitude PA, pulse width PW and the stimulation pulse average power P avg The relationship between the two is established, and the amplitude and pulse width are adjusted in order of priority; The corresponding relationship between the sliding feedback and the electrode stimulation parameter combination is determined according to the following steps: According to the sliding acceleration data detected by the inertial measurement unit, the frequency is dynamically increased within the frequency safety range, and the object sliding speed information is provided by adjusting the frequency change, which is expressed as: f=1+a(f min ≥1Hz,f max ≤50Hz) Where f represents the frequency and a is the sliding acceleration; The corresponding relationship between the slip warning and the electrode stimulation parameter combination is determined according to the following steps: When the inertial measurement unit and pressure sensor jointly detect that an object is about to slip, they adjust the frequency to the set frequency, and adjust the amplitude and pulse width to maximize the parameter range to reach the upper limit of the pain threshold, inducing a painful vibration to provide a slip warning.

8. The method according to claim 1, characterized in that The tactile induction type includes one or more of vibration, sliding, pressing and instantaneous pressing.

9. A computer-readable storage medium having a computer program stored thereon, wherein: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

10. A computer device comprising a memory and a processor, wherein a computer program capable of being run on the processor is stored in the memory, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

Images