Intelligent injection system and method based on graphene heterojunction annular sensor

Abstract

The invention provides an intelligent injection system and method based on a graphene heterojunction annular sensor. The intelligent injection system comprises the annular sensor, a central processing unit and an injection execution mechanism. The annular sensor is made of a graphene / PEDOT: PSS heterojunction material, surrounds the periphery of the injection needle and monitors multi-modal data in real time; an AI algorithm module is arranged in the central processing unit to generate personalized injection parameters; the injection execution mechanism accurately controls injection based on a closed-loop feedback system and a fuzzy PID control algorithm. The intelligent injection method comprises the steps of biomechanical signal collection, injection parameter generation through an AI algorithm, injection control execution, data storage and the like. Accurate control of injection can be achieved, the tissue damage rate is reduced, and an intelligent injection solution is provided for personalized medical treatment.

Description

Technical Field

[0001] The present invention belongs to the field of medical equipment and artificial intelligence technology, and specifically relates to an intelligent injection system and method based on a graphene heterojunction ring sensor. Background Art

[0002] Injection technology is a core medical procedure, widely used in scenarios such as insulin administration, vaccine injections, and targeted therapies. Its accuracy and safety directly impact treatment outcomes and patient experience. Traditional injections rely on medical staff's experience to determine injection depth and speed. Individual differences (such as skin thickness, tissue density, and vascularity) can easily lead to over-deep or shallow injections, causing pain, tissue damage, or inefficient drug absorption.

[0003] In recent years, breakthroughs in nanomaterials and biosensing technologies have provided new avenues for high-precision physiological signal monitoring. For example, graphene-based flexible sensors can detect micron-level displacement, while multimodal data fusion technology can simultaneously analyze displacement, stress, and temperature signals. Furthermore, the introduction of artificial intelligence algorithms (such as deep learning and reinforcement learning) into medical devices has enabled real-time analysis and decision-making of complex biological signals. However, existing technologies have yet to effectively combine high-precision biosensing with AI dynamic control to form a closed-loop intelligent injection system, resulting in technical bottlenecks in achieving precise and personalized injections.

[0004] While auxiliary positioning technologies like radio frequency identification and infrared navigation can improve the automation level of equipment, their anti-interference capabilities and real-time performance in complex biological environments still need to be optimized. Therefore, an intelligent injection system that integrates high-precision sensing, multimodal data fusion, and adaptive AI decision-making is urgently needed to address the inherent shortcomings of traditional methods and achieve safe, accurate, and personalized medical injections.

[0005] A search revealed a Chinese patent application with publication number CN213131232U, which discloses an injection system. This system utilizes a controller, detection devices (such as deformation sensors, pressure sensors, and impedance sensors), and a high-pressure injector. The sensors monitor physical changes at the injection site (such as leakage and needle displacement) and trigger the controller to stop the injection. The system's core mechanism is a passive response mechanism based on electrical signals, relying on preset thresholds to detect abnormalities.

[0006] The present application differs from the aforementioned reference documents in the following ways:

[0007] 1. This application integrates a graphene / PEDOT:PSS heterojunction ring sensor to collect multimodal data (needle displacement, tissue pressure, and temperature) in real time, combines an AI algorithm to dynamically generate personalized injection parameters, and accurately controls the injection process through a fuzzy PID closed-loop feedback system, which is an active intelligent regulation. The above-mentioned comparative document uses a combination of a controller, a detection device (such as a deformation sensor, a pressure sensor, and an impedance sensor) and a high-pressure injector. Its core is a passive response mechanism based on electrical signals, which relies on preset thresholds to judge abnormalities.

[0008] 2. This application emphasizes personalized medical needs, is suitable for subcutaneous, intramuscular and targeted injection scenarios, supports dynamic parameter adjustment, and is suitable for complex medical environments that require high precision and low damage. In contrast, the above-mentioned comparative documents are suitable for scenarios such as high-pressure contrast agent injection, focusing on solving the problems of abnormal detection and emergency stop of venous leakage or needle misentry into subcutaneous tissue, and relying on medical staff to preset thresholds.

[0009] A search revealed patent publication number CN109497960A, which discloses an integrated injection system, an integrated injection device, and a method for monitoring injection events. Sensors on the injection device collect patient physiological data (such as body fat composition and blood pressure) and injection behavior data, and transmit them to an external communication device for compliance analysis. The technology focuses on multi-dimensional health data integration and remote medical monitoring, with sensors primarily used for biometric collection.

[0010] The present application differs from the aforementioned reference documents in the following ways:

[0011] 1. This application focuses on real-time monitoring of physical parameters during the injection process, including needle displacement, tissue pressure, and temperature signals. High-precision displacement detection is achieved through capacitive coupling and piezoelectric effect of graphene heterojunction sensors. The core goal is to monitor injection parameters in real time to reduce the risk of tissue damage. The above-mentioned comparative documents collect the patient's physiological data (such as body fat composition, blood pressure) and injection behavior data through sensors on the injection device, and transmit them to an external communication device for compliance analysis. The technical focus is on multi-dimensional health data integration and remote medical monitoring, and the sensors are mainly used for biometric collection.

[0012] 2. Displacement detection is the key technology of this application, which achieves micron-level displacement resolution through capacitive coupling between the comb-shaped moving electrode laser-etched on the needle surface and the static electrode of the ring-shaped sensor. In contrast, the above-mentioned comparative documents only indirectly infer the injection completion status through the needle position and do not involve high-precision displacement measurement. Summary of the Invention

[0013] Purpose of the invention:

[0014] The purpose of this invention is to provide an intelligent injection system based on a graphene heterojunction ring sensor. By real-time monitoring of biomechanical parameters during the injection process (such as needle displacement, tissue pressure, temperature signals, etc.), combined with AI algorithms, the system dynamically adjusts the injection depth, speed, and pressure, thereby reducing the difficulty of operation for medical staff and improving injection accuracy and safety.

[0015] Technical solution:

[0016] The present invention provides an intelligent injection system based on a graphene heterojunction ring sensor, comprising a ring sensor, a central processing unit and an injection actuator;

[0017] The ring-shaped sensor is a nanomaterial composite structure made of graphene / PEDOT:PSS heterojunction material. It surrounds the needle with a fixed gap and is connected to the housing of the injection actuator through a sliding sleeve mechanism. It stays on the skin surface during the injection process.

[0018] The ring-shaped sensor is composed of a heterojunction formed by alternating graphene layers and PEDOT:PSS conductive polymer layers, with an embedded temperature sensing unit, an integrated microelectrode array on the outer surface, and a symmetrical electrostatic level integrated on the inner surface;

[0019] The annular sensor monitors the needle displacement, tissue pressure and temperature signals of the injection actuator during the injection process in real time, and uses a capacitive encoder as the core detection component to detect the displacement;

[0020] The surface of the non-penetration section of the needle of the injection actuator is laser-etched with micron-scale periodic comb-shaped moving electrodes;

[0021] The central processing unit has a built-in AI algorithm module, including a machine learning model and real-time control logic;

[0022] The injection actuator is equipped with a micro motor and a closed-loop feedback system to dynamically adjust the injection parameters according to AI instructions;

[0023] The AI algorithm module includes an improved LSTM-CRF temporal classification network for dynamically identifying skin tissue layers;

[0024] The closed-loop feedback system is based on a fuzzy PID control algorithm to achieve precise control of the injection depth.

[0025] Furthermore, the system has a built-in abnormality prediction module based on the Transformer model, which can provide early warning of the risk of microbleeding or nerve damage and trigger an injection termination instruction; the various types of data collected in real time by the annular sensor provide an analysis basis for the Transformer model.

[0026] The present invention also provides an intelligent injection method based on a graphene heterojunction ring sensor, comprising the following steps:

[0027] A: Biomechanical signals during injection are collected in real time through a ring sensor;

[0028] Step A1: Activate the ring sensor to simultaneously monitor needle displacement, tissue pressure, and temperature signals. Based on its unique heterojunction structure and operating mechanism, the sensor senses and converts these biomechanical signals into electrical signals in real time.

[0029] Step A2: Using wavelet packet decomposition and Hilbert-Huang transform to extract the time-frequency domain features of the signal, the collected original signal is processed to mine more valuable information;

[0030] B: Dynamically generate injection parameters based on AI algorithm;

[0031] Step B1: Calculate the optimal injection depth, speed, and pressure based on an AI algorithm. Patient-related data collected by the sensor provides the basis for the AI algorithm, enabling it to comprehensively consider multiple factors and generate personalized injection parameters;

[0032] Step B2: If an abnormal stress mutation is detected, the Transformer early warning model is triggered to perform risk assessment. The early warning is triggered based on the stress data monitored in real time by the ring sensor. Once an abnormality occurs, the early warning model immediately starts analysis.

[0033] C: Execute dynamic injection control and complete data storage;

[0034] Step C1: The injection actuator adjusts the needle's trajectory and advancement speed based on AI instructions. The annular sensor provides real-time tissue information, and the AI accurately issues instructions based on the feedback to control the injection actuator's movements.

[0035] Step C2: Correcting injection path deviation in real time through a closed-loop feedback system. The tissue information fed back by the sensor in real time provides a reference for the closed-loop feedback system, which adjusts the injection path in real time based on this information to ensure injection accuracy.

[0036] Step C3: Encrypt and transmit the injection parameters, biomechanical data, and warning records to the cloud medical database; the transmitted data mainly includes various data collected by the annular sensor.

[0037] Furthermore, in step A1, the micron-scale periodic comb-shaped moving electrodes laser-etched on the non-penetrating surface of the syringe needle form a non-contact capacitive coupling with the symmetrical static electrodes fixed on the inner side of the ring-shaped sensor. The change in the overlapping area of the electrodes caused by needle displacement drives the periodic fluctuation of the capacitance value, achieving micron-level displacement resolution.

[0038] By utilizing the deformation characteristics of the graphene / PEDOT:PSS heterojunction, changes in tissue pressure will cause the sensor to deform, and then the tissue pressure will be converted into a measurable electrical signal through capacitance changes or piezoelectric effect, thus realizing tissue pressure monitoring;

[0039] By directly contacting the tissue with the embedded thermistor, changes in skin tissue temperature will cause corresponding changes in the thermistor tissue, thereby enabling monitoring of tissue temperature during the injection process.

[0040] Furthermore, in step A1, the annular sensor synchronously collects capacitance change signals, mechanical stress and temperature data;

[0041] When collecting capacitance change signals, a high-frequency carrier is used to excite the static electrode, and a lock-in amplifier is used to extract the amplitude and phase response of the dynamic electrode. The periodic change of capacitance is converted into displacement using a sub-pixel interpolation algorithm.

[0042] When collecting mechanical stress, the sensor material deforms when subjected to stress, causing its electrical properties to change. Mechanical stress information is obtained by detecting this change.

[0043] When collecting temperature data, the data collection is achieved by utilizing the characteristic that the electrical properties of the sensor material change with temperature;

[0044] Multimodal feature signals are extracted through wavelet packet decomposition and Hilbert-Huang transform, and the collected complex signals are processed to extract more representative and discriminative features, providing more accurate data for subsequent AI analysis.

[0045] Furthermore, in step B1, the AI algorithm module integrates a Bayesian optimization framework, collects force signals, displacement signals and temperature data during the injection process in real time, and constructs a multi-dimensional feature space in combination with the patient's age, skin condition and drug characteristics, and establishes a nonlinear mapping relationship between the input signal and the puncture attribute based on Gaussian process regression; dynamically generates personalized injection parameters through the Bayesian optimization framework, and iteratively optimizes the objective function using the expected improvement acquisition function. At the same time, historical injection data and clinical feedback are introduced to correct the parameter constraint space, and the tissue elasticity changes are analyzed through real-time force-displacement phase difference, and finally generates personalized injection parameters that adapt to individual differences and drug needs.

[0046] Furthermore, in step B2, the Transformer early warning model analyzes historical injection data and real-time signals through an attention mechanism to predict the risk of microbleeding and issue an alarm in advance; the signals collected in real time by the annular sensor are an important source of real-time signals, and combined with historical data, provide a comprehensive analysis basis for the early warning model.

[0047] Furthermore, in step C, the injection actuator supports dynamic adjustment of injection speed and pressure to adapt to subcutaneous, intramuscular and targeted injection scenarios; the tissue information fed back in real time by the annular sensor enables the injection actuator to dynamically adjust the injection parameters according to different tissue characteristics and injection requirements.

[0048] Furthermore, in step C1, the injection actuator adopts a fuzzy PID control algorithm to dynamically adjust the motor torque and step accuracy according to tissue density; information such as tissue pressure fed back by the annular sensor can reflect changes in tissue density and provide a control basis for the fuzzy PID control algorithm.

[0049] Furthermore, in step C3, the cloud medical database supports cross-device retrieval of patients' historical injection records for optimizing personalized injection strategies; the data collected and transmitted by the annular sensor enriches the content of the cloud medical database.

[0050] Beneficial effects:

[0051] The present invention has the beneficial effects of proposing an intelligent injection system and method based on a graphene heterojunction ring sensor. Medical staff only need to initialize the injection parameters, and the subsequent real-time monitoring, dynamic adjustment, and safety protection are automatically completed by the system. Through multimodal sensor fusion and AI dynamic decision-making, precise control of injections is achieved, significantly reducing the rate of tissue damage, while supporting personalized adaptation of injection strategies based on patient age, skin condition, and drug characteristics. In addition, the system uses the Transformer model to provide early warning of injection risks, combined with cloud-based data management, to provide a highly secure and high-precision intelligent medical solution for chronic disease treatment (such as diabetic insulin injections) and targeted drug delivery. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is the overall flow chart of the present invention.

[0053] Figure 2 Schematic diagram of the structure of the ring sensor in an embodiment of the present invention.

[0054] Figure 3 is a line drawing of a ring sensor according to an embodiment of the present invention.

[0055] Figure 4 4 is a control flow chart of a closed-loop feedback system in an embodiment of the present invention.

[0056] In the figure: S100 is a ring sensor, S101 is a graphene layer, S102 is a PEDOT:PSS conductive polymer layer, S103 is an integrated microelectrode array, S104 is a temperature sensing unit, S105 is an integrated symmetrical electrode, S200 is a central processing unit, and S300 is an injection actuator. DETAILED DESCRIPTION

[0057] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings.

[0058] The present invention provides an intelligent injection system and method based on a graphene heterojunction ring sensor. By real-time monitoring of biomechanical parameters during the injection process (such as needle displacement, tissue pressure, and temperature signals), combined with AI algorithms, the system dynamically adjusts the injection depth, speed, and pressure, thereby reducing the operational difficulty for medical staff and improving injection accuracy and safety.

[0059] like Figure 1 As shown, the intelligent injection system based on the graphene heterojunction ring sensor mainly consists of a ring sensor (S100), a central processing unit (S200) and an injection actuator (S300).

[0060] like Figure 2 、 Figure 3 As shown, the ring-shaped sensor (S100) is a nanomaterial composite structure made of a graphene / PEDOT:PSS heterojunction material. Graphene layers (S101) and PEDOT:PSS conductive polymer layers (S102) are alternately stacked to form the heterojunction. A microelectrode array (S103) is integrated on the outer surface to detect changes in capacitance or resistance, improving sensitivity, signal-to-noise ratio, and linearity, enabling monitoring of tissue pressure. The ring-shaped sensor also includes an embedded temperature sensing unit (S104) that monitors tissue temperature changes in real time via a thermistor. A symmetrical electrostatic stage (S105) is integrated on the inner surface to detect periodic fluctuations in capacitance as a needle, which is engraved with micrometer-scale comb-shaped moving electrodes, moves, enabling the simultaneous acquisition of micrometer-level displacement data.

[0061] The ring sensor (S100) surrounds the needle with a fixed gap and is connected to the injection actuator (S300) housing via a sliding sleeve mechanism. It remains on the skin surface during injection. The non-insertion section of the needle of the injection actuator (S300) is laser-etched with micrometer-scale periodic comb-shaped moving electrodes.

[0062] The annular sensor (S100) can monitor needle displacement, tissue pressure and temperature signals during the injection process in real time, and uses a capacitive encoder as the core detection component to detect displacement.

[0063] Graphene, as a ring-shaped sensor material, has excellent electrical properties and high specific surface area, and PEDOT:PSS has good conductivity and biocompatibility. The heterojunction formed by the two can enhance the perception of biological signals.

[0064] The central processing unit has a built-in AI algorithm module, including a machine learning model and real-time control logic; the AI algorithm module includes an improved LSTM-CRF temporal classification network for dynamically identifying skin tissue stratification.

[0065] The injection actuator is equipped with a micro motor and a closed-loop feedback system, which dynamically adjusts the injection parameters according to AI instructions; the closed-loop feedback system is based on a fuzzy PID control algorithm to achieve precise control of the injection depth.

[0066] The intelligent injection system based on the graphene heterojunction annular sensor of the present invention has a built-in abnormality prediction module based on the Transformer model, which can provide early warning of the risk of microbleeding or nerve damage and trigger an injection termination instruction; the various types of data collected in real time by the annular sensor provide an analysis basis for the Transformer model.

[0067] The present invention also provides an intelligent injection method based on a graphene heterojunction ring sensor, which mainly includes the following steps:

[0068] A: Biomechanical signals during injection are collected in real time through a ring sensor;

[0069] Step A1: Activate the ring sensor to simultaneously monitor needle displacement, tissue pressure, and temperature signals. Based on its unique heterojunction structure and operating mechanism, the sensor senses and converts these biomechanical signals into electrical signals in real time.

[0070] Step A2: Using wavelet packet decomposition and Hilbert-Huang transform to extract the time-frequency domain features of the signal, the collected original signal is processed to mine more valuable information;

[0071] B: Dynamically generate injection parameters based on AI algorithm;

[0072] Step B1: Calculate the optimal injection depth, speed, and pressure based on an AI algorithm. Patient-related data collected by the sensor provides the basis for the AI algorithm, enabling it to comprehensively consider multiple factors and generate personalized injection parameters;

[0073] Step B2: If an abnormal stress mutation is detected, the Transformer early warning model is triggered to perform risk assessment. The early warning is triggered based on the stress data monitored in real time by the ring sensor. Once an abnormality occurs, the early warning model immediately starts analysis.

[0074] C: Execute dynamic injection control and complete data storage;

[0075] Step C1: The injection actuator adjusts the needle's trajectory and advancement speed based on AI instructions. The annular sensor provides real-time tissue information, and the AI accurately issues instructions based on the feedback to control the injection actuator's movements.

[0076] Step C2: Correcting injection path deviation in real time through a closed-loop feedback system. The tissue information fed back by the sensor in real time provides a reference for the closed-loop feedback system, which adjusts the injection path in real time based on this information to ensure injection accuracy.

[0077] Step C3: Encrypt and transmit the injection parameters, biomechanical data, and warning records to the cloud medical database; the transmitted data mainly includes various data collected by the annular sensor.

[0078] In step A1, the micron-scale periodic comb-shaped moving electrodes laser-etched on the non-penetrating surface of the injector needle form a contactless capacitive coupling with the symmetrical static electrodes fixed on the inner side of the ring-shaped sensor. The change in the overlapping area of the electrodes caused by needle displacement drives the periodic fluctuation of the capacitance value, achieving micron-level displacement resolution.

[0079] By utilizing the deformation characteristics of the graphene / PEDOT:PSS heterojunction, changes in tissue pressure will cause the sensor to deform, and then the tissue pressure will be converted into a measurable electrical signal through capacitance changes or piezoelectric effect, thus realizing tissue pressure monitoring;

[0080] By directly contacting the tissue with the embedded thermistor, changes in skin tissue temperature will cause corresponding changes in the thermistor tissue, thereby enabling monitoring of tissue temperature during the injection process.

[0081] Micrometer-scale periodic comb-shaped moving electrodes laser-etched on the non-penetrating surface of the syringe needle form a contactless capacitive coupling with the symmetrical static electrodes fixed on the inner side of the ring-shaped sensor. The change in the overlapping area of the electrodes caused by needle displacement drives the periodic fluctuation of the capacitance value, achieving micrometer-level displacement resolution.

[0082] By utilizing the deformation characteristics of the graphene / PEDOT:PSS heterojunction, changes in tissue pressure will cause the sensor to deform, and then the tissue pressure will be converted into a measurable electrical signal through capacitance changes or piezoelectric effect, thus realizing tissue pressure monitoring;

[0083] By directly contacting the tissue with the embedded thermistor, changes in skin tissue temperature will cause corresponding changes in the thermistor tissue, thereby enabling monitoring of tissue temperature during the injection process.

[0084] The ring sensor synchronously collects capacitance change signals, mechanical stress and temperature data;

[0085] When collecting capacitance change signals, a high-frequency carrier is used to excite the static electrode, and a lock-in amplifier is used to extract the amplitude and phase response of the dynamic electrode. The periodic change of capacitance is converted into displacement using a sub-pixel interpolation algorithm.

[0086] When collecting mechanical stress, the sensor material deforms when subjected to stress, causing its electrical properties to change. Mechanical stress information is obtained by detecting this change.

[0087] When collecting temperature data, the electrical properties of the sensor material change with temperature.

[0088] In step A2, multimodal feature signals are extracted through wavelet packet decomposition and Hilbert-Huang transform, and the collected complex signals are processed to extract more representative and discriminative features, providing more accurate data for subsequent AI analysis.

[0089] In step B1, the AI algorithm module integrates the Bayesian optimization framework, collects force signals, displacement signals and temperature data during the injection process in real time, and constructs a multi-dimensional feature space in combination with the patient's age, skin condition and drug characteristics. It also establishes a nonlinear mapping relationship between the input signal and the puncture attribute based on Gaussian process regression. It dynamically generates personalized injection parameters through the Bayesian optimization framework, iteratively optimizes the objective function using the expected improvement acquisition function, introduces historical injection data and clinical feedback to correct the parameter constraint space, and analyzes tissue elasticity changes through real-time force-displacement phase difference, ultimately generating personalized injection parameters that adapt to individual differences and drug needs.

[0090] In step B2, the Transformer early warning model uses an attention mechanism to analyze historical injection data and real-time signals to predict the risk of microbleeding and issue an early warning alarm. The real-time signals collected by the annular sensor are an important source of real-time signals. Combined with historical data, they provide a comprehensive analytical basis for the early warning model.

[0091] In step C, the injection actuator supports dynamic adjustment of injection speed and pressure to adapt to subcutaneous, intramuscular and targeted injection scenarios; the tissue information fed back in real time by the annular sensor enables the injection actuator to dynamically adjust the injection parameters according to different tissue characteristics and injection requirements.

[0092] In step C1, the injection actuator uses a fuzzy PID control algorithm to dynamically adjust the motor torque and step accuracy according to tissue density. Information such as tissue pressure fed back by the annular sensor can reflect changes in tissue density and provide a control basis for the fuzzy PID control algorithm.

[0093] In step C3, the cloud medical database supports cross-device retrieval of patients' historical injection records for optimizing personalized injection strategies; the data collected and transmitted by the ring sensor enriches the content of the cloud medical database.

[0094] Example 1:

[0095] like Figure 4 As shown, the actual operation process of the intelligent injection system and method based on the graphene heterojunction ring sensor of the present application is as follows:

[0096] Before injection, the ring-shaped sensor is activated and begins synchronously collecting capacitance change signals, mechanical stress, and temperature data. The capacitance change signal uses a high-frequency carrier to excite the static electrode, and a lock-in amplifier extracts the amplitude and phase response of the dynamic electrode. Combined with a sub-pixel interpolation algorithm, the periodic capacitance changes are converted into displacement for acquisition. Mechanical stress is acquired by detecting changes in the electrical properties of the sensor material, which deform when subjected to stress. Temperature data is acquired by exploiting the temperature-dependent changes in the sensor material's electrical properties.

[0097] After completing data collection, multimodal feature signals are extracted through wavelet packet decomposition and Hilbert-Huang transform. The collected complex signals are processed to extract more representative and discriminative features, providing more accurate data for subsequent AI analysis.

[0098] After completing data collection and extraction, the AI algorithm integrating the Bayesian optimization framework is used to collect force signals, displacement signals and temperature data during the injection process in real time. A multi-dimensional feature space is constructed in combination with the patient's age, skin condition and drug characteristics, and a nonlinear mapping relationship between the input signal and the puncture attribute is established based on Gaussian process regression. Personalized injection parameters are dynamically generated through the Bayesian optimization framework, and the objective function is iteratively optimized using the expected improvement acquisition function. At the same time, historical injection data and clinical feedback are introduced to correct the parameter constraint space, and the tissue elasticity changes are analyzed through real-time force-displacement phase difference, ultimately generating personalized injection parameters that adapt to individual differences and drug needs.

[0099] Dynamic injection control is performed based on the personalized injection parameters generated by the AI algorithm. The injection actuator adjusts the needle's motion trajectory and propulsion speed according to the AI instructions. The annular sensor provides feedback on real-time tissue information, and the AI accurately issues instructions based on the feedback to control the movement of the injection actuator.

[0100] During the injection process, the injection path deviation is corrected in real time through a closed-loop feedback system based on the fuzzy PID control algorithm; the tissue information fed back by the sensor in real time provides a reference for the closed-loop feedback system, which adjusts the injection path in real time based on this information to ensure the accuracy of the injection.

[0101] The annular sensor provides real-time feedback on the needle position deviation. If the path deviation is greater than 50 μm, the dynamic compensation instruction is triggered, and the injection path is replanned according to the above method. The injection is executed again according to the replanned and adjusted injection path. If the path deviation is ≤ 50 μm, the injection is continued until the injection is completed.

[0102] At the same time, if an abnormal stress mutation is detected, the Transformer early warning model is triggered to perform risk assessment; an early warning is triggered based on the stress data monitored in real time by the annular sensor, and once an abnormality occurs, the early warning model immediately starts analysis.

[0103] After the injection is completed, the injection parameters, biomechanical data and warning records are encrypted and transmitted to the cloud medical database; the transmitted data mainly includes various data collected by the ring sensor, providing important value for subsequent medical analysis and optimization of personalized injection strategies.

[0104] The basic principles, main features, and advantages of the present invention are shown and described above. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are merely illustrative of the principles of the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent injection system based on a graphene heterojunction ring sensor, characterized in that: It includes a ring sensor, a central processing unit and an injection actuator; The ring-shaped sensor is a nanomaterial composite structure made of graphene / PEDOT:PSS heterojunction material. It surrounds the needle with a fixed gap and is connected to the housing of the injection actuator through a sliding sleeve mechanism. It stays on the skin surface during the injection process. The ring-shaped sensor is composed of a heterojunction formed by alternating graphene layers and PEDOT:PSS conductive polymer layers, with an embedded temperature sensing unit, an integrated microelectrode array on the outer surface, and a symmetrical electrostatic level integrated on the inner surface; The annular sensor monitors the needle displacement, tissue pressure and temperature signals of the injection actuator during the injection process in real time, and uses a capacitive encoder as the core detection component to detect the displacement; The surface of the non-penetration section of the needle of the injection actuator is laser-etched with micron-scale periodic comb-shaped moving electrodes; The central processing unit has a built-in AI algorithm module, including a machine learning model and real-time control logic; The injection actuator is equipped with a micro motor and a closed-loop feedback system to dynamically adjust the injection parameters according to AI instructions; The AI algorithm module includes an improved LSTM-CRF temporal classification network for dynamically identifying skin tissue layers; The closed-loop feedback system is based on a fuzzy PID control algorithm to achieve precise control of the injection depth.

2. The intelligent injection system based on graphene heterojunction ring sensor according to claim 1, characterized in that: The system has a built-in abnormality prediction module based on the Transformer model, which can provide early warning of microbleeding or nerve damage risks and trigger an injection termination instruction; the annular sensor collects various types of data in real time, providing an analysis basis for the Transformer model.

3. An intelligent injection method based on a graphene heterojunction ring sensor, characterized in that: The following steps are involved: A: Biomechanical signals during injection are collected in real time through a ring sensor; Step A1: Activate the ring sensor to simultaneously monitor needle displacement, tissue pressure, and temperature signals. Based on its unique heterojunction structure and operating mechanism, the sensor senses and converts these biomechanical signals into electrical signals in real time. Step A2: Using wavelet packet decomposition and Hilbert-Huang transform to extract the time-frequency domain features of the signal, the collected original signal is processed to mine more valuable information; B: Dynamically generate injection parameters based on AI algorithm; Step B1: Calculate the optimal injection depth, speed, and pressure based on an AI algorithm. Patient-related data collected by sensors provides the basis for the AI algorithm, which comprehensively considers multiple factors to generate personalized injection parameters. Step B2: If an abnormal stress mutation is detected, the Transformer early warning model is triggered to perform risk assessment. The early warning is triggered based on the stress data monitored in real time by the ring sensor. When an abnormality occurs, the early warning model immediately starts analysis. C: Execute dynamic injection control and complete data storage; Step C1: The injection actuator adjusts the needle's trajectory and advancement speed based on AI instructions. The annular sensor provides real-time tissue information, and the AI accurately issues instructions based on the feedback to control the injection actuator's movements. Step C2: Correcting injection path deviation in real time through a closed-loop feedback system. The tissue information fed back by the annular sensor in real time provides a reference for the closed-loop feedback system, which adjusts the injection path in real time based on the information fed back by the annular sensor to ensure injection accuracy. Step C3: Encrypt and transmit the injection parameters, biomechanical data, and warning records to the cloud medical database; the transmitted data mainly includes various data collected by the annular sensor.

4. The intelligent injection method based on graphene heterojunction ring sensor according to claim 3, characterized in that: In step A1, a micrometer-scale periodic comb-shaped moving electrode laser-etched on the non-penetrating surface of the syringe needle forms a contactless capacitive coupling with a symmetrical static electrode fixed on the inner side of the ring-shaped sensor. The change in the overlapping area of the electrodes caused by needle displacement drives the periodic fluctuation of the capacitance value, achieving micrometer-level displacement resolution. By utilizing the deformation characteristics of the graphene / PEDOT:PSS heterojunction, changes in tissue pressure will cause the sensor to deform, and then the tissue pressure will be converted into a measurable electrical signal through capacitance changes or piezoelectric effect, thus realizing tissue pressure monitoring; By directly contacting the tissue with the embedded thermistor, changes in skin tissue temperature will cause corresponding changes in the thermistor tissue, thereby enabling monitoring of tissue temperature during the injection process.

5. The intelligent injection method based on graphene heterojunction ring sensor according to claim 3, characterized in that: In step A1, the annular sensor synchronously collects capacitance change signals, mechanical stress and temperature data; When collecting capacitance change signals, a high-frequency carrier is used to excite the static electrode, and a lock-in amplifier is used to extract the amplitude and phase response of the dynamic electrode. The periodic change of capacitance is converted into displacement using a sub-pixel interpolation algorithm. When collecting mechanical stress, the sensor material deforms when subjected to stress, causing its electrical properties to change. Mechanical stress information is obtained by detecting this change. When collecting temperature data, the electrical properties of the sensor material change with temperature. Multimodal feature signals are extracted through wavelet packet decomposition and Hilbert-Huang transform, and the collected complex signals are processed to extract more representative and discriminative features, providing more accurate data for subsequent AI analysis.

6. The intelligent injection method based on graphene heterojunction ring sensor according to claim 3, characterized in that: In step B1, the AI algorithm module integrates a Bayesian optimization framework, collects force signals, displacement signals and temperature data during the injection process in real time, constructs a multi-dimensional feature space in combination with the patient's age, skin condition and drug characteristics, and establishes a nonlinear mapping relationship between the input signal and the puncture attribute based on Gaussian process regression; dynamically generates personalized injection parameters through the Bayesian optimization framework, iteratively optimizes the objective function using the expected improvement acquisition function, introduces historical injection data and clinical feedback to correct the parameter constraint space, and analyzes tissue elasticity changes through real-time force-displacement phase difference, and finally generates personalized injection parameters that adapt to individual differences and drug needs.

7. The intelligent injection method based on graphene heterojunction ring sensor according to claim 3, characterized in that: In step B2, the Transformer early warning model analyzes historical injection data and real-time signals through an attention mechanism to predict the risk of microbleeding and issue an alert in advance; The signals collected in real time by the ring sensor are an important source of real-time signals. Combined with historical data, they provide a comprehensive analysis basis for the early warning model.

8. The intelligent injection method based on graphene heterojunction ring sensor according to claim 3, characterized in that: In step C, the injection actuator supports dynamic adjustment of injection speed and pressure to adapt to subcutaneous, intramuscular and targeted injection scenarios; the tissue information fed back in real time by the annular sensor enables the injection actuator to dynamically adjust the injection parameters according to different tissue characteristics and injection requirements.

9. The intelligent injection method based on graphene heterojunction ring sensor according to claim 3, characterized in that: In step C1, the injection actuator adopts a fuzzy PID control algorithm to dynamically adjust the motor torque and step accuracy according to tissue density; the information such as tissue pressure fed back by the annular sensor can reflect the changes in tissue density and provide a control basis for the fuzzy PID control algorithm.

10. The intelligent injection method based on graphene heterojunction ring sensor according to claim 3, characterized in that: In step C3, the cloud medical database supports cross-device retrieval of patients' historical injection records for optimizing personalized injection strategies; the data collected and transmitted by the annular sensor enriches the content of the cloud medical database.

Images