Nutrition pump drip identification method, nutrition pump and electronic equipment

Abstract

The invention provides a nutrition pump drop identification method, a nutrition pump and electronic equipment, and relates to the technical field of drop identification, and the nutrition pump drop identification method comprises the steps: obtaining liquid drop detection signals collected by multiple infrared receiving channels; performing signal conditioning processing on each path of liquid drop detection signal to obtain multiple paths of pulse signals; combining the multiple paths of pulse signals into a single-path effective pulse sequence by using redundancy decision logic; and calculating the real-time dripping speed according to the pulse frequency of the single-path effective pulse sequence. According to the scheme, the liquid drop detection signals of the multiple infrared receiving channels are collected in parallel, or operation merging is carried out on the multiple pulse signals through the redundancy judgment logic, effective recognition of liquid drop events can still be maintained when part of channel signals fail, and therefore the fault-tolerant capability and robustness of the detection process are improved.

Description

Technical Field

[0001] This application relates to the field of drip recognition technology, and more specifically, to a method for recognizing drips from a nutrient pump, a nutrient pump, and an electronic device. Background Technology

[0002] Enteral nutrition pumps are essential devices for clinical nutritional support, used to deliver water, nutritional solutions, and nutritional emulsions of specific concentrations to patients via nasogastric tubes. They are suitable for patients who cannot eat orally or whose food intake is insufficient. To achieve precise infusion monitoring, modern nutrition pumps commonly employ infrared detection technology for drip counting. This involves emitting infrared light into the drip chamber of the infusion tubing via an infrared emitter, and a receiver tube collecting the light intensity changes as the droplets fall. This signal is then shaped into pulse signals by a processing circuit, and the processor calculates the pulse frequency and converts it into real-time drip rate, thereby achieving closed-loop control of the infusion rate and triggering alarms for abnormalities. This type of drip recognition technology has become a core functional module for nutrition pumps to achieve automated feeding and ensure infusion safety.

[0003] However, existing nutrition pump infusion detection technologies generally employ a single-channel infrared receiver structure, which has certain limitations in clinical applications. When the device is installed at an angle due to ICU bed layout, patient positioning restrictions, or mobility requirements, the droplet's trajectory deviates from the center of the detection area. The detection field of view of the single-channel receiver cannot completely cover the droplet's fall range, resulting in a sharp increase in the false negative rate and failing to meet the high-reliability monitoring requirements for bedside nutritional support in critically ill patients. This problem directly leads to difficulties in timely detection of infusion abnormalities, increasing nursing risks. Summary of the Invention

[0004] The purpose of this application is to provide a method for identifying nutrient pump drips, a nutrient pump, and an electronic device to solve the above-mentioned problems.

[0005] In a first aspect, embodiments of this application provide a method for identifying dripping from a nutrient pump. The method includes: acquiring droplet detection signals collected by multiple infrared receiving channels; wherein the multiple infrared receiving channels are arranged on the same detection section of the nutrient pump drip chamber; performing signal conditioning processing on each of the droplet detection signals to acquire multiple pulse signals; using redundant determination logic to merge the multiple pulse signals into a single valid pulse sequence; wherein the redundant determination logic is configured to determine that a droplet is valid when at least one of the pulse signals meets the droplet determination condition; and calculating the real-time dripping rate based on the pulse frequency of the single valid pulse sequence.

[0006] In the implementation of the above scheme, droplet detection signals from multiple infrared receiving channels are acquired in parallel, and redundant decision logic is used to combine the multiple pulse signals by OR operation. Even when some channel signals fail, the effective identification of droplet events can still be maintained, thereby improving the fault tolerance and robustness of the detection process. On the other hand, based on the collaborative processing of multiple signals on the same detection section, the droplet judgment conditions have a higher tolerance for the deviation of the droplet falling trajectory. The consistency of detection can still be maintained even when the nutrient pump is installed at an angle, which enhances the adaptability to different installation postures of the nutrient pump.

[0007] In one implementation of the first aspect, the method further includes: periodically acquiring ambient light intensity values; using the ambient light intensity values ​​to determine a preset correspondence with a judgment threshold, and determining a judgment threshold for generating a pulse signal.

[0008] In the implementation of the above scheme, by periodically acquiring the ambient light intensity value and dynamically adjusting the judgment threshold based on the preset correspondence, the signal conditioning processing can adaptively optimize the judgment parameters according to changes in ambient light conditions. This suppresses the detection sensitivity imbalance caused by the fixed threshold in weak or strong light environments and improves the adaptability of the method in day and night light fluctuation scenarios. On the other hand, by using the mapping mechanism between ambient light intensity value and judgment threshold, external light interference is incorporated into the feedforward control of the signal processing stage to avoid pulse signal misjudgment caused by sudden changes in light. This enhances the anti-interference ability of the above nutrient pump drip identification method in complex lighting environments such as clinics and operating rooms.

[0009] In one implementation of the first aspect, determining the determination threshold for generating a pulse signal by utilizing the preset correspondence between the ambient light intensity value and the determination threshold includes: when the ambient light intensity value is less than a first light intensity threshold, using a first fixed determination threshold; when the ambient light intensity value is greater than a second light intensity threshold, using a second fixed determination threshold; when the ambient light intensity value is not less than the first light intensity threshold and not greater than the second light intensity threshold, determining the determination threshold for generating a pulse signal by linear interpolation based on the ambient light intensity value, the first light intensity threshold, the second light intensity threshold, the first fixed determination threshold, and the second fixed determination threshold.

[0010] In the implementation of the above scheme, by dividing the ambient light intensity value into three intervals and adopting a differentiated threshold determination strategy, a fixed judgment threshold is used in the low light and high light areas to ensure threshold stability, while linear interpolation is used in the intermediate light intensity area to achieve a smooth transition of the threshold. This avoids threshold jumps at the segmentation points and improves the accuracy and smoothness of the judgment threshold when it changes continuously with the ambient light intensity. On the other hand, by using a processing method that combines the three-interval architecture with linear interpolation, the range of ambient light intensity adaptation is expanded while ensuring computational efficiency. This allows the signal judgment threshold to adaptively switch between the aforementioned fixed threshold interval and the dynamic interpolation interval, enhancing the detection consistency of the above-mentioned nutrient pump drip recognition method in scenarios with gradual changes in light intensity, such as day-night transitions and light switching.

[0011] In one implementation of the first aspect, the determination of the judgment threshold includes: determining the upper limit threshold by using the ambient light intensity value to determine a preset correspondence with the upper limit threshold of the judgment threshold; and determining the lower limit threshold of the judgment threshold based on the upper limit threshold and a preset hysteresis width; wherein the lower limit threshold is obtained by subtracting the preset hysteresis width from the upper limit threshold.

[0012] In the implementation of the above scheme, a lower threshold is generated by the linkage calculation of the upper threshold and the preset hysteresis width. A double threshold window with hysteresis characteristics is formed in the decision logic, so that the input signal must cross the full hysteresis width to trigger the state flip, thereby suppressing false triggering and missed triggering of pulse edges caused by signal jitter or noise interference, and improving the anti-interference capability of the signal conditioning and processing stage. On the other hand, the hysteresis width is used as a fixed parameter in the linkage calculation of the lower threshold to ensure that the hysteresis window of the comparator remains constant when the upper threshold is dynamically adjusted with the change of ambient light intensity. This keeps the decision hysteresis characteristics consistent under different lighting conditions, avoids the introduction of new decision uncertainties due to threshold adaptive adjustment, and enhances the robustness of the above nutrient pump drip identification method in the dynamic adjustment process.

[0013] In one implementation of the first aspect, the method further includes: detecting the fluctuation range of the ambient light intensity value within a preset time window; when the fluctuation range exceeds a preset fluctuation threshold, and the number of times the fluctuation range exceeds the preset fluctuation threshold is continuously detected reaches a preset number threshold, locking the current determination threshold and maintaining it for a preset locking duration.

[0014] In the implementation of the above scheme, by detecting the fluctuation range of ambient light intensity within a preset time window and using a threshold for the number of consecutive detections as a trigger condition, it is possible to identify sources of interference from sudden changes in light intensity. When a scene of transient light change is determined, threshold locking is initiated to prevent the judgment threshold from frequently changing due to short-term strong light disturbances, thereby maintaining the stability and continuity of the signal conditioning process. On the other hand, the threshold locking mechanism maintains the current judgment threshold unchanged within a preset locking time, preventing pulse signal misjudgment or missed judgment caused by sudden changes in light, thus enhancing the anti-interference capability of the above-mentioned nutrient pump drip identification method against transient interference sources such as surgical lights and monitor flashlights.

[0015] In one implementation of the first aspect, the periodic acquisition of ambient light intensity values ​​includes: synchronously sampling the ambient light intensity values ​​in response to an infrared emitting diode conduction signal.

[0016] In the implementation of the above scheme, ambient light intensity value sampling is performed synchronously in response to the infrared emitting tube conduction signal, so that the ambient light detection timing is strictly aligned with the infrared emission timing, thereby suppressing crosstalk of infrared emitted light to the light intensity detection element and improving the purity and accuracy of the acquired ambient light intensity value. On the other hand, the synchronous sampling mechanism ensures that each acquired ambient light intensity data corresponds to the same infrared emission state, eliminating data fluctuations caused by sampling phase deviation, providing input parameters with timing consistency for the dynamic adjustment of the judgment threshold, and enhancing the stability and reliability of the signal conditioning and processing process.

[0017] In one implementation of the first aspect, the method further includes: before performing infusion monitoring, acquiring a baseline value of the received signal and storing it as a calibration reference value; during infusion monitoring, monitoring the current baseline value of the received signal in real time and determining the attenuation degree of the current baseline value of the received signal compared to the calibration reference value; when the attenuation degree exceeds a preset attenuation threshold, determining a drip atomization state, and determining a compensation coefficient based on the attenuation degree; wherein the compensation coefficient is positively correlated with the attenuation degree; and generating an atomization compensation command based on the compensation coefficient to instruct the atomization compensation unit to enhance the infrared emission power of the infrared emitting tube.

[0018] In the implementation of the above scheme, by storing the baseline value of the received signal as a correction reference during the calibration stage and continuously monitoring the attenuation degree of the current baseline value relative to the correction reference during infusion monitoring, an automatic identification mechanism for the atomization state of the drip chamber is constructed. This avoids relying on manual visual inspection to determine atomization and improves the autonomous monitoring capability of the above-mentioned nutrient pump drip identification method in continuous infusion scenarios. On the other hand, based on the positive correlation between the attenuation degree and the compensation coefficient, atomization compensation instructions are dynamically generated. By enhancing the infrared emission power, the infrared light transmission attenuation caused by atomization is actively offset, so that the signal conditioning and processing stage can still maintain effective droplet detection sensitivity under atomization interference. This enhances the adaptability of the above-mentioned nutrient pump drip identification method to the interference of accumulated droplets attached to the inner wall of the drip chamber.

[0019] In one implementation of the first aspect, determining the compensation coefficient based on the attenuation degree includes: determining the corresponding attenuation level based on the attenuation interval to which the attenuation degree belongs; wherein, multiple attenuation intervals are formed by multiple preset attenuation thresholds, and the attenuation level corresponds one-to-one with the attenuation interval; determining the compensation coefficient according to a preset correspondence between the attenuation level and the compensation coefficient; wherein, the compensation coefficient is positively correlated with the attenuation level.

[0020] In the implementation of the above scheme, a hierarchical compensation architecture is constructed by mapping the attenuation degree to discrete attenuation intervals and determining the corresponding attenuation levels. This allows different levels of atomization to correspond to differentiated compensation coefficients, avoiding the problem of insufficient or excessive compensation by a single compensation coefficient during atomization development and improving the precision of the compensation strategy. On the other hand, by utilizing the preset correspondence between attenuation levels and compensation coefficients and their positive correlation characteristics, the compensation strategy is standardized and configurable. This enables the above-mentioned nutrient pump drip identification method to dynamically select the matching compensation intensity according to the severity of atomization, enhancing its adaptability to the gradual evolution of the dripping atomization degree.

[0021] In one implementation of the first aspect, the method further includes: when the attenuation level exceeds a preset severe fogging threshold, determining it as a severe fogging state and performing an alarm operation.

[0022] In the implementation of the above scheme, by setting a heavy atomization threshold and monitoring whether the attenuation exceeds the threshold, an alarm operation is triggered when the atomization exceeds the automatic compensation adjustment range. This combines the autonomous compensation mechanism with manual intervention prompts, avoiding the risk of detection failure that may result from relying solely on automatic compensation. On the other hand, the alarm operation actively outputs a heavy atomization status signal, providing timely status indications for manual cleaning and maintenance, thereby improving the safety margin of the above nutrient pump drip identification method during long-term continuous monitoring.

[0023] In one implementation of the first aspect, the step of calculating the real-time drip rate based on the pulse frequency of the single-channel effective pulse sequence includes: determining the number of pulses used to calculate the real-time drip rate based on the drip rate interval in which the pulse frequency of the single-channel effective pulse sequence falls; and calculating the real-time drip rate based on the latest pulse signal generated in the single-channel effective pulse sequence.

[0024] In the implementation of the above scheme, the number of pulses used to calculate the real-time drip rate is determined based on the drip rate range in which the pulse frequency is located, so that the number of sampling points is dynamically matched with the current drip rate. In the low-speed range, the number of pulses is reduced to shorten the calculation response time, and in the high-speed range, the number of pulses is increased to smooth random fluctuations, thereby balancing the real-time performance and accuracy of drip rate detection. On the other hand, the latest generated pulse signal is used for real-time drip rate calculation, and a sliding time window mechanism is adopted to ensure that the calculation benchmark is always based on the latest sampled data, avoiding the drip rate calculation delay caused by the lag of historical data, and improving the tracking and response capability to sudden changes in drip rate.

[0025] In one implementation of the first aspect, the method further includes: obtaining a preset target drip rate; calculating the deviation between the real-time drip rate and the preset target drip rate; and when the absolute value of the deviation exceeds a preset allowable deviation, generating a motor speed control command based on the deviation to make the real-time drip rate approach the target drip rate.

[0026] In the implementation of the above scheme, by acquiring the preset target drip rate and calculating its deviation from the real-time drip rate, a motor speed adjustment command is generated when the deviation exceeds the allowable range, forming a negative feedback closed-loop control loop. This enables the real-time drip rate to automatically track the preset target, thereby improving the accuracy and stability of the drip rate control. On the other hand, the absolute value of the deviation is used as the speed adjustment trigger condition, and adjustment is only initiated when the deviation reaches the threshold, avoiding over-response to small deviations. At the same time, a speed adjustment command is generated based on the deviation, so that the speed adjustment intensity matches the deviation degree, realizing adaptive dynamic adjustment of the infusion rate.

[0027] Secondly, embodiments of this application provide a nutrient pump, comprising: An infrared detection unit is used to emit infrared light into the drip chamber and receive infrared light signals transmitted through the drip chamber, and output multiple droplet detection signals; wherein, the infrared detection unit includes multiple infrared emitting channels and multiple infrared receiving channels, and the multiple infrared receiving channels are arranged on the same detection section of the drip chamber; The main control unit is used to execute the methods provided by the first aspect or any possible implementation thereof. A light intensity detection unit is used to acquire ambient light intensity values ​​and transmit them to the main control unit; A fogging compensation unit is used to adjust the infrared emission power in response to the fogging compensation command of the main control unit; The motor drive unit is used to respond to the motor speed adjustment command of the main control unit and drive the infusion actuator to adjust the infusion speed.

[0028] Thirdly, embodiments of this application provide an electronic device, including: a processor, a memory, and a communication bus, wherein the processor and the memory communicate with each other through the communication bus; the memory stores computer program instructions that can be executed by the processor, and the computer program instructions are read and executed by the processor to perform the method provided in the first aspect or any possible implementation of the first aspect.

[0029] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer program instructions, which, when read and executed by a processor, perform the method provided in the first aspect or any possible implementation thereof.

[0030] Fifthly, embodiments of this application provide a computer program product, the computer program product including a computer program, which, when executed by a processor, implements the method provided by the first aspect or any possible implementation of the first aspect.

[0031] Other features and advantages of this application will be set forth in the following description and will be apparent in part from the description or may be learned by practicing embodiments of this application. The objectives and other advantages of this application may be realized and obtained by means of the structures particularly pointed out in the written description, claims and drawings. Attached Figure Description

[0032] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0033] Figure 1 This is a schematic diagram of the structure of a nutrient pump provided in an embodiment of this application; Figure 2 A flowchart illustrating the nutrient pump drip identification method provided in this application embodiment; Figure 3 A schematic diagram illustrating the deployment of multiple infrared receiving channels in an application scenario provided in this application embodiment; Figure 4 A schematic diagram of the detection section where the multiple receiving channels are located, as provided in the embodiments of this application; Figure 5 This is a schematic diagram of the detection cross-section when the dripping bucket is tilted, provided in an embodiment of this application. Figure 6 A flowchart illustrating the threshold adaptive adjustment mechanism in a certain application scenario provided in this application embodiment; Figure 7 This is a schematic diagram of a dripping chamber subjected to atomization interference, provided in an embodiment of this application; wherein, Figure 7 (a) is a schematic diagram of a drip chamber at an atomization level of 30% to 50%; Figure 7 (b) is a schematic diagram of a drip chamber at an atomization level of 50% to 70%. Figure 7 (c) in the diagram is a schematic of a dripping chamber with an atomization level of 70% or higher; Figure 8 A flowchart illustrating a fogging interference compensation mechanism in a specific application scenario provided in this application embodiment; Figure 9 A schematic diagram of the architecture of a nutrient pump provided in an embodiment of this application; Figure 10 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0034] The technical solutions of the embodiments of this application will now be described with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of this application, and are therefore merely examples and should not be used to limit the scope of protection of this application.

[0035] Please see Figure 1A nutrition pump is a medical device used to precisely control the infusion of enteral nutrition solutions. It uses a mechanical drive mechanism to apply controllable pressure or create volumetric displacement in a dedicated infusion tubing, continuously and stably pumping nutritional preparations to the patient's gastrointestinal tract at a preset flow rate. A nutrition pump typically consists of a power system (stepper motor or peristaltic pump head), a main control unit (microcontroller and embedded software), a multimodal sensing module (drip sensor, pressure sensor, temperature sensor), a human-machine interface unit (display screen and buttons), and an audible and visual alarm system. It complies with medical electrical equipment safety standards and specific technical requirements for enteral nutrition infusion devices. The procedure for using the nutrition pump includes: First, install the disposable enteral nutrition tubing, select the appropriate tubing model (e.g., 1000mL capacity), fix the drip chamber in the nutrition pump slot, and ensure that the center of the drip chamber is aligned with the infrared detection module; then, press and hold the power button to start the device, and the device will perform a self-test process. After verifying that all modules are functioning normally, the indicator light status will change; set the feeding speed, total feeding volume, and heating temperature through the human-machine interface; after starting the venting mode to drive the tubing to vent until there are no air bubbles, connect the enteral feeder to the patient's nasogastric tube; enter the feeding monitoring stage, which monitors the drip in real time, calculates and displays the drip rate. If the drip rate deviates from the set value by a certain amount, the motor speed will be adjusted immediately and an alarm will be triggered; when the total feeding volume is reached or manually stopped, the device will automatically stop. After pressing and holding the power button to turn off the device, remove the infusion tubing consumables to complete the entire feeding cycle.

[0036] Infrared drip detection technology in enteral nutrition pumps typically employs a single-channel infrared pair architecture. An infrared emitting tube emits infrared light of a specific wavelength into the drip chamber of the infusion tubing, while a receiving tube collects the light intensity signal transmitted through the drip chamber. When a droplet obstructs the light path during its descent, the intensity of the electrical signal output from the receiving tube changes accordingly. After processing by signal conditioning circuits such as amplification and comparison, a corresponding pulse signal is generated. The processor calculates the real-time drip rate by statistically analyzing the pulse frequency. Under ideal conditions—vertical installation of the drip chamber, stable ambient light, and a clean inner wall of the drip chamber—this architecture can achieve basic drip counting functionality. The signal conditioning circuit often uses a fixed-threshold hysteresis comparator to suppress small-amplitude noise interference and ensure the stability of the pulse edge. However, the single-channel infrared pair architecture has a limited detection field of view, facing multiple reliability challenges in clinical applications. When the device is installed at an angle due to ICU bed layout, patient positioning requirements, or mobile care, the droplet's trajectory deviates from the center of the detection channel, resulting in significant attenuation of the signal amplitude output from the receiving tube. The fixed-threshold comparator struggles to effectively identify the pulse edge, leading to an increased rate of missed droplet detection.

[0037] In view of this, this application provides a method for identifying droplet events in a nutrient pump. This method acquires droplet detection signals from multiple infrared receiving channels in parallel and uses redundant decision logic to perform OR operations on the multiple pulse signals to merge them. Even when some channel signals fail, it can still maintain effective identification of droplet events, thereby improving the fault tolerance and robustness of the detection process. On the other hand, based on the collaborative processing of multiple signals on the same detection section, the droplet determination conditions have a higher tolerance for the deviation of the droplet falling trajectory. It can still maintain the consistency of detection even when the nutrient pump is installed at an angle, thus enhancing the adaptability to different installation postures of the nutrient pump.

[0038] Please see Figure 2 The diagram illustrates a flowchart of the nutrient pump drip identification method provided in this application embodiment. The nutrient pump drip identification method provided in this application embodiment can be applied to electronic devices, which may include physical devices such as servers, PCs, tablets, or smartphones, or virtual devices such as virtual machines or containers. The electronic device can be a single device, a combination of multiple devices, or a cluster of a large number of devices. The above-mentioned nutrient pump drip identification method may include: Step S110: Acquire droplet detection signals collected by multiple infrared receiving channels respectively; wherein, the multiple infrared receiving channels are arranged on the same detection section of the nutrient pump drip chamber.

[0039] The aforementioned detection cross-section refers to a geometric reference plane perpendicular to the central axis of the nutrient pump drip chamber, used to calibrate the installation height of the multiple infrared receiving channels. When a droplet crosses this cross-section during its descent, it causes infrared light path obstruction, thereby triggering a detection event. Configuring multiple infrared receiving channels on the same detection cross-section ensures that the droplet detection signals collected by each channel originate from the same spatial location along the droplet's trajectory, providing a spatially consistent signal input basis for subsequent redundant judgment logic. The aforementioned multiple infrared receiving channels refer to a set of parallel detection paths composed of multiple independent infrared receiving tubes and their associated signal conditioning circuits. Each receiving channel contains a complete photoelectric conversion unit, capable of independently receiving the infrared light signal transmitted through the drip chamber and converting it into an electrical signal, which is then amplified, filtered, compared, and other signal conditioning stages to output a pulse signal. Each receiving channel is electrically independent, outputting a pulse sequence characterizing changes in infrared light intensity, providing multiple parallel droplet judgment bases for redundant judgment logic. This multi-channel architecture expands the detection field of view coverage through multiple spatially distributed receiving units, enabling a single droplet event to be effectively captured by at least one channel.

[0040] by Figure 3Taking the illustrated scenario as an example, in this scenario, two infrared emitting tubes 210 and three infrared receiving tubes 220 are set up. In use, the two infrared emitting tubes 210 are symmetrically distributed on both sides of the dripping bucket, forming a fan-shaped emission angle, projecting an infrared light field of the same wavelength into the same detection area inside the dripping bucket. The three infrared receiving tubes 220 are arranged in the same detection section on opposite sides of the dripping bucket, with the photosensitive surface of each receiving tube facing the infrared emission direction, independently receiving the infrared light signal transmitted through the dripping bucket and converting it into an electrical signal output. This setup can expand the spatial coverage of the infrared light field through multiple emission channels. Combined with the redundant receiving capability of multiple receiving channels, it ensures that the droplet can be effectively captured by at least one receiving tube regardless of whether it is at the center or offset position of the detection section during its fall. The droplet detection signals output by each receiving channel are processed by signal conditioning to obtain multiple pulse signals. A valid pulse from any channel is determined to be a droplet event, thus constructing a detection mechanism with spatial and logical redundancy. Furthermore, as... Figure 4 As shown, one of the two infrared emitting tubes 210 can be tilted, so that the infrared light field forms an axially asymmetrical distribution within the cross-section of the droplet. When the detected droplet deviates from the ideal falling trajectory due to equipment tilting or pipeline disturbance, the tilted emitting tube provides lateral detection coverage, ensuring that droplets that have deviated by a certain range can still be effectively irradiated.

[0041] like Figure 4 As shown, the detection section can be set 5-10 mm above the liquid surface in the dropping chamber. This position is selected based on a comprehensive consideration of droplet motion dynamics and detection reliability: on the one hand, the detection section should be higher than the stable height of the liquid surface in the dropping chamber to avoid interference from liquid surface fluctuations, rising bubbles, or liquid surface reflection on the infrared light path; on the other hand, this position should be located in the free fall phase of the droplet but not touching the liquid surface, ensuring that the droplet has completed the infrared blocking process before impacting the liquid, so that the received signal has complete edge characteristics. Figure 5 As shown, when the droplet is installed at a 25° angle, the setting position of the detection section can still ensure that the droplet's falling trajectory remains within the combined field of view of the multiple infrared receiving channels, thus avoiding detection failure due to trajectory deviation.

[0042] Step S120: Perform signal conditioning processing on each droplet detection signal to obtain multiple pulse signals.

[0043] The aforementioned signal conditioning process refers to a series of operations that optimize signal quality and transform the shape of the raw droplet detection signal output from the infrared receiving channel. Its purpose is to convert the analog signal, which contains noise, amplitude attenuation, and edge jitter, into a digital pulse signal with clear edges, stable amplitude, and meeting the requirements of subsequent processing logic. The signal conditioning process can be implemented through either hardware circuits or software algorithms. In the circuit implementation, signal conditioning typically includes a gain amplification stage composed of operational amplifiers to increase signal amplitude, passive or active filter networks to suppress high-frequency noise, and a hysteresis comparator to perform threshold decision and edge shaping, ultimately outputting a pulse sequence. In the algorithm implementation, signal conditioning can obtain discrete signal samples through an analog-to-digital converter, eliminate random noise through digital filtering algorithms (such as moving average or median filtering), and then complete edge detection and pulse generation through programmable hysteresis comparator logic. Regardless of the implementation method, the core function of signal conditioning is to filter out interference, enhance the effective signal characteristics, and generate standardized pulse signals, providing consistent and reliable multi-path pulse inputs for subsequent redundant decision logic.

[0044] Step S130: Using redundant decision logic, multiple pulse signals are merged into a single valid pulse sequence; wherein, the redundant decision logic is configured to determine that the droplet is valid when at least one pulse signal meets the droplet determination condition.

[0045] The aforementioned redundancy decision logic refers to a fault-tolerant decision mechanism. Its function is defined as outputting a valid droplet pulse when at least one of the multiple parallel input pulse signals meets a preset droplet determination condition, without requiring all channels to respond simultaneously. The redundancy decision logic can be implemented using OR gates in hardware circuits or logical OR operations in software programs. Its core function is to decouple the failure risks of multiple independent channels, ensuring that failures in a single channel due to droplet trajectory deviation, device aging, or transient interference do not propagate to the final decision result. The aforementioned droplet determination condition refers to a set of technical criteria used to identify whether a pulse signal is triggered by a real droplet falling event. This condition is constructed based on the pulse characteristic parameters after signal conditioning.

[0046] The optional implementation of step S130 above, which merges multiple pulse signals into a single effective pulse sequence, includes: the pulse signals generated by each infrared receiving channel after signal conditioning are connected in parallel to the input of the redundant judgment logic. The judgment logic monitors and judges each input signal in real time. When any input signal has a pulse edge that meets the droplet judgment condition, the judgment logic immediately generates a corresponding standardized pulse at the output. The timing of this output pulse is synchronized with the original pulse of the trigger channel and is independent of the state of other untriggered channels. Thus, the multiple parallel pulse streams are integrated into a single serial effective pulse sequence for frequency statistics in the subsequent drop rate calculation stage.

[0047] Step S140: Calculate the real-time drip rate based on the pulse frequency of the single effective pulse sequence.

[0048] The above step S140 can calculate the real-time drip rate in one of the following ways: The first method is the fixed time window counting method. In this method, the number of pulses contained in a single effective pulse sequence can be counted within a preset fixed time window. The number of pulses is divided by the window duration to obtain the instantaneous frequency, which is then converted into the real-time drip rate.

[0049] The second method is the fixed pulse count timing method. In this method, a fixed number of pulses can be preset. The time interval between the latest pulses of that number generated in a single effective pulse sequence is measured. The instantaneous frequency is obtained by dividing the number of pulses by the measured time interval, and then converted into the real-time drop rate. In addition, to balance detection accuracy and response speed across different drop rate ranges, an adaptive variable window length strategy can be adopted. This dynamically adjusts the number of pulses or the length of the time window used for calculation based on the drop rate range in which the current pulse frequency falls, resulting in faster response at low speeds and more accurate calculations at high speeds.

[0050] It is understandable that in the detection scheme of enteral nutrition pump dripping without shielding, ambient light, as a stray light component, inevitably couples to the infrared receiving channel and linearly superimposes with the target infrared light signal, causing the received signal baseline to drift with the ambient light intensity. When the ambient light intensity increases significantly, the signal baseline shifts upward. If the judgment threshold remains fixed, the amplitude of the negative pulse caused by droplet obstruction is relatively reduced, which may not be able to cross the threshold to effectively trigger the comparator toggling, leading to an increase in the false negative rate. Conversely, when the ambient light intensity decreases, the signal baseline shifts downward, and the noise component becomes relatively prominent. The fixed threshold may be lower than the noise peak, causing false triggers and alarms. In view of this, the embodiments of this application provide the following solution: Optionally, the above-mentioned nutrient pump drip identification method may further include: periodically acquiring ambient light intensity values; using the ambient light intensity values ​​to determine a preset correspondence with a judgment threshold, and determining a judgment threshold used to generate a pulse signal.

[0051] The aforementioned ambient light intensity value can be obtained through the collaborative operation of a photosensitive detection element and an analog-to-digital converter circuit. The photosensitive detection element can be deployed near the infrared detection unit, and its electrical characteristics change synchronously with changes in ambient light intensity, outputting an analog electrical signal related to the ambient light intensity. This analog signal is sampled and quantized by the analog-to-digital converter circuit, converting it into a digital form of the ambient light intensity value. The acquisition period for the ambient light intensity value needs to be configured based on the combined requirements of the infrared detection period and the threshold adjustment response speed, typically set to an integer multiple of the infrared detection period to ensure that the ambient light intensity data update frequency is coordinated with the droplet detection frequency. Within the acquisition period, the number of samplings of the ambient light intensity value should meet the timeliness requirement of the threshold adjustment frequency's response to changes in illumination, avoiding either sparse acquisition leading to a threshold lagging behind ambient light changes or excessively dense acquisition increasing the processing load. Generally, the acquisition period can be configured as a fixed time interval mode or a trigger-counting mode based on droplet detection. In fixed time interval mode, the acquisition cycle is preset to a millisecond-level time window based on the clinical illumination fluctuation characteristics, enabling threshold adjustment to adapt to routine illumination changes. In trigger counting mode, the acquisition cycle is based on droplet pulse counting, performing ambient light intensity sampling once for every fixed number of droplet events detected, achieving adaptive matching between the acquisition cycle and the infusion rate. The acquisition cycle setting also needs to consider the sampling rate and computing resources of the hardware analog-to-digital conversion, ensuring that the entire process of sampling, conversion, and threshold calculation is completed within the cycle, avoiding data loss or control failure due to cycle timeout.

[0052] The above scheme acquires ambient light intensity values ​​periodically and dynamically adjusts the judgment threshold based on a preset correspondence. This enables the signal conditioning process to adaptively optimize the judgment parameters as ambient light conditions change, thereby suppressing the detection sensitivity imbalance caused by fixed thresholds in low or high light environments and improving the adaptability of the method in scenarios with fluctuating day and night light. On the other hand, by utilizing the mapping mechanism between ambient light intensity values ​​and judgment thresholds, external light interference is incorporated into the feedforward control of the signal processing stage, avoiding misjudgment of pulse signals caused by sudden changes in light intensity. This enhances the anti-interference capability of the above-mentioned nutrient pump drip identification method in complex lighting environments such as clinics and operating rooms.

[0053] The preset correspondence between the aforementioned ambient light intensity value and the judgment threshold can be achieved using one of the following methods: The first method is a segmented fixed threshold mapping method. This method divides the effective range of ambient light intensity into several discrete intensity intervals, each corresponding to a pre-defined fixed judgment threshold. For example, for low-light environments, a light intensity threshold is set. If the real-time sampled ambient light intensity value is lower than this threshold, the first fixed judgment threshold is directly applied. For high-light environments, another light intensity threshold is set. If the ambient light intensity value is higher than this threshold, the second fixed judgment threshold is applied. This mapping method can be implemented through an interval lookup table mechanism. By comparing the sampled ambient light intensity value with the preset boundary threshold, the corresponding interval is located, and the corresponding threshold parameter is extracted, completing the fast mapping from ambient light intensity to the judgment threshold. This method has low computational overhead and fast response speed.

[0054] The second method is linear interpolation. This method establishes a linear function relationship between two adjacent calibrated light intensity points. It uses a first light intensity threshold and its corresponding first fixed judgment threshold as the starting point, and a second light intensity threshold and its corresponding second fixed judgment threshold as the ending point, forming an interpolation interval. When the ambient light intensity value falls within this interval, a continuous threshold parameter between the first and second fixed judgment thresholds is calculated linearly based on the relative position of the ambient light intensity value within the interval. This method generates a smooth threshold that continuously changes with the light intensity through linear calculation, avoiding the abrupt changes at the boundaries of piecewise mapping.

[0055] The third method is a lookup table mapping approach. A pre-calibrated two-dimensional mapping table of ambient light intensity values ​​and threshold values ​​can be stored in the electronic device executing the aforementioned nutrient pump drip identification method. This mapping table describes the correspondence between light intensity and threshold values ​​using discrete sampling points. During runtime, the periodically acquired ambient light intensity values ​​are used as lookup indexes. A linear or binary search algorithm is used to locate the closest sampling point in the mapping table, and the corresponding threshold value is extracted for signal conditioning. This method allows for arbitrarily complex nonlinear relationships to describe the light intensity-threshold mapping. It adapts to different types of infrared devices or varying clinical environments simply by updating the mapping table, offering high flexibility and configurability.

[0056] The fourth approach involves building a machine learning model. This method collects a large amount of paired data between ambient light intensity values ​​and the optimal judgment threshold during offline training. Supervised learning algorithms are used to train a regression model or a shallow neural network, enabling the model to learn the implicit nonlinear mapping between light intensity and the threshold. During online operation, periodically acquired ambient light intensity values ​​are input into the deployed machine learning model. Based on the trained weight parameters, the model infers and outputs a judgment threshold adapted to the current light intensity conditions in real time. This approach possesses self-learning and generalization capabilities, can approximate arbitrarily complex mapping relationships, and can continuously optimize model parameters through incremental training to adapt to mapping drift caused by aging infrared devices or migration in clinical environments.

[0057] The fifth method: a combination of fixed threshold and linear interpolation; Optionally, the above-mentioned determination of the determination threshold for generating a pulse signal by using the ambient light intensity value to determine the preset correspondence with the determination threshold includes: when the ambient light intensity value is less than the first light intensity threshold, using the first fixed determination threshold; when the ambient light intensity value is greater than the second light intensity threshold, using the second fixed determination threshold; when the ambient light intensity value is not less than the first light intensity threshold and not greater than the second light intensity threshold, determining the determination threshold for generating a pulse signal by linear interpolation based on the ambient light intensity value, the first light intensity threshold, the second light intensity threshold, the first fixed determination threshold, and the second fixed determination threshold.

[0058] The aforementioned first and second light intensity thresholds can be determined based on the clinical ambient lighting characteristics and the linear response range of infrared detection. For example, the first light intensity threshold, serving as the boundary between the low-light and intermediate-light-intensity zones, is typically set at the lighting boundary between low-light nighttime scenarios and standard ward lighting scenarios. The selection of the first light intensity threshold must ensure that when the ambient light intensity is below this threshold, the infrared receiver operates at the beginning of the linear region. Using a fixed first threshold in this case can avoid noise-induced false triggering due to oversensitivity. The second light intensity threshold, serving as the boundary between the intermediate-light-intensity and high-light zones, is typically set at the lighting boundary between standard treatment room lighting and high-brightness operating room lighting scenarios. Its selection must ensure that when the ambient light intensity is above this threshold, the received signal approaches the saturation region. Using a fixed second threshold in this case can prevent missed detections due to insufficient signal amplitude. The intermediate range between the first and second light intensity thresholds corresponds to standard treatment scenarios such as day wards and ICU rooms. Within this range, the infrared receiver output signal has sufficient dynamic range and signal-to-noise ratio, making it suitable for smooth adjustment of the threshold using linear interpolation to maintain consistent detection sensitivity. The rules for dividing the first and second light intensity thresholds can be determined by combining clinical lighting environment surveys and infrared device characteristic calibration to ensure that the boundary values ​​of each interval can both cover the light intensity distribution of the actual scene and match the linear range of the electrical parameters of the receiving device.

[0059] The aforementioned linear interpolation uses a first light intensity threshold as the interpolation starting point, corresponding to a first fixed judgment threshold; and a second light intensity threshold as the interpolation ending point, corresponding to a second fixed judgment threshold, thereby establishing a linear mapping relationship between light intensity and threshold. After acquiring the real-time ambient light intensity value, the normalized position coefficient of this value relative to the starting point within the interpolation interval is calculated, which is the difference between the ambient light intensity value and the first light intensity threshold divided by the interval width (the difference between the second light intensity threshold and the first light intensity threshold). Subsequently, this position coefficient is multiplied by the threshold interval width (the difference between the second fixed judgment threshold and the first fixed judgment threshold) to obtain the threshold increment. This threshold increment is then added to the first fixed judgment threshold to finally obtain a dynamic judgment threshold adapted to the current ambient light intensity value.

[0060] It is understandable that in the weak light range where the ambient light intensity is below the first light intensity threshold, the receiver is at the front end of the linear response region, with weak signal amplitude and low signal-to-noise ratio. Using the first fixed decision threshold in this case avoids the additional noise sensitivity introduced by frequent threshold fine-tuning, ensuring detection stability. In the strong light range where the ambient light intensity is above the second light intensity threshold, the received signal approaches saturation, and the signal dynamic range is compressed. Using the second fixed decision threshold prevents saturation distortion and misjudgment caused by excessive pursuit of sensitivity. In the intermediate light intensity range between the two thresholds, the infrared receiver operates within a dynamic range with good linearity, and the signal amplitude changes monotonically with the ambient light intensity. Using linear interpolation allows the decision threshold to transition smoothly with light intensity, eliminating threshold jumps at segmentation points, maintaining consistent detection sensitivity, and reducing computational overhead compared to complex nonlinear mapping, thus balancing real-time performance and accuracy. This combined approach utilizes the robustness of the fixed threshold in the boundary region and leverages the smoothness and efficiency of linear interpolation in the intermediate region, forming a threshold control strategy that balances illumination adaptability and algorithm feasibility.

[0061] The above scheme divides the ambient light intensity value into three intervals and adopts a differentiated threshold determination strategy. In the low light and high light areas, a fixed judgment threshold is used to ensure threshold stability, while linear interpolation is used in the intermediate light intensity area to achieve a smooth transition of the threshold. This avoids threshold jumps at the segmentation points and improves the accuracy and smoothness of the judgment threshold when the ambient light intensity changes continuously. On the other hand, by using a processing method that combines the three-interval architecture with linear interpolation, the range of ambient light intensity adaptation is expanded while ensuring computational efficiency. This allows the signal judgment threshold to adaptively switch between the aforementioned fixed threshold interval and the dynamic interpolation interval, enhancing the detection consistency of the above-mentioned nutrient pump drip recognition method in scenarios with gradual changes in light intensity, such as day-night transitions and light switching.

[0062] Optionally, the above-mentioned method for determining the threshold includes: determining the upper limit threshold by using the ambient light intensity value to determine the preset correspondence between the upper limit threshold and the upper limit threshold in the judgment threshold; and determining the lower limit threshold of the judgment threshold based on the upper limit threshold and the preset hysteresis width; wherein the lower limit threshold is obtained by subtracting the preset hysteresis width from the upper limit threshold.

[0063] The aforementioned upper and lower thresholds together define a decision window with hysteresis characteristics, forming upper and lower boundary constraints within the signal amplitude space. When the input signal crosses the upper threshold from a low level upwards, the decision logic output state flips to an active level; when the signal crosses the lower threshold from a high level downwards, the output state flips back to an inactive level. Since the lower threshold is obtained by subtracting a preset hysteresis width from the upper threshold, and the hysteresis width is a fixed positive value, the upper threshold is always higher than the lower threshold. This creates an amplitude difference between the flip points in the rising and falling directions of the signal, which is the width of the hysteresis window. The hysteresis window ensures that the input signal must cross the full hysteresis width in the opposite direction to achieve state switching, thereby suppressing small jitters or noise interference near the threshold, avoiding repeated triggering of the output pulse edge, and improving the disturbance rejection stability of the signal conditioning process.

[0064] The aforementioned hysteresis width can be set to be greater than the peak amplitude of residual random noise in the signal conditioning process, ensuring that noise fluctuations are insufficient to trigger threshold cut-off, thereby avoiding the generation of false pulses. Simultaneously, the hysteresis width can also be set to be less than the minimum change in signal amplitude when the droplet blocks the infrared optical path, preventing effective pulses from being suppressed because their amplitude does not fully cross the window. In engineering implementation, the baseline noise statistical distribution of the received signal in a droplet-free state can be analyzed to extract the peak-to-peak noise level as a benchmark. An initial value for the hysteresis width is then formed by superimposing a certain margin. Subsequently, in a dynamic test environment simulating droplet fall, the false trigger rate and false negative rate of pulses under different hysteresis widths are statistically analyzed, and the value that minimizes the combined risk of false alarms and false negatives is selected as the final parameter.

[0065] Taking a specific application scenario as an example, in a hospital clinical environment, enteral nutrition pumps need to operate stably under various lighting conditions. At night, the ambient illuminance is approximately 10 lux, in general wards it is approximately 100 lux, in examination rooms and ICUs it is approximately 300 lux, and in operating rooms it can reach 750 lux. Furthermore, the actual illuminance may fluctuate within the range of 0-800 lux. The segmented threshold settings are shown in Table 1. Table 1 Segmented Threshold Setting Scheme

[0066] In the above scenario, the ambient light intensity value is obtained through a photosensitive detection element and converted into a digital quantity by an analog-to-digital converter circuit. The conversion relationship can be expressed by the following formula: L=(AdcVal×3.3 / 4095-1.25)×2000, where L (lux) is the light intensity illuminance detection value, and AdcVal is the ADC sampling value, which ranges from 0 to 4095. The threshold values ​​are calculated in segments based on the ambient light intensity range, with a fixed hysteresis width of 0.1 volts introduced to suppress signal jitter. When the ambient light intensity is in the low-light range of 0-100 lux, the upper threshold is 1.0V, and the lower threshold is the upper threshold minus the hysteresis width. When the ambient light intensity is in the medium-light range of 100-350 lux, the upper threshold is calculated using a linear interpolation formula, and the lower threshold is also the upper threshold minus the hysteresis width. When the ambient light intensity is in the high-light range of 350-800 lux, the upper threshold is fixed at 1.8V, and the lower threshold is fixed at 1.6V. This segmented design ensures that the threshold values ​​maintain compatibility with the signal baseline under different lighting conditions. Figure 6 As shown, in the above application scenario, the threshold adaptive adjustment mechanism is executed cyclically according to a fixed period. The process includes: first, preprocessing the droplet detection signals collected by multiple infrared receiving channels, eliminating high-frequency noise in the original signal and enhancing the effective feature amplitude through filtering and amplification to provide clean input for subsequent decision-making; then, synchronously sampling the ambient light intensity value in response to the infrared emitting tube's conduction signal to ensure strict alignment between the sampling timing and the emission state to suppress crosstalk; and performing interval determination based on the sampled light intensity value. When the ambient light intensity is in the weak light range of 0-100 lux, the first fixed determination threshold is directly used as the output; when the light intensity is in the weak light range of 0-100 lux, the first fixed determination threshold is directly used as the output. When the light intensity is in the middle range of 100-350 lux, the linear interpolation calculation module is activated to generate a continuous threshold parameter between the first and second fixed thresholds in real time based on the relative position of the current light intensity within the range. When the light intensity is higher than 350 lux in the strong light range, the second fixed judgment threshold is used. After the threshold is initially determined, the hysteresis processing stage is entered, where the upper threshold is subtracted from the preset hysteresis width to generate the lower threshold, forming a dual threshold window with anti-jitter capability. The final output judgment threshold parameter is loaded onto the reference terminal of the hysteresis comparator of the signal conditioning circuit. After the threshold update for this cycle is completed, the above process is repeated in the next acquisition cycle.

[0067] The aforementioned method generates a lower threshold by linking an upper threshold with a preset hysteresis width. This forms a dual-threshold window with hysteresis characteristics in the decision logic, ensuring that the input signal must cross the full hysteresis width to trigger a state flip. This suppresses false triggering and missed triggering of pulse edges caused by signal jitter or noise interference, improving the anti-interference capability of the signal conditioning and processing stage. On the other hand, by using the hysteresis width as a fixed parameter in the linkage calculation of the lower threshold, the comparator's hysteresis window remains constant when the upper threshold is dynamically adjusted with changes in ambient light intensity. This ensures that the decision hysteresis characteristics remain consistent under different lighting conditions, avoiding the introduction of new decision uncertainties due to adaptive threshold adjustment, and enhancing the robustness of the above-mentioned nutrient pump drip identification method during dynamic adjustment.

[0068] Optionally, the above-mentioned nutrient pump drip identification method may further include: detecting the fluctuation range of ambient light intensity within a preset time window; when the fluctuation range exceeds a preset fluctuation threshold, and the number of consecutive detections of fluctuation range exceeding the preset fluctuation threshold reaches a preset number threshold, locking the current judgment threshold and maintaining it for a preset lock duration. For example, this implementation sets the preset fluctuation threshold to 500 lux / s, the preset number threshold to 5 times, and the preset lock duration to 500 milliseconds. During operation, the fluctuation range of ambient light intensity within a 10-millisecond time window is continuously detected. When a fluctuation range exceeding 500 lux / s is detected, an over-limit counter is activated; if the subsequent five consecutive samples all detect fluctuation ranges exceeding the threshold, a judgment threshold locking mechanism is triggered, locking the current judgment threshold parameter and maintaining it unchanged for 500 milliseconds. This parameter configuration is based on statistical analysis of clinical light interference characteristics: a fluctuation threshold of 500 lux per second is higher than the normal light intensity change rate caused by daily nursing activities, effectively filtering out minor disturbances such as personnel movement; the requirement of 5 consecutive detections ensures that locking is only activated for strong transient interferences lasting longer than 50 milliseconds (such as the on / off switching of surgical lights), avoiding accidental triggering by single spikes; the 500-millisecond lockout duration covers the typical duration of the surgical light transition process, ensuring threshold stability during interference while avoiding excessively long lockouts that could affect the response to normal light changes. This implementation achieves a balance between effective isolation of transient interference and timely adaptive threshold adjustment through parameter optimization.

[0069] It is understandable that in the clinical application environment of enteral nutrition pumps, in addition to steady-state light changes, there are also transient light source interferences, such as the switching on and off of surgical lights, the triggering of monitor flashlights, or sudden changes in local light intensity caused by personnel movement. These transient light fluctuations are characterized by short duration and drastic amplitude changes. If the judgment threshold continuously adjusts adaptively to follow the transient light intensity, the decision benchmark in the signal conditioning process will frequently jump. At the moment of light intensity change, the threshold adjustment lags behind the signal baseline change, easily leading to misjudgment or missed judgment at the pulse edge. The above solution introduces a threshold locking mechanism. When a transient light intensity fluctuation is detected to exceed a preset condition, the adaptive update of the judgment threshold is temporarily suspended, allowing the threshold to remain stable within the locked duration. This isolates the impact of transient interference on the signal decision process, ensuring the continuity and reliability of droplet detection in scenarios with sudden light changes.

[0070] The fluctuation range of the ambient light intensity value within a preset time window can be determined through statistical analysis. Within the preset time window, an ambient light intensity value sequence is continuously collected at a fixed sampling period, and then the statistical dispersion of this sequence is calculated as the fluctuation range. One implementation method is to calculate the absolute value of the difference between light intensity values ​​at adjacent sampling times and record its maximum value, which represents the maximum instantaneous rate of change within the window period. Another implementation method is to calculate the difference between the maximum and minimum light intensity values ​​within the entire time window, which represents the total range of change within the window period. When using the rate of change method, the absolute value of the difference can be divided by the sampling period to normalize it to the change per unit time; when using the range method, the difference between the maximum and minimum values ​​can be directly used as a quantitative indicator of the fluctuation range.

[0071] The selection of the aforementioned preset fluctuation threshold should be based on statistical analysis of the amplitude of typical transient light source interference. It must be ensured that this threshold is higher than the normal light intensity fluctuations caused by daily nursing activities to avoid frequent triggering of the lock, which could lead to the failure of the threshold adjustment function. Simultaneously, the preset fluctuation threshold should be lower than the minimum amplitude of strong transient interference such as the switching on and off of surgical lights and the flashing of monitors to prevent missed lockouts and detection failure. Furthermore, the setting of the aforementioned preset fluctuation threshold should also consider the noise floor of light intensity detection, allowing for a certain signal-to-noise ratio margin to suppress false triggers caused by random noise. The setting of the aforementioned preset number of times threshold should match the duration of the transient interference and the sampling frequency. Its value can be set to be greater than the number of times a single spike interference might exceed the limit, ensuring that the lockout is only triggered when the interference persists for multiple consecutive sampling periods. At the same time, the preset number of times threshold can also be set to be less than the number of sampling periods corresponding to the total duration of the transient interference to avoid excessively long lockout delays that could miss the optimal opportunity for interference isolation. The aforementioned lock-in duration can be set to be greater than the total duration of a typical transient interference event, including but not limited to the transition time of the operating light's on / off process, the pulse width period of the monitor's flash, or the complete duration of personnel movement blocking the light source. This ensures that the judgment threshold remains stable during the duration of the interference, completely isolating the impact of transient fluctuations on signal decision-making. Simultaneously, the lock-in duration can also be set to be less than the transition time from transient to steady state in the clinical lighting environment. This avoids prolonged threshold locking after the interference disappears, which could lead to an inability to respond to normal light intensity changes and affect adaptability.

[0072] The above-described implementation method for locking the current judgment threshold and maintaining a preset lock duration is as follows: When the fluctuation range of the ambient light intensity value exceeds a preset fluctuation threshold, and the number of consecutive exceedances reaches a preset number threshold, a latch signal is triggered. The latch signal can act on the judgment threshold update logic, loading the judgment threshold parameters (including the upper and lower thresholds) at the current moment into the latch register, and freezing the threshold calculation logic, causing it to pause responding to subsequent changes in ambient light intensity, thereby completing the threshold locking. The maintenance of the latch state can be achieved by an independent hardware timer or a software decrementing counter. When the lock is triggered, the timer is started and a count value corresponding to the preset lock duration is loaded. The timer decrements the count based on the system clock, and the latch signal remains valid until the count value reaches zero. If a software implementation is used, a lock flag is set after the lock is triggered and a lock duration countdown is started. The lock duration variable is decremented during each clock interrupt until the variable reaches zero and the lock flag is cleared. After the lock period ends, the latch register is released, and the threshold calculation logic resumes its periodic sampling and threshold update function for the ambient light intensity value, allowing the judgment threshold to re-enter the dynamic adaptive adjustment mode.

[0073] The above scheme detects the fluctuation range of ambient light intensity within a preset time window and uses a threshold for the number of consecutive detections as a trigger condition to identify sources of instantaneous light intensity change. When a scene of transient light change is identified, threshold locking is initiated to prevent the judgment threshold from frequently changing due to short-term strong light disturbances, thereby maintaining the stability and continuity of the signal conditioning process. On the other hand, the threshold locking mechanism maintains the current judgment threshold unchanged within a preset locking time, preventing pulse signal misjudgment or missed judgment caused by sudden light changes, thus enhancing the anti-interference capability of the above-mentioned nutrient pump drip identification method against transient interference sources such as surgical lights and monitor flashlights.

[0074] Optionally, the above-mentioned periodic acquisition of ambient light intensity value includes: synchronously sampling the ambient light intensity value in response to the infrared emitting diode conduction signal.

[0075] The aforementioned infrared LED turn-on signal is a logic level signal characterizing that the infrared LED is in the on state. At the circuit level, this signal is a control signal applied to the infrared LED driver circuit. When the driver circuit receives a valid level (high or low level, depending on the driver topology), the infrared LED is forward biased and radiates infrared light, at which point the turn-on signal is valid; when the driver level is removed, the infrared LED is cut off, and the turn-on signal is invalid. The infrared LED turn-on signal can be obtained in two ways: hardware sampling or software status monitoring. In hardware sampling, the signal can be directly extracted from the input control terminal of the infrared LED driver circuit (such as the base of a transistor or the gate of a field-effect transistor), or a synchronization signal can be generated by sensing the drive current through an optocoupler isolation device. In software status monitoring, if the infrared LED is driven by the general purpose input / output port (GPIO) of the main control chip, the turn-on signal can be obtained by reading the status of the output data register bit of the GPIO port, or the GPIO can be configured to multiplexed mode and a turn-on signal with a precise phase relationship can be generated through the timer output comparison function.

[0076] The above scheme synchronously performs ambient light intensity sampling in response to the infrared emitting diode's conduction signal, ensuring strict alignment between the ambient light detection timing and the infrared emission timing. This suppresses crosstalk between the infrared emitted light and the light intensity detection element, improving the purity and accuracy of the acquired ambient light intensity values. On the other hand, the synchronous sampling mechanism ensures that each acquired ambient light intensity data corresponds to the same infrared emission state, eliminating data fluctuations caused by sampling phase deviations. This provides timing-consistent input parameters for the dynamic adjustment of the judgment threshold, enhancing the stability and reliability of the signal conditioning process.

[0077] It can be understood that atomization refers to the phenomenon where, during continuous infusion, as nutrient solution droplets fall from the top of the drip chamber and impact the liquid surface, some of their kinetic energy is converted into splash energy, generating tiny droplets that adhere to the inner wall of the drip chamber. Simultaneously, the vapor generated by the evaporation of the nutrient solution itself condenses at the lower temperature on the inner wall of the drip chamber, together forming an atomized layer covering the inner wall of the drip chamber. For example... Figure 7 As shown in (a) to (c), the atomization layer is composed of a large number of tiny droplets, which scatter and absorb infrared light, causing the intensity of the light signal acquired by the infrared receiver to decrease and the signal baseline to continuously shift downward. The accumulation of the atomization layer gradually reduces the effective signal-to-noise ratio of infrared detection. If a fixed signal judgment threshold is maintained, the pulse amplitude caused by droplet obstruction will gradually decrease to the point where it is difficult to trigger threshold flipping, resulting in an increase in the false detection rate. At the same time, the non-uniformity of the atomization layer introduces additional noise components, increasing the probability of false triggering. In view of this, the embodiments of this application provide the following solution: Optionally, the above-mentioned nutrient pump drip identification method may further include: before infusion monitoring, acquiring the baseline value of the received signal and storing it as a calibration reference value; during infusion monitoring, monitoring the current baseline value of the received signal in real time and determining the degree of attenuation of the current baseline value of the received signal compared to the calibration reference value; when the degree of attenuation exceeds a preset attenuation threshold, determining the drip chamber atomization state, and determining a compensation coefficient based on the degree of attenuation; wherein, the compensation coefficient is positively correlated with the degree of attenuation; based on the compensation coefficient, generating an atomization compensation command to instruct the atomization compensation unit to enhance the infrared emission power of the infrared emitting tube. For example, after the device is started, under standard calibration conditions of no dripping, no atomization, and an ambient light intensity of 200 lux before infusion monitoring, the received signal amplitude is continuously collected 10 times and the arithmetic mean is taken. The result is used as the calibration reference value Vbase, typically approximately 1.5 volts; this calibration value is updated every 10 seconds to compensate for the slow baseline change caused by device temperature drift and long-term aging. During infusion monitoring, the baseline value of the received signal is monitored in real time. When the attenuation ratio of the baseline value relative to the calibration reference value reaches or exceeds 30%, it is determined that the inner surface of the drip chamber has entered an atomization state. Subsequently, an atomization compensation command is generated, driving the infrared emitting tube current to increase by 20% from the reference value of 10mA to 14mA, thereby increasing the infrared emission power by approximately 10%. This actively enhances the amplitude of the received signal, offsetting the attenuation effect of the atomization layer on infrared light transmission and maintaining an effective decision margin in the signal conditioning and processing stage.

[0078] The aforementioned baseline value of the received signal refers to the signal amplitude output by the infrared receiving channel under static conditions where no droplets cross the detection optical path. This amplitude characterizes the background light intensity level after the infrared light is transmitted through the drip chamber wall and internal medium. Its value is affected by multiple factors, including ambient light, infrared emission power, drip chamber transparency, and inner wall cleanliness. During the calibration phase before infusion monitoring, under conditions of no droplets, no atomization, and standard ambient light intensity, the received signal amplitude is sampled multiple times consecutively, and the sampling sequence is processed by arithmetic averaging or moving average, with the calculation result stored as the calibration reference value. During infusion monitoring, the real-time acquisition of the received signal baseline value can be achieved by sampling during the silent period between two adjacent droplet pulse events. That is, within the time window when no effective pulse edge is detected, the instantaneous signal amplitude of the receiving channel is periodically collected, and after digital filtering to eliminate random noise, a real-time estimate of the current baseline value is formed. The baseline value of the received signal can be calculated using one of the following methods: (1) Arithmetic average method: Multiple received signal samples are continuously collected within a preset sampling window. The arithmetic average operation is performed on the sample sequence, and the result is used as the current baseline value. This method is simple to calculate but has a large response lag. (2) Moving average method: By maintaining a fixed-length first-in-first-out sample queue, each time a new sample is added to the queue and the oldest sample is removed, the average value of the samples in the queue is recalculated as the updated baseline value. It has the ability to track the gradual change of the baseline. (3) Weighted average method: Differentiated weights are assigned to samples at different times. The more recent the sample, the higher the weight. The baseline is calculated by weighted summation, which improves the response speed to the change of the baseline. (4) Median filtering method: The median value is taken as the baseline after sorting the sample sequence within the sampling window. It can effectively suppress sudden spike interference and is suitable for scenarios with complex noise environments. The above attenuation degree can be calculated using the normalized ratio method. Specifically, the difference between the correction reference value and the current baseline value can be divided by the correction reference value to obtain a quantitative index characterizing the signal attenuation ratio.

[0079] When setting the preset attenuation threshold, one or more of the following factors can be considered: From a hardware perspective, the preset attenuation threshold should be higher than the long-term baseline fluctuations introduced by the power attenuation of the infrared emitter and the sensitivity drift of the receiver, avoiding the technical risk of false triggering due to device aging. The preset attenuation threshold should also be lower than the minimum discernible attenuation of the received signal when the drip chamber is slightly atomized, ensuring effective identification of the early atomization state. From the perspective of nutrient solution characteristics, differences in viscosity, surface tension, and evaporation rate of different nutrient solutions lead to different atomization formation speeds and adhesion strengths. The preset attenuation threshold can be calibrated for the target nutrient solution type to ensure that when the atomization degree exceeds the threshold at a normal infusion rate, it is already in a critical state requiring compensation. From the perspective of environmental conditions, changes in temperature and humidity affect the evaporation rate of the nutrient solution and the condensation effect on the inner wall of the drip chamber. When setting the preset attenuation threshold, a margin for temperature and humidity drift can be reserved to prevent threshold mismatch caused by environmental fluctuations. From the perspective of detection performance, the preset attenuation threshold can balance the false alarm rate and the false alarm rate. If the threshold is too low, the slight baseline disturbance during normal infusion may trigger nebulization judgment, increasing the probability of false compensation. If the threshold is too high, the signal attenuation caused by actual nebulization may not be recognized in time, resulting in detection failure.

[0080] The above scheme establishes an automatic identification mechanism for the atomization state of the drip chamber by storing the baseline value of the received signal as a correction reference during the calibration phase and continuously monitoring the attenuation of the current baseline value relative to the correction reference during infusion monitoring. This avoids relying on manual visual inspection to determine atomization and improves the autonomous monitoring capability of the above-mentioned nutrient pump drip identification method in continuous infusion scenarios. On the other hand, based on the positive correlation between the attenuation degree and the compensation coefficient, atomization compensation instructions are dynamically generated. By enhancing the infrared emission power, the attenuation of infrared light transmission caused by atomization is actively offset, so that the signal conditioning and processing stage can still maintain effective droplet detection sensitivity under atomization interference. This enhances the adaptability of the above-mentioned nutrient pump drip identification method to the interference of accumulated droplets attached to the inner wall of the drip chamber.

[0081] The compensation coefficient for the above scheme can be determined using one of the following methods: The first method involves determining the attenuation level through linear interpolation. This method establishes a linear functional relationship between two adjacent attenuation thresholds, using the attenuation level as the independent variable in the interpolation formula to calculate continuous values ​​between the upper and lower compensation coefficients in real time. Electronic devices implementing the above nutrient pump drip identification method can pre-store the attenuation level and compensation coefficients corresponding to the endpoints of each interval. During operation, the compensation coefficient is calculated proportionally based on the relative position of the current attenuation level within the interval.

[0082] The second method involves determining the attenuation level and compensation coefficient through a lookup table. This method stores a pre-calibrated two-dimensional mapping table of attenuation level and compensation coefficient within the electronic device executing the aforementioned nutrient pump drip identification method. This mapping table describes the nonlinear correspondence between the two through discrete sampling points. During online operation, the calculated attenuation level can be used as an index to locate the closest sampling point in the mapping table using a linear or binary search algorithm, and the corresponding compensation coefficient can be extracted. This method allows for arbitrarily complex nonlinear relationships to describe the attenuation and compensation mapping. Simply updating the mapping table adapts to different types of infrared devices or varying nutrient solution atomization characteristics, offering high flexibility and configurability.

[0083] The third method involves determining the compensation coefficient through function calculation. This method is based on a pre-defined mathematical function model, using the degree of attenuation as an input variable and directly substituting it into the function to calculate the compensation coefficient. The function can be a linear function, a polynomial function, or a piecewise function, and its coefficients are obtained through offline fitting calibration data and stored in the electronic device executing the aforementioned nutrient pump drip identification method. The calculation process in this method is entirely implemented by hardware arithmetic logic units or software floating-point operations, resulting in lower storage overhead and predictable execution time.

[0084] The fourth method is achieved through segmented mapping. This method divides the continuous range of attenuation values ​​into multiple discrete attenuation intervals, each interval corresponding to a pre-calibrated fixed compensation coefficient.

[0085] Optionally, determining the compensation coefficient based on the attenuation level includes: determining the corresponding attenuation level based on the attenuation interval to which the attenuation level belongs; wherein multiple attenuation intervals are formed by multiple preset attenuation thresholds, and the attenuation level corresponds one-to-one with the attenuation interval; determining the compensation coefficient according to a preset correspondence between the attenuation level and the compensation coefficient; wherein the compensation coefficient is positively correlated with the attenuation level. For example, this implementation pre-sets three attenuation thresholds, dividing the continuous range of attenuation level into three discrete intervals: the first interval corresponds to an attenuation level of 30% to 50%, the second interval corresponds to an attenuation level of 50% to 70%, and the third interval corresponds to an attenuation level greater than 70%. When the real-time calculated attenuation level falls into the first interval, it is determined to be a light atomization state, corresponding to attenuation level one; when it falls into the second interval, it is determined to be moderate atomization, corresponding to level two; and when it falls into the third interval, it is determined to be heavy atomization, corresponding to level three. Then, the preset attenuation level and compensation coefficient mapping table is queried to obtain the compensation coefficient that matches the current attenuation level: Level 1 corresponds to a 20% current increase ratio, Level 2 corresponds to a 40% current increase ratio, and Level 3 corresponds to a 60% current increase ratio.

[0086] The above scheme constructs a hierarchical compensation architecture by mapping the degree of attenuation to discrete attenuation intervals and determining the corresponding attenuation levels. This allows different levels of atomization to correspond to differentiated compensation coefficients, avoiding the problem of insufficient or excessive compensation by a single compensation coefficient during the atomization process, and improving the precision of the compensation strategy. On the other hand, by utilizing the preset correspondence between attenuation levels and compensation coefficients and their positive correlation, the compensation strategy is standardized and configurable. This enables the above-mentioned nutrient pump drip identification method to dynamically select the matching compensation intensity according to the severity of atomization, enhancing its adaptability to the gradual evolution of the atomization degree in the drip chamber.

[0087] Optionally, the above-mentioned nutrient pump drip identification method may further include: when the attenuation exceeds a preset severe atomization threshold, determining it as a severe atomization state and executing an alarm operation.

[0088] The above-mentioned scheme adds a severe nebulization warning function to the tiered compensation mechanism. When the attenuation exceeds the preset severe nebulization threshold (e.g., baseline attenuation exceeds 70%), it is automatically identified as a severe nebulization state and an alarm is triggered. Specifically, after each attenuation calculation is completed, the calculation result is compared with the severe nebulization threshold. If the threshold is exceeded for multiple consecutive sampling cycles, it is determined that the nebulization level has exceeded the automatic compensation adjustment range. At this time, an alarm command is generated to drive the audible and visual alarm device, prompting nursing staff to manually clean the drip chamber. The automatic alarm mechanism combines autonomous compensation with manual prompting, solving the problem that simple power compensation under severe nebulization may lead to infrared emitter overload or failure to maintain detection sensitivity, thus avoiding the risk of detection failure. The alarm operation can be implemented through buzzer sounding, LED indicator flashing, or sending alarm messages to the central monitoring system. Its trigger threshold and duration parameters can be adjusted through the configuration interface to adapt to different nutrient solution types and clinical false alarm tolerance requirements, improving the safety margin of long-term continuous monitoring.

[0089] The above solution presets a severe atomization threshold and monitors whether the attenuation exceeds the threshold. When the atomization level exceeds the automatic compensation adjustment range, an alarm operation is triggered. This combines the autonomous compensation mechanism with manual intervention prompts, avoiding the risk of detection failure that may result from relying solely on automatic compensation. On the other hand, the alarm operation actively outputs a severe atomization status signal, providing timely status indications for manual cleaning and maintenance, thereby improving the safety margin of the above nutrient pump drip identification method during long-term continuous monitoring.

[0090] like Figure 8As shown, in a certain application scenario, the fogging interference compensation mechanism mainly includes: First, under conditions of no droplets, no fogging, and stable ambient light, the baseline of the received signal is corrected, and the reference amplitude Vbase is acquired and stored as the reference zero point for subsequent comparisons; after entering the continuous working stage, the current infrared received signal amplitude Vcurrent is periodically monitored, and the attenuation degree K is calculated according to the formula K=(Vbase-Vcurrent) / Vbase. When it is determined that the K value does not exceed 30%, the drip chamber is considered to be in a clean state, the fogging level is 0, and the current power output of the infrared emitting tube is maintained; if the K value falls within the range of 30% to 50%, it is determined to be light fogging, the fogging level is 1, and a compensation command to slightly increase the power of the infrared emitting tube is generated; if the K value falls within the range of 50% to 70%, it is determined to be moderate fogging, the fogging level is 2, and a compensation command to moderately increase the power of the infrared emitting tube is generated; when the K value exceeds 70%, the fogging level is 3, and it is determined whether heavy fogging has been reached. If confirmed, an alarm command is generated to request user intervention; otherwise, a large power increase is executed as an extreme compensation measure. The above process achieves closed-loop control from monitoring and grading to power adjustment by mapping the attenuation degree with the compensation coefficient, ensuring that the droplet detection sensitivity remains within the effective range under atomization interference.

[0091] Optionally, step S140 may include: determining the number of pulses used to calculate the real-time drip rate based on the drip rate range where the pulse frequency of the single-channel effective pulse sequence is located; and calculating the real-time drip rate based on the latest pulse signal generated in the single-channel effective pulse sequence.

[0092] The droplet's inherent randomness and non-uniformity during the dripping process cause statistical fluctuations in the number of pulses per unit time. If a fixed number of pulses is used to calculate the drop rate, in the low-speed range, due to the sparse drop events, it takes a long time to accumulate enough pulses, resulting in a delay in drop rate updates and insufficient real-time performance. In the high-speed range, due to the dense droplets, too few fixed sampling points increase the statistical variance, reducing the repeatability and accuracy of the calculation results. The number of pulses used for calculation is dynamically determined based on the drip rate range in which the pulse frequency falls. This allows for the use of fewer pulses at low speeds to shorten the update cycle and quickly respond to changes in drip rate; and the use of more pulses at high speeds to smooth out random fluctuations and improve statistical accuracy. For example, at the initial stage of infusion, a first sampling interval (e.g., 4 pulses) is used for initial measurement to obtain a rough estimate of the current drip rate. Subsequently, based on the drip rate range in which the initial measurement result falls, the number of pulses used for subsequent measurements is dynamically adjusted: when the drip rate is in the high-speed range (e.g., greater than 27 drops / min), a second sampling interval (e.g., 15 pulses) is used to calculate the real-time drip rate to fully smooth out random fluctuations during the high-speed dripping process; when the drip rate is in the medium-speed range (e.g., 13 to 27 drops / min), a third sampling interval (e.g., 10 pulses) is used to achieve a balance between accuracy and response speed; and when the drip rate is in the low-speed range (e.g., less than 13 drops / min), a fourth sampling interval (e.g., 5 pulses) is used to shorten the update delay at low speeds and quickly respond to changes in drip rate. This tiered strategy guides the dynamic switching of sampling intervals based on initial measurement results, enabling drop rate calculations to achieve statistical accuracy and real-time performance matching the current drop rate across different rate ranges. This adaptive strategy achieves a dynamic balance between statistical error and response latency by matching the number of sampling points with the current drop rate, ensuring acceptable real-time performance and accuracy for drop rate calculations across the entire range.

[0093] The above implementation method for calculating the real-time drip rate based on the latest generated pulse signal in a single valid pulse sequence may include: maintaining a first-in-first-out pulse timestamp queue, the queue length of which is equal to the number of pulses determined by the current drip rate interval. Whenever the redundancy determination logic outputs a new valid pulse, the trigger time of that pulse is recorded and enqueued, while the oldest pulse timestamp in the queue is removed. Subsequently, the time difference between the latest and oldest pulses in the queue is calculated; this difference is the total time interval Δt experienced by the most recently generated N pulses. The real-time drip rate is calculated by dividing N-1 by Δt, where N-1 is the number of pulse cycles within the time interval.

[0094] The above scheme determines the number of pulses used to calculate the real-time drip rate based on the drip rate range in which the pulse frequency is located, so that the number of sampling points is dynamically matched with the current drip rate. In the low-speed range, the number of pulses is reduced to shorten the calculation response time, and in the high-speed range, the number of pulses is increased to smooth random fluctuations, thus balancing the real-time performance and accuracy of drip rate detection. On the other hand, the latest generated pulse signal is used for real-time drip rate calculation, and a sliding time window mechanism is adopted to ensure that the calculation benchmark is always based on the latest sampled data, avoiding drip rate calculation delays caused by historical data lag, and improving the tracking and response capability to sudden changes in drip rate.

[0095] Optionally, the above-mentioned nutrient pump drip identification method may further include: obtaining a preset target drip rate; calculating the deviation between the real-time drip rate and the preset target drip rate; and when the absolute value of the deviation exceeds a preset allowable deviation, generating a motor speed adjustment command based on the deviation to make the real-time drip rate approach the target drip rate.

[0096] The aforementioned preset target drip rate can be a desired infusion rate parameter pre-set by clinical medical staff based on the patient's nutritional support plan requirements. The preset target drip rate can serve as a reference benchmark for closed-loop control. The target drip rate setting is calculated based on the total feeding volume and feeding time required by the doctor's order, and its unit can be milliliters per hour or drops per minute, and it supports continuous adjustment within the range from zero to the maximum flow rate of the device.

[0097] The aforementioned allowable deviation can be set to be greater than the inherent random fluctuation range of the dripping process, avoiding frequent speed adjustments triggered by normal statistical variance, which increases motor wear and power consumption. Furthermore, the allowable deviation is less than the maximum clinically permissible drip rate deviation limit, ensuring that the actual infusion rate deviates from the target value without affecting the effectiveness and safety of nutritional delivery. When setting the allowable deviation, the minimum step angle of the stepper motor and the microstepping settings of the driver can also be matched to ensure that speed control commands can be converted into effective mechanical displacement, preventing ineffective adjustments and system oscillations caused by deviations less than the motor's minimum adjustment resolution. In addition, the allowable deviation setting can also take into account the mechanical inertia delay of the tubing, allowing sufficient transition time for the drip rate response after the speed control command is executed, avoiding repeated adjustments triggered during the transition period due to the instantaneous deviation not converging. Furthermore, different clinical scenarios have different requirements for infusion accuracy. For example, ICU patients have stringent requirements for drip rate stability, and the allowable deviation should be set smaller; while nutritional support in general wards can appropriately relax the deviation range to balance control accuracy and response frequency.

[0098] The generation of the aforementioned motor speed control command can be based on a proportional control algorithm, making the command amplitude proportional to the magnitude of the deviation; the larger the deviation, the larger the speed control amplitude, achieving rapid correction. Alternatively, a proportional-integral-derivative (PID) control algorithm can be used, superimposing an integral term on the proportional term to eliminate accumulated steady-state error and superimposing a derivative term to suppress overshoot and oscillation, ensuring the smoothness of the speed control process. The generated speed control command can be output to the motor drive unit in the form of pulse frequency or duty cycle, driving a stepper motor or DC motor to adjust the operating speed of the infusion actuator, thereby changing the flow rate of the nutrient solution in the infusion tubing. After the speed control command is executed, the real-time drip rate can continue to be monitored, the deviation can be recalculated, and a new speed control command can be generated. This iterative process allows the real-time drip rate to gradually converge to the target drip rate during dynamic adjustment, forming a negative feedback closed loop, ultimately achieving precise tracking and stable maintenance of the drip rate.

[0099] The above scheme obtains a preset target drip rate and calculates its deviation from the real-time drip rate. When the deviation exceeds the allowable range, a motor speed adjustment command is generated, forming a negative feedback closed-loop control loop. This allows the real-time drip rate to automatically track the preset target, thereby improving the accuracy and stability of drip rate control. On the other hand, by using the absolute value of the deviation as the speed adjustment trigger condition, adjustment is only initiated when the deviation reaches a threshold, avoiding over-response to minor deviations. At the same time, the speed adjustment command is generated based on the deviation, matching the speed adjustment intensity with the degree of deviation, thus achieving adaptive dynamic adjustment of the infusion rate.

[0100] To verify the effectiveness of the nutrient pump drip identification method provided in this application embodiment, a comparative experiment with multi-factor coupling interference was also designed. The comparative experiment used a traditional single-channel infrared detection method as a control group, and conducted parallel tests with the nutrient pump drip identification method provided in this application embodiment under the same conditions to quantitatively evaluate the differences in detection reliability indicators between the two methods. The experimental conditions were set as follows: the infusion rate was constant at 300 ml / hour, and the drip type used was standard enteral nutrition solution, whose physical properties conformed to the viscosity and surface tension parameters of commonly used clinical nutrient solutions. Atomization interference was simulated using a 40℃ temperature and humidity environment. This temperature condition can accelerate the evaporation of the nutrient solution and the process of droplet splashing and adhesion, forming a stable atomized layer on the inner wall of the drip chamber. Illumination conditions were simulated using direct surgical light, providing a high-brightness, highly directional light source to reproduce the illumination fluctuation characteristics of operating rooms and other diagnostic and treatment scenarios. The experiment statistically analyzed the total number of droplet detections, false alarms, and missed alarms for both methods within the same test period, calculating the false alarm rate as the core evaluation index. Response delay and stability parameters were also recorded to comprehensively assess the performance improvement of the proposed method under conditions of tilted installation, fogging interference, and strong light irradiation coupling. The experimental results are shown in Table 2. Table 2 Comparison of experimental results

[0101] As shown in Table 2, the traditional single-channel infrared detection scheme exhibits certain false alarm degradation under scenarios involving drip atomization and strong light interference. Its false alarm rate is as high as 28% under drip atomization and rises to 42% under strong light (600 lux). This indicates that the scheme lacks adaptability to atomization attenuation and light fluctuations in a single detection channel architecture, leading to a mismatch between the detection threshold and dynamically changing signal characteristics, resulting in pulse edge misjudgment. In contrast, the nutrient pump drip identification method provided in this application reduces the false alarm rate under drip atomization to no more than 2% and the false alarm rate under strong light to no more than 1%, verifying the effectiveness of the nutrient pump drip identification method provided in this application.

[0102] Please see Figure 9 Based on the same inventive concept, this application also provides a nutrient pump 300, comprising: The infrared detection unit 310 is used to emit infrared light into the drip chamber and receive infrared light signals transmitted through the drip chamber, and output multiple droplet detection signals; wherein, the infrared detection unit includes multiple infrared emitting channels and multiple infrared receiving channels, and the multiple infrared receiving channels are arranged on the same detection section of the drip chamber. The main control unit 320 is used to execute the nutrient pump drip identification method provided in the embodiments of this application; The light intensity detection unit 330 is used to acquire the ambient light intensity value and transmit it to the main control unit; The atomization compensation unit 340 is used to adjust the infrared emission power in response to the atomization compensation command of the main control unit; The motor drive unit 350 is used to respond to the motor speed adjustment command of the main control unit and drive the infusion actuator to adjust the infusion speed.

[0103] It is understood that the main control unit 320 provided in this application embodiment can be used to execute the nutrient pump drip identification method provided in this application embodiment. Its implementation principle and the resulting technical effects have been introduced in the foregoing method embodiments. For the sake of brevity, any part not mentioned in the device embodiment can be referred to the corresponding content in any of the foregoing method embodiments.

[0104] Figure 10 This is a schematic diagram of an electronic device provided in an embodiment of this application. (Refer to...) Figure 10Electronic device 400 includes a processor 410, a memory 420, and a communication interface 430. These components are interconnected and communicate with each other via a communication bus 440 and / or other forms of connection mechanisms (not shown). The memory 420 includes one or more (only one is shown in the figure), which may be, but is not limited to, Random Access Memory (RAM), Read Only Memory (ROM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), etc. The processor 410 and other possible components can access the memory 420, reading and / or writing data therein. The processor 410 includes one or more (only one is shown in the figure), which may be an integrated circuit chip with signal processing capabilities. The processor 410 mentioned above can be a general-purpose processor, including a central processing unit (CPU), a micro controller unit (MCU), a network processor (NP), or other conventional processors; it can also be a special-purpose processor, including a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The communication interface 430 includes one or more (only one is shown in the figure), which can be used to communicate directly or indirectly with other devices to exchange data. For example, the communication interface 430 can be an Ethernet interface; it can be a mobile communication network interface, such as an interface for 3G, 4G, or 5G networks; or it can be other types of interfaces with data transmission and reception functions. One or more computer program instructions can be stored in the memory 420, and the processor 410 can read and execute these computer program instructions to implement the nutrient pump drip identification method provided in this application embodiment and other desired functions. It is understood that... Figure 10 The structure shown is for illustrative purposes only; the electronic device 400 may also include more than [other components]. Figure 10 The more or fewer components shown, or having the same Figure 10 The different configurations shown. Figure 10 The components shown can be implemented using hardware, software, or a combination thereof. For example, electronic device 400 can be a single server (or other device with computing power), a combination of multiple servers, a cluster of a large number of servers, etc., and can be either a physical device or a virtual device.

[0105] This application also provides a computer-readable storage medium storing computer program instructions. These computer program instructions are read and executed by a processor to perform the nutrient pump drip identification method provided in this application. For example, the computer-readable storage medium can be implemented as follows: Figure 10 The memory 420 in the electronic device 400, or a separate storage product (such as a USB flash drive, portable hard drive, etc.).

[0106] This application also provides a computer program product, which includes computer program instructions. These computer program instructions are read and executed by a processor to perform the nutrient pump drip identification method provided in this application. For example, these computer program instructions can be stored in... Figure 10 The memory 420 in the electronic device 400 is located inside the memory, or it is stored in a separate storage product (such as a USB flash drive, portable hard drive, etc.).

[0107] In the embodiments provided in this application, it should be understood that the disclosed apparatus and method can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and there may be other division methods in actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the shown or discussed mutual coupling or direct coupling or communication connection may be through some communication interfaces. The indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms. Additionally, the units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs. Moreover, the functional modules in the various embodiments of this application can be integrated together to form an independent part, or each module may exist independently, or two or more modules may be integrated to form an independent part.

[0108] It should be noted that if the function is implemented as a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0109] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms "comprising" and "having," and any variations thereof, in the specification, claims, and accompanying drawings of this application, are intended to cover non-exclusive inclusion. In the description of embodiments of this application, technical terms such as "first," "second," etc., are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary or secondary relationship of the indicated technical features. In the description of embodiments of this application, "a plurality of" means two or more, unless otherwise expressly and specifically defined. The reference to "embodiment" herein means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments. In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0110] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for identifying nutrient pump drip rates, characterized in that, The method includes: The droplet detection signals collected by multiple infrared receiving channels are acquired; wherein the multiple infrared receiving channels are arranged on the same detection section of the nutrient pump drip chamber. Each droplet detection signal is processed by signal conditioning to obtain multiple pulse signals; Using redundant decision logic, the multiple pulse signals are combined into a single valid pulse sequence; wherein, the redundant decision logic is configured to determine that the droplet is valid when at least one of the pulse signals meets the droplet determination condition; The real-time drip rate is calculated based on the pulse frequency of the single effective pulse sequence.

2. The nutrient pump drip identification method according to claim 1, characterized in that, The method further includes: Periodically acquire ambient light intensity values; The preset correspondence between the ambient light intensity value and the judgment threshold is determined by using the ambient light intensity value to determine the judgment threshold used to generate the pulse signal.

3. The nutrient pump drip identification method according to claim 2, characterized in that, The step of determining the threshold for generating a pulse signal by using the ambient light intensity value to establish a preset correspondence with the threshold value includes: When the ambient light intensity value is less than the first light intensity threshold, the first fixed judgment threshold is used; When the ambient light intensity value is greater than the second light intensity threshold, the second fixed judgment threshold is used; When the ambient light intensity value is not less than the first light intensity threshold and the ambient light intensity value is not greater than the second light intensity threshold, a judgment threshold for generating a pulse signal is determined by linear interpolation based on the ambient light intensity value, the first light intensity threshold, the second light intensity threshold, the first fixed judgment threshold, and the second fixed judgment threshold.

4. The nutrient pump drip identification method according to claim 2, characterized in that, The method for determining the judgment threshold includes: The upper limit threshold is determined by using the ambient light intensity value to establish a preset correspondence with the upper limit threshold in the judgment threshold. Based on the upper threshold and the preset hysteresis width, the lower threshold of the judgment threshold is determined; wherein, the lower threshold is obtained by subtracting the preset hysteresis width from the upper threshold.

5. The nutrient pump drip identification method according to claim 2, characterized in that, The method further includes: The fluctuation range of the ambient light intensity value within a preset time window is detected; When the fluctuation amplitude exceeds a preset fluctuation threshold, and the number of times the fluctuation amplitude exceeds the preset fluctuation threshold is continuously detected reaches a preset number threshold, the current determination threshold is locked and maintained for a preset lock duration.

6. The nutrient pump drip identification method according to claim 2, characterized in that, The periodic acquisition of ambient light intensity values ​​includes: In response to the infrared emitting diode's activation signal, the ambient light intensity value is sampled synchronously.

7. The nutrient pump drip identification method according to claim 1, characterized in that, The method further includes: Before performing infusion monitoring, acquire the baseline value of the received signal and store it as a calibration reference value; In infusion monitoring, the current received signal baseline value is monitored in real time, and the degree of attenuation of the current received signal baseline value compared with the correction reference value is determined. When the attenuation exceeds a preset attenuation threshold, it is determined to be in a drip atomization state, and a compensation coefficient is determined based on the attenuation; wherein, the compensation coefficient is positively correlated with the attenuation. Based on the compensation coefficient, a fogging compensation command is generated to instruct the fogging compensation unit to enhance the infrared emission power of the infrared emitting tube.

8. The nutrient pump drip identification method according to claim 7, characterized in that, Determining the compensation coefficient based on the attenuation level includes: Based on the attenuation range to which the attenuation level belongs, the corresponding attenuation level is determined; wherein, multiple attenuation ranges are formed by multiple preset attenuation thresholds, and the attenuation level corresponds one-to-one with the attenuation range; The compensation coefficient is determined according to the preset correspondence between the attenuation level and the compensation coefficient; wherein the compensation coefficient is positively correlated with the attenuation level.

9. The nutrient pump drip identification method according to claim 7, characterized in that, The method further includes: When the attenuation level exceeds the preset severe atomization threshold, it is determined to be a severe atomization state, and an alarm operation is executed.

10. The nutrient pump drip identification method according to any one of claims 1 to 9, characterized in that, The calculation of the real-time drip rate based on the pulse frequency of the single-channel effective pulse sequence includes: Based on the drip rate range where the pulse frequency of a single effective pulse sequence falls, the number of pulses used to calculate the real-time drip rate is determined. The real-time drip rate is calculated based on the latest pulse signal generated in the single-channel effective pulse sequence.

11. The nutrient pump drip identification method according to any one of claims 1 to 9, characterized in that, The method further includes: Obtain the preset target drip rate; Calculate the deviation between the real-time drip rate and the preset target drip rate; When the absolute value of the deviation exceeds the preset allowable deviation, a motor speed adjustment command is generated based on the deviation to make the real-time dripping speed approach the target dripping speed.

12. A nutrient pump, characterized in that, include: An infrared detection unit is used to emit infrared light into the drip chamber and receive infrared light signals transmitted through the drip chamber, and output multiple droplet detection signals; wherein, the infrared detection unit includes multiple infrared emitting channels and multiple infrared receiving channels, and the multiple infrared receiving channels are arranged on the same detection section of the drip chamber; The main control unit is used to execute the method according to any one of claims 1 to 11. A light intensity detection unit is used to acquire ambient light intensity values ​​and transmit them to the main control unit; A fogging compensation unit is used to adjust the infrared emission power in response to the fogging compensation command of the main control unit; The motor drive unit is used to respond to the motor speed adjustment command of the main control unit and drive the infusion actuator to adjust the infusion speed.

13. An electronic device, characterized in that, include: A processor, a memory, and a communication bus, wherein the processor and the memory communicate with each other via the communication bus; The memory stores program instructions that can be executed by the processor, and the processor can execute the method as described in any one of claims 1 to 11 by calling the program instructions.

Images