A dynamic event awareness method, system and apparatus

Abstract

The application relates to the technical field of intelligent sensing, and provides a dynamic event sensing method, system and device. The dynamic event sensing method comprises the following steps: acquiring an electrical response signal output by a synapse device within a preset time period to obtain an electrical signal sequence; calculating a change rate of the electrical signal sequence within the preset time period according to the electrical signal sequence; acquiring a preset event discrimination threshold, comparing the change rate with the event discrimination threshold to obtain a comparison result; if the comparison result is an effective dynamic event, recording event information corresponding to the electrical response signal; and encapsulating the event information based on the electrical signal sequence to generate an event signal. The application utilizes the characteristics of the response signal evolution of an artificial synapse device with time under external stimulation, realizes self-triggering sensing of a dynamic event, and improves the robustness of a system to complex environmental changes.

Description

Technical Field

[0001] This application relates to the field of intelligent sensing technology, and in particular to a dynamic event sensing method, system and device. Background Technology

[0002] Existing dynamic behavior perception and motion detection technologies primarily rely on continuous signal sampling or frame-based data acquisition methods. For example, they utilize CCD / CMOS image sensors combined with inter-frame difference analysis, optical flow calculation, or deep learning-based video analysis algorithms to process consecutive frames of images to identify the target's motion state. These technologies typically require high-frequency sampling and continuous computation, resulting in high power consumption, significant data redundancy, and high system complexity. Furthermore, some non-imaging motion detection schemes rely on infrared, millimeter-wave, or ultrasonic sensors, determining target motion through fixed thresholds or simple time-difference analysis. However, these schemes are sensitive to environmental noise, struggle to distinguish between real dynamic events and background disturbances, and lack the ability to analyze complex dynamic behaviors. Summary of the Invention

[0003] This application aims to improve at least one technical problem in the background art.

[0004] This application provides a dynamic event sensing method, which includes: The electrical response signal output by the synaptic device within a preset time period is acquired to obtain an electrical signal sequence; Calculate the rate of change of the electrical signal sequence within the preset time period based on the electrical signal sequence. Obtain a preset event discrimination threshold, compare the rate of change with the event discrimination threshold to obtain the comparison result; If the comparison result is a valid dynamic event, then record the event information corresponding to the electrical response signal; Event information is encapsulated based on electrical signal sequences to generate event signals.

[0005] According to some technical solutions of this application, the step of calculating the rate of change of the electrical signal sequence within the preset time period based on the electrical signal sequence specifically includes: Extract the current values ​​and corresponding time intervals of adjacent sampling points from the electrical signal sequence; The current difference between adjacent sampling points is obtained by calculating the current values ​​of adjacent sampling points; The instantaneous rate of change is calculated based on the current difference and time interval between adjacent sampling points.

[0006] According to some technical solutions of this application, the step of obtaining a preset event discrimination threshold and comparing the rate of change with the event discrimination threshold to obtain a comparison result specifically includes: Obtain a preset event discrimination threshold, which includes a stability threshold and a mutation threshold; When the absolute value of the rate of change is less than or equal to the stability threshold, it is determined to be a static background signal; When the absolute value of the rate of change is greater than the stability threshold and less than or equal to the mutation threshold, it is determined to be a non-target weak disturbance. When the absolute value of the rate of change is greater than the mutation threshold, it is determined to be a valid dynamic event. According to some technical solutions of this application, if the comparison result is a valid dynamic event, the event information corresponding to the electrical response signal is recorded, specifically including: If the comparison result is a valid dynamic event, read the spatial coordinates of the synaptic device in the array that received the electrical response signal; Obtain the current time as the event occurrence time and associate it with spatial coordinates to obtain event information.

[0007] According to some technical solutions of this application, the process of encapsulating event information based on an electrical signal sequence to generate an event signal specifically includes: Extract the peak intensity and duration of the event from the electrical signal sequence; Peak intensity, duration, and event information are integrated to obtain event characteristic parameters; The event characteristic parameters are encapsulated according to a preset protocol to generate an event signal that conforms to the preset protocol.

[0008] According to some technical solutions of this application, after encapsulating the event information based on the electrical signal sequence to generate the event signal, the method further includes: Obtain historical response data within a preset time period; Based on historical response data, the environmental noise level is analyzed to obtain the analysis results; Based on the analysis results, the event discrimination threshold was optimized and adjusted. The optimized event discrimination threshold is used as the preset event discrimination threshold.

[0009] According to some technical solutions of this application, the optimization and adjustment of the event discrimination threshold based on the analysis results specifically includes: When the analysis results indicate an increase in ambient background noise, the stability threshold should be increased. Increase the mutation threshold when the analysis results indicate that non-target weak perturbations occur frequently. When the analysis results indicate a need to improve the detection sensitivity for dynamic events of the target, the mutation threshold should be lowered.

[0010] According to some technical solutions of this application, the step of obtaining the electrical response signal output by the synaptic device within a preset time period to obtain an electrical signal sequence specifically includes: Acquire the electrical response signal output by the synaptic device within a preset time period; The electrical response signal is amplified and converted from analog to digital to obtain the converted signal; The converted signal is sampled at a preset sampling rate to obtain a digitized electrical signal sequence.

[0011] This application also provides a dynamic event sensing system, which includes: The acquisition module is used to acquire the electrical response signal output by the synaptic device within a preset time period to obtain an electrical signal sequence; The calculation module is used to calculate the rate of change of the electrical signal sequence within a preset time period based on the electrical signal sequence. The comparison module is used to obtain a preset event discrimination threshold, compare the rate of change with the event discrimination threshold, and obtain the comparison result. The recording module is used to record the event information corresponding to the electrical response signal if the comparison result is a valid dynamic event; The generation module is used to encode event information to generate event signals.

[0012] This application also provides a dynamic event sensing device, the dynamic event sensing device comprising: a memory and at least one processor, wherein the memory stores instructions; At least one of the processors invokes the instructions in the memory to cause the dynamic event sensing device to execute the various steps of the dynamic event sensing method as described above.

[0013] The dynamic event sensing method provided in this application has at least the following beneficial effects: It utilizes the characteristics of the evolution of response signals of artificial synaptic devices over time under external stimuli to achieve self-triggered sensing of dynamic events, reducing the need for continuous high-frequency sampling and full data processing of external stimulus signals. Through the inherent response characteristics of artificial synaptic devices to time changes, it achieves pre-screening of static backgrounds at the device level, reading and processing only response signals that have undergone significant changes, thereby reducing data redundancy and system power consumption. Furthermore, the generation of dynamic events stems from changes in the response behavior of synaptic devices, rather than fixed thresholds or preset rules, improving the system's robustness to complex environmental changes. Attached Figure Description

[0014] Figure 1 A flowchart illustrating the dynamic event sensing method provided in an embodiment of this application; Figure 2 This is a schematic diagram of the structure of the dynamic event sensing system provided in the embodiments of this application; Figure 3 This is a schematic diagram of the structure of the dynamic event sensing device provided in the embodiments of this application; Figure 4This is a schematic diagram illustrating the response characteristics of a synaptic device under static stimulation, as provided in an embodiment of this application. Figure 5 The diagram shows the response characteristics of the synaptic device provided in the embodiments of this application under dynamically changing stimuli. Detailed Implementation

[0015] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0016] In the description of this application, it should be understood that the orientation descriptions, such as up, down, front, back, left, right, etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation or be constructed or operated in a specific orientation. Therefore, they should not be construed as limiting the present invention.

[0017] In the description of this application, unless otherwise expressly defined, terms such as "setup," "installation," and "connection" should be interpreted broadly, and those skilled in the art can reasonably determine the specific meaning of the above terms in this application in conjunction with the specific content of the technical solution.

[0018] The following is combined with Figures 1 to 5 Embodiments of the present invention will be described.

[0019] Existing time-aware detection technologies rely on continuous or periodic sampling, resulting in high data redundancy and high power consumption. The discrimination of dynamic events mainly depends on external algorithms or circuit logic, leading to high system complexity. Static or slowly changing background signals consume computational resources and have a high probability of false triggering. The event generation mechanism does not fully utilize the time response characteristics of the device itself, making it difficult to achieve highly integrated edge sensing. Existing event-based solutions are mostly based on circuit or pixel-level comparison, and their sensing behavior does not originate from the physical time response characteristics of the device itself, making it difficult to achieve deep integration of sensing and computing at the device level.

[0020] Based on the above, this application provides a dynamic event sensing method, which includes: S100: Acquire the electrical response signal output by the synaptic device within a preset time period to obtain an electrical signal sequence. Exemplarily, this embodiment can be applied to dynamic event perception scenarios involving human movement and equipment vibration. The artificial synaptic device is a bottom-gate thin-film transistor with a ZnAlSnO or ZnO oxide semiconductor functional layer, consisting of a substrate, source, drain, and functional layer from top to bottom. A 32×32 pixel array is formed, with the system sampling rate set to 1kHz, i.e., a preset time period of 1ms. By acquiring the electrical response signal output by a single synaptic device in the 32×32 pixel array within the preset time period, and processing it, a digitized electrical signal sequence is obtained. The preset time period is determined to be 1ms by the 1kHz sampling rate, and the artificial synaptic device is a bottom-gate thin-film transistor structure. S200, calculate the rate of change of the electrical signal sequence within the preset time period based on the electrical signal sequence; for example, calculate the instantaneous rate of change of the current-time of the electrical signal sequence within the preset time period, i.e., ΔI / Δt; S300, obtain a preset event discrimination threshold, compare the rate of change with the event discrimination threshold to obtain a comparison result; for example, obtain a preset event discrimination threshold based on the inherent noise of the artificial synaptic device and the dynamic event characteristics of the target, and compare the instantaneous rate of change of current-time with the event discrimination threshold to obtain a comparison result.

[0021] S400, if the comparison result is a valid dynamic event, then record the event information corresponding to the electrical response signal; the event information includes at least the spatial coordinates of the synaptic device in the array and the event occurrence timestamp; It should be noted that an effective dynamic event refers to a stimulus change process caused by the target object, which can reflect changes in the target's behavior and satisfies that ΔI / Δt is greater than the preset mutation threshold δ1; in contrast, a slow-changing signal generated in a stable environment (ΔI / Δt ≤ stability threshold) ) is defined as the background signal, while that caused by non-target factors and ΔI / Δt is between and Signals between these events are defined as non-target disturbances. For example, in a human motion detection scenario, rapid changes in light intensity or electrical signals caused by human movement, passing by, or limb movements can be defined as valid dynamic events, while slow changes in ambient light are background signals, and moderate fluctuations caused by swaying leaves or slight shading can be considered non-target disturbances. In a device vibration monitoring scenario, high-rate-of-change signals generated by abnormal vibrations or impacts of the device can be defined as valid dynamic events, while stable signals under normal device operation are background signals, and fluctuations caused by environmental micro-vibrations or electromagnetic interference are non-target disturbances.

[0022] S500 encapsulates event information based on an electrical signal sequence to generate an event signal; for example, it extracts feature parameters of a dynamic event based on the electrical signal sequence, integrates them with the event information, and encapsulates them according to a preset industrial protocol to generate an event signal, wherein the preset industrial protocol can be the Modbus-RTU protocol.

[0023] For example, in one embodiment, the MCU (Microcontroller Unit) initializes the parameters of each module, reads the stability threshold and mutation threshold stored in the NVM (Non-Volatile Memory), and then the artificial synaptic device array enters a standby sensing state to receive external light or electrical stimulation in real time. Within 10 ms, the current signal output by a certain synaptic device unit is continuously acquired. The continuous analog current signal output by the synaptic device is amplified and converted from analog to digital, and then sampled at fixed time intervals to obtain a series of discrete digital current values. Each current value is associated with its corresponding timestamp and arranged in chronological order to form a sequence, thereby obtaining a set of electrical signal sequences. Then, the rate of change is calculated based on the electrical signal sequence. Specifically, the time interval between adjacent sampling points is extracted, the difference between adjacent currents is calculated, and then the rate of change of current is calculated based on the time interval and the difference between adjacent currents. Finally, the rate of change of current is compared with a preset event discrimination threshold. If the rate of change of current is greater than the event discrimination threshold, it is determined to be a valid dynamic event; otherwise, the signal is ignored. At the same time, if it is determined to be a valid dynamic event, the occurrence time, corresponding pixel coordinates and peak value of the event are recorded and encapsulated as an event signal, which can then be transmitted to the terminal control system.

[0024] It should be noted that, Figure 4 and Figure 5 This diagram illustrates the time-dependent response characteristics of synaptic devices under different stimuli. It demonstrates the differentiated electrical response patterns of artificial synaptic devices to constant static stimuli and dynamically changing stimuli. It also serves as the core device characteristic basis for the entire sensing system to achieve self-triggered perception of dynamic events and self-inhibition of static backgrounds.

[0025] Specifically, Figure 4 The diagram illustrates the time-dependent response characteristics of a synaptic device under static stimulation. The horizontal axis represents time, and the vertical axis represents the device's output current. The overall curve shows a slow and stable change in current over time. That is, when the external stimulus is in a stable state, such as a static background without movement or a slowly changing environmental signal, the carrier distribution, defect state occupancy, or ion migration state within the artificial synaptic device gradually tends towards equilibrium. The time-dependent rate of change of its output current fluctuates very little and remains stable, without transient current spikes. This characteristic allows the device to recognize static background signals, which can then be used to subsequently determine and suppress static signals.

[0026] Figure 5 This diagram illustrates the time-dependent response characteristics of a synaptic device under dynamic changes. Similarly, the horizontal axis represents time, and the vertical axis represents the device's output current. The overall curve shows the transient changes in current over time. That is, when external stimuli undergo dynamic changes, such as human movement or equipment vibration, the internal physical state of the synaptic device is rapidly readjusted, resulting in a significant transient enhancement response in the output current. This manifests as a sudden change in current over a short period, with its rate of change exceeding the fluctuation range under static conditions. This characteristic allows the device to directly convert external dynamic stimuli into identifiable electrical signal mutations, achieving physical-level self-triggering of dynamic events without the need for external continuous sampling or preset logic to distinguish between dynamic events and static backgrounds.

[0027] Thus, the device exhibits a stable and low-fluctuation current response to static or slowly changing stimuli, and a transient and high-amplitude current response to dynamic and sudden stimuli. This differentiated response, determined by the physical characteristics of the device, allows the entire sensing system to complete dynamic event extraction and static background suppression at the device level. Static and slowly changing stimuli automatically decay in the synaptic device, achieving background self-suppression. Sensing and primary computation are completed at the device level, rather than relying on external algorithms or circuit logic, which is conducive to achieving low-power, highly robust dynamic behavior detection and miniaturized integration.

[0028] In some embodiments, step S200, calculating the rate of change of the electrical signal sequence within a preset time period based on the electrical signal sequence, specifically includes: S210, extract the current values ​​of adjacent sampling points and the corresponding time intervals from the electrical signal sequence; specifically, extract the amplified current values ​​of adjacent sampling points and the corresponding time intervals from the digitized electrical signal sequence, wherein the time interval is determined to be 1ms by a 1kHz sampling rate, and the current value is the current value after being amplified 1000 times by a low-noise operational amplifier; in particular, the sampling rate is used to read the state of the synaptic device response, and its sampling object is the electrical signal after the device has completed the initial time response modulation, rather than continuously sampling the original external stimulus signal at high frequency, so it will not introduce the data redundancy problem caused by traditional frame sampling.

[0029] S220, calculate the current value of adjacent sampling points to obtain the current difference between adjacent sampling points; S230 calculates the instantaneous rate of change based on the current difference and time interval between adjacent sampling points. Based on the current difference ΔI between adjacent sampling points and a fixed time interval Δt of 1 ms, the instantaneous rate of change is calculated using the formula... The instantaneous rate of change of current-time at each group of adjacent sampling points is calculated, and then the average instantaneous rate of change within a preset time period is taken as the final rate of change.

[0030] For example, the current values ​​of adjacent sampling points are extracted from an electrical signal sequence, such as... =1.02μA, =1.08μA, time interval Δt=1ms. The current difference ΔI=0.06μA is calculated, and the instantaneous rate of change is obtained as 0.06μA / ms. This rate of change reflects the response intensity of the synaptic device to changes in external stimuli at that moment, which is used for subsequent event discrimination.

[0031] In some embodiments, step S300 involves obtaining a preset event discrimination threshold and comparing the rate of change with the event discrimination threshold to obtain a comparison result, specifically including: S310, Obtain a preset event discrimination threshold, wherein the event discrimination threshold includes a stability threshold and a mutation threshold; wherein the event discrimination threshold includes a stability threshold. and mutation threshold ,in =0.03μA / ms, which is 1.5 times the inherent current noise of 0.02μA / ms for artificial synaptic devices. =0.4μA / ms, which is set based on the minimum effective rate of change of common dynamic scenarios such as human movement and equipment vibration, and > ; In this study, under conditions without external stimuli, the current of all devices in a 32×32 array was continuously monitored for 10 minutes, and the current change of each device within a 1ms time window was statistically analyzed. The overall noise fluctuation amplitude was calculated to be ≤±0.02μA / ms, and 1.5 times this was taken as the preset stability threshold δ0. This avoids false triggering caused by device thermal noise and readout circuit noise while ensuring sensitivity to subtle background changes. For typical dynamic events, such as a human body passing 30cm in front of the sensor at a speed of 0.5m / s, repeated experiments were conducted. The minimum current change rate at the occurrence of the event was 0.38μA / ms, and the maximum was 0.45μA / ms. To ensure no omissions in detection, the statistical median value of 0.4μA / ms was taken as the preset abrupt change threshold. This value is also significantly higher than the noise level (signal-to-noise ratio ≥13dB) to ensure detection reliability.

[0032] S320, when the absolute value of the rate of change is less than or equal to the stability threshold, it is determined to be a static background signal; for example, when the absolute value of the current-time rate of change is less than or equal to 0.03μA / ms, it is determined to be a static background signal, and the system ignores the signal. S330, when the absolute value of the rate of change is greater than the stability threshold and less than or equal to the mutation threshold, it is determined to be a non-target weak disturbance; for example, when the current-time rate of change is greater than 0.03μA / ms and less than or equal to 0.4μA / ms, it is determined to be a non-target weak disturbance, and the system does not trigger event recording and output; S340, when the absolute value of the rate of change is greater than the abrupt change threshold, it is determined to be a valid dynamic event. For example, when the current-time rate of change is greater than 0.4 μA / ms, it is determined to be a valid dynamic event, and the system triggers the subsequent event recording and signal generation process.

[0033] In some embodiments, in step S400, if the comparison result is a valid dynamic event, the event information corresponding to the electrical response signal is recorded, specifically including: S410, if the comparison result is a valid dynamic event, read the spatial coordinates of the synaptic device that outputs the electrical response signal in the array; for example, if the comparison result is a valid dynamic event, read the two-dimensional spatial coordinates (x,y) of the artificial synaptic device that outputs the electrical response signal in the 32×32 pixel array, where the values ​​of x and y are both in the range of 1-32. S420: Obtain the current time as the event occurrence time and associate it with spatial coordinates to obtain event information. Specifically, obtain the millisecond-level real-time timestamp of the system clock as the event occurrence time, associate and store the two-dimensional spatial coordinates with the timestamp to obtain event information containing spatial location and occurrence time. The storage medium is static random access memory.

[0034] In some embodiments, S500 encapsulates event information based on an electrical signal sequence to generate an event signal, specifically including: S510, extract the peak intensity and duration corresponding to the event from the electrical signal sequence; specifically, extract the peak intensity of the current-time change rate and the duration of the event corresponding to the effective dynamic event from the electrical signal sequence. The peak intensity is the maximum value of ΔI / Δt within a preset time period, and the duration is the time span from triggering the dynamic event to restoring it to a static state. It should be noted that the duration is when ΔI / Δt is continuously greater than a preset abrupt change threshold. The length of the time interval, whose starting time is when ΔI / Δt first exceeds [the specified value]. The time of termination is when ΔI / Δt decreases to The following moments can also be obtained by performing continuous interval detection on the sampled ΔI / Δt time series.

[0035] S520 integrates peak intensity, duration, and event information to obtain event feature parameters, i.e., event feature parameters containing multi-dimensional features; for example, when a valid dynamic event occurs in a certain synaptic device unit, such as when a valid dynamic event occurs in the 5th row and 8th column, the spatial coordinates (5,8) of the unit in the 32×32 array are read, and the current system time is recorded. The time is associated with the coordinates to form event information for subsequent trajectory tracking and behavior analysis.

[0036] S530, the event feature parameters are encapsulated according to a preset protocol to generate an event signal that conforms to the preset protocol. For example, the event feature parameters are encapsulated according to the frame structure of the Modbus-RTU industrial protocol to generate an event signal that conforms to the industrial communication standard. The frame structure includes a 1-byte address code, a 1-byte function code, N bytes of data code, and a 2-byte CRC checksum.

[0037] In some embodiments, after encapsulating the event information based on the electrical signal sequence to generate an event signal, step S500 further includes: S600, acquire historical response data within a preset time period; for example, acquire historical device response data for the past 5 minutes from non-volatile memory (NVM). The historical response data includes the average current change rate in a steady state, the current change rate fluctuation range, the peak current change rate during dynamic events, and the event frequency. Here, "steady state" refers to the artificial synaptic device maintaining its output current change rate ΔI / Δt at a stable threshold when it is not stimulated by an effective dynamic event. The system determines the following states by checking whether ΔI / Δt is less than or equal to a preset threshold for N consecutive sampling periods. For example, if ΔI / Δt is less than or equal to a preset threshold for 10 consecutive sampling points. To determine whether the current state is stable.

[0038] S610, based on historical response data, analyze the environmental noise level to obtain analysis results; specifically, based on the historical response data, perform statistical analysis on the environmental background noise level, the frequency of non-target disturbances, and the intensity of events using a microcontroller (MCU) to obtain analysis results; For example, extracting the steady-state fluctuation amplitude from historical response data. Non-target disturbance frequency and the mean of event intensity Based on the above characteristics and reference benchmarks, analysis was conducted and corresponding analysis results were obtained.

[0039] That is, when σ is compared to the system initial calibration value Increase, for example >1.5 When the background noise level increased, the analysis result was "increased"; when Exceeding the preset frequency threshold ,For example When the frequency is greater than 10 times / minute, the analysis result is "frequent occurrence of weak non-target perturbations"; when... Consistently above the historical average ,For example >1.2 When the target dynamic event is enhanced, the analysis result is "target dynamic event enhancement"; otherwise, if... Persistently below If so, the analysis result is "the detection sensitivity needs to be improved".

[0040] It should be noted that the steady-state fluctuation amplitude can be obtained by calculating the standard deviation or maximum fluctuation range of ΔI / Δt within all time intervals judged as "steady-state", reflecting the background noise level of the current environment; the frequency of non-target disturbances can be obtained by statistically analyzing ΔI / Δt within a preset time window when ΔI / Δt is at the stability threshold. With mutation threshold Between (i.e.) <ΔI / Δt< The number of events occurring (instances / minute) is calculated by dividing by the time window length, reflecting the activity level of non-target interference sources in the environment. The initial calibration value reflects the system's background noise level under ideal conditions, serving as a benchmark for subsequent judgment of whether environmental noise has increased. It can be calculated during initial deployment or calibration by collecting steady-state response data over a continuous period without external dynamic stimuli, and calculating the standard deviation or maximum fluctuation range of ΔI / Δt within that period as the initial calibration value. The historical average reflects the normal signal strength level of recent target dynamic events, serving as a benchmark for judging whether the event strength has changed significantly. It can be dynamically updated by periodically calculating the moving average of the ΔI / Δt peak values ​​of historical effective dynamic events during operation. And store for use, in addition to the initial stage It can be obtained through a few valid event data in the early stages of deployment, or temporarily set to the default value when there is no historical data.

[0041] S620, Based on the analysis results, the event discrimination threshold is optimized and adjusted; based on the analysis results, the event discrimination threshold is adaptively optimized and adjusted, wherein the adjustment range in a single instance does not exceed 20% of the original threshold; S630 uses the optimized event discrimination threshold as the preset event discrimination threshold. The optimized event discrimination threshold is written into NVM in real time as the preset threshold for subsequent event discrimination, overwriting the original threshold data.

[0042] For example, the system records historical response data every 5 minutes, including the average ΔI / Δt value, fluctuation range, and event frequency under steady-state conditions. The NVM stores this data for the controller to analyze. If the analysis finds that the ambient noise floor has increased from ±0.02μA / ms to ±0.04μA / ms, the MCU automatically... Adjust to 0.06 μA / ms; if frequent non-target weak perturbations are detected, then... Temporarily increased to 0.45 μA / ms; if the target event signal strength continues to increase, then... The value was reduced to 0.35 μA / ms, with each adjustment not exceeding 20% ​​of the original threshold.

[0043] In some embodiments, step S620, based on the analysis results, optimizes and adjusts the event discrimination threshold, specifically including: S621, when the analysis result indicates an increase in ambient background noise, the stability threshold is increased; for example, when the fluctuation range of the current change rate under static background exceeds 1.5 times the inherent noise of the device, the analysis result indicates an increase in ambient background noise, and the stability threshold δ0 is increased by no more than 20% of the original threshold to maintain the background signal suppression effect. S622, When the analysis results indicate that non-target weak perturbations occur frequently, increase the mutation threshold; for example, when the perturbation frequency is ≥10 times / minute, the analysis results indicate that non-target weak perturbations occur frequently, and the mutation threshold is increased by no more than 20% of the original threshold. This reduces the probability of false system triggering. S623, When the analysis results indicate a need to improve the detection sensitivity for target dynamic events, lower the mutation threshold. For example, if the peak value of ΔI / Δt for the target dynamic event is close to the original mutation threshold. Mean of time or event intensity Continuing below the historical average The analysis results indicate that the detection sensitivity for target dynamic events needs to be improved, that is, the mutation threshold should be reduced by no more than 20% of the original threshold. This enhances the ability to detect weak dynamic events.

[0044] For example, when the ambient background noise increases, for instance, from 0.02 μA / ms to 0.024 μA / ms, then... The threshold was increased from 0.03 μA / ms to 0.036 μA / ms. This adjustment is made because when ambient noise increases, the original stability threshold falls below the noise peak, leading to a higher false trigger rate in the static background. In this case, increasing the stability threshold reduces the false trigger rate. Conversely, when weak, non-target disturbances occur frequently, such as detecting 5 current change events per second with a rate of 0.3 μA / ms, the original abrupt change threshold is insufficient to filter out these disturbances. The sensitivity was increased from 0.4 μA / ms to 0.45 μA / ms to further suppress non-target disturbances and improve the detection rate of target events; while when it is necessary to improve the detection sensitivity of target dynamic events, such as when the intensity of the detected target event is continuously increasing, in order to improve the system response speed and prevent false triggering due to noise, the sensitivity will be increased. Reduced to 0.35 μA / ms.

[0045] In some embodiments, step S100 involves acquiring the electrical response signal output by the synaptic device within a preset time period to obtain an electrical signal sequence, specifically including: S110, acquire the electrical response signal output by the synaptic device within a preset time period; S120 amplifies and performs analog-to-digital conversion on the electrical response signal to obtain the converted signal; the current signal is input to a low-noise operational amplifier for 1000x in-phase amplification, and then the amplified analog current signal is converted into a digital converted signal by the analog-to-digital converter; S130: The converted signal is sampled at a preset sampling rate to obtain a digitized electrical signal sequence. The converted signal is sampled at equal intervals at a preset sampling rate of 1kHz to obtain a digitized electrical signal sequence within a preset time period. The sampled data is temporarily stored in the on-chip buffer of the MCU.

[0046] For example, the weak current signal output by the artificial synaptic device is first amplified by an operational amplifier, and then the amplified signal is input to an ADC (Analog-to-Digital Converter) for analog-to-digital conversion. The converted digital signal is transmitted to the MCU to form a digitized electrical signal sequence, which is used for subsequent rate of change calculation and event discrimination.

[0047] Therefore, this application characterizes dynamic events through transient response changes of artificial synaptic devices. It eliminates the need for continuous high-frequency frame-by-frame sampling and full data processing of external stimuli, instead performing low-frequency and necessary state readings of the electrical signals output by the artificial synaptic devices to obtain their temporal variation characteristics. Simultaneously, by utilizing the device's own time-dependent response characteristics, static background signals are pre-suppressed at the physical level, ultimately achieving low-power, highly robust dynamic behavior detection. Compared to existing technologies that continuously sample and process all raw signals, this application triggers subsequent discrimination and processing only when the synaptic device response undergoes a significant change, thereby significantly reducing data acquisition and computational burden, and achieving a low-redundancy, event-driven sensing mechanism.

[0048] This application also provides a dynamic event sensing system, which includes: The acquisition module 100 is used to acquire the electrical response signal output by the synaptic device within a preset time period to obtain an electrical signal sequence; The calculation module 200 is used to calculate the rate of change of the electrical signal sequence within a preset time period based on the electrical signal sequence. The comparison module 300 is used to obtain a preset event discrimination threshold, compare the rate of change with the event discrimination threshold, and obtain the comparison result. The recording module 400 is used to record the event information corresponding to the electrical response signal if the comparison result is a valid dynamic event; The generation module 500 is used to encode event information to generate event signals.

[0049] This application also provides a dynamic event sensing device, the dynamic event sensing device comprising: a memory and at least one processor, wherein the memory stores instructions; At least one of the processors invokes the instructions in the memory to cause the dynamic event sensing device to execute the various steps of the dynamic event sensing method as described above.

[0050] Figure 3 This is a schematic diagram of the structure of a dynamic event sensing device 700 provided in an embodiment of the present invention. The dynamic event sensing device 700 can vary significantly due to different configurations or performance characteristics. It may include one or more central processing units (CPUs) 710 (e.g., one or more processors) and a memory 720, and one or more storage media 730 (e.g., one or more mass storage devices) for storing application programs 733 or data 732. The memory 720 and storage media 730 can be temporary or persistent storage. The program stored in the storage media 730 may include one or more modules (not shown in the diagram), each module including a series of instruction operations on the dynamic event sensing device 700. Furthermore, the processor 710 may be configured to communicate with the storage media 730 and execute the series of instruction operations in the storage media 730 on the dynamic event sensing device 700 to implement the steps of the methods provided in the above-described method embodiments.

[0051] The dynamic event sensing device 700 may also include one or more power supplies 740, one or more wired or wireless network interfaces 750, one or more input / output interfaces 770, and / or one or more operating systems 731, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art will understand that... Figure 3 The illustrated dynamic event sensing device structure does not constitute a limitation on the electronic device and may include more or fewer components than illustrated, or combine certain components, or have different component arrangements.

[0052] The preferred embodiments of the present invention have been described in detail above, but the present disclosure is not limited to the embodiments described. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention, and these equivalent modifications or substitutions are all included within the scope defined by the claims of the present disclosure.

Claims

1. A dynamic event sensing method, characterized in that: include: The electrical response signal output by the synaptic device within a preset time period is obtained to obtain an electrical signal sequence; wherein, the synaptic device outputs a stable and low-fluctuation current response in response to static or slowly changing stimuli, and the synaptic device outputs a transient and high-amplitude current response in response to dynamic abrupt stimuli. Calculate the rate of change of the electrical signal sequence within the preset time period based on the electrical signal sequence. Obtain a preset event discrimination threshold, compare the rate of change with the event discrimination threshold to obtain the comparison result; If the comparison result is a valid dynamic event, then record the event information corresponding to the electrical response signal; Event information is encapsulated based on electrical signal sequences to generate event signals; The step of obtaining a preset event discrimination threshold and comparing the rate of change with the event discrimination threshold to obtain a comparison result specifically includes: Obtain a preset event discrimination threshold, which includes a stability threshold and a mutation threshold; When the absolute value of the rate of change is less than or equal to the stability threshold, it is determined to be a static background signal; When the absolute value of the rate of change is greater than the stability threshold and less than or equal to the mutation threshold, it is determined to be a non-target weak disturbance. When the absolute value of the rate of change is greater than the mutation threshold, it is determined to be a valid dynamic event.

2. The dynamic event sensing method according to claim 1, characterized in that: The step of calculating the rate of change of the electrical signal sequence within the preset time period specifically includes: Extract the current values ​​and corresponding time intervals of adjacent sampling points from the electrical signal sequence; The current difference between adjacent sampling points is obtained by calculating the current values ​​of adjacent sampling points; The instantaneous rate of change is calculated based on the current difference and time interval between adjacent sampling points.

3. The dynamic event sensing method according to claim 1, characterized in that: If the comparison result is a valid dynamic event, then the event information corresponding to the electrical response signal is recorded, specifically including: If the comparison result is a valid dynamic event, read the spatial coordinates of the synaptic device in the array that received the electrical response signal; Obtain the current time as the event occurrence time and associate it with spatial coordinates to obtain event information.

4. The dynamic event sensing method according to claim 1, characterized in that: The process of encapsulating event information based on electrical signal sequences to generate event signals specifically includes: Extract the peak intensity and duration of the event from the electrical signal sequence; Peak intensity, duration, and event information are integrated to obtain event characteristic parameters; The event characteristic parameters are encapsulated according to a preset protocol to generate an event signal that conforms to the preset protocol.

5. The dynamic event sensing method according to claim 1, characterized in that: After encapsulating the event information based on the electrical signal sequence to generate the event signal, the process further includes: Obtain historical response data within a preset time period; Based on historical response data, the environmental noise level is analyzed to obtain the analysis results; Based on the analysis results, the event discrimination threshold was optimized and adjusted. The optimized event discrimination threshold is used as the preset event discrimination threshold.

6. The dynamic event sensing method according to claim 5, characterized in that: The optimization and adjustment of the event discrimination threshold based on the analysis results specifically includes: When the analysis results indicate an increase in ambient background noise, the stability threshold should be increased. Increase the mutation threshold when the analysis results indicate that non-target weak perturbations occur frequently. When the analysis results indicate a need to improve the detection sensitivity for dynamic events of the target, the mutation threshold should be lowered.

7. The dynamic event sensing method according to claim 1, characterized in that: The process of acquiring the electrical response signal output by the synaptic device within a preset time period to obtain an electrical signal sequence specifically includes: Acquire the electrical response signal output by the synaptic device within a preset time period; The electrical response signal is amplified and converted from analog to digital to obtain the converted signal; The converted signal is sampled at a preset sampling rate to obtain a digitized electrical signal sequence.

8. A dynamic event sensing system, characterized in that: include: The acquisition module is used to acquire the electrical response signal output by the synaptic device within a preset time period to obtain an electrical signal sequence, wherein the synaptic device outputs a stable and low-fluctuation current response in response to static or slowly changing stimuli, and the synaptic device outputs a transient and high-amplitude current response in response to dynamic abrupt stimuli. The calculation module is used to calculate the rate of change of the electrical signal sequence within a preset time period based on the electrical signal sequence. The comparison module is used to obtain a preset event discrimination threshold and compare the rate of change with the event discrimination threshold to obtain a comparison result. Specifically, obtaining the preset event discrimination threshold and comparing the rate of change with the event discrimination threshold to obtain a comparison result includes: obtaining the preset event discrimination threshold, which includes a stability threshold and a mutation threshold; when the absolute value of the rate of change is less than or equal to the stability threshold, it is determined to be a static background signal; when the absolute value of the rate of change is greater than the stability threshold and less than or equal to the mutation threshold, it is determined to be a non-target weak disturbance; when the absolute value of the rate of change is greater than the mutation threshold, it is determined to be a valid dynamic event. The recording module is used to record the event information corresponding to the electrical response signal if the comparison result is a valid dynamic event; The generation module is used to encode event information to generate event signals.

9. A dynamic event sensing device, characterized in that: The dynamic event sensing device includes: a memory and at least one processor, wherein the memory stores instructions; At least one of the processors invokes the instructions in the memory to cause the dynamic event sensing device to perform the steps of the dynamic event sensing method as described in any one of claims 1-7.

Images