Infrared induction low-power control system based on time sequence check

Abstract

The application provides an infrared induction low-power control system based on timing verification, comprising: an event-driven sparse sensing module for generating event pulses with timestamps when changes occur and keeping zero dynamic power consumption standby when there are no changes; a time slot consistency verification module for dividing time into fixed time slots and performing timing verification on the event pulses to confirm valid human activity events and output validity flags and event rhythm characteristic indicators; and an adaptive wake-up module for combining historical wake-up results and power consumption feedback, calculating event idle intervals using the validity flags, representing signal environment complexity using the event rhythm characteristic indicators, adjusting sampling time slot width and sleep duration, and waking up the executing device; the application aims to solve the technical problems of high monitoring power consumption due to continuous sampling, high false trigger rate due to single threshold discrimination, and inability to adapt to environmental dynamic changes due to fixed sleep strategies in existing infrared induction systems.

Description

Technical Field

[0001] This invention relates to the field of low-power control technology, and in particular to an infrared sensing low-power control system based on timing verification. Background Technology

[0002] Passive infrared (PIR) sensing technology is widely used in human presence detection in fields such as smart lighting, security alarms, and building energy conservation due to its advantages such as low cost, low power consumption, and passive detection. However, as IoT sensor nodes develop towards battery-powered and wireless operation, extremely stringent requirements are placed on the standby power consumption of devices, meaning that a single node needs to work continuously for several years powered by a small-capacity lithium battery. Traditional infrared sensing control systems employ a periodic sampling, threshold judgment, and fixed-delay sleep mode. In this mode, the front-end analog circuit and analog-to-digital converter continuously acquire the sensor's output signal at a fixed frequency, and the main control unit performs threshold comparison and logic decision-making within each sampling cycle. However, the power consumption generated by continuous sampling dominates the total power consumption of the system. Even in the dead of night when there are no people or events, the sampling circuit and ADC continue to consume power. Furthermore, the system relies solely on a single threshold level to determine the trigger source, making it unable to distinguish between real human movement and environmental disturbances such as hot air flow, small animals passing by, or intermittent sunlight exposure. The false trigger rate remains high, and the invalid power consumption caused by falsely waking up the downstream lighting or communication modules is often several times that of the standby mode itself. To reduce power consumption, some existing solutions introduce a duty cycle polling mechanism, which reduces the ADC's working time by lowering the sampling frequency. However, such solutions sacrifice real-time response. When the sampling interval is too long, significant detection delays or even missed detections will occur, which are noticeable to the user. Therefore, there is an irreconcilable contradiction between low power consumption, low false triggering and real-time response in existing technologies, and there is an urgent need for a control scheme that can identify human activity with extremely low standby power consumption. Summary of the Invention

[0003] This invention provides a low-power infrared sensing control system based on timing verification, aiming to solve the technical problems of existing infrared sensing systems, such as high listening power consumption due to continuous sampling, high false trigger rate due to single threshold discrimination, and inability to adapt to dynamic environmental changes due to fixed sleep strategies.

[0004] This invention provides a timing-verified infrared sensing low-power control system, comprising: The event-driven sparse sensing module is used to generate time-stamped event pulses when the infrared radiation change gradient exceeds a preset event threshold, and maintain zero dynamic power consumption standby when there is no change. The time slot consistency verification module, which is connected to the event-driven sparse perception module, is used to divide time into fixed time slots and perform timing verification on the event pulses to confirm valid human activity events and output validity flags and event rhythm characteristic indicators. An adaptive wake-up module, connected to the time slot consistency verification module, is used to receive the validity flag and event rhythm characteristic index, and combine historical wake-up results with power consumption feedback. It uses the validity flag to calculate the event idle interval, uses the event rhythm characteristic index to characterize the signal environment complexity, adjusts the sampling time slot width and sleep duration online, and controls the wake-up of the execution device when the validity flag indicates that a valid human activity event has been confirmed.

[0005] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: This invention generates time-stamped event pulses only when the infrared radiation gradient exceeds a threshold, maintaining zero dynamic power consumption standby when there is no change, thus eliminating the continuous monitoring power consumption of periodic ADC sampling in traditional solutions from the source. It divides time into fixed time slots and performs multiple timing checks on the event pulses, using the pulse's temporal persistence, spatial arrival time difference, and human gait rhythm as triple features to jointly identify real human activities, making isolated or non-rhythmic events such as environmental thermal noise, electromagnetic interference, and pet disturbances unable to pass the check. By receiving validity flags and event rhythm characteristic indicators, it calculates the event idle interval in real time and characterizes the signal environment complexity, dynamically adjusting the sampling time slot width and sleep duration online to achieve adaptive power management with high response during busy periods and deep sleep during idle periods.

[0006] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0007] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention, but do not constitute a limitation thereof; in the drawings: Figure 1 This is a schematic diagram of the structure of an infrared sensing low-power control system based on timing verification provided by the present invention. Detailed Implementation

[0008] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention. Example 1:

[0009] This invention provides a timing-verified low-power infrared sensing control system. Please refer to [link to relevant documentation]. Figure 1 ,include: The event-driven sparse sensing module is used to generate time-stamped event pulses when the infrared radiation change gradient exceeds a preset event threshold, and maintain zero dynamic power consumption standby when there is no change. The time slot consistency verification module, which is connected to the event-driven sparse perception module, is used to divide time into fixed time slots and perform timing verification on event pulses to confirm valid human activity events and output validity flags and event rhythm characteristic indicators. The adaptive wake-up module, connected to the time slot consistency verification module, receives validity flags and event rhythm characteristic indicators. It combines historical wake-up results with power consumption feedback, uses validity flags to calculate event idle intervals, uses event rhythm characteristic indicators to characterize signal environment complexity, adjusts sampling time slot width and sleep duration online, and controls the wake-up of the execution device when the validity flag indicates a valid human activity event.

[0010] Specifically, the event-driven sparse sensing module operates at the sensing layer, continuously monitoring the output signal of the infrared sensor. It calculates the rate of change of infrared radiation over time (i.e., the gradient of infrared radiation change) in real time. Only when the absolute value of the gradient exceeds a preset event threshold will the module generate a level transition signal containing precise arrival time information (i.e., timestamp), called an event pulse. When there is no significant change in the infrared radiation field within the field of view, the analog front-end circuit of the module does not generate any dynamic power consumption and is in a standby state with zero dynamic power consumption, thus eliminating the continuous monitoring power consumption of traditional solutions from the source.

[0011] As the core of signal processing and decision-making, the time slot consistency verification module receives the time-stamped event pulse stream output from the event-driven sparse sensing module. Drawing on the concept of time division multiple access in the field of wireless communication, this module divides the continuous time axis into a series of fixed time slots of equal length and performs a multi-level timing verification process on the arriving event pulses in each time slot. This verification process ultimately produces two output signals: one is a validity flag, which clearly indicates whether a real human activity event has been detected in the current time slot window; the other is an event rhythm characteristic index, which quantifies the degree of consistency between the current infrared event pulse sequence and the human gait cycle rhythm.

[0012] The adaptive wake-up module, as the core of the system-level scheduling, connects to the output of the time slot consistency verification module. It receives validity flags and event rhythm characteristic indicators. This module does not statically sleep and wake up; instead, it uses the received validity flags to calculate the idle time elapsed since the end of the last valid event, i.e., the event idle interval. Simultaneously, it uses the received rhythm characteristic indicators to quantitatively assess the signal complexity or interference level of the current monitoring environment. Based on these real-time physical quantities, combined with feedback from the historical wake-up accuracy and overall power consumption accumulated in the system, the adaptive wake-up module dynamically adjusts two core control parameters online: the width of the sampling time slot and the sleep duration of the sensing module. Finally, this module outputs a wake-up signal, controlling the execution device to wake from low-power sleep mode and enter normal operation mode only when the received validity flag is valid; otherwise, it maintains the sleep state of the execution device.

[0013] The actuator refers to any electronic device that uses this low-power control system and includes a main control chip (MCU), a microprocessor (CPU), a load drive circuit, or a wireless communication module; its specific forms include, but are not limited to: battery-powered smart lighting fixtures, smart doorbells / locks, security alarm detectors, automatic door controllers, and occupancy sensors used for energy management in buildings.

[0014] Through the closed-loop architecture of event-driven, timing verification, and adaptive scheduling composed of the above three modules, this system only wakes up the subsequent execution device when an infrared event sequence with human-like rhythm is confirmed, while maintaining a standby state with extremely low power consumption in most targetless scenarios.

[0015] In one implementation, the event-driven sparse awareness module includes: The gradient detection unit is used to acquire the output of the infrared sensor in real time and calculate the gradient of infrared radiation change. The event pulse generation unit is used to generate an event pulse edge when the gradient exceeds the event threshold; The timestamp latch unit is used to latch the arrival time of the event pulse edge and generate a timestamp. The dynamic refractory period suppression unit, connected to the gradient detection unit, is used to suppress the generation of new event pulses within one dynamic refractory period after the generation of an event pulse; Among them, the dynamic refractory period duration According to the Instantaneous value of infrared radiation gradient at the moment the event is triggered And perform nonlinear dynamic calculations based on the recent gradient historical distribution characteristics: In the formula, The maximum refractory period constant. For the first Instantaneous value of the infrared radiation change gradient at the moment the event is triggered. The voltage or current signal output by the infrared sensor. To find the maximum value operator, This represents the median of the absolute values ​​of the gradients at the trigger times of the last M events. This is the preset minimum gradient noise basis constant.

[0016] Specifically, the gradient detection unit is the signal input front end, which continuously receives the analog voltage or current signal output by the infrared sensor. Then, by performing time differentiation, the gradient representing the rate of change of the infrared radiation field energy is obtained. The infrared sensors mentioned here include, but are not limited to, pyroelectric infrared sensors, thermopile sensors, or other uncooled infrared detectors. The event pulse generation unit is connected to the output of the gradient detection unit and is used to perform threshold comparison. When the absolute value of the detected infrared radiation change gradient exceeds a preset reference value (i.e., the event threshold), the unit generates a fast level transition (a transient transition from low to high or from high to low) on the output line. This level transition is the event pulse edge. The event pulse edge not only marks the occurrence of the event but also serves as the trigger signal for the subsequent timestamp latch unit. Unlike the periodic analog-to-digital conversion based on a fixed sampling rate in the prior art, the event pulse generation unit is an asynchronous logic driven by signal change. There is no transition when there is no change, so no dynamic power consumption is generated. The timestamp latch unit is implemented by a high-resolution timer. Its input is connected to the output of the event pulse generation unit. Whenever the edge of the event pulse arrives, the unit immediately captures and stores the current count value of the timer. This count value is the timestamp with clock precision, recording the moment when the infrared radiation change event occurred. The dynamic refractory period suppression unit is used to prevent a series of dense invalid event pulses generated during a single infrared change event caused by real human movement, due to the tiny jitter of the signal near the threshold. This avoids subsequent processing being overwhelmed by high-frequency oscillation signals and consuming unnecessary computational energy. Its working principle simulates a brief refractory period experienced by biological neurons after an excitation pulse; that is, after each valid event pulse is generated, the suppression unit forcibly shuts down the output of the event pulse generation unit for a duration of [duration missing]. ; exist In the calculation formula, It is the absolute value of the gradient of the current i-th triggered event, representing the drastic degree of signal change; It is the median of the absolute values ​​of the gradients of the most recent M triggering events, representing the average fluctuation level of the recent environmental infrared signal; if the current trigger is caused by violent human movement, Much larger The ratio term increases. This will significantly shorten the time, allowing the system to capture continuous, rapid events that may occur during human movement with high sensitivity; if the current trigger is caused by minor environmental thermal noise, Approaching or less , will to Approximation provides long-term event suppression and filters out noise oscillations; It is a preset, extremely small constant, whose function is to prevent A calculation error occurs when the value is zero; This is a preset maximum refractory period constant, representing the longest suppression time of the system under the most unfavorable signal-to-noise ratio.

[0017] In one implementation, the time slot consistency verification module includes: The time slot timing unit is used to generate time slot boundary signals of fixed time length; The feature extraction unit is used to extract pulse features and form an event feature vector for each time slot if an event pulse occurs. The multi-slot continuous matching and verification unit is used to match the event feature vectors of multiple consecutive time slots with the preset human activity template and output a matching pass flag. The spatial cross-verification unit is used to verify the time difference of arrival of event pulses from at least two infrared sensors and output a spatial valid flag. The biological rhythm time window verification unit is used to determine whether the pulse interval of the event sequence that has passed the multi-slot continuous matching verification conforms to the human gait rhythm, and output the event rhythm characteristic index. The validity flag generation unit is used to receive the matching pass flag, the spatial validity flag, and the event rhythm feature index, and combine them to generate the validity flag.

[0018] Specifically, the time-slot timing unit consists of an independently operating timer that is synchronized with the system's master clock. It generates periodic time-slot boundary signals, uniformly dividing the continuous time axis into segments of duration [missing information]. A fixed time slot, the boundary of which is the time reference for all subsequent verification operations, that is, a signal evaluation is performed uniformly at the end of each time slot; The feature extraction unit listens for event pulses from the event-driven sparse sensing module within each time slot window. If an event pulse arrives in the current time slot, the unit extracts its key features, including but not limited to: the pulse amplitude (or intensity level), the pulse duration (pulse width), and the sensor channel identifier from which the pulse originates. These features are organized into a multidimensional array, which is the event feature vector for that time slot. If there are no events in a time slot, the eigenvector of that time slot is denoted as the zero vector. The multi-slot continuous matching verification unit, the spatial mutual verification unit, and the biorhythm time window verification unit constitute three independent judgment channels, which respectively identify the authenticity of the event stream from three dimensions: temporal persistence, spatial consistency, and biorhythm; each unit will output an independent Boolean value or numerical flag. The validity flag generation unit is a logic combinational circuit or an equivalent software decision maker. It receives the matching pass flag, spatial validity flag and event rhythm characteristic index as input, and performs AND / OR combination operations according to preset logic rules to finally generate a single validity flag.

[0019] In one implementation, the multi-slot continuous matching verification unit is configured as follows: Only when consecutive When the event feature vectors of each time slot match the human activity template, the matching is marked as valid, and the event sequence is determined to be a pre-valid sequence. Among them, for the first Each time slot is used for matching and determination based on the event feature vector of that time slot. With template vector The vector product and the normalized similarity measure of their respective magnitudes : In the formula Representing vectors and The inner product, The L2 norm of a vector. For a preset minimum positive number; if If the time slot matches, then the time slot is considered to be a match; otherwise, it is not a match. This is a preset similarity threshold.

[0020] Specifically, when a real human body passes through an infrared detection area, the resulting change in infrared radiation is not instantaneous but continues for multiple time slots. Pulses generated by electromagnetic interference, instantaneous hot air flow fluctuations, etc., are often isolated and transient. Therefore, the multi-time slot continuous matching verification unit maintains a sliding window of length N (N is a positive integer greater than 1, such as N=3). Only when the feature vector of each frame in the N consecutive time slots entering the sliding window is determined to be "matched" will the unit output a valid match pass flag and mark the event sequence as a pre-valid sequence for subsequent verification. If any frame mismatches, the window is immediately cleared, all accumulated event data is discarded, and no output is generated, thus filtering out most non-persistent environmental interferences with extremely low computational cost. For matching determination of a single time slot, a normalized similarity metric was used. The calculation method; this formula calculates the measurement vector. With preset template vector Robust variant of the cosine of the included angle, inner product This reflects the consistency of the two vectors in direction and magnitude. The product of the magnitudes in the denominator has been normalized, making... Theoretically, the value range of is between 0 and 1. The closer to 1, the better the match; template vector The database stores standard characteristic parameters of the infrared event pulses that a typical human body should generate within a single time slot. These are fixed vectors preset based on the mechanical / optical relationships such as the physical size of the human target, the range of movement speed, and the sensor's field of view; similarity threshold. Using the same preset constant, such as 0.7, this method of judgment can more accurately express the degree of conformity between the event and the target compared to simple binary level comparison.

[0021] It should be noted that, and The feature vector dimension is at least three, that is... =[ , [,1], where The normalized amplitude of the event pulses within this time slot. To normalize the pulse width, the third dimension is fixed as a constant 1; =[ , , ],in , The preset normalized amplitude and width constants for the human body activity template. The preset offset constant is preferably set to 0.5. The vector dimension and feature types are not limited to the examples above. Depending on the actual sensor characteristics, they can also be extended to: pulse rising edge slope (reflecting the steepness of infrared changes), sensor channel number (indicating the source in a multi-sensor configuration), number of event pulses (the number of event pulses in this time slot), etc.

[0022] In one implementation, the spatial cross-check unit is configured as follows: When it includes at least a first infrared sensor Second infrared sensor At the same time, extract data from the sensor within the same time slot. and Calculate the arrival time difference between the two event pulses based on their timestamps. ; Using the physical spacing of the sensors Construct illegal event exclusion criteria based on preset physical motion speed boundaries. : In the formula, and These are the preset minimum and maximum equivalent infrared radiation source moving velocity constants corresponding to the physical characteristics of human movement; only when When = 0, the space validity flag is set to valid; when When the value is greater than 0, the space validity flag is set to invalid and the relevant event data is discarded.

[0023] Specifically, in common application scenarios, a real human body is a heat source with a certain volume and a moving speed within a known range (e.g., walking speed of 0.8 m / s to 2.0 m / s); when this human body passes successively between two points with a distance of... When the sensor's field of view is within the range of two sensors, the time difference between the arrival of infrared events at the two sensors is considered. It must be in a state of being arrive Within the physically reasonable window; if it is a large area of ​​hot air introduced by air conditioning, open doors and windows, or the illumination of vehicle headlights from a distance, because the equivalent infrared radiation source moves at a very fast speed (approaching the speed of light or airflow speed) or very slow speed (such as sunlight and shadow), the time difference between the two sensor events caused by it will fall outside this reasonable window. This unit constructs an illegal event exclusion index based on boundary constraint product calculation. Only when It falls exactly within the closed interval. , When both factors are zero, the product Z is zero, and the event is passed; if Greater than the upper limit (If the movement is too slow, or the target is not human), then the first factor is positive. Less than the lower limit (If the movement is too fast and the target is not human), then the second factor is positive; if any factor is positive, Z will be greater than zero, and the event will be judged as "illegal" and discarded. and This is a preset constant derived from publicly available human kinematics research data; its value is not limited in this embodiment.

[0024] In one implementation, the circadian rhythm time window verification unit is configured as follows: After receiving the prepared valid sequence, continue recording subsequent data. The arrival time of each event pulse is calculated, and the interval sequence between adjacent pulses is determined. ; Calculating rhythm violation index based on interval sequence and will Output as an indicator of event rhythm characteristics: In the formula, and These represent the maximum and minimum values ​​of the adjacent pulse interval sequence, respectively. It is the arithmetic mean of the intervals between adjacent pulses. Represents the first pulse in the sequence of adjacent pulse intervals. One element; like If so, the pulse sequence is determined to have rhythmic stability consistent with human gait; This is the preset rhythm violation threshold.

[0025] Specifically, human movement in space is not uniform and continuous, but rhythmic movement constrained by gait. The infrared radiation heat source of the human torso will generate a major infrared change peak with each step, forming a gait cycle of about 1.0 to 2.0 seconds. Environmental disturbances such as swaying curtains, random running of pets, and periodic fluctuations in equipment heat dissipation are difficult to simulate this stable long-term rhythm. This unit uses this deep biological characteristic to distinguish the human body from disturbances. The workflow of this unit consists of two steps: The first step is data acquisition. After receiving the pre-valid sequence output by the multi-slot continuous matching and verification unit, it opens a time window to continuously record the precise arrival time of the next K event pulses from the event-driven sparse sensing module, and calculates the time difference between every two adjacent pulses in these K pulses to form a pulse interval sequence. ; The second step is to calculate the rhythm violation index. and will Output as an indicator of event rhythm characteristics; The formula consists of the product of two factors and is a composite index for measuring the rhythmicity of a sequence. First Factor As relative dispersion, it characterizes the ratio of the range of fluctuation between the maximum and minimum values ​​of a human's gait cycle within the observation window to its average cycle. For a steadily walking human body, this factor value is very small. The second factor is an additive term. The addition from arrive It analyzes the first difference of the interval sequence, that is, the jump variables between two adjacent intervals. It is multiplied by the logarithm of its normalized value. It not only measures the absolute change in the interval, but also uses an information content function. It amplifies those non-periodic, chaotic jumps (high entropy) while relatively suppressing the inherent smooth gradations of periodic signals; this means that for rhythmically stable human gait, the interval jumps are small and regular, and the factor is small, but for irregular random disturbances, the interval jumps are chaotic, and the factor will increase exponentially. Ultimately, the product of the two factors It comprehensively measures the degree of rhythm violation in a sequence in terms of both overall dispersion and local jump disorder. The lower the value, the closer the event flow is to the stable biological rhythms of the human body; while the preset threshold It is a constant determined experimentally, when When the value is less than this threshold, the system can determine with high confidence that the current event sequence originates from real human activity; in a preferred embodiment, =0.10, this value can maintain a pass rate of ≥95% for the rhythmic signals of normal adult walking in typical office and home scenarios, while controlling the false trigger rate of pets, fans and airflow to ≤3%.

[0026] In one implementation, the validity flag generation unit is configured as follows: If and only if the match passes the flag, the spatial flag is valid, and the event rhythm feature index is valid, then the match is valid. When the event occurs, the validity flag is set to valid, confirming that the current event is a valid human activity event; When the match passes through an invalid flag, or the space is valid but the flag is invalid, or When this happens, the validity flag will be set to invalid.

[0027] Specifically, this unit performs a strict "logical AND" operation, meaning it is only valid under the following conditions: persistence (match pass flag is valid), spatial condition (spatial validity flag is valid), and rhythmic condition (rhythmic violation degree). Below the threshold Only when all three conditions are met simultaneously will the system generate a valid validity flag to confirm the detection of a real and valid human activity event. Conversely, if any one of the three conditions is not met, the validity flag will be set to invalid, meaning that the event sequence is considered a false signal generated by environmental interference and is completely filtered out. This redundant decision mechanism enables the false trigger rate of the system to be reduced by orders of magnitude, because different types of interference will be blocked by different checkpoints. For example, short-time electromagnetic pulses will be filtered out in the first checkpoint (multi-timeslot matching) and will not affect the subsequent two checkpoints.

[0028] In one implementation, the adaptive wake-up module includes: An idle time slot counter is used to receive validity flags and, while the validity flag remains invalid, record the number of idle time slots for which no valid human activity events are detected consecutively. ; An event interval estimator receives the flip-off time of the validity flag and calculates the reciprocal of the time interval between the most recent adjacent valid human activity events as the average event interval. The estimate; The scheduling decision unit is used to receive validity flags and event rhythm characteristic indicators. Number of idle time slots and average event interval Calculate the sleep duration of the next event-driven sparse sensing module. Width of a fixed time slot It also controls the execution device to wake up when the validity flag is valid.

[0029] Among them, the scheduling decision unit calculates the sleep duration. The formula is: In the formula, This is the minimum sleep duration constant. Based on the basic time slot width constant; Calculate the width of a fixed time slot The formula is: when The higher the value, the more the rhythm of the current verification signal deviates, and the wider the time slot is to enhance capture; when The value continues to increase, indicating that it is in an idle state for a long time. The time slot width shrinks towards the base value to reduce the energy consumption of single-cycle monitoring, while the sleep duration increases non-linearly.

[0030] Specifically, the idle time slot counter is an incremental counter that is synchronized with the system's time slot clock. One end of the counter is connected to the validity flag signal. Its operation is as follows: at the end of each time slot, the validity flag status is checked; if it is invalid, the count value is reset. Increment by 1; once the validity flag flips to valid, the count value is immediately reset to zero; therefore, The value directly represents the length of continuous idle time since the end of the last valid event, i.e., the event idle interval; The event interval estimator is used to learn the historical activity of a scene. It captures the flip point when the validity indicator changes from invalid to valid, calculates the time difference between two adjacent valid events, and uses simple filtering methods such as moving average or exponentially weighted moving average to obtain an estimate of the average interval of valid events in the scene. ; The size reflects whether the environment is lively or quiet; The scheduling decision unit integrates real-time physical quantities (validity flags, etc.) from the front-end verification module. ), and historical feature quantities statistically derived from within this module ( , Based on a preset nonlinear deterministic formula, closed-form calculations are performed, ultimately outputting two control variables: sleep duration. and sampling time slot width Meanwhile, this unit is also responsible for directly managing the wake-up of the execution device based on the validity flag status, that is, only when the validity flag is valid will a wake-up signal be sent to the main control chip or load circuit of the execution device. hibernation duration In the calculation formula, the molecule The denominator represents the total duration of continuous idle time. The ratio of the two represents the inherent average active interval of the scenario, indicating how many times the current cumulative idle time has increased compared to the usual average waiting time. This ratio is squared, meaning that the increase in sleep depth is non-linear; once the system enters a late-night or prolonged period of inactivity, This will continue to accumulate, and the ratio and its squared term will increase dramatically, driving the dormancy period. Far exceeding The system automatically enters an extended sleep mode, significantly reducing power consumption; and when someone occasionally appears, Reset to zero It immediately fell back to To restore rapid response capabilities; Sampling time slot width The calculation formula is a mathematical model reflecting dual-factor adaptive adjustment, which relaxes restrictions during busy periods to prevent missed detections and tightens restrictions during idle periods to save energy; the formula is based on the combined effect of two multiplicative factors. Based on event rhythm characteristic indicators Adjustments are made to the rhythm deviation of recent events when the environment is harsh and interference is significant. This will increase sharply, with the magnitude of this factor exceeding 1 becoming larger, leading to wider time slots. This allows more event information to be captured within each time slot, improving the fault tolerance and capture rate of the verification, and preventing missed wake-ups due to interference; factor Adjustments are made based on the proportion of idle time. This factor is the complement of an exponentially decaying function with the natural logarithm as the base. When activity is frequent, When the value is very small, the exponential term is close to 1, and the factor is close to 0, making the entire... By drastically compressing the base value to an extremely small level, the highest sampling accuracy and response speed are achieved; when the space is idle for a long period of time... As it continues to increase, the exponential term tends towards 0, and the factor self-saturates and tends towards 1, eventually leading to... Completely by The decision maintains a reasonable sampling window without needlessly expanding the time slots and wasting energy.

[0031] In one implementation, when the execution device deploys only a single infrared sensor, the spatial cross-verification unit is configured in bypass mode, directly fixing the spatial validity flag as valid.

[0032] Specifically, this embodiment defines a fault-tolerance mechanism for the system in a simplified application scenario. Under installation conditions with very low cost or limited physical space, the execution device may only be equipped with a single infrared sensor. In this case, the dual-channel signal input required for the spatial cross-check unit to calculate the arrival time difference is missing. To ensure the integrity of the overall system logic, the spatial cross-check unit is equipped with a pattern detection mechanism or preset switch. When it detects that there is only one valid sensor input, it will automatically switch to bypass state and directly output a valid spatial valid flag. The logical decision-making power is completely handed over to multi-timeslot continuous matching and biological rhythm time window verification. This maintains the unity of the system architecture and enables it to work normally under a single sensor configuration.

[0033] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A low-power infrared sensing control system based on timing verification, characterized in that, include: The event-driven sparse sensing module is used to generate time-stamped event pulses when the infrared radiation change gradient exceeds a preset event threshold, and maintain zero dynamic power consumption standby when there is no change. The time slot consistency verification module, which is connected to the event-driven sparse perception module, is used to divide time into fixed time slots and perform timing verification on the event pulses to confirm valid human activity events and output validity flags and event rhythm characteristic indicators. An adaptive wake-up module, connected to the time slot consistency verification module, is used to receive the validity flag and event rhythm characteristic index, and combine historical wake-up results with power consumption feedback. It uses the validity flag to calculate the event idle interval, uses the event rhythm characteristic index to characterize the signal environment complexity, adjusts the sampling time slot width and sleep duration online, and controls the wake-up of the execution device when the validity flag indicates that a valid human activity event has been confirmed.

2. The infrared sensing low-power control system based on timing verification according to claim 1, characterized in that, The event-driven sparse sensing module includes: The gradient detection unit is used to acquire the output of the infrared sensor in real time and calculate the gradient of infrared radiation change. An event pulse generation unit is used to generate an event pulse edge when the gradient exceeds the event threshold; The timestamp latch unit is used to latch the arrival time of the event pulse edge and generate a timestamp; A dynamic refractory period suppression unit, connected to the gradient detection unit, is used to suppress the generation of a new event pulse within one dynamic refractory period after the event pulse is generated; Among them, the dynamic refractory period duration According to the Instantaneous value of infrared radiation gradient at the moment the event is triggered And perform nonlinear dynamic calculations based on the recent gradient historical distribution characteristics: In the formula, The maximum refractory period constant. For the first Instantaneous value of the infrared radiation change gradient at the moment the event is triggered. The voltage or current signal output by the infrared sensor. To find the maximum value operator, This represents the median of the absolute values ​​of the gradients at the trigger times of the last M events. This is the preset minimum gradient noise basis constant.

3. The infrared sensing low-power control system based on timing verification according to claim 1, characterized in that, The time slot consistency verification module includes: The time slot timing unit is used to generate time slot boundary signals of fixed time length; The feature extraction unit is used to extract pulse features and form an event feature vector for each time slot if an event pulse occurs. The multi-slot continuous matching and verification unit is used to match the event feature vectors of multiple consecutive time slots with the preset human activity template and output a matching pass flag. The spatial cross-verification unit is used to verify the time difference of arrival of event pulses from at least two infrared sensors and output a spatial valid flag. The biological rhythm time window verification unit is used to determine whether the pulse interval of the event sequence that has passed the multi-slot continuous matching verification conforms to the human gait rhythm, and output the event rhythm characteristic index. The validity flag generation unit is used to receive the matching pass flag, the spatial validity flag, and the event rhythm feature index, and combine them to generate the validity flag.

4. The infrared sensing low-power control system based on timing verification according to claim 3, characterized in that, The multi-slot continuous matching verification unit is configured as follows: Only when consecutive When the event feature vectors of each time slot match the human activity template, the match is set to valid by setting a flag, and the event sequence is determined to be a pre-valid sequence; Among them, for the first Each time slot is used for matching and determination based on the event feature vector of that time slot. With template vector The vector product and the normalized similarity measure of their respective magnitudes : In the formula Representing vectors and The inner product, The L2 norm of a vector. For a preset minimum positive number; if If the time slot matches, then the time slot is considered to be a match; otherwise, it is not a match. This is a preset similarity threshold.

5. The infrared sensing low-power control system based on timing verification according to claim 3, characterized in that, The spatial cross-verification unit is configured as follows: When it includes at least a first infrared sensor Second infrared sensor At the same time, extract data from the sensor within the same time slot. and Calculate the arrival time difference between the two event pulses based on their timestamps. ; Using the physical spacing of the sensors Construct illegal event exclusion criteria based on preset physical motion speed boundaries. : In the formula, and These are the preset minimum and maximum equivalent infrared radiation source moving velocity constants corresponding to the physical characteristics of human movement; only when When = 0, the space validity flag is set to valid; when When the value is greater than 0, the space validity flag is set to invalid and the relevant event data is discarded.

6. The infrared sensing low-power control system based on timing verification according to claim 4, characterized in that, The biological rhythm time window verification unit is configured as follows: After receiving the prepared valid sequence, continue recording subsequent data. The arrival time of each event pulse is calculated, and the interval sequence between adjacent pulses is determined. ; Calculate the rhythm violation index based on the interval sequence. and will As an output of the event rhythm characteristic index: In the formula, and These are the maximum and minimum values ​​of the adjacent pulse interval sequence, respectively. It is the arithmetic mean of the adjacent pulse interval sequence. Indicates the first pulse in the adjacent pulse interval sequence One element; like If so, the pulse sequence is determined to have rhythmic stability consistent with human gait; This is the preset rhythm violation threshold.

7. The infrared sensing low-power control system based on timing verification according to claim 3, characterized in that, The validity flag generation unit is configured as follows: The matching is valid if and only if the matching flag is valid, the spatial validity flag is valid, and the event rhythm feature index is valid. When the validity flag is set to valid, the current event is confirmed as a valid human activity event. When the match is invalid via the flag, or the space valid flag is invalid, or When this happens, the validity flag is set to invalid.

8. The infrared sensing low-power control system based on timing verification according to claim 1, characterized in that, The adaptive wake-up module includes: An idle time slot counter is used to receive the validity flag and, while the validity flag remains invalid, record the number of idle time slots for which no valid human activity event is detected consecutively. ; An event interval estimator receives the flip-off time of the validity flag and calculates the reciprocal of the time interval between the most recent adjacent valid human activity events as the average event interval. The estimate; The scheduling decision unit is used to receive the validity flag and the event rhythm characteristic index. The number of idle time slots and the average event interval Calculate the next sleep duration of the event-driven sparse sensing module. Width of the fixed time slot And control the execution device to wake up when the validity flag is valid.

9. The infrared sensing low-power control system based on timing verification according to claim 8, characterized in that, The scheduling decision unit calculates the sleep duration. The formula is: In the formula, This is the minimum sleep duration constant. Based on the basic time slot width constant; Calculate the width of the fixed time slot. The formula is: when The higher the value, the more the rhythm of the current verification signal deviates, and the wider the time slot is to enhance capture; when The value continues to increase, indicating that it is in an idle state for a long time. The time slot width shrinks towards the base value to reduce the energy consumption of single-cycle monitoring, while the sleep duration increases non-linearly.

10. The infrared sensing low-power control system based on timing verification according to claim 3, characterized in that, When the execution device is equipped with only a single infrared sensor, the spatial cross-verification unit is configured in bypass mode, and the spatial validity flag is directly fixed as valid.

Images