Self-adaptive UART (Universal Asynchronous Receiver / Transmitter) awakening method and system based on idle period noise feature learning

Abstract

The invention discloses a self-adaptive UART (Universal Asynchronous Receiver / Transmitter) awakening method and system based on idle period noise feature learning, and the method comprises the steps: periodically triggering and sampling noise data on a receiving pin in an idle period of a UART bus; extracting noise features from the noise data; based on the noise features, using a pre-trained machine learning model to dynamically generate wake-up detection parameters; the wake-up detection parameters are configured to a first-stage circuit, and a pre-wake-up signal is generated; uART start bit protocol verification is carried out through the second-stage circuit. According to the invention, through organic combination of environment perception, intelligent decision and dynamic adjustment, the fundamental problem of insufficient adaptability of a traditional UART wake-up technology in a dynamic electromagnetic environment is successfully solved, and remarkable advantages are shown in the aspects of environment adaptability, system reliability, power consumption optimization, technical perspectiveness and the like; the method has an important practical value and a wide application prospect.

Description

Technical Field

[0001] This invention relates to the field of communication technology for low-power electronic devices, and specifically to an adaptive UART wake-up method and system based on idle period noise feature learning. Background Technology

[0002] With the rapid development of the Internet of Things (IoT) and wireless sensor networks, low-power electronic devices are placing higher demands on the energy efficiency and reliability of communication interfaces. The UART (Universal Asynchronous Receiver / Transmitter) interface, due to its simplicity and reliability, is widely used for asynchronous serial communication between devices. In low-power applications, devices often enter deep sleep mode to save energy, requiring a start bit signal on the UART receive pin to wake the system. While this mechanism effectively reduces standby power consumption, in real-world industrial, automotive, or complex home environments, the UART wake-up circuit faces significant electromagnetic interference challenges.

[0003] Current UART wake-up technology primarily employs a fixed-parameter detection strategy. This static design approach fundamentally contradicts the dynamically changing electromagnetic environment. Electromagnetic interference introduces random background noise pulses onto the receiving pins. This noise is easily misinterpreted as a valid start bit, leading to frequent "false wake-ups," which not only consumes valuable battery power but may also cause system logic malfunctions. More seriously, fixed parameters cannot meet the performance requirements of different environments: high-sensitivity parameters set in quiet environments can cause a surge in false wake-up rates in noisy environments; while conservative parameters set for interference resistance may reduce the sensitivity to valid signals, resulting in missed wake-ups.

[0004] The industry has proposed various technical solutions to improve the reliability of UART wake-up. The most basic is the comparator-based solution with hardware hysteresis. This solution uses a comparator with Schmitt trigger characteristics, suppressing signal jitter through a fixed falling edge trigger threshold and hysteresis voltage. Although this solution is widely introduced in digital circuit textbooks, its parameters are determined by the physical characteristics of the hardware and cannot be adjusted once manufactured, making it unable to adapt to changing environmental noise. A more advanced solution is the pulse width verification solution based on digital sampling. This solution adds software or hardware state machines to pulse width discrimination after hardware detection, filtering out brief noise pulses through a preset minimum effective pulse width. This solution is implemented in many microcontrollers, but its parameter settings rely on the developer's experience and judgment, and it cannot distinguish between valid start bits and continuous noise that occasionally meets the pulse width requirements.

[0005] In recent years, some high-end communication chips have begun to offer software-configurable static parameter schemes, allowing engineers to select sensitivity levels based on the expected environment. While this approach provides some flexibility, its configuration remains static and cannot be dynamically adjusted according to changes in the actual environment during device operation. The fundamental flaw shared by all these existing solutions lies in their open-loop architecture: the system lacks the ability to perceive real-time environmental noise, and its decision-making is disconnected from the current environmental state, making it difficult to maintain optimal wake-up performance in complex and changing electromagnetic environments.

[0006] A thorough analysis of the technical shortcomings of existing technologies reveals three main root causes: First, the inherent static nature of the technical path—hardware parameters are fixed by the characteristics of physical devices, and software parameters are encoded as constants, neither of which can be dynamically adjusted during runtime; second, the open-loop nature of the system design—lacking the three key components of an intelligent adaptive system: environmental perception, feature analysis, and strategy optimization; and finally, the pre-existing and empirical basis of optimization—parameter optimization occurs during product development rather than runtime, failing to achieve the transition from "globally static presets" to "locally dynamic adaptations." These limitations render existing UART wake-up technology ineffective in the face of dynamically changing electromagnetic environments, urgently requiring an adaptive solution capable of autonomously sensing environmental noise characteristics and adjusting detection parameters in real time.

[0007] This technological dilemma is particularly pronounced in the context of the large-scale deployment of IoT devices. With the diversification of device application scenarios, devices of the same model may be deployed in completely different electromagnetic environments, from quiet laboratories to noisy industrial workshops. A wake-up circuit with fixed parameters cannot guarantee reliability in all scenarios. This not only increases the complexity of device debugging but also limits the application of low-power communication technologies in a wider range of fields. Therefore, developing a UART wake-up technology that can intelligently adapt to environmental changes has become a crucial issue that urgently needs to be addressed in the field of low-power electronic device communication. Summary of the Invention

[0008] The technical problem to be solved by this invention is to provide an adaptive UART wake-up method and system based on idle period noise feature learning, which addresses the above-mentioned problems of the prior art. By organically combining environmental perception, intelligent decision-making and dynamic adjustment, it successfully solves the fundamental problem of insufficient adaptability of traditional UART wake-up technology in dynamic electromagnetic environments, and shows significant advantages in terms of environmental adaptability, system reliability, power consumption optimization and technological foresight.

[0009] To solve the above-mentioned technical problems, the technical solution adopted by the method of the present invention includes the following steps: Step S1: During the UART bus idle period, periodically trigger the sampling of noise data on the receive pin; Step S2: Extract noise features from the noise data, including voltage peak value, voltage RMS value, and negative pulse density; Step S3: Based on the noise features, dynamically generate wake-up detection parameters using a pre-trained machine learning model, including dynamic voltage threshold and dynamic minimum pulse width; Step S4: Configure the wake-up detection parameters to the first-level circuit, namely the adaptive pre-wake-up detector. This circuit includes a programmable threshold comparison module, a configurable digital filtering module, and an enable gating module, which are used to perform amplitude and pulse width determination on the input signal. A pre-wake-up signal is generated only when the signal simultaneously meets the current dynamic voltage threshold and the dynamic minimum pulse width. Step S5: The pre-wake signal is verified by the second-level circuit, namely the protocol verification wake-up device, using the UART start bit protocol. If the verification is successful, a global wake-up signal is output; otherwise, the event is recorded and the device returns to sleep state. Steps S1 to S3 are executed by the intelligent control unit, which works in conjunction with the first-level circuit and the second-level circuit to form a closed-loop adaptive process. As a further improvement of the present invention, the periodic trigger sampling in step S1 is implemented by a timer or event-driven method, and the analog-to-digital converter is started to collect noise data after confirming that the UART bus is idle, so as to ensure that the data comes from the communication interval.

[0010] As a further improvement of the present invention, the extraction of noise features in step S2 further includes calculating the statistical features of the noise and inputting the feature vector into the machine learning model; the machine learning model is a lightweight gradient boosting decision tree or a micro neural network, used to map the noise features to the optimal detection parameters.

[0011] As a further improvement of the present invention, the dynamic generation of wake-up detection parameters in step S3 includes: calculating the voltage threshold and minimum pulse width in real time based on noise characteristics, and writing the parameters into the configuration register of the first-level circuit, namely the adaptive pre-wake-up detector, through a digital bus; the dynamic voltage threshold can be adjusted within the logic low level range, and the dynamic minimum pulse width is adjustable in units of bit time.

[0012] As a further improvement of the present invention, the specific operation of the first-stage circuit, namely the adaptive pre-wake detector, in step S4 includes: The programmable threshold comparison module converts the digital code value into an analog reference voltage through a digital-to-analog converter, which is used as a dynamic voltage threshold and compared with the receiving pin signal. The configurable digital filtering module verifies pulse width using a counter. The target value of the counter corresponds to the dynamic minimum pulse width, and the signal is only filtered if the duration of the low level exceeds this value. The enable gating module receives the filtered output and the system enable signal, and controls the generation of the pre-wake-up signal.

[0013] As a further improvement of the present invention, the protocol verification in step S5 includes resampling the signal at the UART communication baud rate to verify whether the duration of the low-level pulse is strictly equal to one bit time; the second-level circuit, namely the protocol verification wake-up device, is activated only when a pre-wake-up signal is received to reduce power consumption.

[0014] As a further improvement of the present invention, a learning optimization step is also included: after successful wake-up, the intelligent control unit continues to collect noise data and update the machine learning model parameters for optimization of the initial detection parameters of subsequent sleep cycles.

[0015] This invention also proposes an adaptive UART wake-up system based on idle period noise feature learning to implement the above-mentioned adaptive UART wake-up method based on idle period noise feature learning, including the following steps: The first-stage circuit, namely the adaptive pre-wake detector, is an adaptive pre-wake detector with dynamically configurable parameters. It includes a programmable threshold comparison module, a configurable digital filtering module, and an enable gating module, which are used to pre-detect the UART receive pin signal according to dynamic parameters. The second-level circuit, namely the protocol verification wake-up unit, is used to verify the UART start bit protocol of the pre-wake-up signal generated by the first-level circuit. The intelligent control unit is used to periodically sample noise data during the UART bus idle period, extract noise features, dynamically generate detection parameters through machine learning models, and configure them to the first-level circuit, namely the adaptive pre-wake detector. The first-level circuit, namely the adaptive pre-wake detector, the second-level circuit, namely the protocol verification wake-up device, and the intelligent control unit are interconnected to form a closed-loop adaptive system, enabling autonomous perception of environmental noise and parameter adjustment.

[0016] As a further improvement to the system of the present invention, the intelligent control unit includes: The sampling module acquires noise data through an ADC and triggering logic; The feature extraction module is used to calculate the voltage peak value, the effective voltage value, and the negative pulse density; The model inference module integrates a pre-trained machine learning model and outputs dynamic voltage thresholds and dynamic minimum pulse widths. The parameter update module writes parameters into the first-stage circuit, namely the adaptive pre-wake detector, via the bus.

[0017] The present invention also proposes a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described adaptive UART wake-up method based on idle period noise feature learning.

[0018] Compared with the prior art, the advantages of the present invention are as follows: This invention successfully solves the fundamental problem of insufficient adaptability of traditional UART wake-up technology in dynamic electromagnetic environments by organically combining environmental perception, intelligent decision-making, and dynamic adjustment. Compared with existing technologies, this solution demonstrates significant advantages in environmental adaptability, system reliability, power consumption optimization, and technological foresight. These advantages are not only reflected in the improvement of technical parameters, but more importantly, they provide a brand-new technical path for the reliable operation of low-power electronic devices in complex environments. It has important practical value and broad application prospects, marking the transition of UART wake-up technology from the static and fixed era to the intelligent and adaptive era, and providing strong technical support for the large-scale reliable deployment of IoT devices. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the method flow of an embodiment of the present invention.

[0020] Figure 2 This is a system structure framework diagram of an embodiment of the present invention.

[0021] Figure 3 This is a schematic diagram of the first-stage circuit in an embodiment of the present invention. Detailed Implementation

[0022] The present invention will be further described below with reference to the accompanying drawings and specific preferred embodiments, but this does not limit the scope of protection of the present invention.

[0023] The technical solution adopted in this embodiment includes the following steps: Step S1: During the UART bus idle period, periodically trigger the sampling of noise data on the receive pin; Step S2: Extract noise features from the noise data, including voltage peak value, voltage RMS value, and negative pulse density; Step S3: Based on the noise features, dynamically generate wake-up detection parameters using a pre-trained machine learning model, including dynamic voltage threshold and dynamic minimum pulse width; Step S4: Configure the wake-up detection parameters to the first-level circuit, namely the adaptive pre-wake-up detector. This circuit includes a programmable threshold comparison module, a configurable digital filtering module, and an enable gating module, which are used to perform amplitude and pulse width determination on the input signal. A pre-wake-up signal is generated only when the signal simultaneously meets the current dynamic voltage threshold and the dynamic minimum pulse width. Step S5: The pre-wake signal is verified by the second-level circuit, namely the protocol verification wake-up device, using the UART start bit protocol. If the verification is successful, a global wake-up signal is output; otherwise, the event is recorded and the device returns to sleep state. Steps S1 to S3 are executed by the intelligent control unit, which works in conjunction with the first-level circuit and the second-level circuit to form a closed-loop adaptive process. In a specific application example, the periodic trigger sampling described in step S1 is implemented through a timer or event-driven mechanism, and the analog-to-digital converter is started to collect noise data after confirming that the UART bus is idle, so as to ensure that the data comes from the communication interval.

[0024] In a specific application example, the extraction of noise features in step S2 also includes calculating the statistical features of the noise and inputting the feature vector into the machine learning model; the machine learning model is a lightweight gradient boosting decision tree or a micro neural network, used to map the noise features to the optimal detection parameters.

[0025] In a specific application example, the dynamic generation of wake-up detection parameters in step S3 includes: calculating the voltage threshold and minimum pulse width in real time based on noise characteristics, and writing the parameters into the configuration register of the first-level circuit, namely the adaptive pre-wake-up detector, through a digital bus; the dynamic voltage threshold can be adjusted within the logic low level range, and the dynamic minimum pulse width is adjustable in units of bit time.

[0026] In a specific application example, the specific operation of the first-stage circuit, namely the adaptive pre-wake detector, in step S4 includes: The programmable threshold comparison module converts the digital code value into an analog reference voltage through a digital-to-analog converter, which is used as a dynamic voltage threshold and compared with the receiving pin signal. The configurable digital filtering module verifies pulse width using a counter. The target value of the counter corresponds to the dynamic minimum pulse width, and the signal is only filtered if the duration of the low level exceeds this value. The enable gating module receives the filtered output and the system enable signal, and controls the generation of the pre-wake-up signal.

[0027] In a specific application example, the protocol verification in step S5 includes resampling the signal at the UART communication baud rate to verify whether the duration of the low-level pulse is strictly equal to one bit time; the second-level circuit, namely the protocol verification wake-up device, is activated only when a pre-wake-up signal is received to reduce power consumption.

[0028] In specific application examples, there is also a learning optimization step: after a successful wake-up, the intelligent control unit continues to collect noise data and update the machine learning model parameters for the initial detection parameter optimization of subsequent sleep cycles.

[0029] This embodiment uses a typical IoT terminal device as the implementation platform. The device is powered by 3.3V, and the main controller is an ARM Cortex-M series low-power microcontroller that integrates a digital-to-analog converter (ADC) module, a digital comparator, and a programmable timer. Figure 1As shown, the specific implementation steps are as follows: Step S1: During the UART bus idle period, periodically trigger the sampling of noise data on the receive pin; The intelligent control unit uses a timer interrupt to trigger periodically, checking the UART bus status every second. Once the bus is confirmed to be idle (the RX pin remains high for more than 10 bits), the ADC is started to sample the RX pin. This step ensures that noise data originates from communication gaps, avoiding interference with the valid signal.

[0030] Step S2: Extract noise features from the noise data, including voltage peak value, voltage RMS value, and negative pulse density; Step S3: Based on noise features, dynamically generate wake-up detection parameters using a pre-trained machine learning model; Steps S1 to S3 are executed by the intelligent control unit, which works in conjunction with the first-level circuit and the second-level circuit to form a closed-loop adaptive process. Step S4: Configure the wake-up detection parameters to the first-stage circuit, i.e., the adaptive pre-wake-up detector, and perform signal filtering to generate a pre-wake-up signal: The first-stage circuit consists of three modules working collaboratively: a programmable threshold comparison module sets the DAC output based on V_th_new, applies an analog voltage threshold to the comparator, and outputs a high level when the RX signal voltage is below the threshold; a configurable digital filtering module sets the counter target value based on T_min_new, performs pulse width verification on the comparator output, and triggers filtering only when the high level duration exceeds T_min_new; and an enable gating module outputs a pre-wake-up signal (PWK) when the system enable signal is valid. This step achieves joint determination of amplitude and pulse width, generating PWK only when the signal simultaneously meets the dynamic parameters.

[0031] The circuit principle of the first-level circuit, namely the adaptive pre-wake detector, is as follows: Figure 3 As shown.

[0032] S5: The pre-wake-up signal is verified using the UART start bit protocol via the second-level circuit, i.e., the protocol verification wake-up unit. The pre-wake signal triggers the activation of the second-stage circuit. This circuit resamples the RX signal at a precise baud rate clock (e.g., 16x oversampling) to verify that the duration of the low-level pulse is strictly equal to one bit time (error tolerance ±5%). If the verification is successful, a global wake-up signal is output to fully wake up the system; otherwise, the verification failure event is recorded and the system returns to sleep mode. The second-stage circuit operates with minimal power consumption only when PWK is active.

[0033] To illustrate the effectiveness of the method described in this embodiment, let's take 115200bps communication as an example: Quiet office environment: Step S1 sampling noise peak is 0.3V, pulse density is low; after feature extraction in step S2, the model outputs V_th=0.8V and T_min=0.9 bit time in step S3; in step S4, weak noise (such as 0.5V / 0.1 bit time) is filtered out due to insufficient pulse width, and the effective start bit (0.2V / 1.0 bit time) passes the screening; step S5 verification is successful.

[0034] In noisy industrial environments: Step S1 samples noise peak of 1.2V with high pulse density; after feature extraction in Step S2, the model outputs V_th=0.4V and T_min=1.2 bit time in Step S3; in Step S4, wide pulse noise (0.3V / 0.8 bit time) is filtered out, and standard start bit (0.2V / 1.0 bit time) may be filtered out due to insufficient pulse width (conservative strategy); Step S5 verifies that it depends on protocol retransmission.

[0035] This implementation details how to rigorously implement each step of this embodiment through the collaboration of hardware modules and software algorithms. The core advantage of this method lies in transforming static detection into dynamic adaptation. Figure 3 The closed-loop process illustrated was validated in a real-world environment, demonstrating improved wake-up reliability and power efficiency. Those skilled in the art can make appropriate modifications based on the above teachings without departing from the scope of this invention.

[0036] The adaptive UART wake-up technology based on idle-period noise feature learning proposed in this embodiment brings revolutionary improvements to the field of low-power electronic device communication. Compared with traditional static parameter wake-up schemes, this method achieves significant technical breakthroughs in multiple dimensions by introducing environmental awareness and intelligent adaptive mechanisms, providing a brand-new solution for optimizing the reliability and energy efficiency of IoT devices. Its advantages include the following: 1. Breakthrough improvement in environmental adaptability The core advantage of this method lies in its ability to adaptively adjust detection parameters to environmental noise, completely overcoming the technical dilemma of inconsistent performance under different noise levels when using fixed parameters. This breakthrough is achieved through a unique "perception-decision-execution" intelligent closed loop: the system periodically samples background noise during UART bus idle periods, extracting key features such as voltage peak value, RMS value, and pulse density, and then dynamically generates optimal detection parameters based on a pre-trained machine learning model. This dynamic adjustment mechanism enables the system to automatically improve its immunity in high-noise environments and restore high sensitivity in low-noise environments, achieving intelligent matching with the environmental noise level.

[0037] 2. Significant improvements in system reliability and ease of deployment This method significantly reduces the system's reliance on preset parameters and manual debugging by establishing a fully automated noise perception parameter generation closed loop, thereby improving the reliability and consistency of the device in different deployment environments. Traditional configurable solutions require engineers to manually preset parameters based on estimated environments, while this method allows the system to autonomously complete the entire optimization process. This design enables the same hardware design to automatically adjust to its optimal operating state in different real-world environments, effectively reducing the risk of performance fluctuations caused by improper parameter configuration. This advantage is particularly important in large-scale IoT deployment scenarios, as the device can autonomously adapt to its local environment, significantly reducing the cost of on-site debugging and maintenance.

[0038] like Figure 1 The flowchart of the adaptive UART wake-up method of this invention clearly illustrates the working mechanism of this intelligent closed loop. The entire system forms a continuously optimized learning loop through periodic noise sampling, feature extraction, and parameter updates, ensuring that the system maintains optimal wake-up performance under various environmental conditions.

[0039] 3. The Art of Balancing Power Consumption Optimization and Performance Maintaining high sensitivity to valid signals while suppressing false wake-ups is a key advantage of this technology in terms of power consumption optimization. This advantage is achieved through an innovative two-stage collaborative wake-up architecture and targeted feature design. The first-stage circuit is responsible for environmental adaptability screening based on dynamic parameters. Its innovation lies in the flexibility of the parameters themselves, which can intelligently adjust according to noise characteristics. The feature parameters used for decision-making (such as pulse density) directly reflect the interference frequency similar to the start bit, enabling the system to automatically increase the pulse width requirement in noisy environments to filter out dense interference, while reducing the requirement in quiet environments to ensure response speed. This refined parameter adjustment mechanism, combined with the second-stage standard protocol verification circuit, achieves an optimal balance between false wake-up rate and wake-up sensitivity.

[0040] 4. Technological foresight and scalability This method also demonstrates good foresight and scalability. The adopted machine learning framework leaves ample room for subsequent performance optimization, allowing for continuous improvement of the system's adaptability through model updates and algorithm improvements. Furthermore, the modular design enables easy integration into different hardware platforms, laying a solid foundation for the future development of low-power communication technologies.

[0041] like Figure 2 As shown, this embodiment also proposes an adaptive UART wake-up system based on idle period noise feature learning to implement the above-mentioned adaptive UART wake-up method based on idle period noise feature learning, including the following steps: The first-stage circuit, the adaptive pre-wake detector, is a dynamically configurable adaptive pre-wake detector. It includes a programmable threshold comparison module, a configurable digital filtering module, and an enable gating module, used to pre-detect the UART receive pin signal based on dynamic parameters. The principle of the first-stage circuit is as follows: Figure 3 As shown; The second-level circuit, namely the protocol verification wake-up unit, is used to verify the UART start bit protocol of the pre-wake-up signal generated by the first-level circuit. The intelligent control unit is used to periodically sample noise data during the UART bus idle period, extract noise features, dynamically generate detection parameters through machine learning models, and configure them to the first-level circuit, namely the adaptive pre-wake detector. The first-level circuit, namely the adaptive pre-wake detector, the second-level circuit, namely the protocol verification wake-up device, and the intelligent control unit are interconnected to form a closed-loop adaptive system, enabling autonomous perception of environmental noise and parameter adjustment.

[0042] In a specific application example, the intelligent control unit includes: The sampling module acquires noise data through an ADC and triggering logic; The feature extraction module is used to calculate the voltage peak value, the effective voltage value, and the negative pulse density; The model inference module integrates a pre-trained machine learning model and outputs dynamic voltage thresholds and dynamic minimum pulse widths. The parameter update module writes parameters into the first-stage circuit, namely the adaptive pre-wake detector, via the bus.

[0043] This embodiment also proposes a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described adaptive UART wake-up method based on idle period noise feature learning.

[0044] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-readable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process. Figure 1 One or more processes and / or boxes Figure 1 The functions specified in one or more boxes. These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable apparatus for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0045] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. An adaptive UART wake-up method based on idle period noise feature learning, characterized in that, Includes the following steps: Step S1: During the UART bus idle period, periodically trigger the sampling of noise data on the receive pin; Step S2: Extract noise features from the noise data, including voltage peak value, voltage RMS value, and negative pulse density; Step S3: Based on the noise features, dynamically generate wake-up detection parameters using a pre-trained machine learning model, including dynamic voltage threshold and dynamic minimum pulse width; Step S4: Configure the wake-up detection parameters to the first-level circuit, namely the adaptive pre-wake-up detector. This circuit includes a programmable threshold comparison module, a configurable digital filtering module, and an enable gating module, which are used to perform amplitude and pulse width determination on the input signal. A pre-wake-up signal is generated only when the signal simultaneously meets the current dynamic voltage threshold and the dynamic minimum pulse width. Step S5: The pre-wake signal is verified by the second-level circuit, namely the protocol verification wake-up device, using the UART start bit protocol. If the verification is successful, a global wake-up signal is output; otherwise, the event is recorded and the device returns to sleep state. Steps S1 to S3 are executed by the intelligent control unit, which works in conjunction with the first-level circuit and the second-level circuit to form a closed-loop adaptive process.

2. The adaptive UART wake-up method based on idle period noise feature learning according to claim 1, characterized in that, The periodic trigger sampling described in step S1 is implemented through a timer or event-driven mechanism, and the analog-to-digital converter is started to collect noise data after confirming that the UART bus is idle, so as to ensure that the data comes from the communication interval.

3. The adaptive UART wake-up method based on idle period noise feature learning according to claim 1, characterized in that, The noise feature extraction in step S2 also includes calculating the statistical features of the noise and inputting the feature vector into the machine learning model; the machine learning model is a lightweight gradient boosting decision tree or a micro neural network, used to map the noise features to the optimal detection parameters.

4. The adaptive UART wake-up method based on idle period noise feature learning according to claim 1, characterized in that, The dynamic generation of wake-up detection parameters in step S3 includes: calculating the voltage threshold and minimum pulse width in real time based on noise characteristics, and writing the parameters into the configuration register of the first-level circuit, i.e., the adaptive pre-wake-up detector, through a digital bus; the dynamic voltage threshold can be adjusted within the logic low level range, and the dynamic minimum pulse width is adjustable in units of bit time.

5. The adaptive UART wake-up method based on idle period noise feature learning according to claim 1, characterized in that, The specific operation of the first-stage circuit, namely the adaptive pre-wake detector, in step S4 includes: The programmable threshold comparison module converts the digital code value into an analog reference voltage through a digital-to-analog converter, which is used as a dynamic voltage threshold and compared with the receiving pin signal. The configurable digital filtering module verifies pulse width using a counter. The target value of the counter corresponds to the dynamic minimum pulse width, and the signal is only filtered if the duration of the low level exceeds this value. The enable gating module receives the filtered output and the system enable signal, and controls the generation of the pre-wake-up signal.

6. The adaptive UART wake-up method based on idle period noise feature learning according to claim 1, characterized in that, The protocol verification in step S5 includes resampling the signal at the UART communication baud rate to verify whether the duration of the low-level pulse is strictly equal to one bit time; the second-level circuit, namely the protocol verification wake-up device, is activated only when a pre-wake-up signal is received to reduce power consumption.

7. The adaptive UART wake-up method based on idle period noise feature learning according to claim 1, characterized in that, It also includes a learning optimization step: after a successful wake-up, the intelligent control unit continues to collect noise data and update the machine learning model parameters for the initial detection parameter optimization of subsequent sleep cycles.

8. An adaptive UART wake-up system based on idle period noise feature learning, used to implement the steps of the adaptive UART wake-up method based on idle period noise feature learning as described in any one of claims 1 to 7, characterized in that, include: The first-stage circuit, namely the adaptive pre-wake detector, is an adaptive pre-wake detector with dynamically configurable parameters. It includes a programmable threshold comparison module, a configurable digital filtering module, and an enable gating module, which are used to pre-detect the UART receive pin signal according to dynamic parameters. The second-level circuit, namely the protocol verification wake-up unit, is used to verify the UART start bit protocol of the pre-wake-up signal generated by the first-level circuit. The intelligent control unit is used to periodically sample noise data during the UART bus idle period, extract noise features, dynamically generate detection parameters through machine learning models, and configure them to the first-level circuit, namely the adaptive pre-wake detector. The first-level circuit, namely the adaptive pre-wake detector, the second-level circuit, namely the protocol verification wake-up device, and the intelligent control unit are interconnected to form a closed-loop adaptive system, enabling autonomous perception of environmental noise and parameter adjustment.

9. An adaptive UART wake-up system based on idle period noise feature learning according to claim 8, characterized in that, The intelligent control unit includes: The sampling module acquires noise data through an ADC and triggering logic; The feature extraction module is used to calculate the voltage peak value, the effective voltage value, and the negative pulse density; The model inference module integrates a pre-trained machine learning model and outputs dynamic voltage thresholds and dynamic minimum pulse widths. The parameter update module writes parameters into the first-stage circuit, namely the adaptive pre-wake detector, via the bus.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the adaptive UART wake-up method based on idle period noise feature learning as described in any one of claims 1 to 7.

Images