A wearable motion recognition system based on brain-like principles

Abstract

This invention discloses a wearable motion recognition system based on brain-like principles. The system includes a multimodal sensor module, a microcontroller module, and status indicator lights. The multimodal sensor module collects geomagnetic field and acceleration sensing signals and outputs them to the microcontroller module. The microcontroller module converts these two sensing signals into sensing pulse signals using a pulse coding method and processes them through a programmed artificial neuron model. The artificial neuron model receives the sensing pulse signals and generates biological-like neuronal membrane potential signals and neuronal pulse signals. The microcontroller module determines the motion state by judging the firing rate of the sensing pulse signals, the firing rate of the neuronal pulse signals, and the amplitude of the neuronal membrane potential signals, and outputs the result to the status indicator lights. This invention enables intelligent, low-power, wearable motion state recognition for wearers or robots.

Description

Technical Field

[0001] This invention relates to the field of motion state sensing technology, and specifically to a wearable motion recognition system based on brain-like principles. Background Technology

[0002] Motion recognition systems based on advanced intelligent sensing technology can perceive the wearer's posture and the robot's intentions in real time, making them particularly suitable for cutting-edge applications such as wearable devices, health monitoring, and intelligent robots. With the development of artificial intelligence and sensing technologies, motion recognition systems integrating modules such as LiDAR, high-definition cameras, and gyroscopes are constantly emerging. These systems possess a certain ability to perceive motion information and can perform basic motion state recognition functions such as step counting and fall warning. However, most existing motion recognition systems use data fusion to process massive amounts of high-dimensional data collected by multimodal sensors, and rely on complex algorithms for computation and inference on local computing platforms or cloud servers, resulting in high demands on computing resources, data bandwidth, and power consumption. In summary, developing low-power, intelligent wearable motion recognition systems and achieving accurate identification and classification of complex motion states is a crucial problem that urgently needs to be solved in the field of motion sensing.

[0003] Artificial neuron models, which have garnered widespread attention in recent years, are computational models that simulate the network structure and asynchronous pulse information processing mode of biological nervous systems. They can be used to efficiently process large amounts of complex sensor data in parallel. Brain-inspired technologies based on artificial neuron models and artificial neural networks are of great significance for breaking through the traditional von Neumann computing architecture and overcoming the bottlenecks of the memory wall and power wall. The leak-integral-release (LIF) model, widely used in artificial neuron research, can achieve efficient parsing of complex information by simulating the information processing mechanism of biological neurons, and can achieve advanced functions such as information recognition, memory, and learning by simulating the weight control mechanism of biological synapses. Therefore, introducing artificial neuron models into wearable motion recognition systems to encode and process sensor signals is expected to improve the efficiency of motion state detection and recognition, providing a new technical path to solve existing technological bottlenecks. Summary of the Invention

[0004] This invention proposes a wearable motion recognition system based on brain-like principles, which is used to achieve high-efficiency, intelligent, and low-power wearable motion recognition functions.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: A wearable motion recognition system based on brain-like principles includes, The system includes a multimodal sensor module, a microcontroller module, and status indicator lights, as well as an artificial neuron model. The multimodal sensor module is used to collect geomagnetic field sensing signals and acceleration sensing signals and transmit them to the microcontroller module; The microcontroller module converts the geomagnetic field sensing signal and the acceleration sensing signal into geomagnetic field sensing pulse signal and acceleration sensing pulse signal respectively using a pulse coding method, and encodes the sensing pulse signal using a programmed artificial neuron model. The artificial neuron model simulates the spatiotemporal integration and pulse firing mechanism of biological neurons, generating neuronal membrane potential signals and neuronal pulse signals based on the timing, frequency, and duration of the input pulse signals. The microcontroller module is also used to determine the firing rate of the sensing pulse signal, the amplitude of the neuronal membrane potential signal, and the firing rate of the neuronal pulse signal, thereby completing the identification of the motion state of the wearer or robot, and transmitting the identified motion state information to the status indicator light; wherein, the system as a whole has the structural characteristics of miniaturization, low power consumption and being able to be fixed to the surface of the human body or robot to meet the needs of wearable motion recognition applications.

[0006] The system generates at least two sensing pulse signals in total, including at least one geomagnetic field sensing pulse signal and at least one acceleration sensing pulse signal; The geomagnetic field sensing signal uses differential quantization pulse coding, and the acceleration sensing signal uses amplitude quantization pulse coding.

[0007] Furthermore, the differential quantization pulse coding step includes, within one coding cycle, comparing the absolute value of the change in the sensing signal with at least two change level thresholds arranged in ascending order; when the absolute value of the change in the sensing signal is less than the smallest change level threshold, no sensing pulse signal is output; when the absolute value of the change in the sensing signal is greater than or equal to a certain change level threshold and less than the next higher change level threshold, a sensing pulse signal with a specific frequency is output, the specific frequency being the level frequency corresponding to the current change level threshold interval, and the level frequency being positively correlated with the set value of the change level threshold interval (the larger the set threshold, the higher the set frequency of the corresponding level, and vice versa).

[0008] Further, the amplitude quantization pulse coding step includes: within one coding period, comparing the absolute value of the amplitude of the sensing signal with at least two amplitude grading thresholds arranged in ascending order; when the absolute value of the amplitude of the sensing signal is less than the smallest amplitude grading threshold, not outputting a sensing pulse signal; when the absolute value of the amplitude of the sensing signal is greater than or equal to a certain amplitude grading threshold and less than the next higher amplitude grading threshold, outputting a sensing pulse signal with a specific frequency, wherein the specific frequency is a grading frequency corresponding to the current amplitude grading threshold interval, and the grading frequency is positively correlated with the set value of the amplitude grading threshold interval.

[0009] Furthermore, the artificial neuron model is a Leaky, Integrate and Fire (LIF) model, which is implemented through programming of the microcontroller module and is used to perform spatiotemporal integration and pulse firing of the input sensing pulse signal.

[0010] Furthermore, the number of status indicator lights is n, each status indicator light has two states: on and off. The n status indicator lights can indicate a maximum of 2... n Different states of motion.

[0011] Furthermore, the system's workflow includes the following steps: Step 1: Initialize the microcontroller module by setting the amplitude grading threshold or change grading threshold for pulse coding, as well as the corresponding grading frequency of the sensing pulse signal. Step 2: Within a preset encoding cycle, the microcontroller module performs pulse encoding on the acquired geomagnetic field sensor signal and acceleration sensor signal. Subsequently, the artificial neuron model processes the encoded geomagnetic field pulse signal and acceleration pulse signal and calculates the neuron membrane potential signal and neuron pulse signal. Step 3: Repeat step 2 until a preset detection cycle is reached. During this process, the microcontroller module counts the number of pulses of the geomagnetic field sensor pulse signal, the acceleration sensor pulse signal, and the neuron pulse signal, respectively. Step four: At the end of the detection cycle, the microcontroller module comprehensively judges the firing rate of the geomagnetic field sensor pulse signal, the firing rate of the acceleration sensor pulse signal, the amplitude of the neuronal membrane potential signal, and the firing rate of the neuronal pulse signal to obtain the corresponding motion state category. Step 5: The microcontroller module outputs the identified motion state category information to the status indicator light for display.

[0012] Furthermore, the duration of the detection period is a positive integer multiple of the duration of the encoding period, and the detection period is greater than or equal to 2 seconds.

[0013] Compared with the prior art, the present invention has the following beneficial effects: From a hardware architecture perspective, this invention breaks away from the reliance of traditional motion recognition systems on high-performance processors and cloud servers, proposing a lightweight hardware architecture based on a microcontroller. This architecture integrates multimodal sensing, pulse coding, neuromorphic processing, and status indication functions onto a single microcontroller platform, achieving highly energy-efficient, real-time integration and processing of motion information (acceleration) and orientation information (geomagnetic field). Compared to the traditional separate "sensor-analog-digital converter-processor" structure, this system offers significant advantages in hardware cost, integration level, and system power consumption.

[0014] From an information processing perspective, this invention avoids traditional software processing methods based on massive data computation (such as digital signal processing and deep learning inference), and proposes an information processing paradigm based on brain-like principles. This paradigm converts continuous sensor signals into sparse pulse sequences through hierarchical thresholding and pulse coding, and integrates and processes the pulse sequences using an artificial neuron model. This biological neuron-like processing method fundamentally reduces the computational load and power consumption of data processing.

[0015] In terms of application potential, this invention benefits from the aforementioned hardware and software innovations, achieving a significant reduction in overall system power consumption and a substantial improvement in real-time performance. This makes it suitable for wearable and resource-constrained applications requiring long-term operation, including monitoring daily human activities, rehabilitation training assessment, and posture perception for small robots. The system architecture employed in this invention is simple, consumes little power, and does not rely on complex algorithms or artificial neural network models, thus overcoming the technical challenges of existing technologies that, due to high power consumption and computational complexity, make it difficult to achieve long-term, stable, and autonomous operation on lightweight portable devices. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the structure of a wearable motion recognition system based on brain-like principles, provided in an embodiment.

[0017] Figure 2 A flowchart illustrating the workflow of a wearable motion recognition system based on brain-like principles, provided in an embodiment of the present invention.

[0018] Figure 3 Schematic diagrams of two pulse coding methods provided in embodiments of the present invention; Figure 4 A schematic diagram of an artificial neuron model provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of the geomagnetic field sensing signal tested in an embodiment of the present invention; Figure 6 This is a schematic diagram of the geomagnetic field sensing pulse signal tested in an embodiment of the present invention; Figure 7 This is a schematic diagram of the neuronal membrane potential signal of the artificial neuron model tested in an embodiment of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0020] As shown in Figure 1, this embodiment provides a wearable motion recognition system based on brain-like principles. The system includes an integrated multimodal sensor module, a microcontroller module, and a status indicator light. It can encode, integrate, and process motion information and orientation information, and can identify motion states.

[0021] The purpose of the multimodal sensor module is to measure orientation and motion information, including geomagnetic field sensing signals and acceleration sensing signals. Preferably, the multimodal sensor module includes an acceleration sensing unit for measuring acceleration sensing signals along at least one spatial axis and a magnetic sensing unit for measuring geomagnetic field sensing signals along at least one spatial axis. Preferably, the multimodal sensor module is selected as a miniaturized, highly integrated, and low-power module.

[0022] The uses of the microcontroller module include: (1) The two sensing signals collected by the multimodal sensor module are converted into sensing pulse signals by pulse coding and output to the artificial neuron model; (2) Equipped with an artificial neuron model, including two or more pulse sensing signal input channels, it imitates the transmission, processing and integration functions of biological synapses for peak pulse signals, and can generate neuronal membrane potential signals and neuronal pulse signals according to the timing, frequency and duration of the pulse signals, thereby simulating the response characteristics and firing behavior of biological neurons. (3) The firing rate of the pulse-encoded sensor pulse signal, the amplitude of the neuronal membrane potential signal, and the firing rate of the neuronal pulse signal are determined to identify the motion state. (4) Output the identified motion state information to the status indicator light. Preferably, the microcontroller module is selected based on a low-power control chip; preferably, the acceleration sensing pulse signal and the geomagnetic field sensing pulse signal are both voltage pulse signals with fixed pulse amplitude (5V) and fixed pulse width; preferably, the motion state recognition adopts a decision tree-based judgment method; preferably, the system as a whole has miniaturization and wearability characteristics, and can be fixed on the surface of a human body or robot for motion state recognition.

[0023] The purpose of the status indicator light is to indicate the motion status recognition result. The number of all motion states that the motion status indicator light can indicate should be greater than or equal to the number of all motion states that the wearable motion recognition system can recognize.

[0024] The system generates at least two sensing pulse signals in total, including at least one geomagnetic field sensing pulse signal and at least one acceleration sensing pulse signal; The geomagnetic field sensing signal uses differential quantization pulse coding, and the acceleration sensing signal uses amplitude quantization pulse coding.

[0025] The differential quantization pulse coding step includes: within one coding cycle, comparing the absolute value of the change in the sensing signal with at least two change level thresholds arranged in ascending order; when the absolute value of the change in the sensing signal is less than the smallest change level threshold, not outputting a sensing pulse signal; when the absolute value of the change in the sensing signal is greater than or equal to a certain change level threshold and less than the next higher change level threshold, outputting a sensing pulse signal with a specific frequency, wherein the specific frequency is a level frequency corresponding to the current change level threshold interval, and the level frequency is positively correlated with the set value of the change level threshold interval.

[0026] The amplitude quantization pulse coding step includes: within one coding period, comparing the absolute value of the amplitude of the sensing signal with at least two amplitude grading thresholds arranged in ascending order; when the absolute value of the amplitude of the sensing signal is less than the smallest amplitude grading threshold, not outputting a sensing pulse signal; when the absolute value of the amplitude of the sensing signal is greater than or equal to a certain amplitude grading threshold and less than the next higher amplitude grading threshold, outputting a sensing pulse signal with a specific frequency, wherein the specific frequency is the grading frequency corresponding to the current amplitude grading threshold interval, and the grading frequency is positively correlated with the set value of the amplitude grading threshold interval.

[0027] The artificial neuron model is a Leaky, Integrate and Fire (LIF) model, implemented through programming of the microcontroller module, and is used to perform spatiotemporal integration and pulse firing of the input sensing pulse signal.

[0028] The number of status indicator lights is n, and each status indicator light has two states: on and off. The n status indicator lights can indicate a maximum of 2... n Different states of motion.

[0029] As shown in Figure 2, the workflow of the wearable motion recognition system based on brain-like principles is as follows: Step 1: Initialize the microcontroller module by setting the threshold values ​​for changes in the geomagnetic field sensing signal from small to large (Magnetic Field Sensing Threshold). th1 Mag th2 The frequency gradients of the geomagnetic field pulse signal (20Hz, 40Hz); the amplitude gradient thresholds of the acceleration sensing signal are set from small to large (Acc). th1 , Acc th2 ), the graded frequencies of the acceleration pulse signal (20Hz, 40Hz).

[0030] Step 2: Within one encoding cycle (T1 = 50ms), the microcontroller module performs pulse encoding on the geomagnetic field sensing signal and acceleration sensing signal measured by the multimodal sensor module. Subsequently, the artificial neuron model processes the encoded geomagnetic field pulse signal and acceleration pulse signal and calculates the neuron membrane potential signal and neuron pulse signal.

[0031] Step 3: Repeat step 2 until the time reaches a preset detection period (T2 = 10s). During this process, the microcontroller module counts the number of pulses of the geomagnetic field sensing pulse signal, the acceleration sensing pulse signal, and the neuron pulse signal, respectively. Step four: At the end of the detection cycle, the microcontroller module comprehensively judges the emission rate of the geomagnetic field sensor signal, the emission rate of the acceleration sensor signal, the amplitude of the neuronal membrane potential signal, and the emission rate of the neuronal pulse signal to obtain the corresponding motion state category. Step 5: The microcontroller module transmits the motion status category information to the motion status indicator light.

[0032] The following are examples: like Figure 3 As shown, the pulse coding method (differential quantization pulse coding) for geomagnetic field sensing signals is as follows: Within one encoding cycle (T1 = 50ms), the absolute value of the change in the geomagnetic field sensing signal measured by the multimodal sensor module (|ΔMag|) is compared with the first change threshold (Mag) of the geomagnetic field sensing signal set by the microcontroller module. th1 ) for comparison; when the absolute value of the change in the geomagnetic field sensing signal is less than the first change threshold of the geomagnetic field sensing signal (|ΔMag| <Mag th1 The microcontroller module does not output geomagnetic field sensing pulse signals; when the absolute value of the change in the geomagnetic field sensing signal is greater than or equal to the first change threshold of the geomagnetic field sensing signal (|ΔMag| ≥ Mag), the microcontroller module does not output geomagnetic field sensing pulse signals. th1 When the change is less than the second threshold value of the geomagnetic field sensing signal (|ΔMag|), <Mag th2The microcontroller module outputs geomagnetic field sensing pulse signals at corresponding graded frequencies (20Hz); when the absolute value of the change in the geomagnetic field sensing signal is greater than or equal to the second graded threshold of the geomagnetic field sensing signal (|ΔMag| ≥ Mag), the signal is considered to be indicative of a change in the geomagnetic field sensing signal. th2 The microcontroller module outputs geomagnetic field sensing pulse signals at corresponding graded frequencies (40Hz).

[0033] like Figure 3 As shown, the pulse coding method (amplitude quantization pulse coding) for acceleration sensing signals is as follows: Within one encoding cycle (T1 = 50ms), the absolute value of the amplitude of the acceleration sensing signal measured by the multimodal sensor module (|Acc|) is compared with the first amplitude classification threshold (Acc) of the acceleration sensing signal set by the microcontroller module. th1 The comparison is performed; when the absolute value of the amplitude of the acceleration sensing signal is less than the first amplitude grading threshold of the acceleration sensing signal (|Acc|), the comparison is performed. <Acc th1 The microcontroller module does not output acceleration sensing pulse signals; when the absolute value of the acceleration sensing signal amplitude is greater than or equal to the first amplitude grading threshold of the acceleration sensing signal (|Acc|≥ Acc), the microcontroller module does not output acceleration sensing pulse signals. th1 When the amplitude is less than the second amplitude grading threshold of the acceleration sensing signal (|Acc|), <Acc th2 The microcontroller module outputs acceleration sensing pulse signals at corresponding graded frequencies (20Hz); when the absolute value of the acceleration sensing signal amplitude is greater than or equal to the second amplitude graded threshold of the acceleration sensing signal (|Acc| ≥ Acc), the signal is considered to be in the range of 20Hz. th2 The microcontroller module outputs acceleration sensing pulse signals at corresponding graded frequencies (40Hz).

[0034] like Figure 4 As shown, the principle by which the artificial neuron model simulates the processing of pulse signals by biological neurons is as follows: Artificial neuron models, as mathematical abstractions of the functions of biological neurons, primarily receive and accumulate pulse signal inputs, and generate corresponding outputs based on set conditions. Taking the leak-integral-fire model as an example:

[0035] Artificial neuron models learn by processing input impulses. The output membrane potential is updated by performing time integration. Simultaneously, membrane potential It will be at a fixed time constant It decays naturally until it reaches the resting potential. When the membrane potential When the firing threshold is reached, the neuron generates a neuronal impulse and changes the membrane potential. Reset to initial state.

[0036] like Figure 5 As shown, this is the geomagnetic field sensing signal collected by the multimodal sensor module during the motion process in this embodiment.

[0037] like Figure 6 As shown, the geomagnetic field sensing signal collected in this embodiment is converted into a geomagnetic field sensing pulse signal after differential quantization pulse coding.

[0038] like Figure 7 As shown, the encoded geomagnetic field sensing pulse signal in this embodiment is processed by an artificial neuron model to obtain the neuron membrane potential signal.

[0039] In summary, the wearable motion recognition system provided in this embodiment of the invention employs highly integrated multimodal sensors and low-power control chips in terms of hardware. This system boasts strong information processing capabilities while requiring minimal data processing. Furthermore, at the information processing level, the system utilizes pulse coding technology to convert continuous sensing signals into sparse pulse sequences, combining this with an artificial neuron model to complete signal integration and processing, resulting in minimal demands on computing resources, storage resources, and software algorithms. Moreover, by setting reasonable detection and encoding cycles, the system can achieve continuous and stable recognition of motion states under low-power operation. This system can be applied to wearable applications with high portability requirements, such as health monitoring, human-computer interaction, and small robots.

[0040] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0041] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0042] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. 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. Such 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 wearable motion recognition system based on brain-like principles, comprising a multimodal sensor module, a microcontroller module, and status indicator lights; characterized in that, It also includes artificial neuron models; The multimodal sensor module is used to collect geomagnetic field sensing signals and acceleration sensing signals and transmit them to the microcontroller module; The microcontroller module converts the geomagnetic field sensing signal and the acceleration sensing signal into sensing pulse signals, namely geomagnetic field sensing pulse signals and acceleration sensing pulse signals, respectively, using a pulse coding method, and performs calculations using an artificial neuron model implemented through programming. The artificial neuron model simulates the spatiotemporal integration and pulse firing mechanism of biological neurons, generating neuronal membrane potential signals and neuronal pulse signals based on the timing, frequency, and duration of the input pulse signals. The microcontroller module is also used to determine the firing rate of the sensing pulse signal, the amplitude of the neuronal membrane potential signal, and the firing rate of the neuronal pulse signal, thereby completing the identification of the wearer's or robot's motion state, and transmitting the identified motion state information to the status indicator light.

2. The wearable motion recognition system based on brain-like principles according to claim 1, characterized in that, The system generates at least two sensing pulse signals in total, including at least one geomagnetic field sensing pulse signal and at least one acceleration sensing pulse signal; The geomagnetic field sensing signal uses differential quantization pulse coding, and the acceleration sensing signal uses amplitude quantization pulse coding.

3. The wearable motion recognition system based on brain-like principles according to claim 2, characterized in that, The differential quantization pulse coding process is as follows: Within a coding cycle, the absolute value of the change in the sensing signal is compared with at least two change thresholds arranged in ascending order. When the absolute value of the change in the sensing signal is less than the minimum change grading threshold, no sensing pulse signal is output. When the absolute value of the change in the sensing signal is greater than or equal to a certain change level threshold and less than the next higher change level threshold, a sensing pulse signal with a set frequency is output. The set frequency is the level frequency corresponding to the current change level threshold interval, and the level frequency is positively correlated with the set value of the change level threshold interval.

4. The wearable motion recognition system based on brain-like principles according to claim 2, characterized in that, The amplitude quantization pulse encoding process is as follows: Within one encoding cycle, the absolute value of the amplitude of the sensing signal is compared with at least two amplitude grading thresholds arranged from smallest to largest. When the absolute value of the amplitude of the sensing signal is less than the minimum amplitude grading threshold, no sensing pulse signal is output. When the absolute value of the amplitude of the sensing signal is greater than or equal to a certain amplitude grading threshold and less than the next higher amplitude grading threshold, a sensing pulse signal with a set frequency is output. The set frequency is the grading frequency corresponding to the current amplitude grading threshold interval, and the grading frequency is positively correlated with the set value of the amplitude grading threshold interval.

5. The wearable motion recognition system based on brain-like principles according to claim 1, characterized in that, The artificial neuron model is a Leaky, Integrate and Fire (LIF) model, implemented through programming of the microcontroller module, and is used to perform spatiotemporal integration and pulse firing of the input sensing pulse signal.

6. The wearable motion recognition system based on brain-like principles according to claim 1, characterized in that, The number of status indicator lights is n, and each status indicator light has two states: on and off. The n status indicator lights can indicate a maximum of 2... n Different states of motion.

7. The wearable motion recognition system based on brain-like principles according to any one of claims 1–6, characterized in that, The workflow includes the following: S1. Initialize the microcontroller module, set the amplitude grading threshold and change grading threshold for pulse coding, and the corresponding grading frequency of the sensing pulse signal. S2. Within a preset encoding cycle, the microcontroller module performs pulse encoding on the acquired geomagnetic field sensor signal and acceleration sensor signal. Subsequently, the artificial neuron model processes the encoded geomagnetic field pulse signal and acceleration pulse signal and calculates the neuron membrane potential signal value and neuron pulse signal value. S3. Repeat S2 until the time reaches a preset detection cycle. During this process, the microcontroller module counts the number of pulses of the geomagnetic field sensor pulse signal, the acceleration sensor pulse signal, and the neuron pulse signal, respectively. S4. At the end of the detection cycle, the microcontroller module comprehensively judges the firing rate of the geomagnetic field sensor pulse signal, the firing rate of the acceleration sensor pulse signal, the amplitude of the neuronal membrane potential signal, and the firing rate of the neuronal pulse signal to obtain the corresponding motion state category. S5. The microcontroller module outputs the identified motion state category information to the status indicator light for display.

Images