A system and method for reducing the power consumption of an electroencephalogram acquisition device

Abstract

The present invention discloses a system and a method for reducing the power consumption of an electroencephalogram (EEG) acquisition device. The system includes a visual stimulation module, a tag synchronization module, an EEG signal classification module, and a host computer module. The system further includes a low-power EEG acquisition module, which completes the acquisition and wireless transmission of EEG signals through a multi-core wake-up working mode. In the multi-core wake-up working mode, the calculation and setting of time parameters including the wireless communication connection interval ensure the real-time performance and integrity of data. Among them, the low-power EEG acquisition module is composed of an EEG acquisition electrode unit, an analog acquisition front-end unit, a main control and wireless transmission chip unit, a first clock, a second clock, and a power management unit. The main control and wireless transmission chip unit is composed of a main processor, a radio frequency processor, and a coprocessor. The present invention reduces the power consumption of a wearable electroencephalograph and enhances the battery life of the electroencephalograph under a small-capacity battery.

Description

Technical Field

[0001] The present invention relates to the field of electroencephalogram (EEG) signal acquisition, and more particularly, to a system and method for reducing the power consumption of EEG acquisition devices. Background Art

[0002] A brain-computer interface (BCI) is a system that directly converts central nervous system activities into artificial outputs. It can replace, repair, enhance, supplement, or improve the normal outputs of the central nervous system, thereby improving the interaction between the central nervous system and the internal and external environments. A typical BCI system includes four parts: signal acquisition, signal processing, control devices, and feedback links. As the main part of the BCI, the EEG signal acquisition system has become a hot research direction in current instrumentation science. Currently, there are various EEG signal mapping technologies such as electroencephalogram (EEG), magnetoencephalogram (MEG), and near-infrared spectroscopy (NIRS). Among them, EEG is a non-invasive method for measuring brain electrical activities.

[0003] Wearable EEG devices are the focus of research on EEG devices. They integrate amplifiers, analog-to-digital converters, and wireless transceivers, and can collect signals in real time. Wireless EEG acquisition devices have design requirements for long battery life and portability. Long battery life means that the device can work continuously for a long time without frequent charging, which is crucial for application scenarios that require long-term monitoring of EEG activities. However, to meet the demand for long battery life, most existing wireless EEG machines use large-capacity lithium batteries as power sources. Although large-capacity lithium batteries can provide enough energy to support long-term EEG acquisition, this also brings the problem of an increase in the size of the device. The volume and weight of large-capacity lithium batteries are usually large, resulting in an increase in the volume of the entire EEG machine, thereby reducing the portability of the device. Summary of the Invention

[0004] Aiming at the existing technical problems, the present invention provides a method for reducing the power consumption of EEG acquisition devices. This method addresses the current situation where traditional EEG acquisition devices usually rely on large-capacity batteries to ensure continuous operation of the device, and solves the problems of large volume of traditional EEG acquisition devices, which limit their portability and application scenarios. The present invention conducts research on the key technologies of low-power EEG acquisition devices, proposes a method for reducing power consumption by using heterogeneous communication in EEG acquisition devices. Taking the proposed method as the core, based on the low-power Bluetooth communication protocol, low-power chips are used to design low-power EEG acquisition devices. The present invention is expected to reduce the power consumption of wearable EEG machines, enhance the battery life of EEG machines with small-capacity batteries, replace large-capacity and large-volume batteries to reduce the volume of EEG machines, broaden application scenarios, and provide a new design idea for the practical application of expandable multi-channel brain-computer interfaces.

[0005] To solve the problems of the prior art, the present invention adopts the following technical solutions:

[0006] A system with reduced power consumption of an electroencephalogram (EEG) acquisition device, the system includes a visual stimulation module, a tag synchronization module, an EEG signal classification module, and a host computer module; the system also includes a low-power EEG acquisition module, and the low-power EEG acquisition module completes the acquisition and Bluetooth transmission of EEG signals through a multi-core wake-up working mode. In the multi-core wake-up working mode, the calculation and setting of time parameters including the Bluetooth communication connection interval are involved; wherein: the low-power EEG acquisition module is composed of an EEG acquisition electrode unit, an analog acquisition front-end unit, a main control and Bluetooth transmission chip unit, a first clock, a second clock, and a power management unit. The analog acquisition front-end unit amplifies, samples, and performs AD conversion processing on the EEG signals transmitted by the EEG acquisition electrode unit. The main control and Bluetooth transmission chip caches and transmits the digital EEG type signals transmitted by the analog acquisition front-end unit, and completes the calculation and setting of time parameters in the multi-core wake-up work; the power management unit is used to supply power to the low-power EEG acquisition module; wherein: the main control and Bluetooth transmission chip unit is composed of a main processor, a radio frequency processor, and a coprocessor, including:

[0007] The main processor manages the Bluetooth protocol and the EEG acquisition task, and completes the calculation and setting of time parameters in the multi-core wake-up work;

[0008] The coprocessor interacts with the analog acquisition front-end unit in an ultra-low-power form, controls the EEG acquisition task, and realizes the initialization and data transmission of the analog acquisition front-end unit;

[0009] The radio frequency processor completes the Bluetooth transmission task and transmits the data collected by the coprocessor through the underlying Bluetooth communication task.

[0010] Further, the process of calculating and setting time parameters including the communication connection interval for the low-power EEG acquisition module to complete the acquisition and Bluetooth transmission of EEG signals through the multi-core wake-up working mode includes:

[0011] Step 201. Select the sampling rate of the low-power EEG acquisition module as Fs;

[0012] Step 202. The main processor obtains the connection interval T through the following formula during the initialization of the main control and Bluetooth transmission chip unit;

[0013] T = t 1 +t 2 = 7.5 + 1.25×α (ms) (1)

[0014] where α is an unknown, t 1 is 7.5ms representing the minimum connection interval specified by the Bluetooth protocol, t2 1.25×α is the calculation step. Both addends in the formula are integer multiples of 1.25 ms, meeting the requirement that the Bluetooth connection interval stipulated by the Bluetooth protocol is an integer multiple of 1.25 ms.

[0015] For α, when the sampling rate is Fs, it is necessary to satisfy (t 1 +t 2 )×F s is an integer, that is

[0016]

[0017] Step 203. The main processor sets the coprocessor cache threshold N = T×Fs during the initialization of the main control and Bluetooth transmission chip unit.

[0018] Step 204. The radio frequency processor completes the establishment of the connection with the Bluetooth host during the Bluetooth communication initialization.

[0019] Step 205. The radio frequency processor modifies the connection interval with the Bluetooth host to T during the Bluetooth connection determination.

[0020] Step 206. After the connection interval modification is completed for 1 / Fs / 10, the coprocessor controls the start of the EEG acquisition task.

[0021] Furthermore, the first clock is a 24 MHz crystal oscillator; the second clock is a 32.768 KHz crystal oscillator.

[0022] The present invention can also adopt the following technical solutions:

[0023] A method for reducing the power consumption of an EEG acquisition device, including:

[0024] Step 101. The main processor initializes the main control and Bluetooth transmission chip unit.

[0025] Step 102. The radio frequency processor performs Bluetooth communication initialization.

[0026] Step 103. After the initialization of the main control and Bluetooth transmission chip unit and the communication initialization are completed, wait for the establishment of the Bluetooth communication connection.

[0027] Step 104. After the Bluetooth connection is determined, the main processor and the radio frequency processor go to sleep.

[0028] Step 105. The coprocessor initializes the analog acquisition front-end unit, configures the acquisition parameters, and controls the EEG acquisition task.

[0029] Step 106. After the EEG acquisition task is started, wait for the analog acquisition front-end unit to collect data at the sampling points.

[0030] Step 107: The coprocessor periodically reads and caches the data of the analog acquisition front-end unit;

[0031] Step 108: Wait for the data cached by the coprocessor to reach the cache threshold;

[0032] Step 109: After the amount of data cached by the coprocessor reaches the data threshold N, wake up the main processor;

[0033] Step 110: The main processor wakes up the Bluetooth sending task, that is, wakes up the radio frequency processor, and then the main processor enters the sleep mode again;

[0034] Step 111: The radio frequency processor sends the collected EEG data via Bluetooth and enters the sleep mode after the data transmission is completed;

[0035] When the data cached by the coprocessor reaches the predetermined threshold, the steps (106) to (111) are repeated and looped.

[0036] In the process of waking up the main processor after caching N specified data in the data cached by the coprocessor in Step 109, the quantity N of the data cached in the coprocessor and the two connection intervals T between the radio frequency processors should satisfy the basic relationship of N = T × Fs. The connection interval T is the interval for a data transmission between the two Bluetooth communication parties; the sampling rate Fs is the basis for selecting the connection interval T.

[0037] Beneficial effects

[0038] Compared with the traditional technical solution, the beneficial effects brought by the present invention are:

[0039] 1. The multi-core wake-up method for EEG acquisition of the present invention designs a low-power EEG acquisition module, which can reduce the power consumption of the EEG acquisition module, and further reduce the volume of the battery and the whole of the EEG acquisition module.

[0040] 2. Through this method, the present invention can collect EEG data with lower power consumption in the data acquisition stage and perform Bluetooth transmission of EEG data with lower power consumption in the TX stage. When the sampling rate of the EEG acquisition module is 1000 Hz, the Bluetooth connection interval is 10 ms, and it is powered by a 45 mAh battery, using this method, the working time of the device is increased by about 80 min compared with the device without using this method. At this time, the volume of the EEG acquisition module is <8 cm 3 , and the total volume is <8 g.

[0041] 3. The low-power EEG acquisition module of the present invention has the characteristics of quick wearing and portability, and can meet the requirements of long-term recording, which can solve the problems of traditional desktop EEG acquisition devices. Traditional desktop EEG acquisition devices often use cap-type acquisition electrodes, which are large in size and powered by mains electricity, and can only be applied in laboratories. Portable EEG devices are still relatively large in size and weight, and use fixing methods such as headbands and helmets, resulting in poor concealment and wearability, and are not easily accepted by consumers. Description of the Drawings

[0042] Figure 1 Multi-core wake-up system process for EEG acquisition;

[0043] Figure 2 Multi-core wake-up method based on CC2640R2F;

[0044] Figure 3 Composition of the low-power EEG acquisition module;

[0045] Figure 4 Calculation and setting of time parameters in the multi-core wake-up system process;

[0046] Figure 5 Working current diagram of the low-power EEG acquisition module;

[0047] Figure 6 Online experiment architecture diagram of the low-power EEG acquisition system for SSVEP. Detailed Embodiment

[0048] The following will describe the present invention in conjunction with the attached Figure 1 ~Attached Figure 6 The present invention will be described as follows:

[0049] The present invention provides a module for reducing the power consumption of an EEG acquisition device, and designs a multi-core wake-up system process for EEG acquisition. The low-power EEG acquisition module is composed of an EEG acquisition electrode unit, an analog acquisition front-end unit, a main control and Bluetooth transmission chip unit, a first clock, a second clock, and a power management unit. The main processor serves as the control center of the main control and Bluetooth transmission chip unit and undertakes tasks such as system control. The coprocessor is independent of the main processor and undertakes the task of EEG signal acquisition. The radio frequency processor undertakes the task of Bluetooth data communication. Among them, the main processor and the radio frequency processor are in the low-power mode most of the time and are only awakened when necessary to perform key task processing. The coprocessor enters the low-power mode when not performing signal acquisition and is only awakened when EEG data needs to be acquired.

[0050] The specific working process is as follows, as Figure 3 shown:

[0051] 101 The module is powered on and the main processor initializes the main control and Bluetooth transmission chip unit;

[0052] 102. Initialize Bluetooth communication by the RF processor;

[0053] 103. After the master control and the Bluetooth transmission chip unit are initialized and the communication is initialized, wait for the establishment of a Bluetooth communication connection;

[0054] 104. After the Bluetooth connection is established, the main processor and the RF processor enter the sleep state;

[0055] 105. Initialize the analog acquisition front-end unit by the coprocessor, configure the acquisition parameters, and control the EEG acquisition task;

[0056] 106. After the EEG acquisition task is started, wait for the analog acquisition front-end unit to collect data at the sampling points;

[0057] 107. Periodically read and cache the data of the analog acquisition front-end unit by the coprocessor;

[0058] 108. Wait for the data cached by the coprocessor to reach the cache threshold;

[0059] 109. When the amount of data cached by the coprocessor reaches the data threshold N, wake up the main processor;

[0060] 110. Wake up the Bluetooth sending task by the main processor, that is, wake up the RF processor, and then the main processor enters the sleep mode again;

[0061] 111. The RF processor sends the collected EEG data via Bluetooth and enters the sleep mode after the data transmission is completed;

[0062] When the data cached by the coprocessor reaches the predetermined threshold, repeat and loop through the steps (106) to (111).

[0063] As Figure 1 、 Figure 2 shown, under the condition that the sampling rate of the low-power EEG acquisition module is Fs, that is, the sampling rate of the analog acquisition front-end unit in the low-power EEG acquisition module is Fs, the number N of the data cached in the coprocessor and the two connection intervals T between the RF processors should satisfy the basic relationship of N = T×Fs. At the same time, the number N of the cached data should be the number of integer-point acquisitions of Fs within T. For example, when Fs is 1000Hz, T can be any integer multiple of 1ms, such as 10ms. At this time, N is 10, but T cannot be 1.1ms. If T is 1.1ms, then every 10ms, a situation where the sampling frequency does not match the cache interval will occur, resulting in data loss. At the same time, the connection interval T affects the real-time performance of the Bluetooth transmission of the data of the low-power EEG acquisition module. Therefore, the patent specifies the selection of T based on the sampling rate Fs and further stipulates the selection of the time-related parameters in the EEG acquisition multi-core wake-up system process.

[0064] The Bluetooth communication protocol stipulates that there should be a Bluetooth host and a Bluetooth slave during Bluetooth communication. After each connection interval between the Bluetooth host and the Bluetooth slave, a connection event occurs, and a Bluetooth data transmission task is completed. In this system, the upper computer module acts as the host role, and the low-power EEG acquisition module acts as the slave role. After the connection interval T, a connection event occurs, and N data cached by the coprocessor in the low-power EEG acquisition module are sent to the upper computer module.

[0065] At the same time, the Bluetooth protocol requires that when a single Bluetooth slave is connected, that is, when there is a low-power EEG acquisition module connected to the upper computer module in the system, the minimum value of the connection interval T is 7.5ms - 17.5ms, and T should also be an integer multiple of 1.25ms. Usually, in the Bluetooth protocol, the Bluetooth connection interval T is determined through negotiation between the two communication parties. The real-time performance of the low-power EEG acquisition module expects the EEG acquisition module to send the data as soon as a set of data is completed. Therefore, in order to ensure that the data sampled by the analog acquisition front-end module at different sampling rates Fs can be sent in a timely manner, the present invention designs a calculation method for the Bluetooth connection interval T, providing a way to calculate the connection interval of the low-power EEG acquisition module at different sampling rates:

[0066] T = t 1 + t 2 = 7.5 + 1.25×α (ms) (1)

[0067] In the formula, α is an unknown, t 1 represents the minimum connection interval in the case of a single slave, which is 7.5ms, and t 2 is 1.25×α, which is the calculation step. Both addends in the formula are integer multiples of 1.25ms, meeting the requirement of the Bluetooth protocol that the Bluetooth connection interval time is an integer multiple of 1.25ms. The following gives the derivation formula of α.

[0068] For α, when the sampling rate is Fs, it is necessary to satisfy that (t 1 + t 2 )×F s is an integer, that is

[0069]

[0070] Meanwhile, since there should be no user event, i.e., EEG acquisition task, when a Bluetooth connection event occurs, the start time of the EEG acquisition task is set to 1 / Fs / 10 after the connection interval modification is completed. That is, after the connection interval modification is completed for 1 / Fs / 10, the analog acquisition front-end unit implements the first sampling. The method for selecting time-related parameters in the EEG acquisition multi-core wake-up system process specified in the present invention is applicable to the calculation of the connection interval in the low-power EEG acquisition module of the analog acquisition front-end unit at different sampling rates, and stipulates the start time of the EEG acquisition task, ensuring the real-time performance of the system. The specific process is as follows, as Figure 4 shown:

[0071] Step 201. Select the sampling rate of the low-power EEG acquisition module as Fs;

[0072] Step 202. During the initialization of the main processor for the main control and Bluetooth transmission chip unit, derive the connection interval T through Formulas 1 and 2;

[0073] Step 203. During the initialization of the main processor for the main control and Bluetooth transmission chip unit, set the coprocessor cache threshold N = T×Fs;

[0074] Step 204. The radio frequency processor completes the establishment of the connection with the Bluetooth host during the Bluetooth communication initialization;

[0075] Step 205. The radio frequency processor modifies the connection interval with the Bluetooth host to T during the Bluetooth connection determination;

[0076] Step 206. After the connection interval modification is completed for 1 / Fs / 10, the coprocessor controls the start of the EEG acquisition task.

[0077] When the analog acquisition front-end module operates at a sampling rate of 1 KHz, the main processor of the EEG acquisition module derives α as 2 through Formula 2, and the calculation formula 1 of the Bluetooth connection interval is 10 (ms). The coprocessor of the EEG acquisition module starts the EEG acquisition task after 0.1 ms of the connection interval modification, and the number of caches is 10. When the analog acquisition front-end operates at a sampling rate of 250 Hz, the main processor of the EEG acquisition module derives α as 10 through Formula 2, and the calculation formula 1 of the Bluetooth connection interval is 20 (ms). The coprocessor of the EEG acquisition module starts the EEG acquisition task after 0.4 ms of the connection interval modification, and the number of caches is 20.

[0078] If CC2640R2F is used as the main control and Bluetooth transmission chip, its main processor is an M3 core, the radio frequency processor is an M0 core, and it has a coprocessor. The M3 core completes the calculation and setting of time parameters in the multi-core wake-up workflow. When the coprocessor starts the EEG acquisition task, the M3 and M0 cores are in the sleep mode, and most clock signals stop. At this time, only the 32.768KHz RTC low-power clock is running, and the coprocessor works with this RTC clock. When the analog acquisition front end works at the set sampling rate, the coprocessor completes the acquisition and caching at each sampling point. When the storage threshold N is reached, an interrupt request is generated to wake up the M3 main processor. After receiving the interrupt request, the M3 processor clears the interrupt request flag, wakes up the M0 core to complete the Bluetooth data transmission, and then enters the sleep state again. This multi-core wake-up working method enables the M3 main processor and the M0 radio frequency processor not to run continuously, thus reducing the power consumption.

[0079] The proposed method for reducing the power consumption of EEG acquisition devices is used to design a low-power EEG acquisition module to complete the acquisition and transmission of EEG signals. This module consists of EEG acquisition electrodes, an analog acquisition front end, a main control and Bluetooth transmission chip, a power management unit, etc. The ADS1291 chip serves as the analog acquisition front end to complete the amplification, sampling, and AD conversion of EEG signals. The CC2640R2F main control and Bluetooth transmission chip caches and transmits the EEG type digital signals transmitted by the analog acquisition front end ADS1291, and completes the calculation and setting of time parameters in the multi-core wake-up workflow. The 24MHz crystal oscillator and the 32.768KHz crystal oscillator provide clocks for the chip to work. The LM27762 power management chip completes the supply of the chip working voltage, and the MCP73831 completes the battery charging.

[0080] Each low-power EEG acquisition module consists of a group of reference electrodes, active electrodes, and ground electrodes. The gold cup electrodes are selected for the reference electrodes and the active electrodes, and are fixed with Ten20 conductive paste. The reference electrode is fixed in the middle of the head, the active electrode is placed at the occipital area of the head, and the ear clip electrode is selected for the ground electrode and fixed at the earlobe position. The EEG signals after suppressing part of the common-mode signals through the ground electrode are connected to the preprocessing circuit in a differential manner through the reference electrode and the active electrode, filtered by the filter circuit and then connected to the ADS1291. The EEG signals after AD conversion by the ADS1291 are transmitted through the CC2640R2F, and finally the EEG signals are sent to the host computer module through Bluetooth communication and further processed in the host computer module.

[0081] Meanwhile, this low-power EEG acquisition module also has the function of input impedance detection. This function helps users judge the contact situation between the EEG acquisition electrode and the human head. Whenever the EEG acquisition function and the input impedance detection function are switched, operations in steps 104 and 105 will be performed once, that is, the connection interval will be updated again, and the analog acquisition front-end unit will be re-initialized, so that the EEG acquisition task is executed after the connection interval modification is completed by 1 / Fs / 10. When the EEG acquisition module is in the EEG signal acquisition function, the differential input of ADS1291 comes from the external electrode. When in the input impedance detection function, a DC signal of 250Hz and 6μA is generated inside ADS1291. At this time, the data collected by ADS1291 is the impedance of the acquisition electrode in contact with the human body after calculation. The smaller this value is, the better the contact between the electrode and the human body.

[0082] This example designs a low-power EEG acquisition module for EEG data acquisition. The acquisition circuit includes a low-power Bluetooth MCU for logic control and data transmission, a low-power controllable gain analog acquisition front-end chip for EEG signal acquisition, a charge pump and an LDO with positive and negative voltage outputs for power supply to each chip, and a lithium battery for device power supply; the low-power EEG acquisition module completes the acquisition and transmission of EEG signals, and consists of an EEG acquisition electrode, an analog acquisition front-end, a main control and Bluetooth transmission chip, a power management unit, etc. The ADS1291 chip serves as the analog acquisition front-end to complete the amplification, sampling, and AD conversion of EEG signals. The CC2640F2F main control and Bluetooth transmission chip completes the transmission of EEG data, and calculates and sets the time parameters in the multi-core wake-up workflow. The 24MHz crystal oscillator and the 32.768KHz crystal oscillator provide clocks for the chip to work. The LM27762 power management chip provides the working voltage for the chip, and the MCP73831 completes battery charging.

[0083] During the EEG acquisition process, this EEG acquisition module acts as a Bluetooth slave device and communicates with the Bluetooth master device, that is, the upper computer module. T1 is the current during the EEG acquisition process, and T2 is the current during data transmission in a Bluetooth connection interval. Several small time periods are divided in T2. t2.1 is the protection time for the module's real-time operating system wake-up, Bluetooth setting, and crystal oscillator; t2.2 is the time for the module to turn on the Bluetooth radio frequency and set it to the receiving mode; t2.3 is the time for the EEG acquisition module to receive host data as a slave; t2.4 is the time for the module to switch from the radio frequency receiver to the device transmitter; t2.5 is the time for the module to send the collected EEG data to the host; t2.6 is the time for the module's Bluetooth protocol stack to process the received data packet and set the sleep timer, and then enter the standby state.

[0084] Figure 5G1 and G2 in the figure are the time-current diagrams of the same EEG acquisition module without / with using the method proposed in this patent to reduce the power consumption of EEG acquisition. In the T1 stage, in G1, the M3 main processor executes the acquisition task, while in G2, the M3 main processor is in the sleep state, and the acquisition and caching of EEG data are completed by the coprocessor. By comparing the currents in the T1 stage of G1 and G2, it can be seen that lower power consumption can be achieved during the acquisition process based on this method. In the t2.5 stage, the M3 processor transmits the data sent to the M0 radio frequency core. Based on this method, after the main processor coordinates the acquisition and transmission times, it enters the sleep state. If the EEG acquisition module uses a sampling rate of 1000Hz and the Bluetooth transmission interval is 10ms, within this interval, the EEG data to be sent cached by the coprocessor and the protocol header and trailer of the Bluetooth protocol stack total 320bit. When the transmission rate of Bluetooth transmission is 1Mbps, the time difference Δt between the M3 main processor entering the sleep state and the M0 core completing the data transmission is 320us. By comparing the currents in the t2.5 stage of G1 and G2, it can be seen that lower power consumption can be achieved during the EEG data transmission stage based on this method. Ultimately, the overall power consumption of the EEG acquisition module is reduced.

[0085] This example is a low-power EEG acquisition system used for SSVEP online experiments. In this example, it mainly consists of five parts: a visual stimulation module, a label synchronization module, a low-power EEG acquisition module, an EEG signal classification module, and a host computer module. The low-power EEG acquisition module works based on multi-core wake-up and is used to complete the acquisition and transmission of EEG signals; the visual stimulation module is used for the presentation of visual stimuli; the label synchronization module is used for the alignment of EEG data and label data; the EEG signal classification module is used to realize the classification of EEG signals; the host computer module is used to provide users with the application of the low-power EEG acquisition system.

[0086] The low-power EEG acquisition module communicates with the host computer module via Bluetooth; the visual stimulation module, the label synchronization module, and the host computer module achieve hardware serial communication through the USB port; the EEG signal classification module communicates with the host computer module via TCP.

[0087] The SSVEP EEG experimental application scenario of the low-power EEG acquisition module refers to the visual stimulation module running stimuli at different frequencies. The host computer sends the EEG data collected by the low-power EEG acquisition module and the label data collected by the label synchronization module to the EEG signal classification module together. After classification, the flashing frequency of the current stimulus can be recognized, and the result is feedback to the user. The experimental architecture is as Figure 6 shown, and the specific process is as follows:

[0088] Set the visual stimulation module as an LCD display screen, and display a total of 40 stimulus blocks in 5 rows and 8 columns on the LCD display screen through Psychopy. The frequency range of the stimulus blocks is 10 - 18Hz, and the step of the stimulus blocks is 0.2Hz;

[0089] Turn on the low-power EEG acquisition module, turn on the host computer module, and select the sampling rate Fs of the EEG acquisition module;

[0090] Set the EEG acquisition module to the impedance detection function, install the EEG electrodes, and observe the impedance value at this time;

[0091] After the impedance drops below 10K, use the host computer module to switch the EEG acquisition module to the EEG acquisition function and run the visual stimulation module;

[0092] When the stimulation of a certain frequency of the visual stimulation module starts to flash, the stimulation label will be written to the host computer module through the label synchronization module;

[0093] The host computer module identifies the label event type and parses the position in the EEG data when the visual stimulation module starts to flash;

[0094] The host computer module acts as a TCP server, and the EEG signal classification module acts as a TCP client to complete the TCP transmission of data between the two processes;

[0095] The classification module receives the data to complete the recognition and stores the classification result in the result queue;

[0096] After the visual stimulation module flashes for a certain period of time, it stops flashing and waits for the classification result;

[0097] When there is data in the result queue, the stimulation display module takes out the result, displays it, and starts the next round of stimulation;

[0098] Although the present invention has been described above, the present invention is not limited to the above specific embodiments. The above specific embodiments are merely illustrative rather than restrictive. Under the inspiration of the present invention, those of ordinary skill in the art can also make many variations without departing from the purpose of the present invention, and these all fall within the protection scope of the present invention.

Claims

1. A system for reducing power consumption of an electroencephalogram (EEG) acquisition device, characterized in that: The system also includes a low-power EEG acquisition module, which is composed of an EEG acquisition electrode unit, an analog acquisition front-end unit, a main control and wireless transmission chip unit, a first clock, a second clock and a power management unit. The main control and wireless transmission chip unit is composed of a main processor, a radio frequency processor and a coprocessor, including: the main processor manages wireless protocols and application layer tasks, and completes the calculation and setting of time parameters in multi-core wake-up work; the coprocessor interacts with external sensors in an ultra-low power form, controls the EEG acquisition task, and realizes the initialization and data transmission of the analog acquisition front-end unit; the radio frequency processor completes the wireless transmission task, and sends the data collected by the coprocessor through the underlying wireless communication task, wherein: the low-power EEG acquisition module completes the collection of EEG signals and the calculation and setting process of time parameters including the communication connection interval of wireless transmission through a multi-core wake-up working mode, including: Step 201. Select a low power consumption EEG acquisition module sampling rate of Fs; Step 202. The main processor obtains the connection interval T by the following formula during the initialization of the main control and wireless transmission chip units; T=t1+t2=7.5+1.25×α(ms) (1) Where α is an unknown number, t1 is 7.5ms, which represents the minimum connection interval specified by the wireless protocol, and t2 is 1.25×α, which is the calculation step size. The two addends in the formula are both integer multiples of 1.25ms, which meets the requirement that the wireless connection interval time specified by the wireless protocol is an integer multiple of 1.25ms; For α, (t1+t2)×F must be satisfied s is an integer, that is Step 203: The main processor sets the coprocessor cache threshold N=T×Fs during the initialization of the main control and wireless transmission chip units; Step 204. The RF processor completes the establishment of a connection with the wireless host during wireless communication initialization; Step 205. The RF processor modifies the connection interval with the wireless host to T in the wireless connection determination; Step 206: After the connection interval is modified to 1 / Fs / 10, the coprocessor controls the EEG acquisition task to start executing.

2. A system for reducing power consumption of an EEG acquisition device according to claim 1, characterized in that: The first clock is a 24MHz crystal oscillator; the second clock is a 32.768KHz crystal oscillator.

3. A method for reducing power consumption of an electroencephalogram (EEG) acquisition device, characterized in that: The method realizes the collection and transmission of EEG signals in a multi-core wake-up working mode based on the system according to any one of claims 1-2, comprising: Step 101: The main processor initializes the main control and wireless transmission chip units; Step 102: The radio frequency processor initializes wireless communication; Step 103: After the master control unit and the wireless transmission chip unit are initialized and the communication is initialized, wait for the wireless communication connection to be established; After the wireless connection is confirmed in step 104, the main processor and the radio frequency processor are in sleep mode; Step 105: The coprocessor initializes the analog acquisition front-end unit, configures acquisition parameters, and controls the EEG acquisition task; Step 106: After the EEG acquisition task is started, wait for the analog acquisition front-end unit to acquire data at the sampling point; Step 107: The coprocessor periodically reads and caches the data of the analog acquisition front-end unit; Step 108 waits for the data in the coprocessor cache to reach the cache threshold; Step 109: When the amount of data cached by the coprocessor reaches a data threshold N, the main processor is awakened; Step 110: The main processor wakes up the wireless transmission task, that is, wakes up the radio frequency processor, and then the main processor enters the sleep mode again; Step 111: The radio frequency processor wirelessly transmits the collected EEG data and enters a sleep mode after the data transmission is completed; When the coprocessor cache data reaches a predetermined threshold, steps 106 to 111 are repeated and executed cyclically.

4. The method for reducing power consumption of an EEG acquisition device according to claim 3, characterized in that: In the step 109, after the data cached by the coprocessor specifies N data and caches them, during the process of waking up the main processor, the number of data cached in the coprocessor, N, and the two connection intervals T between the RF processor should satisfy the basic relationship of N=T×Fs, where the connection interval T is the interval between two wireless communication parties for data transmission; the sampling rate Fs is the basis for selecting the connection interval T.

Images