Concentration degree estimation device, instrument provided with concentration degree estimation device, and air conditioner

Abstract

This concentration degree estimation device is provided with a Doppler sensor for detecting a user's pulse waves and an analysis unit for analyzing the pulse waves detected by the Doppler sensor. The analysis unit is provided with: a one-point concentration degree calculation unit for generating index values quantifying the pulse waves on the basis of chaos analysis in which the analysis source is the pulse wave shape displacement, which is the time-series displacement of the pulse wave waveform shape, and calculating one-point concentration degrees on the basis of the index values; a short-interval concentration degree calculation unit for averaging time-series data of one-point concentration degrees for short intervals and calculating short-interval concentration degrees; a long-interval concentration degree calculation unit for averaging time-series data of short-interval concentration degrees or one-point concentration degrees for a long interval longer than the short intervals and calculating a long-interval concentration degree; and a concentration degree change estimation unit for estimating a change in the user's concentration degree on the basis of the short-interval concentration degrees and the long-interval concentration degree.

Description

Concentration level estimation device, equipment equipped with concentration level estimation device, and air conditioning device

[0001] The present disclosure relates to a concentration level estimation device that estimates a concentration level, which is the degree of concentration of a user, and to an appliance and an air conditioning device that include the concentration level estimation device.

[0002] There is a concentration estimation device that determines the number of blinks, the amount of hand or body movement, the degree of pupil dilation, etc. from moving images acquired by an imaging means or a series of still images acquired, and estimates the user's level of concentration based on the determination results (see, for example, Patent Document 1). This concentration estimation device estimates the level of concentration based on the assumption that, when the level of concentration increases, the user's actions and states include increased hand movement, decreased body movement, decreased blinking, and slightly dilated pupils. Furthermore, the estimation results of this concentration estimation device are transmitted to an air conditioning system and used to control the airflow direction and temperature of the air conditioning system.

[0003] Japanese Patent Application Laid-Open No. 2019-082311

[0004] The concentration level estimation device in Patent Document 1 estimates the concentration level based on pre-determined assumptions, such as assuming that a decrease in the number of blinks indicates an increase in the concentration level. However, a decrease in the number of blinks does not necessarily mean that the concentration level has increased, and the concentration level estimation device in Patent Document 1 has room for improvement in estimation accuracy.

[0005] Furthermore, when the estimation results of the concentration level estimation device are used to control equipment such as an air conditioner, it is important for the equipment to grasp not only the current concentration level but also the decrease or increase in the concentration level over time, i.e., the change in the concentration level. However, in Patent Document 1, there is room for improvement in the accuracy of estimating the concentration level to begin with, and therefore there is also room for improvement in the accuracy of estimating the change in the concentration level.

[0006] The present invention is intended to solve the above-mentioned problems and aims to provide a concentration level estimation device that can estimate concentration levels and changes in concentration levels with high accuracy, as well as equipment and air conditioning equipment that are equipped with a concentration level estimation device.

[0007] The concentration level estimation device according to the present disclosure includes a Doppler sensor that detects a user's pulse wave, and an analysis unit that analyzes the pulse wave detected by the Doppler sensor. The analysis unit generates an index value that quantifies the pulse wave based on chaos analysis using pulse wave shape displacement, which is a time-series displacement of the waveform shape of the pulse wave, and calculates a one-point concentration level based on the index value. The analysis unit also includes a one-point concentration level calculation unit that calculates a short-interval concentration level by averaging time-series data of the one-point concentration level over a short interval, a long-interval concentration level calculation unit that calculates a long-interval concentration level by averaging time-series data of the one-point concentration level or the short-interval concentration level over a long interval that is longer than the short interval, and a concentration level change estimation unit that estimates changes in the user's concentration level based on the short-interval concentration level and the long-interval concentration level.

[0008] The equipment disclosed herein comprises the above-mentioned concentration estimation device and a control device that acquires concentration information indicating one of a single-point concentration level, a short-term concentration level, a long-term concentration level, and a change in concentration level from the concentration estimation device, and controls the operation of the equipment itself based on the acquired concentration level information.

[0009] The air conditioning system of the present disclosure comprises the above-mentioned concentration estimation device, an air conditioning unit that conditions the indoor space, and a control device that acquires concentration information indicating one of a single-point concentration level, a short-term concentration level, a long-term concentration level, and a change in concentration level from the concentration estimation device, and controls the air conditioning unit based on the acquired concentration level information.

[0010] According to the present disclosure, a concentration estimation device, a device, and an air conditioner equipped with the concentration estimation device can calculate a single-point concentration level, i.e., estimate a concentration level at a certain point in time, based on an index value generated based on a time-series change in the waveform shape of a pulse wave detected by a Doppler sensor in an analysis unit. Here, a pulse wave is vital data related to the pulsation of the heart and, ultimately, the activity of the nervous system in the brain. The concentration estimation device estimates a single-point concentration level based on an index value quantified based on such a pulse wave, thereby enabling a highly accurate estimation of the single-point concentration level. Furthermore, the analysis unit calculates a short-term concentration level and a long-term concentration level, and estimates a change in concentration level based on the short-term concentration level and the long-term concentration level. The short-term concentration level and the long-term concentration level are calculated based on the single-point concentration level, and the single-point concentration level is estimated with high accuracy, so that a change in concentration level can also be estimated with high accuracy.

[0011] 1 is a block diagram showing a configuration of a concentration estimation device according to embodiment 1 and a configuration for using the concentration estimation device. FIG. 1 is a schematic diagram of an antenna surface of a Doppler sensor according to embodiment 1. FIG. 2 is a schematic diagram of a board component mounting surface of a Doppler sensor according to embodiment 1. FIG. 3 is a diagram showing an example of a pulse wave detected by a Doppler sensor according to embodiment 1. FIG. 4 is a diagram showing time-series data of pulse wave shape displacement in the concentration estimation device according to embodiment 1. FIG. 5 is a conceptual diagram of an attractor in chaos analysis of the concentration estimation device according to embodiment 1. FIG. 6 is a conceptual diagram of Lyapunov exponent generation in the concentration estimation device according to embodiment 1. FIG. 7 is a bar graph showing the relationship between the Lyapunov exponent and various actions that result in different brain activities of the central nervous system. FIG. 8 is a diagram showing an example of changes in the Lyapunov exponent when a user sequentially performs actions that result in different brain activities and therefore different levels of concentration. FIG. 9 is a flowchart of a concentration estimation process in the concentration estimation device according to embodiment 1. FIG. 10 is a block diagram of an analysis unit of the concentration estimation device according to embodiment 1. FIG. 11 is a conceptual diagram of a first-order short-term concentration level, a second-order short-term concentration level, and a long-term concentration level calculated in the concentration estimation device 1 according to embodiment 1. Fig. 1 is a flowchart showing an outline of the processing of an analysis unit of the concentration degree estimation device according to embodiment 1. Fig. 2 is an explanatory diagram of a WORF technique used in a publicly known example. Fig. 3 is a diagram showing the configuration of an air conditioning device according to embodiment 2. Fig. 4 is a block diagram of an air conditioning device according to embodiment 2. Fig. 5 is a flowchart showing the operation of an air conditioning device according to embodiment 2.

[0012] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings. In each drawing, the same or corresponding parts are denoted by the same reference numerals, and their description will be omitted or simplified as appropriate.

[0013] Embodiment 1. Fig. 1 is a block diagram showing the configuration of a concentration estimation device 1 according to embodiment 1 and a usage configuration of the concentration estimation device 1. Fig. 2 is a schematic diagram of the antenna surface of a Doppler sensor 10 according to embodiment 1. Fig. 3 is a schematic diagram of the board component mounting surface of the Doppler sensor 10 according to embodiment 1. The concentration estimation device 1 is a device that estimates the concentration level of a user. The concentration estimation device 1 can objectify the concentration level by quantifying the concentration level.

[0014] The concentration level estimation device 1 includes a Doppler sensor 10 and an analysis unit 103. The Doppler sensor 10 emits a constant sine wave of approximately 24 GHz, known as a microwave or quasi-millimeter wave, toward a user whose concentration level is to be estimated. The user's body surface is displaced by the movement of blood vessels due to changes in blood flow caused by the pulsation of the heart. When the distance between the user's body surface and the Doppler sensor 10 changes, the waves reflected from the user's body surface change due to the Doppler effect.

[0015] The Doppler sensor 10 receives reflected waves from the user in response to blood vessel movement, and detects the pulse wave of the user's central nervous system based on the frequency difference between the reflected waves and the transmitted waves emitted from the Doppler sensor 10. A pulse wave is a waveform that indicates changes in the movement of a person's body surface due to cardiac pulsation, and includes waveforms of changes in blood vessel movement and waveforms of changes in the body surface of the heart. Blood vessels run throughout the user's body, and the Doppler sensor 10 can detect blood vessel movement not only in the heart but also in parts of the user's body, such as parts of the head or arms.

[0016] For distance measurement, a measurement frequency of 60 to 79 GHz is often used due to its high resolution. However, unlike large body movements, the purpose here is to detect pulse waves, and it is necessary to analyze minute fluctuations with ultra-low frequency characteristics of approximately 1 Hz. For this reason, analog detection using the 24 GHz Doppler method is suitable for detecting pulse waves.

[0017] As described above, the Doppler sensor 10 has the advantage of being able to detect the user's pulse wave without contact by using radio waves. Therefore, the Doppler sensor 10 can measure a wide area of ​​the user's body surface. The Doppler sensor 10 can measure vital data such as pulse rate, respiratory rate, body movement, sleep state, and autonomic nervous balance by analyzing the interval between peaks of the pulse wave, but in the first embodiment, it is used to detect the pulse wave.

[0018] Pulse wave detection sensors include non-contact Doppler sensors and contact sensors that detect by contacting the user. While there are many contact-type devices for measuring pulse, photoelectric pulse wave sensors are commonly used in wearable devices. Pulse wave sensors capture the changes in blood vessel volume that occur as the heart pumps blood as a waveform, and are equipped with a detector that monitors this volume change. Pulse wave sensors can obtain the pulse interval by counting the interval between peaks of the pulse wave obtained. The number of pulse beats per minute can be calculated by reciprocalizing the pulse interval. For example, if the average pulse interval is 800 ms (0.8 seconds), the pulse rate is 60 / 0.8, or 75 beats per minute.

[0019] Pulse wave sensors are divided into transmission and reflection types, which differ in their measurement method. Transmission pulse wave sensors measure pulse waves by irradiating the body surface with infrared or red light and measuring the change in blood flow that changes with the heartbeat as the amount of light that passes through the body. However, transmission pulse wave sensors are limited to areas that can be measured easily by infrared or red light, such as the fingertips or earlobes.

[0020] Reflective pulse wave sensors irradiate a living body with infrared light, red light, or light with a green wavelength around 550 nanometers and measure the light reflected from the body using a photodiode or phototransistor. Oxygenated hemoglobin is present in arterial blood, which has the property of absorbing incident light. Therefore, reflective pulse wave sensors can measure pulse wave signals by sensing the blood flow (changes in vascular pressure) that changes with the heartbeat in a time series. Specifically, pulse wave sensors include a light-emitting element and a light-receiving element. The light-emitting element emits light and the light-receiving element detects the reflected light from the finger. The intensity of the light reflected from the finger represents the increase or decrease in hemoglobin flowing through the capillaries in the fingertip. Reflective pulse wave sensors can obtain pulse wave time series data corresponding to the increase or decrease in hemoglobin (blood flow). Because reflective pulse wave sensors measure using reflected light, there is no need to limit the measurement location as with transmissive sensors.

[0021] However, contact-type sensors such as pulse wave sensors require wear for measurement, which is inconvenient and makes it difficult to measure from a distance. For this reason, the concentration estimation device 1 uses a non-contact Doppler sensor 10 as the sensor for measuring the user's pulse wave. By using the Doppler sensor 10, the concentration estimation device 1 can acquire the user's pulse wave in a non-contact manner and does not require the user to wear a sensor, thereby obtaining the data necessary for estimating the concentration level without the user having to go through the hassle of wearing a sensor. Furthermore, by using the Doppler sensor 10, the concentration estimation device 1 can measure the pulse wave over a wide area of ​​the user's body surface.

[0022] The Doppler sensor 10 mixes an oscillation signal with a received signal, extracts fluctuation components due to the Doppler effect, and generates a pulse wave IQ signal. Specifically, the Doppler sensor 10 has an antenna unit 100, a radio unit 101, an analog circuit unit 102, and a board unit 10a. The antenna unit 100 acquires the user's pulse wave, which is central nervous activity. The antenna unit 100 has a TX, which is an oscillation unit, and an RX, which is a reception unit. As shown in FIG. 2, the antenna unit 100 has multiple (12 in this example) antennas 100a. Each of the TX and RX has six antennas 100a.

[0023] The radio unit 101 generates 24 GHz radio waves called RF (Radio Frequency), transmits the radio waves from the TX, and receives the reflected waves from the RX. The analog circuit unit 102 performs IQ detection on the reflected waves, converts them into IQ signals, and converts the frequency components that represent the Doppler shift of the reflected waves. The analog circuit unit 102 also includes an analog amplifier filter unit (OPAMP) that extracts and amplifies the necessary frequency bands, as shown in FIG. 3, and an analog-to-digital converter (LDO) that enables numerical analysis.

[0024] The substrate unit 10a includes a connector unit and a memory for outputting information to the analysis unit 103 or a device equipped with the concentration estimation device 1. The wireless unit 101, the analog circuit unit 102, and the analysis unit 103 are covered with a metal shielding case as shown in Fig. 3. In Fig. 3, the shielding case portion is indicated by dots.

[0025] The analysis unit 103 is a part that analyzes the pulse wave detected by the Doppler sensor 10. The analysis unit 103 generates an index value that quantifies the pulse wave based on chaos analysis using pulse wave shape displacement, which is a time-series displacement of the waveform shape of the pulse wave (hereinafter referred to as pulse wave shape), as the analysis source, and estimates the concentration level based on the index value. The concentration level estimation by the analysis unit 103 will be described later. The analysis unit 103 outputs concentration level information that indicates the concentration level estimation result. The concentration level information is information related to the concentration level. The concentration level information output from the analysis unit 103 is input to the control content determination unit 104 or the cloud unit 106, which will be described later.

[0026] The analysis unit 103 is configured with a microprocessor unit. The analysis unit 103 includes a CPU (Central Processing Unit), a ROM (Read Only Memory), and a RAM (Random Access Memory), and the ROM stores a control program and the like. The analysis unit 103 is not limited to a microprocessor unit. For example, the analysis unit 103 may be configured with updatable firmware or the like. The analysis unit 103 may also be a program module that is executed by commands from a CPU (not shown). The analysis unit 103 may be provided outside the Doppler sensor 10 as a separate entity from the Doppler sensor 10, or may be provided on a board within the Doppler sensor 10 and edge processed within the sensor.

[0027] The concentration level estimated by the analysis unit 103 can be used to control equipment equipped with the concentration level estimation device 1. A specific example of equipment equipped with the concentration level estimation device 1 is an air conditioning device, the details of which will be described again in embodiment 4 below. The equipment equipped with the concentration level estimation device 1 has a control content determination unit 104 and an equipment control unit 105. The control content determination unit 104 determines the control content of the equipment based on the input concentration level information, generates control data, and outputs it to the equipment control unit 105. The equipment control unit 105 controls various actuators of the equipment based on the control data from the control content determination unit 104.

[0028] Furthermore, as a form of use of the concentration estimation device 1, the concentration level information output from the analysis unit 103 may be input to the cloud unit 106. The cloud unit 106 accumulates the concentration level information input from the analysis unit 103. The analysis unit 103 may further output vital data itself, which is pulse wave data obtained by the Doppler sensor 10, and accumulate it in the cloud unit 106. The concentration level information and vital data accumulated in the cloud unit 106 can be visualized by displaying them on the display unit 107. Furthermore, the concentration level information and vital data accumulated in the cloud unit 106 are not limited to being visualized, and can also be provided to the data collection unit 108 from another cloud via the cloud unit 106 and used for various purposes.

[0029] The display unit 107 is a display such as a liquid crystal panel for displaying information. The display unit 107 may be a display unit of a smartphone, or may be one that displays and visualizes the concentration level information on an application installed on the smartphone.

[0030] Conventional methods for estimating a person's level of concentration from images acquired by imaging devices estimate the level of concentration based on pre-defined assumptions, such as assuming that a decrease in the number of blinks indicates an increase in the level of concentration. However, a decrease in the number of blinks does not necessarily mean an increase in the level of concentration, and this method leaves room for improvement in estimation accuracy. Furthermore, while there are conventional devices that estimate people's emotions using electroencephalographs, quantifying alpha and beta brain waves is not easy. Furthermore, electroencephalograph-based estimation devices have issues with practicality and analysis time, such as the need for measurement data over a certain period of time for estimation processing, which makes it impossible to perform estimation in real time, the need to wear the electroencephalograph on the head, and the complexity of the system.

[0031] In contrast, the concentration level estimation device 1 of the first embodiment can estimate the concentration level in a short time by chaos analysis, which will be described below, from the pulse wave detected in a non-contact manner using the Doppler sensor 10 .

[0032] Chaos analysis is now described in detail. Physiological psychology involves estimating a person's physiological and psychological states from physiological changes observed in biosignals. In conventional physiological psychology, various biosignals, such as electroencephalograms (EEGs), electrocardiograms (ECGs), heartbeat intervals, blood pressure, and respiratory fingertip plethysmograms, have been analyzed using various techniques, resulting in significant insights. However, most analyses have relied on linear theory. However, biosignals contain nonlinear properties, which are known to fluctuate due to a nonlinear property called chaos. Chaos refers to a phenomenon in which, despite the system's deterministic state transition rules, the nonlinearity of the system itself creates complexity equivalent to that of a stochastic system. While the state of an object can be deterministically described using equations, the state of the object lacks any discernible pattern, exhibiting highly complex behavior similar to randomness.

[0033] However, chaotic phenomena are phenomena that appear chaotic at first glance, but in fact have definite rules behind them. In other words, the next phenomenon that occurs is not determined by probability, but is determined deterministically according to certain rules. The reason why an object appears chaotic despite following rules is that even if the movements of each of the elements that make up the object are simple, they become complex when they behave as a group. It has been determined that there is a high possibility that chaotic information exists in the biological signals generated by such complex systems.

[0034] In recent years, the effectiveness of chaos analysis for estimating human physiological and psychological states has been proven through various experiments. In conventional chaos analysis, information that cannot be univocally determined, such as a person's sense of warmth or cold or their psychological state, is treated as a chaos object, and by performing chaos analysis, correlations are found among information that at first glance appear unrelated. In chaos analysis, it is important to determine what is treated as a chaos object. If the object changes, it becomes a completely different concept.

[0035] The concentration level estimation device 1 targets brain activity of the central nervous system as a chaos target. Conventionally, no technology targets brain activity of the central nervous system as a chaos target. Concentration level is closely related to brain activity of the central nervous system. Therefore, the concentration level estimation device 1 targets brain activity of the central nervous system as a chaos target and estimates the concentration level based on the results of chaos analysis. The concentration level estimation device 1 performs chaos analysis using pulse wave shape fluctuations, which focus on the movement of the waveform itself, as the analysis source. The pulse wave shape fluctuations are expressed as time-series data of the displacement of the waveform shape of the pulse wave, in other words, time-series displacement data of the waveform shape of the pulse wave. The concentration level estimation device 1 generates an index value that quantifies the pulse wave based on chaos analysis using pulse wave shape displacement, which is the time-series displacement of the waveform shape of the pulse wave, as the analysis source, and estimates the user's concentration level based on the index value.

[0036] Chaos analysis involves the sequential implementation of the following three steps (1) to (3). Details of each step will be explained later. (1) The first step is to calculate a vector specified from the time series data of pulse waveform changes and a preset delay time. (2) The second step is to generate an attractor by arranging the vectors in chronological order in a three-dimensional state space. (3) The third step is to calculate the Lyapunov exponent, which is an index value, based on the trajectory of the attractor.

[0037] The heart's pulsation interval varies from beat to beat, but this originates in the brain and is transmitted to the heart via the autonomic nervous system. Therefore, the inventors of the present invention have come up with the idea that, in estimating the level of concentration, the heart's pulsation is related not only to the autonomic nervous system but also to the central nervous system, and that by finding a correlation between the heart's pulsation and the degree of brain activity, the level of concentration can ultimately be estimated.

[0038] Furthermore, there is a conventional technology that estimates the level of brain arousal using changes in pulse rate. The level of brain arousal can be considered an index similar to the level of concentration. For example, a technology for estimating the level of brain arousal determines that a large change in pulse rate indicates a high level of arousal, and a small change in pulse rate indicates a low level of arousal and drowsiness. The change in pulse rate is the time variation in the interval between peaks of the pulse wave, and is detected by pinpointing only the pulse wave peak information. Hereinafter, the time variation in the interval between peaks of the pulse wave may be referred to as one-dimensional pattern pulse displacement or pulse fluctuation. The term one-dimensional is used because the time variation in the interval between peaks of the pulse wave is not related to the height of the pulse wave, and vital data is obtained by frequency conversion of only the displacement of the pulse interval.

[0039] While the conventional techniques for measuring arousal levels use pulse rate changes, the inventors realized that pulse rate changes alone are insufficient to fully assess the complex neural activity of a living body. Therefore, the inventors sought a method different from pinpoint detection based solely on pulse wave peak information, and focused on pulse waveform displacement. However, when comparing the complexity of peak-to-peak fluctuations with the two-dimensional waveform pattern fluctuations of the pulse wave itself, the two-dimensional waveform pattern fluctuations (i.e., pulse waveform fluctuations) are incomparably more complex than the one-dimensional pulse interval fluctuations. This is thought to be because pulse waveform fluctuations are related to neuronal activity in the six layers of the cerebral cortex. For this reason, the inventors decided to use chaos instead of conventional analytical methods to analyze brain activity and, therefore, concentration.

[0040] Next, we will explain the chaos analysis performed by the analysis unit 103 of the concentration level estimation device 1. First, the analysis unit 103 obtains a pulse wave from the Doppler sensor 10. The following FIG. 4 shows an example of a pulse wave detected by the Doppler sensor 10.

[0041] FIG. 4 is a diagram showing an example of a pulse wave detected by the Doppler sensor 10 according to the first embodiment. The horizontal axis of FIG. 4 represents time, and the vertical axis represents analog pulse wave height. Here, pulse wave height refers to power. Pulse wave height does not refer to the power of only the peak portion of the pulse wave, where power is highest, but rather refers to the power at each time point in a time series including the peak portion of the pulse wave. The Doppler sensor 10 obtains the analog waveform shown in FIG. 4 and outputs it to the analysis unit 103. The pulse wave height decreases as the distance between the Doppler sensor 10 and the user being measured increases. However, since the chaos analysis of the first embodiment analyzes the variation in pulse wave height, the absolute value of the pulse wave height is not necessary information for quantifying the pulse wave. However, a larger pulse wave height results in a clearer pulse wave shape and improved accuracy in attractor generation, as described below. For this reason, a shorter distance between the Doppler sensor 10 and the user is preferable.

[0042] When analyzing time-series data of the pulse waveform, the analyzer 103 can obtain the pulse wave height necessary to ensure accuracy by doing the following: The analyzer 103 calculates the deviation of the pulse wave height, and if the deviation is smaller than a preset threshold, automatically adjusts the input signal amplification factor to artificially increase the analog pulse waveform. This clarifies the shape of the pulse wave, allowing the analyzer 103 to perform a highly accurate analysis even if the user is far from the Doppler sensor 10.

[0043] For example, when the distance from the Doppler sensor 10 to the user is short, the input signal magnification may be 1x. On the other hand, as the distance from the Doppler sensor 10 to the user increases, the pulse wave itself becomes smaller, making the pulse waveform displacement less visible and the deviation also decreases. Therefore, the analysis unit 103 can determine the magnification based on the smallness of the deviation. Specifically, for example, when the deviation is smaller than a first threshold, the analysis unit 103 sets the input signal magnification to 2x, and when the deviation is smaller than a second threshold that is smaller than the first threshold, the analysis unit 103 sets the input signal magnification to 3x. In other words, the farther the distance from the Doppler sensor 10 to the user, the greater the input signal amplification factor, thereby increasing the analog waveform. This allows the analysis unit 103 to analyze pulse waveform fluctuations (time-series deviation data of the pulse waveform) even when the distance from the Doppler sensor 10 to the user is long.

[0044] Furthermore, even if the distance from the Doppler sensor 10 to the user is the same, if the user's surface area and the power of the blood flow sent to the blood vessels are small, such as in the case of an infant, the pulse wave level is small and analysis becomes difficult. Therefore, the smaller the user's surface area and the smaller the deviation in pulse wave height, the larger the input signal amplification factor automatically increases, thereby increasing the analog waveform. This allows the analysis unit 103 to solve the problem of analysis being difficult when the pulse wave level is small.

[0045] Next, the analysis unit 103 generates an attractor, described below, from the pulse waveform. The pulse waveform is a two-dimensional waveform pattern of the pulse wave. Although it is difficult to find a pattern in the time series data of the pulse waveform from one waveform to another, when the time series data of the pulse waveform is converted into an attractor, a certain pattern exists. An attractor is a set toward which a dynamical system evolves over time. When a dynamical system moves from a point sufficiently close to an attractor, the system continues to remain sufficiently close to the attractor. There are no restrictions on the trajectory contained in an attractor other than that it must remain within the attractor.

[0046] Next, the first to third steps of the chaos analysis performed by the analysis unit 103 will be described in order. This first embodiment is characterized in that it targets brain activity of the central nervous system as a chaos target, and uses pulse wave shape displacement, which is a time-series displacement of the waveform shape of a pulse wave, as the analysis source for chaos analysis, and the chaos analysis method itself uses a conventionally known method. Therefore, the following description of the chaos analysis will be an outline.

[0047] (First Step) The first step is a step of calculating a vector specified from the time series data of the pulse waveform variation and a preset delay time, as described above.

[0048] FIG. 5 is a diagram showing time-series data of pulse waveform displacement in the concentration level estimation device 1 according to the first embodiment. The horizontal axis of FIG. 5 represents time, and the vertical axis represents analog pulse wave height. x(i) (i = 1, 2, ..., n) represents time-series data of the pulse waveform based on sensor data obtained by the Doppler sensor 10. The difference between the pulse wave height at x(i) and the pulse wave height at x(i + 1) represents the pulse waveform displacement. Note that x(0) is the initial value of the sensor data obtained over a period of the calculation window length. The calculation window length is an arbitrarily set time length, such as 20 seconds. A shorter calculation window length enables faster calculation, but a longer calculation window length provides more pulse waveform data and higher accuracy.

[0049] Using this time series data, in order to embed two-dimensional time series changes into a d-dimensional state space, in other words, to draw a trajectory in the d-dimensional state space, the analysis unit 103 sets a time delay τ that is an appropriate delay time and creates a vector. Specifically, the analysis unit 103 creates a vector X(i) = {x(i), x(i + τ), x(i + 2τ), ..., x(i + (d - 1) τ)}. For example, if d is set to 3 and two-dimensional time series changes are embedded in a three-dimensional state space, the analysis unit 103 creates a vector with three state variables. In the case of a three-dimensional state space, the vector X(i) is X(i) = {x(i), x(i + τ), x(i + 2τ)}. Since i = 1, 2, ..., n, n vectors X are created. Here, τ is a parameter called an embedding delay time.

[0050] (Second Step) The second step is a step of generating an attractor in which vectors are arranged in time series in a multidimensional state space of three or more dimensions.

[0051] FIG. 6 is a conceptual diagram of an attractor in the chaos analysis of the concentration estimation device 1 according to the first embodiment. By sequentially plotting this vector X(i) on the coordinate axes x(i), x(i+τ), x(i+2τ), ..., x(i+(d-1)τ), a trajectory is obtained. FIG. 6 shows a three-dimensional state space with three coordinate axes: x(i), x(i+τ), and x(i+2τ). For example, with d=3 and a delay time of 0.05 seconds, the vectors obtained in the first step are plotted (arranged) in chronological order in the three-dimensional state space, resulting in a trajectory like that shown in FIG. 6. This trajectory is the attractor trajectory. The shape of the attractor trajectory reveals a spiral trajectory, demonstrating the presence of chaos information in the fluctuations of the two-dimensional pulse waveform pattern.

[0052] (Third Step) The third step is to calculate the Lyapunov exponent, which is an index value based on the attractor trajectory. The Lyapunov exponent is determined by evaluating the instability or divergence of the attractor trajectory.

[0053] FIG. 7 is a conceptual diagram of Lyapunov exponent calculation in the concentration estimation device 1 according to the first embodiment. FIG. 7 illustrates a case where the multidimensional state space is a three-dimensional state space. To ultimately estimate the concentration level, it is necessary to quantify the time series displacement of the pulse waveform. The Lyapunov exponent is used as this quantified index value. Trajectory instability exists in attractors in a multidimensional state space of three or more dimensions. Instability can also be described as divergence. The Lyapunov exponent quantifies this trajectory instability. The Lyapunov exponent is a measure of how far two trajectories starting from two adjacent points move apart. In other words, the Lyapunov exponent is a value that represents the degree to which very close trajectories move apart in a dynamical system. Here, the larger the Lyapunov exponent, the greater the fluctuation range of the attractor, and the greater the width of the fluctuation.

[0054] To calculate the Lyapunov exponents, a tiny sphere (hypersphere) of radius ε is first assigned as an initial value to the orbit of the attractor in a three-dimensional chaotic dynamical system, as shown in Figure 7. Specifically, a sphere of infinitesimal radius ε, centered on a point on the attractor's orbit, is set as the hypersphere. In chaos analysis, points on the attractor's orbit that fall within this hypersphere are found, and the expansion rates in the e1, e2, and e3 directions are calculated by linearly approximating the change from each point to the point after the sliding time. The sliding time is the time interval from the previous sphere to the next sphere. The hypersphere is initially a sphere, but after a predetermined sliding time S, it is mapped once, stretching in the e1 direction, remaining almost unchanged in the e2 direction, and squashing in the e3 direction, resulting in an ellipsoid. Here, if the logarithms of the expansion rates per unit time in the e1, e2, and e3 directions are λ1, λ2, and λ3, respectively, these λ1, λ2, and λ3 are the Lyapunov exponents.

[0055] This set of Lyapunov exponents is called a Lyapunov spectrum. In chaos analysis, the sphere is repeatedly stretched for each slide time S, and the expansion rate is calculated. The overall Lyapunov spectrum is calculated by summing and averaging these values. The largest of the calculated Lyapunov exponents is called the maximum Lyapunov exponent. The maximum Lyapunov exponent can be used to determine the degree to which a dynamical system deviates from its initial state. In other words, the Lyapunov exponent indicates how far the nth state is from the initial state. In terms of brain activity, the further away the nth state is from the initial state, the more active the human brain is, and the less active the brain is. Of course, although accuracy will be reduced, it is also possible to calculate the Lyapunov exponent by comparing two states over a single slide time interval, rather than calculating multiple times for each slide time S.

[0056] In chaos analysis, shortening the slide time improves calculation accuracy but increases calculation time. Also, if there are too many points inside a sphere, calculation time increases, so to shorten calculation time, you can set the number of neighboring points as an upper limit on the number of points that can be inside a sphere. If more points than the set upper limit are counted inside a sphere, the calculation moves on to the next sphere.

[0057] Here, the case where the multidimensional state space is a three-dimensional state space has been described, but the analysis unit 103 can generate attractors in a state space of three or more dimensions. However, if the multidimensionality is five or more dimensions, the accuracy increases but the calculation time becomes longer and the practicality decreases. Therefore, taking into consideration the ability to obtain chaos and the calculation time, it is preferable that the multidimensionality is three or four dimensions.

[0058] Specifically, the Lyapunov exponents are calculated as follows. For example, assume that the radius of the sphere is 0.08, the calculation window length is 20 seconds, the slide time is 1 second, and the measurement frequency is 200 Hz. Because the measurement frequency is 200 Hz, 200 pieces of data are obtained per second, and 4,000 pieces of data are obtained from the start of calculation through the calculation window length. When the number of dimensions of the state space is three, a Lyapunov spectrum consisting of λ1, λ2, and λ3 for three dimensions is obtained using 4,000 pieces of data 20 seconds after the start of measurement. Then, a Lyapunov spectrum consisting of λ1, λ2, and λ3 is similarly obtained over the next second.

[0059] If the output window length is set to 60 seconds, the above process is repeated for 60 seconds. That is, the first Lyapunov spectrum is obtained in the first 20 seconds, and then a Lyapunov spectrum is obtained every second of the slide time for the next 40 seconds, resulting in a total of 40 Lyapunov spectra. The 40 Lyapunov spectra are then summed and averaged for λ1, λ2, and λ3 to determine the averaged λ1, λ2, and λ3. The largest Lyapunov exponent among these is the maximum Lyapunov exponent, and this maximum Lyapunov exponent is one piece of data obtained with the output window length. That is, the maximum Lyapunov exponent is used as the degree of deviation from the initial value of the system.

[0060] Here, the number of measurements of pulse wave data, which is vital data, will be considered. For example, when the measurement frequency is set to 200 Hz and the output window length is set to 60 seconds, 12,000 pieces of data are obtained in 60 seconds. The analysis unit 103 creates an attractor trajectory using the 12,000 pieces of data and calculates the Lyapunov exponents. Note that the output window length is the time required to output one piece of data, as described above, and is the time required to calculate one Lyapunov exponent. Furthermore, when the measurement frequency is set to 500 Hz and the output window length is set to 60 seconds, for example, 30,000 pieces of data are obtained in 60 seconds. In this case, the analysis unit 103 creates an attractor trajectory using the 30,000 pieces of data and calculates the Lyapunov exponents.

[0061] The more data there is, the greater the accuracy is, which is an advantage. However, when there is a large amount of data, there are disadvantages such as longer measurement time, an inability to follow changes in concentration level, and longer CPU processing time or inability to perform calculations. Conversely, the fewer data there is, the lower the accuracy is. However, when there is a small amount of data, there are advantages such as shorter measurement time, shorter CPU processing time, and better tracking.

[0062] Therefore, the measurement frequency is preferably approximately 100 Hz to 1000 Hz. Furthermore, the output window length is preferably 30 seconds or more and less than 5 minutes. Because the central nervous system has a fast reaction time, the output window length should be shorter than when measuring the autonomic nervous system, preferably 30 seconds or more and 60 seconds or less. Furthermore, the analysis unit 103 may be configured as follows to improve accuracy: The analysis unit 103 may collect and average one piece of output data every set time. For example, if the output window length is 30 seconds and the set time is 5 seconds, the analysis unit 103 collects one piece of output data every 5 seconds based on vital data from the 30 seconds prior to that time. Once 30 pieces of data have been collected, the analysis unit 103 may average the 30 pieces of data to obtain one piece of data. In this case, the analysis unit 103 can obtain one piece of highly accurate data (Lyapunov exponent) after spending 3 minutes on digitization.

[0063] Next, the correlation between the "Lyapunov exponent" and the "brain activity of the central nervous system" that is closely related to the concentration level, which was discovered by the present inventors, will be described with reference to Fig. 8. Fig. 8 is a diagram created based on the results of an experiment in which a user was made to perform various actions that resulted in different brain activities, and the Lyapunov exponent was calculated based on the pulse waves acquired from the user performing the actions.

[0064] Figure 8 is a bar graph showing the relationship between the Lyapunov exponent and various behaviors that involve different levels of central nervous system brain activity. The values ​​on the vertical axis are normalized with resting as 1. In other words, the values ​​on the vertical axis for behaviors other than resting are the Lyapunov exponent for that behavior divided by the Lyapunov exponent for resting. The horizontal axis shows human behaviors with different levels of brain activity. The following four behaviors were used for the horizontal axis: resting, reading an article, typing an email, and typing out documents. Reading an article refers to the behavior of reading documents on a computer or smartphone, for example. Typing out documents refers to the behavior of typing out the characters written in the documents into a computer.

[0065] Figure 8 shows that heavy tasks such as typing emails and copying documents, which are subjectively reported by users as requiring higher brain activity, have larger values ​​on the vertical axis than light tasks such as resting and reading articles. In other words, it was confirmed that the higher the level of brain activity, the larger the Lyapunov exponent. Therefore, it was confirmed that there is a relationship between the level of brain activity in the central nervous system and the Lyapunov exponent.

[0066] Furthermore, when resting, the level of concentration is lower than when typing documents. This is also confirmed by the user's subjective report. Therefore, there is a correlation between behavior, Lyapunov exponent, and concentration level, and it is believed that the Lyapunov exponent can be treated as a mental index of concentration. Specifically, it is believed that a large Lyapunov exponent indicates a high level of concentration, and a small Lyapunov exponent indicates a low level of concentration. Therefore, the vertical axis in Figure 8 can also be seen as an index of concentration level. In other words, a larger value on the vertical axis indicates a higher level of concentration, and a smaller value on the vertical axis indicates a lower level of concentration. Therefore, it is presumed that the value on the vertical axis can be regarded as a concentration index indicating, for example, the level of concentration.

[0067] FIG. 9 shows an example of changes in the Lyapunov exponents when a user sequentially performs actions that result in different brain activities and different levels of concentration. The horizontal axis represents the elapsed time (minutes) of the action, and the vertical axis on the left represents the Lyapunov exponents for that action. FIG. 9 also shows changes in the Lyapunov exponents when the action changes from a pre-work state (resting state) to a calculation / typing state, and then to a stop of the calculation / typing state. It was confirmed that the Lyapunov exponents, which were small before the work, increased during the calculation / typing state, and decreased when the work was interrupted. Here, the pre-work, calculation / typing, and interruption states of concentration are respectively low, high, and low. This indicates a relationship between the Lyapunov exponents and the level of concentration. In other words, a large Lyapunov exponent indicates a high level of concentration, while a small Lyapunov exponent indicates a low level of concentration. The vertical axis on the right in FIG. 9 represents the result of the concentration decline estimation, expressed as 0 or 1, which will be explained later.

[0068] The above measurement results confirmed that there is a correlation between the Lyapunov exponent and the level of concentration. Since a decrease in concentration during work leads to a decrease in work efficiency, the measurement results can be displayed on the display unit 107 to alert the user and encourage them to take breaks according to their level of concentration. The measurement results may be displayed as work efficiency measurement results. Note that "work efficiency" is not limited to the words or phrase "work efficiency," and may also be rephrased as productivity, production efficiency, work ability, progress, or smoothness, as long as it refers to actions that activate a person's emotions, mental state, or brain activity.

[0069] The Lyapunov exponents on the left vertical axis of FIG. 9 are values ​​obtained by calculation using the concentration estimation device 1 temporarily as an experimental device. Therefore, FIG. 9 shows that the Lyapunov exponents are obtained at one-minute intervals by the concentration estimation device 1. Furthermore, FIG. 9 proves that the process of calculating the Lyapunov exponents from the pulse wave can be performed in one minute, which is generally considered a short time. Note that while FIG. 9 shows experimental results obtained using the concentration estimation device 1, it merely shows experimental results measuring the correlation between the Lyapunov exponents and the concentration level, and may be generated using a dedicated experimental device instead of the concentration estimation device 1.

[0070] From the above, it can be said that there is a direct correlation between the Lyapunov exponent and concentration level, and that the Lyapunov exponent can also be used as a mental indicator of concentration level. Specifically, the larger the Lyapunov exponent, the higher the concentration level, and the smaller the Lyapunov exponent, the lower the concentration level. Furthermore, this method does not require linking to large amounts of past data or complex memorization, and can easily evaluate concentration level in real time.

[0071] Taking the above relationship into consideration, the concentration level estimation device 1 estimates the concentration level based on the Lyapunov exponent.

[0072] FIG. 10 is a flowchart of the concentration level estimation process in the concentration level estimation device 1 according to the first embodiment. The concentration level estimation device 1 includes a step of acquiring pulse waveform data (step S1) and a step of performing chaos analysis based on the pulse waveform data (step S2). As described above, the step of performing chaos analysis includes three steps (steps S21 to S23). The first step is a step of calculating a vector identified from time-series data of pulse waveform variations and a preset delay time (step S21). The second step is a step of generating an attractor in which the vectors are arranged in chronological order in a three-dimensional state space (step S22). The third step is a step of calculating a Lyapunov exponent, which is an index value, based on the trajectory of the attractor (step S23).

[0073] The analysis unit 103 estimates the concentration level based on the Lyapunov exponents (step S3) and outputs concentration level information indicating the estimation result. The analysis unit 103 estimates that a larger Lyapunov exponent indicates a higher concentration level, and a smaller Lyapunov exponent indicates a lower concentration level. The analysis unit 103 may output the Lyapunov exponents themselves as the concentration level information, or may convert the Lyapunov exponents into a scaled numerical value (e.g., 1 to 10) indicating the concentration level and output the scaled numerical value. The scaled numerical value indicates, for example, a low to high concentration level, in order from small to large. Furthermore, the analysis unit 103 may store, for example, Lyapunov exponents at rest and output a numerical value indicating the concentration level as a percentage (%) of the Lyapunov exponent analyzed by the analysis unit 103 relative to the Lyapunov exponent at rest. Furthermore, the analysis unit 103 is not limited to outputting the concentration level as a numerical value as described above, and may also output the concentration level as, for example, a scaled image. The staged image refers to, for example, an image in which the facial expression changes in stages according to the concentration level. As described above, the analysis unit 103 converts the Lyapunov exponents into concentration level information indicating the concentration level and outputs the information.

[0074] The concentration level information output from the analysis unit 103 is input to the display unit 107 via the cloud unit 106 and displayed. In this way, the concentration level estimation device 1 outputs and visualizes the concentration level information from the analysis unit 103, allowing the user to understand the concentration level. The concentration level information output from the analysis unit 103 may be accumulated in the data collection unit 108 via the cloud unit 106. The concentration level information output from the analysis unit 103 may also be input to and displayed directly on the display unit 107 without going through the cloud unit 106. The concentration level information output from the analysis unit 103 may also be input to the control content determination unit 104 and used to determine the control content of an appliance equipped with the concentration level estimation device 1.

[0075] Furthermore, the concentration level information may be information indicating the current concentration level, the time-series concentration level, or both.

[0076] As described above, when the subjective report indicates a low level of concentration, brain activity is low and the Lyapunov exponent is low. Therefore, the concentration level estimation device 1 preferably outputs from the analysis unit 103 a value indicating a lower level of concentration as the Lyapunov exponent is smaller.

[0077] Conventionally, biometric information such as body temperature, body movement, heart rate, and autonomic nervous activity has been used to estimate a user's concentration level. Examples of autonomic nervous activity include tone values ​​and entropy values. Because these biometric information are physical quantities completely different from concentration levels, it is highly likely that they change due to factors other than concentration levels. Therefore, methods for estimating concentration levels using these biometric information have the drawback of low resolution and prone to variations in the correlation between biometric information and concentration levels. Furthermore, to obtain reliable results, autonomic nervous activity requires outputting a single point based on a rate of change over at least five minutes, which requires at least ten minutes or more for time-series assessment, which is extremely time-consuming. Therefore, these methods require additional repetitions and time to accurately assess a decrease or increase in concentration levels.

[0078] In contrast, the concentration level estimation device 1 can calculate the Lyapunov exponents at short intervals, such as one minute, as described above, estimate the concentration level, and output the concentration level information.

[0079] In the above description, the analysis unit 103 estimates and outputs the user's concentration level at a certain time point (hereinafter referred to as one-point concentration level) at short intervals based on the Lyapunov exponent. Here, when the concentration level estimation device 1 is provided in a device such as an air conditioner, if the device controls the device based on the concentration level information output at short intervals from the analysis unit 103, the following problem occurs. For example, when the device is an air conditioner, if the concentration level fluctuates at short intervals and the fluctuations in concentration level are directly reflected in the air conditioning control or the warnings on the display, the air conditioning control may be changed or the warnings on the display may be issued frequently in a short period of time, which may cause discomfort to the user.

[0080] Taking into account the control performed by devices equipped with such concentration level estimation device 1, analysis unit 103 calculates and outputs, in addition to the one-point concentration level, a primary short-interval concentration level B obtained by averaging the time-series data of the one-point concentration level, and a secondary short-interval concentration level C obtained by averaging the time-series data of the primary short-interval concentration level B. In other words, in addition to the one-point concentration level, concentration level estimation device 1 calculates the average user concentration level for a certain interval and outputs concentration level information for each of the certain intervals. This enables device control to be performed at an appropriate frequency in devices equipped with concentration level estimation device 1, reducing the number of unnecessary device controls.

[0081] Furthermore, although the analysis unit 103 can quantify the concentration level at a certain point in time or over a certain period, the magnitude of the numerical value varies for each user. Therefore, when considering the versatility of a device intended for an unspecified number of users, it is further required to be able to estimate whether the concentration level has decreased or increased. By knowing whether the concentration level has decreased or increased, a device equipped with the concentration level estimation device 1 can accurately control the device in response to the decrease or increase in the concentration level, or can issue alerts using an app screen, sound, or light at the correct timing. Therefore, the analysis unit 103 can also estimate changes in the concentration level, as described in detail below.

[0082] The specific configuration of the analysis unit 103 will be described below.

[0083] FIG. 11 is a block diagram of the analysis unit 103 of the concentration estimation device 1 according to the first embodiment. The analysis unit 103 includes a single-point concentration level calculation unit 103a, a short-interval concentration level calculation unit 103g, a long-interval concentration level calculation unit 103d, a concentration level change estimation unit 103e, and a concentration level information output unit 103f. The short-interval concentration level calculation unit 103g calculates the short-interval concentration level by averaging short-interval time-series data of the single-point concentration level. The short-interval concentration level calculation unit 103g includes a primary calculation unit 103b that calculates the primary short-interval concentration level B and a secondary calculation unit 103c that calculates the secondary short-interval concentration level C. These calculation units are functionally configured by a CPU and a control program that constitute the analysis unit 103. The operation of each calculation unit, the primary short-interval concentration level B, the secondary short-interval concentration level C, and the long-interval concentration level D will be described below with reference to FIG. 11 and the following FIG. 12.

[0084] FIG. 12 is a conceptual diagram of the primary short-interval concentration degree B, secondary short-interval concentration degree C, and long-interval concentration degree D calculated by the concentration estimation device 1 according to the first embodiment. The one-point concentration degree calculation unit 103a calculates the Lyapunov exponents through the processing of the first to third steps described above, and estimates and outputs the one-point concentration degree A based on the Lyapunov exponents. Specifically, the one-point concentration degree calculation unit 103a first generates an attractor based on vectors X(0) to X(n) and calculates the Lyapunov exponents based on the trajectory of the attractor. The one-point concentration degree calculation unit 103a calculates and outputs the one-point concentration degree A, which is an index of the concentration at a single point in time, based on the calculated Lyapunov exponents. The Lyapunov exponents used in calculating the one-point concentration degree A may be raw values ​​of the Lyapunov exponents or values ​​normalized by, for example, reducing the raw values ​​of the Lyapunov exponents to 100% or scaling them. In any case, the one-point concentration degree calculation unit 103a calculates and outputs the one-point concentration degree A, which is one time point, based on the Lyapunov exponents.

[0085] As an example, when the time T1, which is the output window length, is 60 seconds and the sampling time interval is 1 second, the one-point concentration degree calculation unit 103a generates an attractor from 60 pieces of time series data from X(0) to X(60) obtained over 60 seconds, and outputs one one-point concentration degree A 60 seconds after the start of measurement. As another example, when the time T1 is 60 seconds and the sampling time interval is 2 seconds, the one-point concentration degree calculation unit 103a generates an attractor from 30 pieces of time series data obtained over 60 seconds, and outputs one one-point concentration degree A 60 seconds after the start of measurement.

[0086] The time T1 and sampling time interval can be selected appropriately depending on the accuracy, calculation time, and memory capacity. However, because human concentration levels can change even within a minute, it is optimal for the time T1 to be within one minute. Conversely, if the time T1 is longer, such as five minutes, the concentration level may change in different directions, such as decreasing or increasing, during the measurement, making it impossible to accurately depict an attractor, which is unsuitable. Furthermore, if the time T1 is long, such as five minutes, the single-point concentration level A will be smoothed over that long period, which has the disadvantage of making it difficult to see changes in concentration level during the measurement and resulting in reduced estimation accuracy. Therefore, it is best to set the time T1 to less than five minutes.

[0087] The one-point concentration degree calculation unit 103a repeats the process of outputting the one-point concentration degree A during measurement by the Doppler sensor 10. The one-point concentration degree calculation unit 103a uses time-series vector data of the vector X used to calculate the one-point concentration degree A, shifted by the sampling time interval. In the following description, it is assumed that the sampling time interval in each calculation unit of the analysis unit 103 is one second.

[0088] As a specific example, a case where time T1 is 60 seconds will be described. In this case, the one-point concentration degree calculation unit 103a generates an attractor based on 60 pieces of time-series vector data, X(1) to X(60), obtained over 60 seconds from the start of measurement, and calculates and outputs the first one-point concentration degree A(1). When calculating the next one-point concentration degree A(2), the one-point concentration degree calculation unit 103a shifts the time-series vector data and uses the time-series vector data, X(2) to X(61). In other words, the one-point concentration degree calculation unit 103a generates an attractor based on the time-series vector data, X(2) to X(61), and calculates and outputs the one-point concentration degree A(2). Similarly, the one-point concentration degree calculation unit 103a generates an attractor based on the time-series vector data, X(3) to X(62), and outputs the one-point concentration degree A(3). After outputting the first one-point concentration degree A(1), the one-point concentration degree calculation unit 103a repeats the process of outputting the one-point concentration degree A every second until the measurement is completed.

[0089] The primary calculation unit 103b acquires the time series data of the one-point concentration degree A output from the one-point concentration degree calculation unit 103a. The primary calculation unit 103b calculates the primary short-interval concentration degree B by performing averaging processing such as moving average processing on the time series data of the one-point concentration degree A acquired over the short interval. In FIG. 12 , the short interval in the primary calculation unit 103b is time T2. That is, the primary calculation unit 103b calculates the primary short-interval concentration degree B by averaging multiple one-point concentration degrees A over time T2. The primary calculation unit 103b outputs the calculated primary short-interval concentration degree B. The primary calculation unit 103b uses time series data of the one-point concentration degree A shifted by the sampling time interval to calculate the primary short-interval concentration degree B.

[0090] As a specific example, a case will be described where time T2 is 90 seconds, i.e., n in FIG. 12 is 90. In this case, the primary calculation unit 103b averages 90 pieces of time-series data A(1), ..., A(90) to calculate and output the primary short-interval concentration degree B(1). When calculating the next, second primary short-interval concentration degree B(2), the primary calculation unit 103b shifts the time-series data and uses the time-series data A(2), ..., A(91). In other words, the primary calculation unit 103b calculates and outputs the primary short-interval concentration degree B(2) using the time-series data A(2), ..., A(91). After outputting the first primary short-interval concentration degree B(1), the primary calculation unit 103b repeats the process of outputting the primary short-interval concentration degree B every second until the measurement is completed.

[0091] Since the primary short-interval concentration level B is calculated by averaging the time-series data of the one-point concentration level A, it is data that absorbs the ups and downs of the one-point concentration level A. In other words, the primary short-interval concentration level B is an index value that can accurately grasp the concentration level even if the time-series data of the one-point concentration level A fluctuates up and down. Therefore, when controlling equipment equipped with the concentration level estimation device 1 according to the concentration level, using the primary short-interval concentration level B as the concentration level can provide more stable control than using the one-point concentration level A. Human concentration levels tend to decrease at least once every 15 to 30 minutes, and decreases can be observed even within about one minute. For this reason, time T2, like time T1, is optimally set to within one minute. Setting time T2 too long can result in the concentration level being overly smoothed, potentially overlooking changes in concentration level. Therefore, it is preferable to set it to within three minutes.

[0092] Various methods are used for averaging, such as arithmetic averaging, weighted averaging, geometric averaging (geometric averaging), and harmonic averaging. Since the concentration has a time-series trend, it is preferable to use a moving average method for averaging.

[0093] The time series data of primary short interval concentration degree B is represented by a smooth curve on a graph with the vertical axis representing primary short interval concentration degree B, and ultimately time series data is obtained that reflects the trend of change in concentration degree for each interval of time T2. This method is effective for removing events that fluctuate frequently, such as brain activity or pulse waves, and for removing noise superimposed on the input signal.

[0094] Here, when it is desired to further reduce the calculation time or amount of calculations when estimating the concentration level, the concentration level estimation device 1 preferably uses a simple average. For example, if the time T2 is 90 seconds, the primary short-interval concentration level B may be a value obtained by simply averaging the 90 single-point concentration levels A, A(1), ..., A(90). In this case, although accuracy decreases, the advantage is that the amount of calculations is lighter.

[0095] The secondary calculation unit 103c acquires the time series data of the primary short interval concentration degree B output from the primary calculation unit 103b. The secondary calculation unit 103c calculates and outputs a secondary short interval concentration degree C by further averaging the time series data of the primary short interval concentration degree B acquired in the short interval. In FIG. 12 , the short interval in the secondary calculation unit 103c is time T3. That is, the secondary calculation unit 103c calculates the secondary short interval concentration degree C by averaging multiple primary short interval concentration degrees B for time T3. The secondary calculation unit 103c outputs the calculated secondary short interval concentration degree C. The secondary calculation unit 103c uses time series data of the primary short interval concentration degree B used to calculate the secondary short interval concentration degree C, shifted by the sampling time interval.

[0096] As a specific example, a case will be described where time T3 is 60 seconds, i.e., m in FIG. 12 is 60. In this case, the secondary calculation unit 103c averages 60 pieces of time-series data B(1), ..., B(60) to calculate and output the secondary short-interval concentration degree C(1). When calculating the next second secondary short-interval concentration degree C(2), the secondary calculation unit 103c shifts the time-series data and uses the time-series data B(2), ..., B(61). In other words, the secondary calculation unit 103c calculates and outputs the secondary short-interval concentration degree C(2) using the time-series data B(2), ..., B(61). After outputting the first secondary short-interval concentration degree C(1), the secondary calculation unit 103c repeats the process of outputting the secondary short-interval concentration degree C every second until the measurement is completed.

[0097] The one-point concentration degree A output from the one-point concentration degree calculation unit 103a, the primary short-interval concentration degree B output from the primary calculation unit 103b, and the secondary short-interval concentration degree C output from the secondary calculation unit 103c are input to the concentration degree information output unit 103f. The concentration degree information output unit 103f may output any one of the one-point concentration degree A, the primary short-interval concentration degree B, and the secondary short-interval concentration degree C as concentration degree information, or may output some or all of them.

[0098] The secondary short-interval concentration degree C is an index value that is smoothed further and has a more stable numerical value or numerical trend than the primary short-interval concentration degree B. On the other hand, the secondary short-interval concentration degree C is less likely to show small increases in concentration along the way. Therefore, when the concentration level information output unit 103f outputs one of the one-point concentration degree A, the primary short-interval concentration degree B, and the secondary short-interval concentration degree C, it can select and output the one according to the frequency, responsiveness, and estimated time at which control changes are desired in the equipment equipped with the concentration level estimation device 1.

[0099] Here, if it is desired to reduce the calculation time and amount while improving accuracy, the analysis unit 103 can perform the following process. The analysis unit 103 first sets time T2 to a time sufficiently shorter than time T1, and performs moving average processing on the time-series data of the one-point concentration degree A within that short time to calculate the primary short-interval concentration degree B. This improves the accuracy of the concentration estimation compared to when a simple average is used. For example, T1 is set to 30 seconds, and T2 is set to 10 to 15 seconds. The analysis unit 103 then averages the time-series data of the primary short-interval concentration degree B during time T3 using a simple average to calculate the secondary short-interval concentration degree C. If T3 is 30 seconds, the analysis unit 103 calculates the secondary short-interval concentration degree C by simply averaging 30 primary short-interval concentration degrees B. Using a simple average, the analysis unit 103 has the advantage of reducing the amount of calculation, although the accuracy decreases. By performing the above-described processing, the analysis unit 103 can reduce the calculation time and amount of calculation while improving accuracy.

[0100] The long-interval concentration level calculation unit 103d calculates the long-interval concentration level D. The long-interval concentration level calculation unit 103d calculates the long-interval concentration level D using time-series data of the single-point concentration level A or the short-interval concentration level acquired in a long interval longer than the short interval, using a method similar to that for the short-interval concentration level. In Figure 12, the long interval is time T4. Time T4 is set to a time longer than time T2, which is the acquisition time of the time-series data used to calculate the primary short-interval concentration level B.

[0101] The long-term concentration level calculation unit 103d calculates the long-term concentration level D by performing an averaging process such as a moving average process on the time-series data of the one-point concentration level A for the time T4. The long-term concentration level calculation unit 103d may also calculate the long-term concentration level D by performing an averaging process on the time-series data of the short-term concentration level. When the long-term concentration level calculation unit 103d uses the short-term concentration level to calculate the long-term concentration level D, it may use either the primary short-term concentration level B or the secondary short-term concentration level C. When the long-term concentration level calculation unit 103d uses the short-term concentration level when performing averaging by moving average process, the moving average time can be shorter than when using the one-point concentration level A, thereby reducing the storage area or calculation time. The averaging process may be a moving average process, or, if the number of points is small, a simple averaging process of multiple past short-term concentration levels may be used.

[0102] The long-term concentration level D is defined as an index value that represents the user's normal concentration level. The long-term concentration level D is used when estimating changes in concentration level in the concentration level change estimation unit 103e, which will be described later. To estimate whether the concentration level is changing, the current concentration level at the time of estimation and a concentration level to be compared are required. The long-term concentration level D is used as the comparison level. The current concentration level at the time of estimation is determined using the primary short-term concentration level B or the secondary short-term concentration level C.

[0103] As described above, time T4 is set to a time longer than time T2. Time T2 is, for example, approximately 1 minute, and at most approximately 5 minutes. A suitable value for time T4 is, for example, between 3 and 30 minutes. If time T4 is shorter than 3 minutes, there is no temporal difference from time T2, making it difficult to differentiate between the long-interval concentration level D and the primary short-interval concentration level B, making comparison difficult. If time T4 is three times or more the time T2, the long-interval concentration level D and the primary short-interval concentration level B can be differentiated, making comparison easier. However, if time T4 is longer than 30 minutes, the calculation result of the long-interval concentration level D may include abnormal behavior or state of the user, making it difficult to say that the long-interval concentration level D is an index value that indicates the user's normal concentration level. In this case, the accuracy of the estimation result of changes in concentration level may decrease. For this reason, it is preferable that time T4 be at most 60 minutes.

[0104] The concentration level change estimation unit 103e estimates a change in the user's concentration level. The concentration level change estimation unit 103e estimates a decrease or increase in the concentration level as the change in the concentration level. Furthermore, the concentration level change estimation unit 103e estimates the degree of change in the concentration level as the change in the concentration level. The concentration level change estimation unit 103e can estimate both an increase and a decrease in the concentration level, but the estimation concepts are basically the same for both, so the following will describe the case where a decrease in the concentration level is estimated.

[0105] The concentration level change estimation unit 103e estimates whether the concentration level is decreasing based on the short-interval concentration level calculated by the short-interval concentration level calculation unit 103g and the long-interval concentration level D calculated by the long-interval concentration level calculation unit 103d. The concentration level change estimation unit 103e may use either the primary short-interval concentration level B or the secondary short-interval concentration level C as the short-interval concentration level, but the following description will continue assuming that the primary short-interval concentration level B is used.

[0106] The concentration level change estimation unit 103e estimates a decrease in concentration level using the current primary short-interval concentration level B at the time of estimation and the past long-interval concentration level D. The current primary short-interval concentration level B at the time of estimation refers to the real-time primary short-interval concentration level B, including the latest one, input from the primary calculation unit 103b to the concentration level change estimation unit 103e at the time of estimation. The concentration level change estimation unit 103e uses the long-interval concentration level D from a past time period as the long-interval concentration level D. Because the long-interval concentration level D is used as a comparison target for the current concentration level, it is preferably calculated from measurement data from a time period in which the concentration level has not decreased relatively. Therefore, the concentration level change estimation unit 103e uses the long-interval concentration level D from a time period with the highest concentration level or a time period with a relatively high concentration level from the time-series data of the long-interval concentration level D from before the current time of estimation, calculated by the long-interval concentration level calculation unit 103d.

[0107] Furthermore, it is desirable that the concentration level change estimation unit 103e use, as the long-interval concentration level D, a long-interval concentration level D calculated using time-series data of one-point concentration levels A going back in time, which does not include the one-point concentration level A used to calculate the current primary short-interval concentration level B at the time of estimation. In the example of Figure 12, when the concentration level change estimation unit 103e estimates a decrease in concentration level at the output timing of the primary short-interval concentration level B(1), it is desirable that the concentration level change estimation unit 103e estimates the decrease in concentration level using the primary short-interval concentration level B(1) and the long-interval concentration level Da. The long-interval concentration level Da is calculated using time-series data of one-point concentration levels A for a time T4a going back in time, which does not include the one-point concentration levels A(1), ..., A(n) used to calculate the primary short-interval concentration level B(1).

[0108] If the long-term concentration degree D used to estimate a decrease in concentration degree is calculated by including the one-point concentration degrees A(1), ..., A(n) used in calculating the primary short-term concentration degree B(1), the current concentration degree will affect the calculated long-term concentration degree D. For this reason, it is desirable to calculate the long-term concentration degree D using time-series data of the one-point concentration degree A going back in time and not including the one-point concentration degrees A(1), ..., A(n) used in calculating the primary short-term concentration degree B(1). However, the long-term concentration degree D may also be calculated by including the one-point concentration degrees A(1), ..., A(n), although this will result in lower accuracy.

[0109] Here, an example has been described in which the long-term concentration degree D is calculated based on the one-point concentration degree A, but the long-term concentration degree D may also be calculated based on the short-term concentration degree of the primary short-term concentration degree B or the secondary short-term concentration degree C, as described above. When the long-term concentration degree D is calculated based on the short-term concentration degree, the long-term concentration degree D can be calculated more accurately if it is calculated without including the current short-term concentration degree at the time of estimation, but it may also be calculated including the current short-term concentration degree at the time of estimation.

[0110] Based on observations of users working, the inventors have determined that the optimal way to wake up a user whose concentration level is declining is to provide airflow or suggest a break once or twice every 5 to 30 minutes, and to repeat this process. For this reason, the long-term concentration level D used to estimate changes in concentration level is the long-term concentration level D calculated during measurement on the same day as the estimation, and is a long-term concentration level D that dates back to the past and does not include the current short-term concentration level at the time of estimation. For example, if the current short-term concentration level at the time of estimation is based on measurement data for a certain one minute, the long-term concentration level D used is calculated based on measurement data from an earlier period, not including that one minute, preferably from three to 30 minutes immediately preceding that minute.

[0111] Furthermore, the concentration level change estimation unit 103e repeatedly estimates and outputs changes in the concentration level at appropriate timing, and the following value may be used for the long-term concentration level D used for the second and subsequent re-estimations. The concentration level change estimation unit 103e may use the long-term concentration level D used the previous time that it was estimated that the concentration level had not decreased, as the long-term concentration level D used for the second and subsequent re-estimations. In this way, the concentration level change estimation unit 103e has the effect of preventing erroneous estimation that the concentration level has not decreased when in fact it has.

[0112] A method for estimating a decrease in concentration level using the primary short-interval concentration level B and the long-interval concentration level D will be described below. A concentration level decrease index value is used for this estimation. The concentration level decrease index value is the ratio of the primary short-interval concentration level B to the long-interval concentration level D, and is calculated as (primary short-interval concentration level B / long-interval concentration level D) x 100. The lower the concentration level, the smaller the value of each of the primary short-interval concentration level B and the long-interval concentration level D. The concentration level decrease index value is 100% when the long-interval concentration level D and the primary short-interval concentration level B are the same and the concentration level does not change, and approaches 0% as the degree of decrease in concentration level increases. A concentration level decrease index value greater than 100% indicates that the concentration level is increasing.

[0113] The concentration level change estimation unit 103e estimates that the concentration level has decreased when the concentration level decrease index value is equal to or less than a threshold value. The optimum threshold value is, for example, any value between 70% and 90%. If the threshold value is less than 70%, there is a higher risk that it will not be possible to estimate that the concentration level has decreased unless the concentration level has decreased significantly. If it is desired to reduce the frequency of device control, the threshold value may be set lower, but it is preferable that the threshold value be 50% or higher in order to at least estimate a decrease in the concentration level.

[0114] Furthermore, if the threshold is 70% or higher, the device equipped with the concentration level estimation device 1 can estimate a decrease in concentration level when the decrease in concentration level is not significant and control the device to promote wakefulness, thereby enabling the current concentration level to be maintained.The concentration level estimation device 1 uses a method that highly accurately estimates concentration level by indexing the degree of brain activity, but if the threshold is greater than 90%, there is a possibility that the device may erroneously estimate that the concentration level has decreased when it is not actually decreasing.For this reason, the threshold is set to 90% at most.In other words, if the threshold is set to any value between 70% and 90%, the risk of obtaining an estimation result indicating a decrease in concentration level only when the concentration level decreases significantly can be reduced, and erroneous estimation can be avoided.

[0115] The concentration level change estimation unit 103e outputs 1 when the concentration level decline index value is equal to or less than the threshold value and estimates that the concentration level has declined, and outputs 0 when the concentration level decline index value is greater than the threshold value and estimates that the concentration level has not declined. The concentration level change estimation unit 103e may output the concentration level decline index value itself as the estimation result. These values ​​output from the concentration level change estimation unit 103e are input to the concentration level information output unit 103f. Note that the concentration level change estimation unit 103e can also estimate that the concentration level has increased when the concentration level decline index value is greater than a threshold value that exceeds 100%.

[0116] The vertical axis on the right side of Fig. 9 shows an example of the result of estimating a decrease in concentration. The example in Fig. 9 shows the result of estimating a decrease in concentration when time T2 is 1 minute, time T4 is 5 minutes, and the threshold value of the concentration decrease index value is set to 80%. Fig. 9 shows that during a calculation typing task lasting approximately 20 minutes, the concentration decrease index value fell below the threshold twice, and 1 was output twice.

[0117] When estimating a decrease in concentration level, the concentration level change estimation unit 103e uses the long-term concentration level D calculated based on measurement data as a comparison target, and does not use a preset value. The reason for this is that, because the concentration level under normal circumstances differs for each user, using a preset value to estimate a decrease in concentration level would not result in an accurate estimation when targeting an unspecified number of users. If a decrease in concentration level cannot be accurately estimated, a device equipped with the concentration level estimation device 1 may over- or under-control the device. Therefore, the concentration level change estimation unit 103e uses the long-term concentration level D to estimate a decrease in concentration level. This allows the concentration level change estimation unit 103e to make an appropriate estimation even when measuring an unspecified number of users.

[0118] As described above, the concentration level information output unit 103f may output one of the one-point concentration level A, the primary short-interval concentration level B, and the secondary short-interval concentration level C as the concentration level information, or may output some or all of them. Furthermore, when the concentration level information output unit 103f receives a 1 from the concentration level change estimation unit 103e, it outputs concentration level information indicating that the concentration level has decreased. When the concentration level information output unit 103f receives a 0 from the concentration level change estimation unit 103e, it outputs concentration level information indicating that the concentration level has not decreased. As an additional method, in order to further improve estimation accuracy, the concentration level information output unit 103f may set a higher threshold and count the number of 1s input from the concentration level change estimation unit 103e, and determine that the concentration level has decreased when the count reaches multiple times. This allows a device equipped with the concentration level estimation device 1 to improve control accuracy while suppressing frequent control changes.

[0119] Furthermore, when the concentration level decline index value is input from the concentration level change estimation unit 103e, the concentration level information output unit 103f may output the concentration level decline index value itself as concentration level information. As described above, the concentration level decline index value is 1 when the concentration level does not change, and approaches 0 as the degree of decline in the concentration level increases. Therefore, the concentration level information output unit 103f can output the degree of decline in the concentration level as concentration level information using the concentration level decline index value.

[0120] FIG. 13 is a flowchart showing an outline of the processing performed by the analysis unit 103 of the concentration level estimation device 1 according to the second embodiment. The analysis unit 103 calculates a single-point concentration level A (step S11). Next, the analysis unit 103 calculates a primary short-interval concentration level B based on the time-series data of the single-point concentration level A calculated in step S11 (step S12). The analysis unit 103 calculates a long-interval concentration level D based on the time-series data of the single-point concentration level A calculated in step S11 or the primary short-interval concentration level B calculated in step S12 (step S13). The analysis unit 103 estimates an increase in concentration level based on the primary short-interval concentration level B and the long-interval concentration level D as described above (step S14) and outputs concentration level information (step S15). Note that the flowchart in FIG. 13 omits the calculation of the secondary short-interval concentration level C. However, if the secondary short-interval concentration level C is used to estimate an increase in concentration level, the analysis unit 103 naturally calculates the secondary short-interval concentration level C.

[0121] The analysis unit 103 may output some or all of the one-point concentration degree A, the primary short-interval concentration degree B, and the secondary short-interval concentration degree C as concentration level information, as described above, or may output information on whether the concentration level is decreasing. The analysis unit 103 can change the information that can be output as appropriate depending on the equipment in which the concentration level estimation device 1 is equipped.

[0122] Furthermore, the analysis unit 103 may output concentration level information indicating whether the concentration level is low or high based on any one of the one-point concentration level A, the primary short-interval concentration level B, and the secondary short-interval concentration level C. The analysis unit 103 may estimate whether the concentration level is low or high, for example, as follows: When the concentration level is expressed as a numerical scale and takes smaller values ​​as the concentration level decreases, the analysis unit 103 may estimate that the concentration level is low if the estimated concentration level is smaller than a threshold value, and may estimate that the concentration level is high if the estimated concentration level is larger than the threshold value.

[0123] The analysis unit 103 described above can output concentration level information every second after the Doppler sensor 10 starts measuring the user's pulse wave and calculates the initial one-point concentration level A. The analysis unit 103 may output the concentration level information every second or at a set timing.

[0124] For example, when repeatedly outputting concentration level information indicating whether the concentration level has decreased, the analysis unit 103 may set the timing to time T2 and output the concentration level information every time T2. As described above, the short-interval concentration level is required to estimate whether the concentration level has decreased, and time T2 is the time required to acquire the time-series data of the one-point concentration level A required to calculate the primary short-interval concentration level B. In other words, time T2 is the minimum time required to calculate the primary short-interval concentration level B. Therefore, by outputting the concentration level information every time T2, the analysis unit 103 can output an estimated result of a change in concentration level that is estimated based on the primary short-interval concentration level B calculated by replacing all of the time-series data of the one-point concentration level A each time. When repeatedly estimating a decrease in concentration level, the analysis unit 103 can accurately estimate the decrease in concentration level by using the primary short-interval concentration level B calculated by replacing all of the time-series data of the one-point concentration level A each time.

[0125] In the above, it is stated that the analysis unit 103 adjusts the output timing of the concentration level information, but it may also be configured so that it is output each time the concentration level is estimated, and the device that acquires the concentration level information selects the necessary concentration level information and uses it for device control.

[0126] It has been described above that the concentration estimation device 1 can output concentration information according to the results of estimating brain activity from index values. However, other words or characters may be used to express a person's emotions, mental state, or behavior that activates brain activity, as long as they have the same meaning. For example, "concentration" can be expressed as immersion, enthusiasm, enthusiasm, alertness, engrossment, excitement, elation, activity, hyperactivity, single-mindedness, freshness, etc. Therefore, information obtained from brain activity may be expressed as immersion, enthusiasm, enthusiasm, alertness, engrossment, excitement, elation, activity, hyperactivity, single-mindedness, freshness, etc., in addition to "concentration."

[0127] As described above, the concentration level estimation device 1 of the first embodiment includes a non-contact Doppler sensor 10, which is a sensor that uses radio waves to detect pulse waves over a wide area of ​​the user's body surface, and an analysis unit 103 that analyzes the pulse waves detected by the Doppler sensor 10. The analysis unit 103 includes a one-point concentration level calculation unit 103a, a short-interval concentration level calculation unit 103g, a long-interval concentration level calculation unit 103d, and a concentration level change estimation unit 103e. The one-point concentration level calculation unit 103a generates an index value that quantifies the pulse wave based on chaos analysis using pulse wave shape displacement, which is a time-series change in the waveform shape of the pulse wave, and calculates the one-point concentration level based on the index value. The short-interval concentration level calculation unit 103g calculates the short-interval concentration level by averaging the short-interval time-series data of the one-point concentration level. The long-term concentration level calculation unit 103d calculates the long-term concentration level by averaging time-series data of the single-point concentration level or the short-term concentration level for a long period longer than the short period. The concentration level change estimation unit 103e estimates changes in the user's concentration level based on the short-term concentration level and the long-term concentration level. For example, a moving average process can be used for the averaging process.

[0128] In this way, the concentration estimation device 1 can estimate the one-point concentration level A based on an index value generated based on the time-series change in the waveform shape of the pulse wave. The pulse wave is vital data related to the pulsation of the heart and, ultimately, the activity of the nervous system in the brain. Because the concentration estimation device 1 estimates the one-point concentration level A based on an index value calculated based on such a pulse wave, it can estimate the one-point concentration level A with higher accuracy than conventional estimation methods using eye blinking, etc. Furthermore, the analysis unit 103 calculates the short-term concentration level and the long-term concentration level and estimates changes in the concentration level based on the short-term concentration level and the long-term concentration level. Since the short-term concentration level and the long-term concentration level are calculated based on the one-point concentration level and the one-point concentration level is estimated with high accuracy, changes in the concentration level can also be estimated with high accuracy. Here, if the long-term interval is three or more times longer than the short-term interval, the long-term concentration level D can be differentiated from the short-term concentration level D, making comparison easier. Furthermore, considering the accuracy of the estimation result, a long-term interval of 3 to 30 minutes is preferable.

[0129] The concentration level estimation device 1 uses a Doppler sensor 10 as a sensor for detecting the user's pulse wave. Below, we will explain the advantages of the concentration level estimation device 1 estimating the concentration level based on the pulse wave detected using the Doppler sensor 10 rather than a contact-type pulse wave sensor.

[0130] Techniques for detecting a user's physical and mental state using the user's pulse wave include those disclosed in Japanese Patent Application Laid-Open Nos. 4-208136, 2009-195384, and 2015-16273. Each of these known techniques detects the pulse wave using a contact-type pulse wave sensor. Specifically, each of these known techniques uses a pulse wave sensor attached to the fingertip to measure the pulse wave by detecting the movement of hemoglobin flowing through capillaries using a photodiode or the like.

[0131] Contact-type pulse wave sensors, like the Doppler sensor 10, detect pulse vibrations as waves, but are limited to being worn on the fingertips or other locations and are farther away from the brain, which is the main part of brain activity used to estimate concentration. For this reason, the information acquired by the pulse wave sensor is more limited than that acquired by the Doppler sensor 10, making it less susceptible to the effects of brain activity, which is central nervous system activity.

[0132] The concentration level estimation device 1 uses a non-contact Doppler sensor 10 that uses radio waves of 24 GHz or higher to detect vibrations caused by blood vessel pulses. By using radio waves, the Doppler sensor 10 can simultaneously detect not only vibrations on the body surface caused by pulses, but also head information within the range of the radio waves, or vibrations of blood vessels inside the body, not just on the body surface. The Doppler sensor 10 can also acquire pulse wave information for the entire body within the radio wave area. This provides the advantage that the concentration level estimation device 1 can obtain more information about the autonomic or central nervous system, thereby improving the correlation accuracy of analysis of changes in pulse wave shape.

[0133] In this way, the concentration level estimation device 1 can obtain pulse wave information for the entire body, including the head, shoulders, and chest, as well as information from the finger parts that are farthest from the brain, using the Doppler sensor 10. Therefore, by using the Doppler sensor 10, the concentration level estimation device 1 can measure pulse waves that are more deeply influenced by brain activity, which is central nervous activity, and thus the level of concentration, and as a result, can estimate the level of concentration with high accuracy.

[0134] The Doppler sensor 10 can detect minute vibrations throughout the body caused by heartbeats. Internally, the Doppler sensor 10 mixes the received signal with the transmitted signal, extracts fluctuation components due to the Doppler effect, and generates pulse wave IQ signals. Strictly speaking, the pulse wave information generated by the Doppler sensor 10 is pulse wave information created using radio wave IQ signals, unlike the volume pulse wave generated by a pulse wave sensor attached to a fingertip.

[0135] The techniques disclosed in the above-mentioned Japanese Patent Laid-Open Nos. 4-208136, 2009-195384, and 2015-16273 use chaos analysis of pulse waves to detect the mental and physical state of a user. However, the information ultimately obtained by these known techniques differs from that of the concentration level estimation device 1 that estimates the concentration level of a user.

[0136] Japanese Patent Application Laid-Open No. 4-208136 obtains the subject's physical or psychological state, such as relaxing, reading, looking at a manga book, or admiring a beautiful painting. Japanese Patent Application Laid-Open No. 2009-195384 obtains the degree of fatigue as a measure of a person's fatigue. Japanese Patent Application Laid-Open No. 2015-16273 obtains the subject's mental balance, for example, whether the subject's mental state falls into the "ideal zone," "too tense," "depressed mood," "too relaxed," "following instincts," or "quasi-ideal zone."

[0137] In this way, the information that is ultimately obtained in each of the above-mentioned known examples is different, and they differ from the technology for estimating the user's concentration level.

[0138] Chaos analysis is a process that sequentially performs the following steps: calculating a vector identified from time-series data of pulse waveform displacement and a preset delay time; generating an attractor in which the vectors are arranged in chronological order in a multidimensional state space of three or more dimensions; and giving a hypersphere as an initial state to the trajectory of the attractor, and repeatedly stretching the hypersphere based on the trajectory for each preset sliding time, thereby calculating the expansion rate at which the hypersphere becomes an ellipsoid as a Lyapunov exponent, which is an index value.

[0139] In this way, the concentration level estimation device 1 can calculate the Lyapunov exponent, which is an index value, by chaos analysis.

[0140] Incidentally, the method used by the analysis unit 103 to calculate the Lyapunov index from the attractor in the chaos analysis of the pulse wave is significantly different from the WORF method used in the publicly known examples of the above-mentioned Japanese Patent Publication Nos. 4-208136, 2009-195384, and 2015-16273.

[0141] In Japanese Patent Laid-Open No. 4-208136, the next point is searched from the starting point of the finally obtained two-dimensional data, and the Lyapunov exponent is calculated from the ratio of time to the moving distance. In Japanese Patent Laid-Open No. 4-208136, the WORF method is used, in which the Lyapunov exponent calculated from the moving distance from one point to another is repeatedly averaged over the entire attractor to obtain the Lyapunov exponent.

[0142] In JP 2009-195384 A, an attractor is reconstructed for continuous data calculation values ​​at a predetermined time interval, and the time interval is slid by one second at a time to plot the value of the maximum Lyapunov exponent every second. Therefore, the method used in JP 2009-195384 A is also the WORF method.

[0143] In Japanese Patent Application Laid-Open No. 2015-16273, the WORF method is also used for detecting mental flexibility.

[0144] The WORF technique used in the prior art will be briefly described below.

[0145] 14 is an explanatory diagram of the WORF technique used in known examples. Here, as the WORF technique, the contents described in Alan WOLF et al., DETERMINING LYAPUNOV EXPONENTS FROM A TIME SERIES, Physica 16D (1985) 285-317 will be described in detail.

[0146] The WORF method is simpler than the Lyapunov exponent calculation method performed by the analysis unit 103, and is ０ and calculate the distance L between them after a unit time Δt. Δｔ and L ０ The WORF method calculates the logarithm of the ratio between Pi and . The WORF method performs this logarithm calculation for the entire chaotic attractor while shifting Pi, and the Lyapunov exponent λ1 is the simple average of the calculated logarithms. If the object of analysis has periodicity, λ1 is 0, and if it is chaotic, it is a positive value. The WORF method creates a two-dimensional image of the attractor and then calculates the Lyapunov exponent. The WORF method calculates the expansion or contraction of the distance between two points, making it easy to obtain information or behavior of the dynamical system.

[0147] On the other hand, the Lyapunov exponents are calculated in the analysis unit 103 in the same way as described above. In a three- or higher-dimensional trajectory, the expansion rate is calculated from the time change in each multidimensional direction of a microsphere (hypersphere) given as the initial state, and the expansion and contraction of the solid is calculated. Since the expansion rate is calculated from the way the multidimensional hypersphere continuously evolves over time, the amount of information is large and the calculation accuracy is improved. This is a major difference from WORF.

[0148] Furthermore, in known techniques, the Lyapunov exponents obtained by the WOLF method have a weak correlation with brain activity, and therefore with concentration, and the inventors' experiments and studies have not found any correlation with brain activity.

[0149] In contrast, the analysis unit 103 uses the above-mentioned method of giving microspheres (hyperspheres) to a multidimensional chaotic dynamical system and converting it into Lyapunov exponents, making it possible to instantly measure each state of different brain activity levels, and to obtain Lyapunov exponents as estimated indices that are highly correlated and linear with brain activity levels and, in turn, concentration levels.

[0150] Furthermore, the prior art differs from the present disclosure in the relationship between the Lyapunov exponents obtained based on user test results and brain activity. Japanese Patent Application Laid-Open No. 4-208136 describes that the Lyapunov exponents decrease as the level of concentration increases, and as brain information processing becomes more active. In other words, in Japanese Patent Application Laid-Open No. 4-208136, the Lyapunov exponents decrease as brain activity increases. In contrast, the Lyapunov exponents, which are index values ​​obtained by the analysis unit 103, increase as brain activity increases, in other words, as the level of concentration increases, and their meaning as index values ​​is completely opposite to that of the prior art.

[0151] Furthermore, Japanese Patent Laid-Open Publication No. 4-208136 describes that an attractor is output on a two-dimensional screen, and that the smaller the attractor, the more relaxed the user's state. Furthermore, Japanese Patent Laid-Open Publication No. 4-208136 describes the sparseness and density of the spiral local structure of the attractor, as the attractor is two-dimensional, and that the local structure changes from sparse to dense as the concentration of consciousness increases. However, the relationship between the sparseness and density of the attractor has absolutely no relation to the Lyapunov exponents obtained by the analysis unit 103.

[0152] Furthermore, the user test results in JP 2009-195384 A are for fatigue levels when seated on a wooden chair and a urethane seat chair, not for concentration levels. JP 2015-16273 A describes calculating Lyapunov exponents, but does not describe user test results related to Lyapunov exponents and concentration levels.

[0153] As described above, the relationship between the Lyapunov exponent and brain activity, and therefore the concentration level, in all of the above-mentioned known examples differs from that of the present disclosure.

[0154] The concentration level change estimation unit 103e uses the short-term concentration level currently in use at the time of estimation as the short-term concentration level used for estimation.

[0155] This allows the concentration level estimation device 1 to estimate changes in the user's concentration level at the current time in real time.

[0156] The concentration level change estimation unit 103e uses, as the long-term concentration level to be used for estimation, a long-term concentration level calculated using time series data of single-point concentration levels going back in time, which does not include the time series data of single-point concentration levels used to calculate the short-term concentration level to be used for estimation.

[0157] This allows the concentration level estimation device 1 to estimate changes in concentration level with high accuracy.

[0158] The short-term concentration degree and the long-term concentration degree D each take a smaller value as the concentration degree decreases, and the concentration degree change estimation unit 103e estimates that the concentration degree has decreased when the ratio of the short-term concentration degree to the long-term concentration degree D is below a predetermined threshold.

[0159] This allows the concentration level estimation device 1 to estimate that the concentration level has decreased, and by setting the threshold to a value between 70% and 90%, the risk of obtaining an estimation result that the concentration level has decreased only when the concentration level has decreased significantly can be reduced, and erroneous estimation can be avoided.

[0160] The short-interval concentration degree calculation unit 103g has a first-time calculation unit 103b that calculates the short-interval concentration degree calculated by averaging the time-series data of the one-point concentration degree A as a primary short-interval concentration degree B, and a secondary calculation unit 103c that calculates the secondary short-interval concentration degree C by averaging the time-series data of the primary short-interval concentration degree B. The short-interval concentration degree used in the long-interval concentration degree calculation unit 103d and the concentration degree change estimation unit 103e is the primary short-interval concentration degree B or the secondary short-interval concentration degree C.

[0161] In this way, the concentration estimation device 1 may use either the primary short-term concentration degree B or the secondary short-term concentration degree C as the short-term concentration degree used in the long-term concentration degree calculation unit 103d and the concentration degree change estimation unit 103e.

[0162] Second Embodiment The second embodiment relates to an apparatus equipped with the concentration level estimation device 1 of the first embodiment, and in particular, the case where the apparatus is an air conditioner will be described here.

[0163] <Configuration of Air Conditioner 201> Fig. 15 is a diagram showing the configuration of an air conditioner 201 according to embodiment 2. The air conditioner 201 is equipment that conditions an indoor space 271, which is an air-conditioned space. Air conditioning refers to adjusting the temperature, humidity, cleanliness, airflow, etc. of the air in the air-conditioned space, and specifically includes heating, cooling, dehumidification, humidification, and air purification.

[0164] 15 , an air conditioner 201 is installed in a house 203. The air conditioner 201 is a heat pump type air conditioning facility that uses, for example, HFC (hydrofluorocarbon) as a refrigerant. The air conditioner 201 is equipped with a vapor compression type refrigerant circuit and operates by obtaining power from a commercial power source, a power generation facility, or a power storage facility (not shown).

[0165] 15 , the air conditioner 201 includes an outdoor unit 211 provided outside a house 203, an indoor unit 213 provided inside the house 203, and a remote controller 255 operated by a user. The outdoor unit 211 and the indoor unit 213 are connected via a refrigerant pipe 261 through which a refrigerant flows and a communication line 263 through which various signals are transferred. The air conditioner 201 cools an indoor space 271 by blowing conditioned air, for example, cool air, from the indoor unit 213, and heats the indoor space 271 by blowing warm air.

[0166] The outdoor unit 211 includes a compressor 221, a four-way valve 222, an outdoor heat exchanger 223, an expansion valve 224, an outdoor blower 231, and an outdoor unit control unit 251. The indoor unit 213 includes an indoor heat exchanger 225, an indoor blower 233, and an indoor unit control unit 252. A refrigerant pipe 261 connects the compressor 221, the four-way valve 222, the outdoor heat exchanger 223, the expansion valve 224, and the indoor heat exchanger 225 in a ring shape. The air conditioning device 201 has a refrigerant circuit configured by connecting the compressor 221, the four-way valve 222, the outdoor heat exchanger 223, the expansion valve 224, and the indoor heat exchanger 225 by the refrigerant pipe 261. The refrigerant circuit circulates refrigerant to perform a refrigeration cycle operation.

[0167] The compressor 221 compresses the refrigerant and circulates it through the refrigerant pipe 261. Specifically, the compressor 221 compresses a low-temperature, low-pressure refrigerant and discharges the high-pressure, high-temperature refrigerant to the four-way valve 222. The compressor 221 is equipped with an inverter circuit that can change its operating capacity according to its drive frequency. The operating capacity is the amount of refrigerant that the compressor 221 delivers per unit time. The compressor 221 changes its operating capacity according to instructions from the outdoor unit control unit 251.

[0168] The four-way valve 222 is installed on the discharge side of the compressor 221. The four-way valve 222 switches the flow direction of the refrigerant in the refrigerant pipe 261 depending on whether the air conditioner 201 is operating in cooling operation, dehumidification operation, or heating operation. The expansion valve 224 is installed between the outdoor heat exchanger 223 and the indoor heat exchanger 225, and reduces the pressure of the refrigerant flowing through the refrigerant pipe 261 to expand it. The expansion valve 224 is an electronic expansion valve whose opening degree can be controlled variably. The expansion valve 224 changes its opening degree in accordance with instructions from the outdoor unit control unit 251 to adjust the pressure of the refrigerant.

[0169] The outdoor heat exchanger 223 exchanges heat between the refrigerant flowing through the refrigerant pipe 261 and the air in an outdoor space (external space) 272 that is outside the indoor space 271. The outdoor blower 231 is provided near the outdoor heat exchanger 223, draws in air from the outdoor space 272, and sends the drawn-in air to the outdoor heat exchanger 223. The air sent to the outdoor heat exchanger 223 exchanges heat with the refrigerant flowing through the refrigerant pipe 261, and is then blown out into the outdoor space 272.

[0170] The indoor heat exchanger 225 exchanges heat between the refrigerant flowing through the refrigerant pipe 261 and the air in the indoor space 271. The indoor blower 233 is provided near the indoor heat exchanger 225, draws in air from the indoor space 271, and sends the drawn air to the indoor heat exchanger 225. The air sent to the indoor heat exchanger 225 exchanges heat with the refrigerant flowing through the refrigerant pipe 261, and is then blown out into the indoor space 271. The air that has undergone heat exchange in the indoor heat exchanger 225 is supplied to the indoor space 271 as conditioned air. In this way, the indoor space 271 is conditioned.

[0171] The outdoor unit control unit 251 controls the operation of the outdoor unit 211. The indoor unit control unit 252 controls the operation of the indoor unit 213.

[0172] A remote controller 255 is disposed in the indoor space 271. The remote controller 255 transmits and receives various signals to and from an indoor unit control unit 252 included in the indoor unit 213. The remote controller 255 includes a display unit 255a, as shown in FIG. 16 (described later). The display unit 255a includes a touch screen, a liquid crystal display, and LEDs (light-emitting diodes). The remote controller 255 also includes a push button (not shown). The remote controller 255 functions as a command receiving unit that receives various commands from the user and a display unit that displays various information to the user. The user inputs commands to the air conditioner 201 by operating the remote controller 255. The commands include, for example, a command to switch between operation and stop, or a command to switch the operation mode, set temperature, set humidity, airflow, air direction, or timer. The air conditioner 201 operates in accordance with the input commands.

[0173] 16 is a block diagram of an air conditioner 201 according to embodiment 2. The air conditioner 201 includes a control device 250, an air conditioning unit 280, and the concentration level estimation device 1 according to embodiment 1. An information device 290 operated by a user is connected to the air conditioner 201 via a network N.

[0174] The control device 250 controls the entire air conditioning device 201. The control device 250 also controls the operation of the device itself, in other words, the air conditioning unit 280, based on the concentration level information output from the concentration level estimation device 1. The control device 250 includes the outdoor unit control unit 251 and indoor unit control unit 252 described above. Although not shown in FIG. 16 , the control device 250 also includes the control content determination unit 104 and device control unit 105 described in the first embodiment. The control content determination unit 104 and the device control unit 105 may be provided in either the outdoor unit control unit 251 or the indoor unit control unit 252.

[0175] The outdoor unit control unit 251 includes a control unit 251a, a storage unit 251b, a timer unit 251c, and a communication unit 251d. These units are connected via a bus (not shown).

[0176] The control unit 251a controls the entire outdoor unit. The storage unit 251b is composed of memory such as RAM or ROM, and stores data necessary for control. The timing unit 251c is a unit that measures time. The communication unit 251d is an interface for communicating with the indoor unit control unit 252 via a communication line 263 (see FIG. 15).

[0177] 15 , the outdoor unit control unit 251 is connected to the indoor unit control unit 252 via a communication line 263. The outdoor unit control unit 251 cooperates with the indoor unit control unit 252 by receiving various signals from the indoor unit control unit 252 via the communication line 263.

[0178] The indoor unit control unit 252 includes a communication unit 252a that communicates with the outdoor unit control unit 251 and the remote controller 255. The communication unit 252a is an interface for communicating with the outdoor unit control unit 251 and the remote controller 255. The communication unit 252a is further connected to an information device 290 via a network N. The communication unit 252a receives various commands from the user from the remote controller 255 and transmits the various commands received from the remote controller 255 to the indoor unit control unit 252. The communication unit 252a also transmits notification information to the remote controller 255 to notify the user.

[0179] The outdoor unit control unit 251 and the indoor unit control unit 252 are configured with microprocessor units. The outdoor unit control unit 251 and the indoor unit control unit 252 are equipped with a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), etc., and control programs and the like are stored in the ROM. Note that the outdoor unit control unit 251 and the indoor unit control unit 252 are not limited to microprocessor units. For example, the outdoor unit control unit 251 and the indoor unit control unit 252 may be configured with updatable firmware, etc. Furthermore, the outdoor unit control unit 251 and the indoor unit control unit 252 may be program modules that are executed in response to commands from a CPU (not shown) or the like. In addition, here, an example is shown in which the control device 250 has an outdoor unit control unit 251 and an indoor unit control unit 252, and is configured separately as an outdoor unit 211 and an indoor unit 213, but it may also be configured as a single control unit that has both functions.

[0180] The air conditioning unit 280 is a part that conditions the air in the indoor space 271, and corresponds to the refrigerant circuit, the outdoor fan 231, and the indoor fan 233 in FIG.

[0181] The air conditioning device 201 configured as described above controls the operation of the air conditioning unit 280 based on the concentration level information output from the concentration level estimation device 1. Specifically, if the concentration level information indicates a low level of concentration, the control device 250 of the air conditioning device 201 controls the air conditioning unit 280 to perform an operation to wake up the user. It has been conventionally recognized that exposure to wind wakes up users. Therefore, when the user's level of concentration is low, the control device 250 performs an awakening operation to wake up the user by, for example, increasing the rotation speed of the indoor blower 233 to increase the amount of air blown.

[0182] The air conditioner 201 also includes a motion sensor 256 that detects the user's position. When the user's level of concentration is low, the control device 250 controls the air conditioning unit 280 based on the user's position detected by the motion sensor 256 to perform the following awakening action to wake up the user. The control device 250 swings an up-and-down air deflector (not shown) provided on the indoor unit 213 up and down so that air intermittently hits the user, or swings a left-and-right air deflector (not shown) left and right so that air intermittently hits the user. Through the above control, the air conditioner 201 can wake up a user who has a low level of concentration.

[0183] It is also said that a room temperature about 1°C lower than the temperature that the user feels is comfortable has the effect of cooling the brain and improving work efficiency. In particular, when heating, it is said that a set temperature that is too high reduces concentration. For this reason, when the concentration level is low, the control device 250 controls the room set temperature to be slightly lower than the current temperature.

[0184] Conversely, when the level of concentration is high, the control device 250 reduces the rotation speed of the indoor blower 233 to reduce the amount of air blown, or controls the vertical air deflectors to point as upward as possible to prevent the wind from blowing on the user. The control device 250 also controls one or both of the vertical air deflectors (not shown) and the horizontal air deflectors (not shown) to prevent the wind from blowing on the user. Through the above control, the air conditioner 201 can prevent the user's attention from being directed toward the wind due to the wind blowing on the user, which can lead to a decrease in the level of concentration. In other words, the air conditioner 201 has the effect of allowing a user who is highly concentrated to work without being conscious of the wind.

[0185] Here, the concentration levels of the concentration level information output from the concentration level estimation device 1 include a one-point concentration level A, a primary short-interval concentration level B, and a secondary short-interval concentration level C. When the air conditioner 201 controls the operation of the air conditioning unit 280 based on the one-point concentration level A, i.e., the user's concentration level at a certain point in time, there is a possibility that control changes of the air conditioner 201 will be implemented frequently. In this case, the user may be bothered by the operation of the air conditioner 201 or may frequently feel uncomfortable due to changes in temperature or airflow. For this reason, by using the primary short-interval concentration level B or the secondary short-interval concentration level C when performing control based on the user's concentration level, the air conditioner 201 has the advantage of being able to determine control that accurately reflects trends and at an appropriate frequency.

[0186] However, because the secondary short-interval concentration degree C is a value calculated by performing averaging twice, when the air conditioner 201 performs control based on the secondary short-interval concentration degree C, it becomes more difficult to perform control that responds quickly to small decreases in concentration compared to when the air conditioner 201 performs control based on the primary short-interval concentration degree B. Furthermore, because the secondary short-interval concentration degree C is a value calculated by performing averaging twice, it takes more time to calculate it than the primary short-interval concentration degree B. For this reason, when the air conditioner 201 uses the secondary short-interval concentration degree C as the user's concentration degree, it takes more time to estimate a control change. Therefore, it is preferable for the air conditioner 201 to select either the primary short-interval concentration degree B or the secondary short-interval concentration degree C depending on the frequency, responsiveness, and estimated time for the desired control change.

[0187] As a wake-up action, the air conditioner 201 performs a swing action to send airflow toward the user or automatically changes the target temperature, but while such control has the effect of increasing the level of concentration, there are also cases where the airflow causes discomfort.In order to reduce the number of unnecessary wake-up airflows, the air conditioner 201 can also control the operation of the air conditioning unit 280 so as to perform a wake-up action when there is a change in the level of concentration that decreases, as shown in the flowchart in Figure 17.

[0188] 17 is a flowchart showing the operation of air conditioner 201 according to embodiment 2. Time T2 used in the following description is the same as time T2 in embodiment 1. Air conditioner 201 acquires concentration information output from concentration estimation device 1 (step S31). Air conditioner 201 acquires concentration information output from concentration estimation device 1 at every time T2. In the following description, concentration information R1, concentration information R2, concentration information R3, and concentration information R4 refer to the first, second, third, and fourth pieces of concentration information output from concentration estimation device 1 at every time T2, respectively.

[0189] The air conditioner 201 determines whether the concentration level information R1 acquired in step S31 indicates that the user's concentration level is decreasing (step S32). If the concentration level information R1 indicates that the user's concentration level is decreasing, the air conditioner 201 goes through step S33 (described later) and controls the air conditioning unit 280 to start a wake-up action to wake the user up (step S34). The wake-up action is a swing action that moves the vertical air deflectors up and down, as described above. The air conditioner 201 then counts the number of times the wake-up action has been performed (step S35). Here, the air conditioner 201 counts to 1.

[0190] The air conditioner 201 continues the awakening action until it acquires the next concentration level information R2. When the air conditioner 201 acquires new concentration level information R2 (step S36), it determines whether the concentration level information R2 indicates a decrease in the user's concentration level (step S37). If the concentration level information R2 indicates a continued decrease in the user's concentration level, the air conditioner 201 determines whether the count is less than the set number (step S38). If the set number is 3, the current count is 1, which is less than the set number. Therefore, the air conditioner 201 returns to step S34 and continues the awakening action. The air conditioner 201 may continue this awakening action with the same airflow volume as when it started the previous awakening action in step S34, or it may increase the airflow volume to further promote awakening. Increasing the airflow volume allows the air conditioner 201 to increase the user's concentration level more quickly. Then, the air conditioner 201 counts the number of times the awakening action has been performed to 2 (step S35).

[0191] The air conditioner 201 continues the awakening action until it acquires the next concentration level information R3. When the air conditioner 201 acquires new concentration level information R3 (step S36), it determines whether the concentration level information R3 indicates a decrease in the user's concentration level (step S37). If the concentration level information R3 indicates a continued decrease in the user's concentration level, the air conditioner 201 determines whether the count is less than the set number of times (step S38). The current count is 2, which is less than the set number of times (3). Therefore, the air conditioner 201 returns to step S34 and continues the awakening action. During this awakening action, the air conditioner 201 may continue the same airflow volume as when the awakening action was last performed in step S34, or may increase the airflow volume to further promote awakening. The air conditioner 201 then counts the number of times the awakening action has been performed to 3 (step S35).

[0192] The air conditioner 201 continues the awakening action until it acquires the next concentration level information R4. Then, when the air conditioner 201 acquires new concentration level information R4 (step S36), it determines whether the concentration level information R4 indicates that the user's concentration level is decreasing (step S37). If the concentration level information R4 indicates that the concentration level is continuing to decrease, the air conditioner 201 determines whether the counted number is less than the set number (step S38). The counted number is 3, which is not less than the set number. Therefore, the air conditioner 201 stops the awakening action (step S39).

[0193] In other words, if the user's concentration level does not increase even after performing the awakening action a set number of times, the air conditioner 201 stops the awakening action because there is a possibility that the user may feel uncomfortable due to the airflow. Although the set number is set to 3 here, the set number is not limited to 3. Also, although the air conditioner 201 performs the swing action as the awakening action, it may also perform control to lower the room temperature to cool the user's brain, although this takes time to wake the user. In other words, the above awakening action includes not only the swing action but also an action to lower the room temperature.

[0194] Furthermore, in step S37, if the concentration level information R2 or concentration level information R3 does not indicate that the user's concentration level is decreasing, that is, if the user's concentration level has increased and the user has woken up, the air conditioning device 201 stops the awakening operation without returning to the awakening operation of step S34 (step S39).

[0195] Here, step S33 will be described. When the air conditioner 201 estimates that the concentration level has decreased in step S32, it does not immediately start executing the wake-up action, but starts the wake-up action when it determines that an inactive time has elapsed since the previous wake-up action. In other words, the control device 250 of the air conditioner 201 performs the wake-up action once it estimates that the concentration level has decreased, and then has an inactive time during which it does not perform the wake-up action for a certain period of time. This is because it is thought that the user will maintain their wakeful state after being woken up by the wake-up action of the air conditioner 201.

[0196] If there were no inactive time, the air conditioner 201 would frequently continue to blow air onto the user, which could reduce comfort. By providing an inactive time, the air conditioner 201 can avoid causing discomfort to the user due to airflow or discomfort caused by hot or cold. The inactive time is preferably set to, for example, 10 to 15 minutes, which is the time when a decrease in concentration is often observed.

[0197] As described above, the air conditioner 201 can accurately perform a wake-up action at an appropriate frequency according to the user's concentration level by controlling the air conditioner 280 using the concentration level information output from the concentration level estimation device 1. As a result, the air conditioner 201 can maintain the user's work or study efficiency. Alternatively, the air conditioner 201 has the effect of suppressing a decrease in the user's concentration level and improving work efficiency.

[0198] Meanwhile, the remote controller 255 is positioned as one of the components of the air conditioner 201, while the information device 290 is positioned as a device owned by the user. The information device 290 has a display unit 291 such as a liquid crystal panel, and various information is displayed on the display unit 291. The information device 290 is configured as, for example, a smartphone or a tablet. An application for displaying the user's concentration level and the like is installed in the information device 290. Furthermore, the information device 290 can also be used in place of the remote controller 255 by installing an application for air conditioning control.

[0199] When operated by the user, the information device 290 starts an application, acquires the concentration level information output from the concentration level estimation device 1 via the network N, and displays it on the display unit 291. As a specific control, the control device 250 of the air conditioner 201 transmits the concentration level information output by the concentration level estimation device 1 to the information device 290 via the communication unit 252a in response to a request from the information device 290, and performs processing to display the information on the information device 290. Note that although the concentration level information is displayed on the information device 290 here, it may also be displayed on the display unit 255a of the remote controller 255.

[0200] In this way, the air conditioner 201 displays and visualizes the concentration level information on the information device 290 or the display unit 255a of the remote controller 255, allowing the user to visually confirm the concentration level.

[0201] Although the device equipped with the concentration estimation device 1 has been described as being an air conditioner 201, the device is not limited to the air conditioner 201 and can be incorporated into various devices such as electrical equipment, cars, or entertainment equipment. The concentration estimation device 1 can also be incorporated into, for example, a labor management device or a learning management device.

[0202] Furthermore, the concentration estimation device 1 can be incorporated into devices in a wide range of fields, such as healthcare, labor, education, sleep, mindfulness, meditation, customer service, marketing, or sports mental training. When the concentration estimation device 1 is incorporated into these devices, the devices can control the devices using the concentration estimation results and present the concentration estimation results to the user. In this way, by applying the concentration estimation device 1 to various devices, the concentration level can be visualized and presented to the user of the device.

[0203] 1 Concentration estimation device, 10 Doppler sensor, 10a Board unit, 100 Antenna unit, 100a Antenna, 101 Wireless unit, 102 Analog circuit unit, 103 Analysis unit, 103a One-point concentration calculation unit, 103b Primary calculation unit, 103c Secondary calculation unit, 103d Long-term concentration calculation unit, 103e Concentration change estimation unit, 103f Concentration information output unit, 103g Short-term concentration calculation unit, 104 Control content determination unit, 105 Equipment control unit, 106 Cloud unit, 107 Display unit, 108 Data collection unit, 201 Air conditioning device, 203 House, 211 Outdoor unit, 213 Indoor unit, 221 Compressor, 222 Four-way valve, 223 Outdoor heat exchanger, 224 Expansion valve, 225 Indoor heat exchanger, 231 outdoor blower, 233 indoor blower, 250 control device, 251 outdoor unit control unit, 251a control unit, 251b memory unit, 251c timing unit, 251d communication unit, 252 indoor unit control unit, 252a communication unit, 255 remote controller, 255a display unit, 256 human presence sensor, 261 refrigerant piping, 263 communication line, 271 indoor space, 272 outdoor space, 280 air conditioning unit, 290 information equipment, 291 display unit.

Claims

1. A Doppler sensor that detects the user's pulse wave; an analysis unit that analyzes the pulse wave detected by the Doppler sensor, The analysis unit is a one-point concentration degree calculation unit that generates an index value by quantifying the pulse wave based on chaos analysis using a pulse wave shape displacement, which is a time series displacement of the waveform shape of the pulse wave, as an analysis source, and calculates a one-point concentration degree based on the index value; a short-interval concentration calculation unit that calculates a short-interval concentration by averaging time-series data of the one-point concentration degree in a short interval; a long interval concentration degree calculation unit that calculates a long interval concentration degree by averaging time series data of the one-point concentration degree or the short interval concentration degree in a long interval that is longer than the short interval; a concentration level change estimation unit that estimates a change in the concentration level of the user based on the short-term concentration level and the long-term concentration level, The concentration degree change estimation unit is The long-term concentration used in the estimation is A concentration estimation device that adopts a long-term concentration level calculated using time series data of the one-point concentration level going back in time, which does not include the time series data of the one-point concentration level used to calculate the short-term concentration level.

2. The concentration degree change estimation unit A concentration estimation device as described in claim 1, wherein the time series data of the one-point concentration degree going back in time, which does not include the time series data of the one-point concentration degree used to calculate the short-term concentration degree, is the time series data of the one-point concentration degree from the same day and immediately before the calculation of the short-term concentration degree.

3. The analysis unit, The concentration level estimation device according to claim 1 or 2, which estimates the degree of brain activity of the user during a task or behavior based on the index value, and estimates that the higher the index value, the higher the degree of brain activity during the task or behavior.

4. The chaos analysis is calculating a vector specified from the time series data of the pulse waveform change and a preset delay time; generating an attractor by arranging the vectors in a time-series order in a multi-dimensional state space of three or more dimensions; a step of giving a hypersphere as an initial state to the orbit of the attractor, repeating the process of stretching the hypersphere for each preset sliding time based on the orbit, and calculating the expansion ratio at which the hypersphere becomes an ellipsoid as the Lyapunov exponent, which is the index value.

5. The concentration degree change estimation unit is The short-term concentration used in the estimation is 3. The concentration degree estimation device according to claim 1, wherein a current short-term concentration degree is adopted at the time of estimation.

6. The shorter the concentration level is, the smaller the value of each of the short-term concentration level and the long-term concentration level is. The concentration level estimation device according to claim 1 or 2, wherein the concentration level change estimation unit estimates that the concentration level has decreased when a ratio of the short-term concentration level to the long-term concentration level is equal to or less than a preset threshold value.

7. The concentration level estimation device according to claim 6 , wherein the threshold value is any value between 70% and 90%.

8. The concentration level estimation device according to claim 1 or 2, wherein the long interval is at least three times as long as the short interval.

9. The concentration level estimation device according to claim 7 , wherein the long interval is any value between 3 minutes and 30 minutes.

10. The short interval concentration calculation unit a first calculation unit that calculates the short-term concentration degree calculated by averaging time-series data of the one-point concentration degree as a first-order short-term concentration degree; a secondary calculation unit that averages the time series data of the first short-term concentration degree to calculate a second short-term concentration degree, 3. The concentration level estimation device according to claim 1, wherein the short interval concentration level used in the long interval concentration level calculation section and the concentration level change estimation section is the first short interval concentration level or the second short interval concentration level.

11. The concentration level estimation device according to claim 1 , wherein the averaging process is a moving average process.

12. The concentration estimation device according to claim 1 or 2, An apparatus comprising: a control device that acquires concentration information indicating any one of the one-point concentration level, the short-term concentration level, the long-term concentration level, or a change in the concentration level from the concentration level estimation device, and controls the operation of the apparatus main body based on the acquired concentration level information.

13. The concentration estimation device according to claim 1 or 2, an air conditioning unit that conditions the indoor space; An air conditioning system comprising: a control device that acquires concentration information indicating any one of the one-point concentration degree, the short-term concentration degree, the long-term concentration degree, or a change in the concentration degree from the concentration degree estimation device, and controls the air conditioning unit based on the acquired concentration degree information.

14. The air conditioner according to claim 12 , wherein the control device acquires the concentration degree information indicating a change in the concentration degree for each short interval, and controls the air conditioner based on the acquired change in the concentration degree.

15. The air conditioning device according to claim 13 , wherein the control device, when the acquired concentration level information indicates that the user's concentration level has decreased, controls the air conditioning unit to perform a wake-up action to wake up the user.

16. The air conditioning device according to claim 14, wherein the control device continues the awakening operation if the user's level of concentration remains low after starting the awakening operation, or increases the air volume to continue the awakening operation, and stops the awakening operation if the user's level of concentration increases.

17. The air conditioner according to claim 14 , wherein the control device has a non-operation time during which the control device does not perform the awakening action for a certain period of time after performing the awakening action.