TWI747108B - Reminder method and reminder device using the same - Google Patents

Abstract

A reminder method includes the following steps. A reminder device determines a first current user state, selects a first rule state according to the first current user state, and determines a notification behavior of a reminding issue according to the maximum utility value corresponding to the first rule state. Then, the reminder device determines a second current user state. Then, the reminder device selects a second rule state that according to the second current user state. Then, the reminder device obtains an adjustment weight according to a relationship between a return signal transmitted by the electronic device and a notification action. Finally, the reminder device obtains an updated utility value according to the adjustment weight and the maximum utility value corresponding to the second rule state, and replaces the maximum utility value corresponding to the first rule state with the updated utility value.

Description

Reminder method and reminder device using the same

The present invention relates to a reminding method and a reminding device using the same, and more particularly to a reminding method that determines whether to remind based on the current user status and a reminding device using the same.

At present, most reminders only consider the time factor. For example, when the current time reaches the set time of the reminder event set by the user, the reminder device sends a notification message to remind the user that the set time is up. However, the habits of users are constantly changing. Therefore, it is usually difficult to meet the habits of users to decide whether to send notification messages only based on time.

The embodiment of the present invention provides a reminding method and a reminding device using the same, which can improve the aforementioned conventional problems.

An embodiment of the present invention provides a reminding method. The reminding method is used to determine whether to send a notification to a user, and the user is equipped with a sensor. The reminding method includes the following steps. A reminder device judges the first current user status of one of the users according to a perception signal from the sensor; the reminder device selects a first rule state that matches the first current user status from a number of rule states ; The reminder device determines a first notification behavior for a reminder according to one of the notification utility value corresponding to the first rule state and the one of the non-notification utility value corresponding to the first one; The sensing signal from the sensor determines the second current user state of one of the users; the reminding device selects a second rule state that matches the second current user state from these rule states; the reminding device is based on an electronic The relative relationship between a report signal from the device and one of the first notification behaviors is obtained, and an adjustment weight is obtained, where the report signal is the actual behavior of the user to the reminder; the reminder device corresponds to the adjustment weight and the second rule state The second largest one of the notification utility value and the non-notification utility value obtains an updated utility value; and the updated utility value replaces the first largest one corresponding to the first rule state.

Another embodiment of the present invention provides a reminder device. The reminding device is used for judging whether to send a notification to a user, and the user is equipped with a sensor. The reminding device includes a judgment unit, a selection unit, a notification behavior determination unit, and a calculation unit. The judging unit is used for judging the first current user status of one of the users. The selection unit is used for selecting a first rule state that matches the first current user state from the plurality of rule states. The notification behavior determination unit is used for determining a first notification behavior for a reminder according to the first largest one of the notification utility value and the non-notification utility value corresponding to the first rule state. Among them, the judging unit is further used to: after a period of time, judge the second current user state of one of the users according to the sensing signal from the sensor; Two current user status matches one of the second rule status. The calculation unit is used for: obtaining an adjustment weight according to the relative relationship between a report signal from an electronic device and one of the first notification behaviors, wherein the report signal is an actual behavior of the user to the reminder; according to the second rule The second largest one of the notification utility value and the non-notification utility value corresponding to the state and the adjustment weight are used to obtain an updated utility value; and the updated utility value replaces the first largest one corresponding to the first rule state.

In order to have a better understanding of the above and other aspects of the present invention, the following specific examples are given in conjunction with the accompanying drawings to describe in detail as follows:

In order to have a better understanding of the above-mentioned and other aspects of the present invention, the following specific embodiments are described in detail in conjunction with the accompanying drawings.

Please refer to FIG. 1, which shows a schematic diagram of a reminder system 10 according to an embodiment of the present invention. The reminding system 10 includes a reminding device 100, at least one sensor 200 and an electronic device 300. The reminding device 100 is, for example, a portable electronic device, such as a mobile phone and a notebook computer, but it can also be other types of electronic devices, such as a desktop computer. The electronic device 300 is, for example, a medicine tank, sports equipment (such as treadmills, etc.), medical examination equipment (such as blood pressure monitors, weight scales, blood glucose detectors, etc.), or any other electronic devices that users need to use within their range of activities .

As shown in FIG. 1, at least one sensor 200 is disposed on the user to sense the location of the user and/or the physiological state of the user. The sensor 200 is, for example, a Global Positioning System (GPS) module, which can sense the location of the user. Alternatively, the sensor 200 is, for example, a physiological signal sensing module, which can sense the physiological state of the user, such as heartbeat and/or blood pressure. Alternatively, the sensor 200 is, for example, an acceleration sensor (G-sensor), which can sense the user's actions, such as the user is in a static state (such as lying down or standing) or dynamic (such as walking). In addition, the sensor 200 can be integrated into a wearable device suitable for the user to wear, such as a bracelet, a watch, an earphone, an earring, a necklace, or a ring. In another embodiment, the sensor 200 and the reminder device 100 can also be integrated into a single wearable device.

The reminding device 100 is used to determine whether to send a notification to the user. For example, the reminding device 100 judges the first current user state of the user according to the sensing signal S1 from the sensor 200. Then, the reminding device 100 selects the first rule state that matches the first current user state from the plurality of rule states. Then, the reminding device 100 determines the first notification behavior for a reminder according to the first largest _{of the notification utility value Q Y} and the non-notification utility value Q _{N corresponding to the first rule state.} Then, after a period of time, the reminding device 100 determines the second current user state of the user according to the sensing signal S1 from the sensor 200. Then, the reminding device 100 selects a second rule state that matches the second current user state from these rule states. Then, the reminder device 100 obtains an adjustment weight according to the relative relationship between the report signal S2 from the electronic device 300 and the first notification behavior, where the report signal S2 is the actual behavior of the user on the reminder. Then, the reminding device 100 obtains an updated utility value Q _n+1 (s, a) according to _{the second largest of the notification utility value Q Y} and the non-notification utility value Q _N corresponding to the adjustment weight and the second rule state . Then, the reminding device 100 replaces the first largest one corresponding to the first rule state _{with the updated utility value Q n+1 (s, a).}

The reminder device 100 includes a controller 110, a judgment unit 120, a selection unit 130, a notification behavior determination unit 140, a calculation unit 150, and a reminder communication unit 160. The controller 110 is electrically connected to the judgment unit 120, the selection unit 130, the notification behavior determination unit 140, the calculation unit 150, and the reminder communication unit 160 to uniformly control these units. In an embodiment, at least two of the controller 110, the judgment unit 120, the selection unit 130, the notification behavior determination unit 140, the calculation unit 150, and the reminder communication unit 160 may be integrated into one unit. The controller 110, the judgment unit 120, the selection unit 130, the notification behavior determination unit 140, the calculation unit 150, and the reminder communication unit 160 may adopt a physical circuit structure formed by a semiconductor manufacturing process. The reminding device communication unit 160 is, for example, a wireless communication module, such as a Wi-Fi communication module, a Bluetooth communication module, or a communication module supporting other communication protocols. In addition, at least one of the judgment unit 120, the selection unit 130, the notification behavior determination unit 140, and the calculation unit 150 may also be firmware or software, which is built in the controller 110 or loaded and executed by the controller 110.

The following is an example of the specific implementation of the reminding device 1001 of the embodiment of the present invention with FIG. 2. Fig. 2 shows a flowchart of the reminding method of the reminding device 100 of Fig. 1.

In step S110, the determining unit 120 determines the first current user state of the user according to the sensing signal S1 transmitted from the sensor 200. As shown in Table 1, the current user status includes "user's location", "user's physiological state (such as walking, still or sleeping)", "current time", "whether the current time belongs to a working day", and "whether the user is During the meeting", "the user’s actual behavior in the first time interval T1 (for example, whether the electronic device 300 has been used)" and/or "the first notification behavior in the second time interval T2 (for example, reminding the device 100) Has a notice been issued)". However, the embodiment of the present invention does not limit the type of the current user state, depending on the actual application, the current user state may include other types.

Table I i Rule status R Notification behavior location Physiological state Working day Current time in meeting Used or not in time interval T1 Has been notified in the time interval T2 Utility value Area A, Area B, or Area C Walking, resting or sleeping Yes or no A day is divided into 24 periods Yes or no Yes or no Yes or no Notification utility value Q _Y No notification utility value Q _N 1 A walk Yes 6:00 no no no 0 1 2 B still no 7:00 no no no 1 0 3 C Sleep no 8:00 no Yes no 1 0 4 A still Yes 9:00 no no no 1 0 5 B still Yes 10:00 no no Yes 0 1 6 C still Yes 11:00 no no no 1 0 7 A walk Yes 12:00 no no no 1 0 8 B Sleep Yes 13:00 Yes no Yes 1 0 9 C Sleep Yes 14:00 no no no 0 1 10 A walk Yes 15:00 no no Yes 0 1 11 B walk Yes 16:00 Yes no no 1 0 12 C walk Yes 17:00 Yes no no 1 0 13 A walk Yes 18:00 no no no 1 0 14 B walk Yes 19:00 no no no 1 0

In one embodiment, the reminder device 100 may adopt a reinforcement learning method (such as Q-learning) to learn the habits of the user, and Table 1 may also be referred to as a Q-Table.

In an embodiment, the determining unit 120 can be used to: (1) learn the location of the user based on the signal from the GPS module; (2) based on the sensing signal from the physiological signal sensing module Know the physiological state of the user; (3). Know the current time from the clock of the portable device worn by the user or the clock of the reminder device 100 itself; (4). Determine whether the current date belongs to a working day; (5). Basis Reminding the calendar, current time and current date in the device 100, whether the user is in the meeting; (6). Determine whether the user has used the electronic device 300 in the first time interval T1 according to the report signal S2 of the electronic device 300; And, (7). According to the signal from the notification behavior determining unit 140, determine whether the reminder device 100 has issued a notification in the second time interval T2.

The electronic device 300 includes a main body 305, a sensor 310, and an electronic device communication unit 320. The sensor 310 and the electronic device communication unit 320 are disposed on the main body 305, and the sensor 310 can sense whether the main body 305 has been used. When the main body 305 is used, the sensor 310 transmits a signal to the electronic device communication unit 320, and the electronic device communication unit 320 then transmits a report signal S2 to the reminder communication unit 160 of the reminder device 100. For example, when the electronic device 300 is a medicine tank, the sensor 310 may be disposed on the lid of the medicine tank. When the sensor 310 senses that the lid of the medicine canister is opened, the electronic device communication unit 320 transmits a report signal S2 to the reminder communication unit 160 of the reminder device 100. For another example, when the electronic device 300 is a sports equipment, the sensor 310 may be electrically connected to or integrated with a processor of the sports equipment. When the sensor 310 senses that the sports equipment is used, the electronic device communication unit 320 transmits a report signal S2 to the reminder communication unit 160 of the reminder device 100. For another example, when the electronic device 300 is a medical examination instrument, the sensor 310 may be configured on the body of the medical examination instrument or electrically connected to the processor of the medical examination instrument. When the sensor 310 senses that the medical examination instrument is used, the electronic device communication unit 320 transmits a report signal S2 to the reminder communication unit 160 of the reminder device 100.

The sensor 310 is, for example, a position sensor, an acceleration sensor, or other sensors capable of sensing that the electronic device 300 is used. The electronic device communication unit 320 is, for example, a wireless communication module, such as a Wi-Fi communication module, a Bluetooth communication module, or a communication module supporting other communication protocols.

In this embodiment, the "user's location" of the first current user state is "located in area A" as an example, "the user's physiological state" is "walking" as an example, and the "current time" is "6: 00" as an example, "Is the current time on a working day" as an example, "Is the user in a meeting" as an example, "No" as an example, and "Whether the user has used the electronic device in the first time interval T1 300" takes "No" as an example, and "reminds whether the device 100 has issued a notification in the second time interval T2" takes "No" as an example.

In step S115, refer to a table at the same time, the selection unit 130 state from a number of rules R _i, the user selects the first state match the current state of the first rule.

The state-based rule R _i where all possible permutations of the various types of user status. For example, as shown in Table 1, there are 3 situations for "user's location", 3 situations for "user's physiological state", 2 situations for "whether the current date belongs to a working day", and "current time" There are 24 situations (for example, 1 day is divided into 24 time slots), there are 2 situations for "whether the user is in a meeting", and 2 situations for "whether the user has used the electronic device 300 in the first time interval T1", and There are two situations of "reminding device 100 whether a notification has been sent in the second time interval T2". According to this, a total of 3456 (calculation formula: 3×3×2×24×2×2×2=3456) permutations and combinations can be arranged in the _{rule state R i.} Table 1 lists only 14 of them. However, depending on the user status type and the number of possible situations, the rule status R _i can be arranged in other combinations.

In this embodiment, as shown in Table 1, the first current user state matches the rule state R ₁ , so the selection unit 130 selects the rule state R ₁ as the first rule state.

In step S120, the notification behavior determining unit 140 determines the notification based on the first maximum of the notification utility value Q _Y and the non-notification utility value Q _{N corresponding to the first rule state (taking the rule state R 1 as an example)} The first notification act of the matter. For example, as shown in Table 1, the notification behavior determining unit 140 determines the size of the notification utility value Q _Y and the non-notification utility value Q _{N in the first rule state (taking the rule state R 1} as an example), and selects the two The largest one, such as the non-notification utility value Q _N (as shown in Table 1, the value of the non-notification utility value Q _N is 1, which is greater than _{the value 0 of the notification utility value Q Y} ), and the first notification behavior is determined accordingly . For example, since the first largest one is the non-notification utility value Q _N , the notification behavior determination unit 140 does not issue any notification signal.

In another embodiment, if the first largest one is the notification utility value Q _Y , the notification behavior determining unit 140 sends a notification signal N1. Although not shown, the reminding device 100 further includes a notifier, which sends a notification according to the notification signal N1. When the notifier is, for example, a speaker, the notification is a voice or a sound. When the notifier is, for example, a display screen, the notification is a text or symbol on the display screen. When the notification device is, for example, a vibrator, the notification is a vibration. When the notification device is, for example, a light-emitting element, the notification is light.

After the user receives the notification signal N1 (based on the first notification behavior), if the timing of the notification signal N1 matches the user's habit of using the electronic device 300, the user will use the electronic device based on the notification signal N1 for a period of time 300, such as taking medicine, using sports equipment to exercise, using medical examination equipment to measure physical conditions, etc. When the timing of the notification signal N1 is not in line with the user's habit of using the electronic device 300, the user usually ignores the notification signal N1, that is, the electronic device 300 is not or cannot be used. In another case, it is also possible to remind the device 100 not to send the notification signal N1 (based on the first notification behavior), but within a period of time, the user uses the electronic device 300, such as taking medicine, using sports equipment to exercise, and using medical examinations. Equipment to measure physical condition, etc.

Therefore, after a period of time, the reminding device 100 updates the value of the non-notification utility value Q _N or the value of the notification utility value Q _Y according to the current user status of the user in a certain period of time, so that the reminder device 100 reminds The model is more in line with the user's original or changed habits. Further examples are given below.

In step S125, after a period of time, the determining unit 120 determines the second current user state of the user according to the sensing signal S1 transmitted from the sensor 200. This period of time is, for example, several seconds, several minutes (such as 10 minutes, 15 minutes), or several hours.

For the second current user state of this embodiment, the "user's location" is "located in area C" as an example, the "user's physiological state" is "sleep" as an example, and the "current time" is "8:00". For example, "Is the current time on a working day" take "No" as an example, "Is the user in a meeting" take "No" as an example, "Whether the user has used the electronic device 300 in the first time interval T1" Take "Yes" as an example, and "Remind whether the device 100 has issued a notification in the second time interval T2" takes "No" as an example.

In step S130, the selection unit 130 state from a number of rules R _i, the user selects the second current state matches the state of the second rule.

In the present embodiment, as shown in Table I, the current user state and second state rule R ₃ because it matches the state selection unit 130 selects the rule R ₃ as a second rule state.

In step S135, the calculation unit 150 obtains the adjustment weight r according to the relative relationship between the first notification behavior and the report signal S2 transmitted from the electronic device 300.

For example, as shown in Table 2 below, when the relationship between the first notification behavior and the report signal S2 is "notify the reminder and the actual behavior is an action on the reminder", the calculation unit 150 sets the adjustment weight r as the first Weights. When the first notification behavior is "not notify the reminder" and there is no report signal S2 (indicating that the user's actual behavior is "do not act on the reminder (such as not using the electronic device 300)"), the calculation unit 150 sets the adjustment weight r Is the second weight value, where the first weight value is greater than the second weight value. In an embodiment, the first weight value is, for example, 0.5, and the second weight value is, for example, 0. As shown in Table 2, when the relationship between the first notification behavior and the report signal S2 is "not notify the reminder and the actual behavior is an action on the reminder", the calculation unit 150 sets the adjustment weight r to the third weight value; When the first notification behavior is "notify the reminder" and there is no report signal S2 (indicating that the user's actual behavior is "do not act on the reminder (such as not using the electronic device 300)"), the calculation unit 150 sets the adjustment weight r as The fourth weight value. In an embodiment, the third weight value is equal to the fourth weight value, such as -0.5, and is smaller than the first weight value and the second weight value. In another embodiment, the third weight value and the fourth weight value may be different. In addition, as long as the first weight value, the second weight value, the third weight value, and the fourth weight value conform to the foregoing relationship, the embodiment of the present invention does not limit the value range of the weight value.

Table II Remind the actual behavior of the device 100 Notice No notice Is there an electronic device 300 Used The first weight value (e.g. 0.5) The third weight value (e.g. -0.5) did not use The fourth weight value (e.g. -0.5) The second weight value (such as 0)

By adjusting the weight r, the subsequent notification behavior determined by the reminder device 100 can be more in line with the user's habits. For example, the relationship between the first notification behavior and the report signal S2 is "notify the reminder and the actual behavior is an action on the reminder", which means that in the current user state, the reminder device 100 determines the first The notification behavior does conform to the user's usage habit of the electronic device 300, so the calculation unit 150 increases the proportion of the adjustment weight r, for example, the first weight value is greater than the second weight value, the third weight value, and the fourth weight value. For the first notification behavior is "not notify reminders" but the actual user behavior is "action on reminders" and the first notification behavior is "notify reminders" but the actual user behavior is "remind The two cases of "item inactive" indicate that in the current user state, the first notification line determined by the reminder device 100 does not conform to the user's habit of using the electronic device 300, so the calculation unit 150 reduces the adjustment weight r The proportion, for example, the third weight value and the fourth weight value are both smaller than the first weight value and the second weight value.

In summary, during the actual operation of the reminder device 100, by continuously setting the adjustment weight r corresponding to the current user state, the accuracy of the notification behavior determined by the reminder device 100 can be increased, so that the reminder device 100 is in the current user state. The notification behavior determined below conforms to the user's usage habit of the electronic device 300 as much as possible.

After using the reminder device 100 for a long time, the notification behavior determined by the reminder device 100 usually meets more than 80% of the user's usage habits. That is, when the notification behavior determined by the reminder device 100 is to send a notification, the user has an 80% probability (Probability or habit) The electronic device 300 will be used. Similarly, when the notification behavior determined by the reminder device 100 is not to send a notification, there is an 80% probability (probability or habit) that the user will not use the electronic device 300.

Since the reminder device 100 will continue to increase the consistency between the determined notification behavior and the user's habits, even in the early stage of use of the reminder device 100, the notification utility value Q _{Y of each rule state R i} in Table 1 and no notification are sent The utility value Q _N is any preset value, such as any preset value between 0 and 1 (including endpoint values). In an embodiment, one of the notification utility value Q _Y and the non-notification utility value Q _{N of the rule state R i} can be preset to 0, and the other is preset to 1. With the reminder device 100 of the embodiment of the present invention constantly updating the utility value, after the reminder device 100 is used for a long time, the notification behavior determined by the reminder device 100 conforms to more than 80% of the user's usage habits.

In step S140, as shown in the following formula (1), the calculation unit 150 adjusts the weight W and the second rule state (in this embodiment, the rule state R _{3 is taken} as an example) corresponding to the notification utility value Q _Y and non-notification The second largest utility value Q _N obtains the updated utility value Q _n+1 (s, a).

Formula 1):

表示第二最大者(本例係以規則狀態R₃ 的發通知效用值Q_Y 為例)，而Q_n+1 (s, a)表示更新效用值。In formula (1), r represents the adjustment weight, α represents the learning rate, γ represents the discount coefficient, s represents the first rule state (in this example, the rule state R _{1 is taken} as an example), and a represents the first notification behavior ( This example uses no notification), Q _n (s, a) represents the first largest one (this example uses the non-notification utility value Q _{N of the rule state R 1} as an example), and s'represents the second rule state ( This example takes the rule state R ₃ as an example), a'represents the second largest one (this example takes the notification utility value Q _{Y of the rule state R 3} as an example) corresponding to the second notification behavior (this example uses Send out the notification signal N1 as an example),

Represents the second largest one (in this example, the notification utility value Q _{Y of the rule state R 3} is taken as an example), and Q _n+1 (s, a) represents the update utility value.

In an embodiment, the learning rate α is a preset value, and its value is, for example, between 0 and 1, and includes an endpoint value. The discount coefficient γ is a preset value, and its value is, for example, between 0 and 1, and includes the endpoint value.

、調整權重r及Q_n (s, a)的數值可透過前述步驟得知，此些數值在代入式(1)後，即可取得(計算出) 更新效用值Q_n+1 (s, a)。Due to the learning rate α, the discount coefficient γ, the second largest

, The value of adjusting the weight r and Q _n (s, a) can be known through the aforementioned steps. After these values are substituted into equation (1), the updated utility value Q _n+1 (s, a) can be obtained (calculated) ).

In step S145, the calculation unit 150 replaces the first rule state (in this example, the rule state R ₁ ) corresponding to the first _{rule state (in this example, the rule state R 1) with the updated utility value Q n+1 (s, a)} Take the non-notification utility value Q _{N of the} rule state R ₁ as an example). For example, as shown in Table 1, if the updated utility value Q _n+1 (s, a) is 0.8, then 0.8 is used to replace the non-notification utility of the _{first rule state (in this example, the rule state R 1 is taken as an example)} The value Q _N. After the replacement, Table 1 is updated as Table 3 below. Except for the non-notification utility value Q _N of the rule state R ₁ , the other rule states R _i remain unchanged, so they are not shown in Table 3.

Table Three i Rule status R Notification behavior location Physiological state Working day Current time in meeting Used or not in the time interval T1 Has been notified in the time interval T2 Utility value Area A, Area B, or Area C Walking, resting or sleeping Yes or no A day is divided into 24 periods Yes or no Yes or no Yes or no Notification utility value Q _Y No notification utility value Q _N 1 A walk Yes 6:00 no no no 0 0.8

Next, after a period of time, the reminding device 100 can obtain the updated utility value _{for the rule state R 3 in accordance with similar or the same steps S110 to S145 described above.} In this principle, after a period of time, the state of such rule R _i, one who issued a notice utility value Q _Y values, without notice or utility value Q _N value is replaced with the updated utility values.

In one embodiment, when the alert device 100 that a user is accustomed to sudden changes, the rule table can be reset state R _i a content of the user to re-learn new habits change. For example, when the reminding device 100 determines that the user's location has changed based on the current user status (for example, it is changed to a new location due to moving, and the new location is hundreds of kilometers away from the original location), it can send a notification signal, Ask the user whether to re-table the _{contents of the rule status R i} , such as resetting the notification utility value Q _Y of the rule status R _i and the non-notifying utility value Q _N , such as the notification utility value Q _{Y of the rule state R i One of the utility value Q N} and the non-notification utility value Q N is reset to 0, and the other is reset to 1. When the user agrees to reset, the device 100 is reminded to re-state the contents _{of the rule status R i.} When the user disagrees with the reset, the reminding device 100 can suspend the learning function, and when the function suspension time is over or the user returns to the original area, the reminding device 100 will resume the learning function. In another embodiment, when the alert device 100 that a user is accustomed to sudden changes, regardless of whether the user agrees, the alert device 100 always reset state table rules R _i a content or suspend the learning function.

In summary, although the present invention has been disclosed in the above embodiments, it is not intended to limit the present invention. Those with ordinary knowledge in the technical field of the present invention can make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the scope of protection of the present invention shall be subject to those defined by the attached patent scope.

:第二最大者10: Reminder system 100: Reminder device 110: Controller 120: Judgment unit 130: Selection unit 140: Notification behavior determination unit 150: Calculation unit 160: Reminder device communication unit 200: Perceptron 300: Electronic device 305: Body 310: Perception Device 320: electronic device communication unit a: first notification behavior a': second notification behavior r: adjustment weight N1: notification signal Q _n : (s, a) the first largest one Q _n+1 : (s, a) update utility value Q _Y: send notification utility value Q _N: do not send notification utility value _{R i, R 1, R 3} : status rules s: a first rule state s': a second rule state S1: sensing signals S2: return signal S110~S145: Step T1: First time interval T2: Second time interval α: Learning rate γ: Discount coefficient

: The second largest

Figure 1 is a schematic diagram of a reminder system according to an embodiment of the invention. Fig. 2 shows a flowchart of the reminding method of the reminding device of Fig. 1.

10: Reminder system

100: reminder

110: Controller

120: Judgment unit

130: select unit

140: Notification Behavior Decision Unit

150: calculation unit

160: reminder communication unit

200: Perceptron

300: Electronic device

305: body

310: Perceptron

320: Electronic device communication unit

S1: Sensing signal

S2: report signal

N1: Notification signal

Claims (12)

A reminding method for judging whether to send a notification to a user, the user is equipped with a sensor, and the reminding method includes: A reminding device determines the first current user status of one of the users according to at least one sensing signal from the sensor; The reminding device selects a first rule state that matches the first current user state from a plurality of rule states; The reminding device determines a first notification behavior for a reminder according to the first largest one of a notification utility value and a non-notification utility value corresponding to the first rule state; After a period of time, the reminding device determines a second current user status of one of the users according to the at least one sensing signal from the sensor; The reminding device selects a second rule state that matches the second current user state from the rule states; The reminding device obtains an adjustment weight according to a relative relationship between a report signal sent from an electronic device and one of the first notification behaviors, wherein the report signal is an actual behavior of the user for the reminder; The reminding device obtains an updated utility value according to the second largest one of the notification utility value and the non-notification utility value corresponding to the adjustment weight and the second rule state; and Replace the first largest one corresponding to the first rule state with the updated utility value. For example, the reminding method described in item 1 of the scope of patent application, wherein the first current user status includes the location of the user, a physiological state of the user, a current time, whether the current time belongs to a working day, and the usage Whether the person is in the meeting, the actual behavior of the user in a first time interval, and/or the first notification behavior in a second time interval. For example, in the reminding method described in item 1 of the scope of patent application, the step of obtaining the updated utility value is completed according to formula (1). In formula (1), r represents the adjustment weight, α represents a learning rate, and γ Represents a discount coefficient, s represents the first rule state, a represents the first notification behavior, Q _n (s, a) represents the first largest one, s'represents the second rule state, and a'represents the first rule state. One of the second notification behaviors corresponding to the two largest ones,

Represents the second largest one, and Q _n+1 (s, a) represents the updated utility value;

. For example, in the reminding method described in item 1 of the scope of patent application, the steps in obtaining the updated utility value include: Calculating the adjustment weight, the sum of the first maximum value and the second maximum value; and Use this sum value as the updated utility value. For example, in the reminding method described in item 1 of the scope of patent application, the steps in obtaining the adjustment weight include: When the first notification behavior is a notification to the reminder and the actual behavior is an action on the reminder, the adjustment weight is set to a first weight value; and When the first notification behavior is not to notify the reminder and the actual behavior is not to act on the reminder, set the adjustment weight to a second weight value; Wherein, the first weight value is greater than the second weight value. For example, the reminding method described in item 5 of the scope of patent application, wherein the steps of obtaining the adjustment weight include: When the first notification behavior is not to notify the reminder and the actual behavior is an action on the reminder, set the adjustment weight to a third weight value; and When the first notification behavior is a notification to the reminder and the actual behavior is not to act on the reminder, the adjustment weight is set to a fourth weight value; Wherein, the third weight value is equal to the fourth weight value and less than the first weight value and the second weight value. A reminding device for judging whether to send a notification to a user, the user is equipped with a sensor, and the reminding device includes: A judging unit for judging the first current user status of one of the users; A selection unit for selecting a first rule state that matches the first current user state from a plurality of rule states; A notification behavior determination unit for determining a first notification behavior for a reminder according to the first largest one of the notification utility value and the non-notification utility value corresponding to the first rule state; unit; Wherein, the judging unit is further used to: after a period of time, judge a second current user state of the user according to the at least one sensing signal from the sensor; the selecting unit is further used to: Among the rule states, a second rule state that matches the second current user state is selected; the calculation unit is used for: according to a relative relationship between a report signal from an electronic device and the first notification behavior, Obtain an adjustment weight, where the report signal is an actual behavior of the user to the reminder; the second largest one of the notification utility value and the non-notification utility value corresponding to the second rule state, and The adjustment weight obtains an updated utility value; and the updated utility value replaces the first largest one corresponding to the first rule state. Such as the reminder device described in item 7 of the scope of patent application, wherein the first current user state includes the user's position, a physical movement of the user, a current period, whether the current period belongs to a working day, and the use Whether the person is in the meeting, the actual behavior of the user in a first time interval, and/or the first notification behavior in a second time interval. For example, the reminder device described in item 7 of the scope of patent application, wherein the calculation unit is further used to obtain the updated utility value according to formula (1); wherein, in formula (1), r represents the adjustment weight, and α represents a learning Rate, γ represents a discount coefficient, s represents the first rule state, a represents the first notification behavior, Q _n (s, a) represents the first largest one, s'represents the second rule state, a' Represents one of the second notification behaviors corresponding to the second largest one,

Represents the second largest one, and Q _n+1 (s, a) represents the updated utility value;

Such as the reminder device described in item 7 of the scope of patent application, wherein the calculation unit is further used for: Calculating the adjustment weight, the sum of the first maximum value and the second maximum value; and Use the sum value as the update utility value. Such as the reminder device described in item 7 of the scope of patent application, wherein the calculation unit is further used for: When the first notification behavior is a notification to the reminder and the actual behavior is an action on the reminder, the adjustment weight is set to a first weight value; When the first notification behavior is not to notify the reminder and the actual behavior is not to act on the reminder, set the adjustment weight to a second weight value; Wherein, the first weight value is greater than the second weight value. Such as the reminder device described in item 11 of the scope of patent application, wherein the calculation unit is further used for: When the first notification behavior is not to notify the reminder and the actual behavior is an action on the reminder, set the adjustment weight to a third weight value; and When the first notification behavior is a notification to the reminder and the actual behavior is not to act on the reminder, the adjustment weight is set to a fourth weight value; Wherein, the third weight value is equal to the fourth weight value and less than the first weight value and the second weight value.

Images