An intelligent reminding triggering method and system based on multi-source evidence fusion and electronic equipment

Abstract

The application discloses a kind of intelligent reminding trigger method, system and electronic equipment based on multi-source evidence fusion.The method comprises: obtaining the reminding task created by user, the semantic analysis of reminding task is carried out, and the reminding action, reminding object and trigger scene are determined;According to trigger scene, the target reminding task in triggerable state is filtered;At least two kinds of positioning evidence, wireless signal evidence, motion state evidence, vehicle connection evidence, time context evidence and user operation state evidence related to target reminding task are collected;The evidence confidence of each trigger evidence is calculated, and the fusion trigger result is generated based on evidence fusion rule;According to fusion trigger result, determine trigger stage, output the reminding notice of corresponding stage when meeting reminding output condition;Record trigger state, and carry out repeat control or stage upgrade control.The scheme can improve the timeliness and accuracy of reminding trigger, reduce false triggering, missed triggering and repeated reminding.

Description

I. Technical Field

[0001] This invention relates to the fields of smart terminals, mobile computing, task reminders, location services, and sensor data fusion technologies, and in particular to a smart reminder triggering method, system, and electronic device based on multi-source evidence fusion. II. Background Technology

[0002] With the widespread use of smartphones, smartwatches, in-vehicle systems, and other mobile electronic devices, users are increasingly relying on electronic devices for schedule management, reminders, and daily life assistance. Existing reminder systems are typically triggered based on fixed times, locations, or simple conditions. For example, users can set reminders such as "remind me to attend a meeting at 9:00 AM," "remind me to clock in at the office," or "remind me to bring my keys when leaving home."

[0003] However, existing reminder systems still have the following shortcomings.

[0004] First, traditional alerts often rely on a single trigger condition. For example, time alerts depend only on the system time, and location alerts depend only on geofences or location distance. Single trigger conditions are easily affected by factors such as location drift, system backend restrictions, signal obstruction, building structure, and changes in vehicle environment, resulting in alerts being triggered too early, too late, or not at all.

[0005] Second, traditional location alerts typically treat "entering a location" or "leaving a location" as a one-time event, lacking a layered understanding of the user's actual behavioral process. For example, a user leaving home, entering an elevator, exiting the building, entering an underground parking lot, connecting to a vehicle, and driving away from the community are often simplified into a single "leaving the geofence" event in traditional alert systems. Because mobile operating systems have energy-saving and anti-shake mechanisms for background location and geofencing, the actual alert may only appear after the user has already traveled a considerable distance from the location, failing to meet the need for an "alert as soon as the user leaves the house" response.

[0006] Third, traditional alert systems lack the ability to fuse multi-source evidence. Whether a user has left home, is getting into a vehicle, or has left home or work cannot often be accurately determined by a single signal. For example, a disconnected Wi-Fi network may indicate that the user has left home, or it may simply be a network fluctuation; an increased location distance may indicate that the user has left, or it may be location drift; a vehicle Bluetooth connection may indicate that the user has gotten into the car, or it may simply be a temporary device connection. Without synthesizing evidence from different sources, with varying intensities, and in different time sequences, alert systems struggle to balance timeliness and accuracy.

[0007] Fourth, traditional reminder systems lack the ability to provide phased reminders and control repetition. In many real-world scenarios, users want to receive multiple reminders at different stages. For example, a reminder when leaving home, another when getting into a vehicle, and a backup reminder when the task is still incomplete after leaving the location. However, existing systems often either only provide one reminder or are prone to generating repeated reminders, lacking a balance mechanism between "allowing escalating reminders at different stages" and "suppressing repeated reminders at the same stage."

[0008] Fifth, traditional reminder systems have limited ability to understand natural language. Users typically prefer to input natural language expressions such as "Don't forget to bring your contract tomorrow morning," "Remind me to pick up my medicine when I get in the car," or "Remind me to submit my documents when I get to the office," rather than manually selecting complex time, location, trigger direction, vehicle conditions, and reminder stage. Existing systems struggle to accurately convert natural language expressions into structured reminder conditions.

[0009] Sixth, in shared reminders, dispatched reminders, or multi-person collaborative tasks, traditional reminder systems typically lack cross-device status synchronization mechanisms. For example, after one person completes a task, another person's device may continue to send reminders, causing interference.

[0010] Therefore, a new intelligent reminder triggering method is needed that can combine multi-source evidence such as user natural language input, task status, location changes, wireless signals, motion status, vehicle connection status, time context, and user operation status, and output reminder notifications in stages, with low power consumption and deduplication through evidence confidence calculation and fusion decision-making, so as to improve the timeliness, accuracy and usability of reminders. III. Summary of the Invention

[0011] (a) Purpose of the invention

[0012] The purpose of this invention is to provide an intelligent reminder triggering method, system, and electronic device based on multi-source evidence fusion, so as to solve the problems of existing reminder systems relying on a single triggering condition, untimely triggering, false triggering or missed triggering, lack of phased reminders, lack of repetitive control, and insufficient natural language understanding ability.

[0013] This invention determines the reminder action, reminder object, and triggering scenario by performing semantic parsing on the reminder task; it collects location evidence, wireless signal evidence, motion state evidence, vehicle connection evidence, time context evidence, and user operation state evidence through a multi-source evidence collection module; it generates a fused triggering result by calculating evidence confidence and using fusion rules; and it determines the triggering stage based on the fused triggering result and outputs the corresponding stage's reminder notification, thereby achieving more timely, accurate, low-power, and intelligent reminder triggering.

[0014] (II) Technical Solution

[0015] To achieve the above objectives, this invention provides a smart reminder triggering method based on multi-source evidence fusion, applied to electronic devices, the method comprising:

[0016] Obtain user-created reminder tasks, wherein the reminder task includes reminder content and triggering conditions associated with the reminder content;

[0017] The reminder task is semantically parsed to determine the reminder action, reminder object, and at least one triggering scenario corresponding to the reminder task. The triggering scenario includes at least one of time-triggered scenario, location arrival triggering scenario, location departure triggering scenario, and vehicle-related triggering scenario.

[0018] Based on the triggering scenario, determine the target reminder task that is currently in a triggerable state from the candidate reminder task set;

[0019] When the target reminder task is in a triggerable state, multi-source triggering evidence associated with the target reminder task is collected. The multi-source triggering evidence includes at least two of the following: location evidence, wireless signal evidence, motion state evidence, vehicle connection evidence, time context evidence, and user operation state evidence.

[0020] The confidence level of each of the multi-source triggering evidences is determined, and a fusion triggering result is generated based on the preset evidence fusion rules.

[0021] The triggering stage corresponding to the target reminder task is determined based on the fusion triggering result;

[0022] When the triggering stage meets the reminder output conditions of the target reminder task, a reminder notification corresponding to the triggering stage is output.

[0023] Record the triggering status of the target reminder task, and control the repetition or escalation of subsequent reminder notifications based on the triggering status.

[0024] In one optional implementation, semantic parsing of the reminder task includes: performing intent recognition on the natural language text or speech recognition text input by the user; extracting at least one of time information, location information, departure intent, arrival intent, vehicle usage intent, item information, and event information; and generating structured reminder conditions based on the extraction results.

[0025] In one alternative implementation, the multi-source triggering evidence includes at least two of the following: location evidence, wireless signal evidence, motion state evidence, vehicle connection evidence, temporal context evidence, and user operation state evidence.

[0026] In one alternative implementation, the confidence level of the evidence is determined based on at least one of the following: the source of the triggering evidence, the time of collection, the location accuracy, the signal strength, the distance to the associated location, and the degree of matching with historical behavioral patterns.

[0027] In one alternative implementation, when the triggering scenario is a location departure triggering scenario, the triggering phase includes at least two of the following: an early departure phase, a dynamic departure phase, a vehicle departure phase, and a fallback departure phase.

[0028] In one alternative implementation, the dynamic departure phase is determined based on a dynamic fence, which is generated according to at least one of the following: the center location of the associated location, the boundary, the user's current location, the user's orientation, the road direction, the building exit direction, or the historical departure direction.

[0029] In one alternative implementation, the vehicle departure phase is determined based on at least one of the following vehicle evidence: in-vehicle Bluetooth connection, in-vehicle audio routing, vehicle-mounted system connection, screen mirroring connection, or vehicle pairing device connection.

[0030] In one optional implementation, a trigger cycle identifier is generated for the target reminder task, and the triggering stages that have been triggered under the trigger cycle identifier are recorded; when the same triggering stage under the same trigger cycle identifier has already output a reminder notification, duplicate reminders are suppressed; when different triggering stages meet the reminder output conditions, the output stage is allowed to upgrade the reminder notification.

[0031] This invention also provides an intelligent reminder triggering system based on multi-source evidence fusion, comprising:

[0032] The task acquisition module is used to acquire reminder tasks created by the user.

[0033] The semantic parsing module is used to perform semantic parsing on reminder tasks to determine the reminder action, the reminder object, and at least one triggering scenario;

[0034] The candidate task filtering module is used to determine the target reminder task that is currently in a triggerable state from the candidate reminder task set based on the triggering scenario;

[0035] The multi-source evidence collection module is used to collect multi-source triggering evidence associated with the target reminder task;

[0036] The evidence fusion module is used to determine the corresponding evidence confidence level based on multi-source triggering evidence, and generate fusion triggering results based on preset evidence fusion rules;

[0037] The triggering phase decision module is used to determine the triggering phase corresponding to the target reminder task based on the fusion triggering results;

[0038] The notification output module is used to output a corresponding reminder notification when the reminder output conditions are met during the triggering phase.

[0039] The status control module is used to record the trigger status of the target reminder task and to control the repetition or escalation of subsequent reminder notifications based on the trigger status.

[0040] The present invention also provides an electronic device or a computer-readable storage medium. The electronic device includes a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the above-described method. The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the above-described method.

[0041] (III) Beneficial Effects

[0042] Compared with the prior art, the present invention has at least the following beneficial effects.

[0043] This invention uses multi-source trigger evidence fusion to reduce false triggers and missed triggers caused by location drift, signal fluctuations, and system backend limitations, instead of relying on a single time, location, or geofence event.

[0044] This invention can generate structured reminder conditions based on the natural language expression of reminder tasks, so that users do not need to manually configure complex trigger rules, thus improving the efficiency of reminder creation.

[0045] This invention enables the setting of an early departure phase, a dynamic departure phase, a vehicle departure phase, and a fallback departure phase in location departure scenarios, allowing the reminder system to cover the user's continuous behavioral process from indoors, doorway, parking lot, vehicle to distant locations, thereby improving the timeliness of reminders.

[0046] This invention combines multiple weak pieces of evidence into a valid trigger result through evidence confidence and evidence fusion rules, and also allows a single strong piece of evidence to trigger directly, thus balancing the accuracy and real-time nature of the alert.

[0047] This invention uses trigger cycle identifiers and trigger stage records to suppress repeated reminders in the same stage and upgrade reminders in different stages, avoiding meaningless repeated disturbances while retaining multi-stage reminder capabilities.

[0048] This invention can suppress subsequent reminders based on task completion status, list item completion status, user confirmation operation, or shared task synchronization status, thereby improving the accuracy of reminders in multi-person collaborative scenarios.

[0049] This invention can selectively enable dynamic fences, short-term high-precision positioning, or vehicle evidence collection based on whether the target reminder task is in a triggerable state and whether the electronic device is in the vicinity of the associated location, thereby reducing device power consumption. IV. Description of the attached drawings

[0050] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments are briefly described below. Obviously, the following drawings are only some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without creative effort.

[0051] Figure 1 This is a flowchart illustrating an intelligent reminder triggering method based on multi-source evidence fusion, provided as an embodiment of the present invention.

[0052] Figure 2 This is a schematic diagram of the structure of an intelligent reminder triggering system provided in an embodiment of the present invention.

[0053] Figure 3 This is a schematic diagram of a multi-stage reminder process in a location departure trigger scenario provided by an embodiment of the present invention.

[0054] Figure 4 This is a schematic diagram illustrating the generation and judgment of a dynamic fence, as provided in an embodiment of the present invention.

[0055] Figure 5 This is a schematic diagram of a vehicle departure phase triggering process provided in an embodiment of the present invention.

[0056] Figure 6 This is a schematic diagram of a trigger state recording, repeat control, and stage upgrade control process provided in an embodiment of the present invention. V. Detailed Implementation Methods

[0057] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings and embodiments. It should be understood that the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of the present invention.

[0058] In this specification, expressions such as "one embodiment," "some embodiments," "optionally," and "exemplary" indicate that the described features, structures, or steps may be included in at least one embodiment of the present invention. Features in the various embodiments may be combined with each other without conflict.

[0059] The electronic device in this invention can be a smartphone, tablet computer, smartwatch, vehicle terminal, wearable device, portable computing device, smart earphone, vehicle system, or other device with computing, communication, positioning, or sensor data acquisition capabilities.

[0060] The reminder tasks in this invention can be time reminders, location arrival reminders, location departure reminders, vehicle-related reminders, sharing reminders, delivery reminders, list reminders, or other reminders. Reminder content can include items, matters, actions, people, locations, times, list items, or user-defined text.

[0061] Example 1: Overall Method Flow

[0062] like Figure 1 As shown, this embodiment provides a smart reminder triggering method based on multi-source evidence fusion, which is applied to electronic devices and includes the following steps.

[0063] S101, retrieve reminder tasks created by the user.

[0064] Users can create reminders through text input, voice input, template selection, shortcuts, sharing from external applications, importing from calendars, or other methods. A reminder includes the reminder content and the triggering conditions associated with that content.

[0065] For example, if a user enters "Don't forget to bring your contract when you leave tomorrow morning," the system can retrieve that natural language text as a reminder task. The reminder task's content is "bring the contract," and the triggering conditions may include "tomorrow morning," "leave home," and "leave your current location or a frequently visited place," etc.

[0066] For example, if a user enters "Remind me to pick up my medicine when I get in the car", the reminder content of the task will be "pick up medicine", and the triggering conditions may include "vehicle-related triggering scenarios" or "location departure triggering scenarios".

[0067] For example, if a user enters "Remind me to submit materials when I arrive at the company", the reminder task will have the reminder content "submit materials", and the triggering condition may include "arrive at the company".

[0068] S102, perform semantic parsing on the reminder task.

[0069] Electronic devices perform semantic analysis on natural language text or speech recognition text input by users to determine the reminder action, reminder object, and at least one triggering scenario corresponding to the reminder task.

[0070] Semantic parsing can include intent recognition, slot extraction, time parsing, location parsing, action recognition, object recognition, and scene mapping.

[0071] Among them, intent recognition is used to determine whether the user's expression belongs to the intent of reminder creation, and whether it belongs to one or more of the following: time reminder, location reminder, departure reminder, arrival reminder, vehicle reminder, sharing reminder, or distribution reminder.

[0072] Slot extraction is used to extract time information, location information, departure intention, arrival intention, vehicle usage intention, item information, event information, person information, and effective time range, etc.

[0073] Scene mapping is used to map natural language expressions to structured triggering scenarios. For example, expressions such as "go out," "leave home," "when leaving," and "when going downstairs" can be mapped to location departure triggering scenarios; expressions such as "go to the company," "go to the farm," and "go to the customer" can be mapped to location arrival triggering scenarios; expressions such as "get in the car," "drive," "connect to the car's Bluetooth," and "go to the car" can be mapped to vehicle-related triggering scenarios; and expressions such as "tomorrow morning," "3 pm," and "before leaving get off work on Friday" can be mapped to time triggering scenarios or time constraints.

[0074] In some implementations, semantic parsing can be performed locally on the electronic device or assisted by a server or cloud-based model. Semantic parsing can employ rule engines, dictionary matching, machine learning models, natural language understanding models, large language models, or a combination of these methods.

[0075] S103, determine the target reminder task based on the triggering scenario.

[0076] Electronic devices can maintain a set of candidate reminder tasks. The set of candidate reminder tasks includes reminder tasks that may be triggered at the current time, current location, user state, or device state.

[0077] Electronic devices determine the target reminder task that is currently triggerable from the set of candidate reminder tasks based on information such as the triggering scenario, task validity period, task completion status, associated location, current distance, user's current behavior, whether it has been triggered, and whether it is in a paused state.

[0078] For example, when a user is near their "home" and there is an unfinished reminder such as "Don't forget to bring your contract when you go out," this reminder task can be added to the set of candidate reminder tasks for the location departure trigger scenario.

[0079] For example, when a user is not near any associated location, the system can avoid initiating high-frequency location checks related to the user's departure location, thereby saving power consumption.

[0080] S104, Collect multi-source triggering evidence.

[0081] When the target reminder task is in a triggerable state, the electronic device collects multi-source triggering evidence associated with the target reminder task.

[0082] Multi-source triggering evidence may include at least two of the following: location evidence, wireless signal evidence, motion state evidence, vehicle connection evidence, temporal context evidence, and user operation state evidence.

[0083] Location evidence can include satellite positioning, base station positioning, wireless LAN positioning, system geofence events, location coordinates, positioning accuracy, speed, direction, distance from associated locations, whether inside or outside the fence, etc.

[0084] Evidence of wireless signals may include wireless LAN connection status, wireless LAN disconnection status, Bluetooth connection status, Bluetooth signal strength, near-field wireless broadcast signals, short-range communication signals, and connection status of home or office networks.

[0085] Evidence of motion status can include accelerometer data, gyroscope data, step count data, motion coprocessor results, stationary status, walking status, running status, cycling status, driving status, or other system motion status recognition results.

[0086] Evidence of vehicle connectivity may include in-vehicle Bluetooth connectivity, in-vehicle audio router connectivity, in-vehicle infotainment system connectivity, screen mirroring connectivity, vehicle pairing device connectivity, vehicle-related peripheral connectivity, or motion status related to vehicle use.

[0087] Time context evidence can include the current time, reminder validity period, schedule time, user's frequently used travel times, historical trigger times, weekdays or rest days, morning and evening rush hours, etc.

[0088] Evidence of user operation status can include reminder task completion status, list item completion status, triggered stage, user confirmed operation, user ignored operation, user postponed operation, shared task completion status, or dispatched task receipt status.

[0089] S105, calculate the confidence level of the evidence and generate the fusion trigger result.

[0090] The electronic device calculates the confidence level of the triggering evidence based on at least one of the following: source of each type of triggering evidence, collection time, positioning accuracy, signal strength, distance from the associated location, and degree of matching with historical behavioral patterns.

[0091] For example, location evidence with high accuracy and continuously increasing distance from the associated location can have a high confidence level; when the wireless LAN is disconnected but the location is still within the associated location, it can have a medium or low confidence level; when the vehicle's Bluetooth is connected and the movement state is driving, it can have a high confidence level of vehicle departure; when the near-field wireless broadcast signal is lost and accompanied by changes in walking state, it can have a high confidence level of early departure.

[0092] Electronic devices can assign different weights to different categories of triggering evidence based on the triggering scenario. For example, in a location departure triggering scenario, wireless signal evidence, location evidence, and motion state evidence can have higher weights; in a vehicle-related triggering scenario, vehicle connection evidence and motion state evidence can have higher weights; and in a time-triggered scenario, time context evidence can have higher weights.

[0093] Evidence fusion can employ weighted calculation, threshold judgment, temporal combination judgment, rule matching, state machine judgment, Bayesian inference, machine learning models, or a combination of the above methods.

[0094] In one implementation, if a single strong piece of evidence satisfies a preset strong triggering condition, the fusion triggering result is determined to satisfy the reminder output condition. For example, if the in-vehicle device is connected and the target reminder task is "getting on the vehicle reminder," the vehicle stage reminder condition can be directly satisfied.

[0095] In another implementation, if multiple weak pieces of evidence simultaneously meet the triggering conditions within a preset time window, the fused triggering result is determined to meet the alert output condition. For example, weak evidence such as Wi-Fi disconnection, commencement of walking mode, loss of nearby wireless broadcast signal, and slight increase in positioning distance occurring simultaneously within a certain time period can collectively constitute an early departure triggering result.

[0096] S106, Determine the triggering stage based on the fusion triggering result.

[0097] The electronic device determines the triggering stage corresponding to the target reminder task based on the fusion triggering results.

[0098] The triggering phase can be a time-triggered phase, an arrival-triggered phase, an early departure phase, a dynamic departure phase, a vehicle departure phase, a last-minute departure phase, an alighting phase, or another phase.

[0099] For location departure triggering scenarios, the triggering phase may include at least two of the following: early departure phase, dynamic departure phase, vehicle departure phase, and fallback departure phase.

[0100] S107, output reminder notification.

[0101] When the triggering phase meets the reminder output conditions of the target reminder task, the electronic device outputs a reminder notification corresponding to the triggering phase.

[0102] The reminder notification can be a system notification, lock screen notification, banner notification, sound alert, vibration alert, in-app pop-up, desktop widget notification, smartwatch reminder, in-vehicle system reminder, voice broadcast, instant messaging message, or other forms of reminder.

[0103] Different trigger stages can correspond to different reminder text, reminder intensity, or reminder methods. For example, in the early departure stage, the message could be "You're leaving home, don't forget your contract"; in the vehicle departure stage, the message could be "You're on the vehicle now, please confirm that you have your contract"; and in the final departure stage, the message could be "You've left home a certain distance, have you taken your contract?"

[0104] S108 records the trigger status and performs repeated control or stage upgrade control.

[0105] Electronic devices generate trigger cycle identifiers for target reminder tasks. Trigger cycle identifiers can be generated based on task identifier, associated location, date, time window, departure cycle, arrival cycle, or user behavior cycle.

[0106] The electronic device records the triggering phase that the target reminder task has been triggered under the triggering cycle identifier.

[0107] When it is detected that a reminder notification has already been output under the same trigger period identifier and the same trigger phase, the repeated reminder notifications for that trigger phase are suppressed.

[0108] When different trigger stages under the same trigger cycle identifier meet the reminder output conditions, the corresponding stage upgrade reminder notification is allowed to be output.

[0109] When it is detected that the target reminder task has been completed by the user, marked as completed by the system, or all items in the list corresponding to the target reminder task have been completed, stop or suppress the subsequent triggering of the target reminder task.

[0110] For shared or dispatched reminder tasks, the trigger or completion status can be synchronized across multiple user devices. For example, after the first user completes a shared reminder task, the second user device receives the completion status and suppresses subsequent reminder triggers.

[0111] Example 2: Natural Language Reminder Task Analysis

[0112] This example illustrates the process of semantic parsing for reminder tasks.

[0113] The user inputs, "Don't forget to bring a tent, water, a power bank, and medicine when you leave tomorrow morning." The electronic device can then interpret this information as follows:

[0114] Time information: Tomorrow morning;

[0115] Intention to leave: to go out;

[0116] Reminder: Don't forget to bring it;

[0117] Items to be reminded: tents, water, power banks, medicines;

[0118] Triggering scenario: Leaving a location within a time limit;

[0119] Structured conditions: Trigger an alert when the user leaves the associated location within the valid time frame tomorrow morning.

[0120] The associated location can be determined based on the user's current location, the user's default location, the user's frequently used locations, or the user's selection. For example, when the user creates the reminder task at home, the system can use "home" as the default associated location; when the current location is not saved as a frequently used location, the system can prompt the user to create a new location or use a temporary location.

[0121] The user inputs "Remind me to pick up my medicine when I get on the bus." The electronic device can parse this information as follows:

[0122] Vehicle usage intention: To get in the vehicle;

[0123] Reminder action: Remind me to take it;

[0124] Target of reminder: Medicine;

[0125] Triggering scenarios: Vehicle-related triggering scenarios, or location-based departure triggering scenarios;

[0126] Structured conditions: An alert is triggered when in-vehicle connectivity or vehicle-related evidence is detected.

[0127] The user inputs "Remind me to submit the report when I get to the company." The electronic device can parse this information as follows:

[0128] Location information: Company;

[0129] Intention to arrive: to;

[0130] Notification action: Submit;

[0131] Target of reminder: Reports;

[0132] Triggering scenario: Arrival at the location triggers the scenario;

[0133] Structured conditions: An alert is triggered when an electronic device enters the company's associated scope and meets the confidence conditions.

[0134] Example 3: Multi-stage reminders in location departure scenarios

[0135] like Figure 3 As shown in the figure, this embodiment illustrates the multi-stage reminder process in a location departure trigger scenario.

[0136] The user creates a reminder task: "Don't forget to bring your contract when you go out." The system parses this reminder task as a location departure trigger scenario and associates it with the user's frequently used location, "home."

[0137] When the user is near their home, the reminder task is added to the candidate reminder task set. The system initiates low-frequency evidence collection based on power consumption strategies, such as monitoring system geofencing, Wi-Fi connection status, nearby wireless broadcast signals, and changes in motion.

[0138] When the system detects evidence such as a disconnection of the home Wi-Fi, loss of nearby radio signals, a user changing from stationary to walking, or a slight increase in the distance of electronic devices from the home, it calculates the early departure confidence level. If multiple weak pieces of evidence jointly meet the triggering conditions within a preset time window, the system determines that the early departure phase meets the alert output conditions and outputs an early departure alert.

[0139] If the user continues to move, the system can generate a dynamic fence based on the current location, home location, user orientation, building exit direction, or historical departure direction. The system can briefly activate higher-precision location detection when the user is near home and the target alert task is not yet completed. If the system detects that the user has crossed the dynamic fence, moved away from home, or the distance from home meets a preset departure threshold, the system determines that the dynamic departure phase meets the alert output conditions and outputs a dynamic departure alert.

[0140] If a vehicle Bluetooth connection, vehicle audio route switching, or vehicle infotainment system connection is subsequently detected, and the target reminder task is still not completed, the system determines that the vehicle departure phase meets the reminder output conditions and outputs a vehicle departure reminder.

[0141] If the system fails to issue a reminder in the aforementioned stages, or if the user ignores the aforementioned reminder, and the system's geofence departure event, continuous moving away trend, or normal location results indicate that the user has already left home a certain distance, then the system determines that the fallback departure stage meets the reminder output conditions and outputs a fallback departure reminder.

[0142] During the above process, repeated reminders at the same triggering stage are suppressed. For example, if an early departure reminder has already been issued, and a wireless LAN disconnection or loss of nearby wireless broadcast signal occurs again within a short period, the early departure reminder will not be issued again. However, if the vehicle departure stage or the last-minute departure stage subsequently begins, a stage escalation reminder can continue to be issued.

[0143] Example 4: Dynamic Fence Generation and Judgment

[0144] like Figure 4 As shown in the figure, this embodiment illustrates the dynamic fence generation and judgment process.

[0145] Traditional geofencing typically generates a circular fence with a fixed radius centered on the associated location. However, in real-world scenarios, users' departure paths are directional. For example, the direction of an exit from a residential building, the direction of an underground parking lot, the direction of a company entrance, or the direction of a user's historically frequently traveled routes can all influence the optimal trigger area.

[0146] In this embodiment, the electronic device can obtain at least one of the following: the center location of the associated location, the boundary of the associated location, the user's current location, the user's orientation, the road direction, the building exit direction, or the historical departure direction.

[0147] The electronic device generates a fenced area offset relative to the associated location based on the above information. This fenced area can be circular, elliptical, polygonal, fan-shaped, strip-shaped, or other shapes. The fenced area may not be centered solely on the center of the associated location, but may be offset based on the departure direction, road direction, exit direction, or historical departure direction.

[0148] In one implementation, the electronic device determines whether the user is moving away from the location based on the directional relationship between the user's orientation and the center of the associated location. When the user is within or near the associated location, the dynamic fence may not be triggered temporarily; when the user crosses the offset boundary of the dynamic fence and moves in the same direction as away from the associated location, the dynamic departure phase is determined to meet the alert output condition.

[0149] In another implementation, the electronic device calculates common departure directions based on the user's historical departure trajectory and sets a dynamic fence near these common departure directions to improve the lead time and accuracy of departure reminders.

[0150] In another implementation, the electronic device only enables dynamic fence or short-term high-precision positioning judgment when the target reminder task is in a triggerable state, the electronic device is within a preset proximity range of the associated location, and within the effective reminder time range; when the target reminder task is completed, expires, the user moves away from the associated location, or the judgment ends, the high-precision positioning judgment is stopped, thereby reducing power consumption.

[0151] Example 5: Triggering the vehicle departure phase

[0152] like Figure 5 As shown, this embodiment illustrates the triggering process of the vehicle departure phase.

[0153] Users create reminder tasks such as "Remind me to pick up my medicine when I get in the car" or "Don't forget to bring your contract when you go out." The system parses this reminder task as a vehicle-related trigger scenario, or as a candidate task for the vehicle departure phase within a location departure trigger scenario.

[0154] When an electronic device detects at least one type of vehicle evidence, such as in-vehicle Bluetooth connection, in-vehicle audio routing switch, in-vehicle system connection, screen mirroring connection, or vehicle pairing device connection, the system generates vehicle connection evidence.

[0155] The system determines whether the target reminder task is associated with a location departure trigger scenario, a vehicle-related trigger scenario, or travel-related semantics. If associated, it further determines whether the confidence level of the vehicle connection evidence meets the vehicle trigger threshold.

[0156] For example, in-vehicle Bluetooth connections can have a high confidence level of vehicle evidence; when in-vehicle audio routing switching and motion state occur simultaneously with driving, the confidence level of vehicle evidence can be further improved; if the in-vehicle connection occurs within a preset time window after the user leaves home or work, the vehicle evidence can form a temporal combination with the departure scenario, improving the trigger credibility.

[0157] When the confidence level of the vehicle connection evidence meets the vehicle trigger threshold, and the target reminder task is not in a completed state or the same vehicle departure phase has not been triggered, the system determines the vehicle departure phase as meeting the reminder output condition and outputs a vehicle departure reminder.

[0158] In one implementation, when the electronic device enters the background, locks the screen, or switches application states, the system does not immediately stop collecting vehicle evidence. Instead, it continues to collect vehicle connection evidence or audio routing evidence within a preset grace period. If vehicle evidence is detected within the grace period, the vehicle departure phase is triggered. If no vehicle evidence is detected after the grace period ends, vehicle evidence collection is stopped or the system switches to a low-power monitoring mode.

[0159] This method can solve the problem of missing the vehicle trigger opportunity when the user has just left the house and entered the underground parking lot, the application has switched to the background, and then the vehicle device is connected, due to prematurely stopping the listening.

[0160] Example 6: Fusion of Evidence Confidence

[0161] This embodiment illustrates a specific method for fusing evidence confidence.

[0162] For each piece of triggering evidence, the system can generate an evidence object. The evidence object can include evidence type, evidence source, collection time, associated task, associated location, original data, confidence level, validity period, and evidence status.

[0163] For example, evidence of a wireless LAN disconnection may include:

[0164] Evidence type: Wireless signal evidence;

[0165] Source of evidence: Home Wi-Fi disconnected;

[0166] Data collection time: T1;

[0167] Related location: Home;

[0168] Confidence level: 0.5;

[0169] Validity period: 60 seconds.

[0170] A piece of evidence that the location is far away can include:

[0171] Type of evidence: Location evidence;

[0172] Source of evidence: High-precision positioning;

[0173] Data collection time: T2;

[0174] Distance from associated location: 45 meters;

[0175] Positioning accuracy: 20 meters;

[0176] Direction of movement: Move away from associated locations;

[0177] Confidence level: 0.7;

[0178] Validity period: 30 seconds.

[0179] Evidence of a vehicle connection may include:

[0180] Evidence type: Vehicle connection evidence;

[0181] Source of evidence: In-vehicle Bluetooth connection;

[0182] Data collection time: T3;

[0183] Confidence level: 0.85;

[0184] Validity period: 120 seconds.

[0185] The system can select fusion rules based on the triggering scenario. For example, in a location departure triggering scenario, one of the following rules can be used:

[0186] If the confidence level of wireless signal evidence is greater than the first threshold, and the confidence level of motion state evidence is greater than the second threshold, and the time difference between their acquisition is less than a preset time window, then it is determined that the early departure phase meets the alert output conditions.

[0187] If the confidence level of the location evidence is greater than the third threshold, and the dynamic fence judgment result is that the boundary has been crossed, then it is determined that the dynamic departure stage meets the reminder output condition;

[0188] If the confidence level of the vehicle connection evidence is greater than the fourth threshold and the target reminder task has not been completed, then it is determined that the reminder output condition is met during the vehicle departure phase.

[0189] If a geofence departure event occurs and the task is not completed, then the fallback departure phase is determined to meet the alert output conditions.

[0190] In another implementation, the system can perform a weighted calculation of the confidence levels of multiple pieces of evidence:

[0191] Fusion confidence = Location evidence confidence × Location weight + Wireless signal evidence confidence × Wireless weight + Motion state evidence confidence × Motion weight + Vehicle connectivity evidence confidence × Vehicle weight + Temporal context evidence confidence × Time weight.

[0192] When the fusion confidence level is greater than the threshold of the corresponding triggering stage, it is determined that the triggering stage meets the reminder output condition.

[0193] Different trigger scenarios or different users can have different weights and thresholds. The system can also adaptively adjust the weights and thresholds based on the user's historical confirmation, ignoring, completion, or postponement of operations.

[0194] Example 7: Triggering state recording, repeat control, and stage escalation

[0195] like Figure 6 As shown, this embodiment illustrates trigger state recording, repetition control, and stage upgrade control.

[0196] The system generates a trigger cycle identifier for the target reminder task. For one-time tasks, the trigger cycle identifier can be generated from the task identifier and the task's effective time window. For recurring tasks, the trigger cycle identifier can be generated from the task identifier, date, associated location, and departure cycle. For location-departure tasks, the trigger cycle identifier can be generated from the task identifier, associated location, and the current departure cycle.

[0197] The system records the stages that have been triggered under each trigger cycle identifier. For example:

[0198] Task A, cycle C1, has triggered the early exit phase;

[0199] Task A, Cycle C1, vehicle departure phase has been triggered;

[0200] Task A, cycle C1, did not trigger the fallback exit phase.

[0201] When new triggering evidence arrives, the system first determines whether the target notification task has been completed. If it has, subsequent notifications are suppressed.

[0202] If the target reminder task is not completed, the system determines whether the current triggering phase has already been triggered within the same cycle. If it has been triggered, it suppresses duplicate reminders in the same phase.

[0203] If the current triggering stage has not yet been triggered, the system further determines whether the stage meets the conditions for outputting a reminder. If it does, a reminder is output, and the stage is recorded as triggered.

[0204] In one implementation, the system allows reminders to be output separately for different triggering stages. For example, if the user has not completed the task after the early departure stage reminder, a reminder can still be output during the vehicle departure stage; if the user still has not completed the task after the vehicle departure stage reminder, a reminder can still be output during the fallback departure stage.

[0205] In another implementation, the system can adjust the escalation strategy based on user settings or task importance. For example, a regular reminder may only be issued once; an important reminder may be issued in multiple stages; and an emergency reminder may be issued with sound, vibration, or continuous alerts.

[0206] Example 8: Sharing and Distributing Reminders

[0207] This example illustrates the scenarios of sharing reminders and distributing reminders.

[0208] Shared reminders refer to multiple users sharing the same reminder task. For example, user A and user B share the reminder "Don't forget your camping gear when you go out." Once any user completes the reminder, the other users' devices synchronize the completion status and suppress subsequent reminders.

[0209] Distributing reminders refers to one user assigning a task to another user. For example, user A sends a reminder to user B to "submit materials upon arriving at the office." User A can check whether the task is completed, but the reminder is primarily sent to user B. After user B completes the task, the task completion status is synchronized to user A's device or server.

[0210] In shared or dispatched reminders, the system can synchronize task status, trigger status, completion status, list item completion status, and user confirmation status.

[0211] For example, User A and User B share a list of reminders, including items such as "tent, water, power bank, and medicine." When User B completes the "tent" and "water" tasks, User A's device can display the corresponding completion status. Once all items on the list are completed, both devices suppress subsequent reminders.

[0212] This mechanism can prevent the problem of one device completing a task while other devices continue to send notifications in multi-user scenarios.

[0213] Example 9: Low Power Consumption Control

[0214] This embodiment illustrates the low-power control method of the present invention.

[0215] Since continuous high-precision positioning, frequent scanning of wireless signals, or continuous monitoring of vehicle evidence may increase the power consumption of electronic devices, this invention adopts a candidate task screening and hierarchical acquisition strategy.

[0216] When there is no target reminder task in a triggerable state, the system maintains a low power consumption state and does not start high-frequency positioning or high-frequency sensor acquisition.

[0217] When a target reminder task exists but the user is not within the vicinity of the associated location, the system can retain only low-frequency time judgments or system-level event listening.

[0218] When a user enters the vicinity of an associated location and the target notification task is within its valid time frame, the system initiates evidence collection related to the triggering scenario. For example, for a location departure task, it initiates monitoring of Wi-Fi status, nearby wireless broadcast signals, movement status, or geofence.

[0219] When early evidence of departure is detected, the system can activate high-precision positioning or dynamic fence detection within a short time window. If departure is not confirmed within this time window, high-precision positioning is stopped, and the system returns to a low-power state.

[0220] When an electronic device goes into the background, is locked, or the application status changes, the system can determine whether to continue retaining the short-term vehicle evidence collection window based on the task status. If vehicle evidence is detected within the window, an alert is triggered; otherwise, vehicle evidence collection is stopped.

[0221] Through the above methods, the present invention can reduce power consumption while ensuring timely reminders.

[0222] Example 10: System Structure

[0223] like Figure 2 As shown, this embodiment provides an intelligent reminder triggering system based on multi-source evidence fusion, including a task acquisition module, a semantic parsing module, a candidate task filtering module, a multi-source evidence collection module, an evidence fusion module, a triggering stage decision module, a notification output module, and a status control module.

[0224] The task acquisition module is used to retrieve reminder tasks created by the user. Reminder tasks can be generated through text input, voice input, template creation, importing from external applications, or sharing tasks.

[0225] The semantic parsing module is used to perform semantic parsing on reminder tasks to determine the reminder action, the reminder object, and at least one triggering scenario.

[0226] The candidate task filtering module is used to determine the target reminder task that is currently in a triggerable state from the candidate reminder task set based on the triggering scenario, task status, task validity period, associated location, current location, user status, and device status.

[0227] The multi-source evidence acquisition module is used to collect multi-source triggering evidence associated with the target reminder task. The multi-source evidence acquisition module may include a positioning acquisition unit, a wireless signal acquisition unit, a motion state acquisition unit, a vehicle connectivity acquisition unit, a time context acquisition unit, and a user operation state acquisition unit.

[0228] The evidence fusion module is used to determine the corresponding evidence confidence level based on multi-source triggering evidence, and generate fusion triggering results based on preset evidence fusion rules.

[0229] The trigger phase decision module is used to determine the trigger phase corresponding to the target reminder task based on the fusion trigger results.

[0230] The notification output module is used to output a reminder notification corresponding to the triggering stage when the reminder output conditions are met during the triggering stage.

[0231] The status control module is used to record the trigger status of the target reminder task and to control the repetition or escalation of subsequent reminder notifications based on the trigger status.

[0232] The modules described above can be deployed in the same electronic device, or partially deployed in electronic devices, servers, vehicle terminals, wearable devices, or other devices. Each module can be implemented through software, hardware, firmware, or a combination thereof.

[0233] Example 11: Electronic Devices and Storage Media

[0234] This embodiment provides an electronic device, including a processor, a memory, a communication module, a positioning module, a sensor module, a display module, and a notification module.

[0235] The memory stores a computer program. When the processor executes the computer program, it can implement the method described in any embodiment of the present invention.

[0236] The communication module can be used to communicate with wireless LANs, Bluetooth devices, in-vehicle devices, servers, or other user devices.

[0237] The positioning module can be used to acquire the location data of electronic devices.

[0238] The sensor module may include an accelerometer, gyroscope, magnetometer, motion coprocessor, or other sensors.

[0239] The notification module can be used to output system notifications, sounds, vibrations, in-app notifications, or other reminders.

[0240] This embodiment also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the method described in any embodiment of the present invention.

[0241] VI. Supplementary Optional Implementation Methods

[0242] In some implementations, the system can personalize the triggering rules based on the user's historical behavior. For example, if a user repeatedly confirms completion of a certain type of reminder only during the vehicle departure phase, the system can increase the priority of reminders during the vehicle departure phase; if a user frequently ignores early departure reminders, the system can adjust the reminder method or threshold for the early departure phase.

[0243] In some implementations, the system can adjust the alert intensity based on the importance of the alert content. For example, highly important items such as medicines, certificates, contracts, and keys can trigger multi-stage alerts; ordinary matters can trigger only one alert.

[0244] In some implementations, the system can expand the reminder content based on weather, schedule, traffic, or other contextual information. For example, when the reminder task is related to outdoor travel and the weather may rain, the system can add a suggestion to bring rain gear to the reminder content.

[0245] In some implementations, the system can store trigger evidence and trigger status in a local database, or synchronize them to a server or multiple user devices.

[0246] In some implementations, the system can desensitize, minimize storage of, or compute triggering evidence only on the device side, while protecting user privacy.

[0247] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, or combinations made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for triggering intelligent reminders based on multi-source evidence fusion, characterized in that, Applied to electronic devices, the method includes: Obtain user-created reminder tasks, wherein the reminder task includes reminder content and triggering conditions associated with the reminder content; The reminder task is semantically parsed to determine the reminder action, reminder object, and at least one triggering scenario corresponding to the reminder task. The triggering scenario includes at least one of time-triggered scenario, location arrival triggering scenario, location departure triggering scenario, and vehicle-related triggering scenario. Based on the triggering scenario, determine the target reminder task that is currently in a triggerable state from the candidate reminder task set; When the target reminder task is in a triggerable state, multi-source triggering evidence associated with the target reminder task is collected. The multi-source triggering evidence includes at least two of the following: location evidence, wireless signal evidence, motion state evidence, vehicle connection evidence, time context evidence, and user operation state evidence. The confidence level of each of the multi-source triggering evidences is determined, and a fusion triggering result is generated based on the preset evidence fusion rules. The triggering stage corresponding to the target reminder task is determined based on the fusion triggering result; When the triggering stage meets the reminder output conditions of the target reminder task, a reminder notification corresponding to the triggering stage is output. Record the triggering status of the target reminder task, and control the repetition or escalation of subsequent reminder notifications based on the triggering status.

2. The method according to claim 1, characterized in that, The reminder task is semantically parsed to determine the reminder action, reminder object, and at least one triggering scenario corresponding to the reminder task, including: Perform intent recognition on natural language text or speech recognition text input by the user; Extract at least one of the following from the natural language text or speech recognition text: time information, location information, departure intention, arrival intention, vehicle usage intention, item information, and event information; Based on the extraction results, structured reminder conditions are generated, which include at least one of the following: trigger type, associated location, trigger direction, reminder content, and effective time range; Specifically, when the reminder task contains the semantics of leaving, exiting, getting on a vehicle, arriving, getting off a vehicle, or similar, the reminder task is mapped to the corresponding location departure trigger scenario, vehicle-related trigger scenario, or location arrival trigger scenario.

3. The method according to claim 1, characterized in that, The multi-source triggering evidence includes at least two of the following categories: Location evidence generated by satellite positioning, base station positioning, wireless LAN positioning, or system geofencing; Evidence of wireless signals generated by wireless LAN connection status, wireless LAN disconnection status, Bluetooth signals, near-field wireless broadcast signals, or short-range communication signals; Evidence of motion state generated from accelerometer, gyroscope, step count data, motion coprocessor, or system motion state recognition results; Vehicle connectivity evidence generated by in-vehicle Bluetooth connection, in-vehicle audio router, in-vehicle system connection, screen mirroring connection, or vehicle pairing device connection; Temporal contextual evidence generated by current time, scheduled time, reminder validity period, user's frequently traveled time periods, or historical trigger times; Evidence of user action status generated by reminder task completion status, list item completion status, triggered stage, user confirmation action, or user ignored action.

4. The method according to claim 1, characterized in that, Based on the multi-source triggering evidence, the corresponding evidence confidence level is determined, and a fusion triggering result is generated based on a preset evidence fusion rule, including: The confidence level of each triggering evidence is calculated based on at least one of the following: source, collection time, location accuracy, signal strength, distance from the associated location, and degree of matching with historical behavior patterns. Different weights are assigned to different categories of triggering evidence based on the triggering scenario; The confidence levels of multiple pieces of evidence are weighted, thresholded, time-series combination, or rule-matched to obtain the fusion trigger result. Specifically, when multiple weak pieces of evidence jointly meet the triggering conditions within a preset time window, the fusion triggering result is determined to meet the reminder output condition; when a single strong piece of evidence meets the preset strong triggering condition, the fusion triggering result is determined to meet the reminder output condition.

5. The method according to claim 1, characterized in that, When the triggering scenario is a location departure triggering scenario, the triggering stage includes at least two of the following: an early departure stage, a dynamic departure stage, a vehicle departure stage, and a fallback departure stage; The early departure phase is determined based on at least one of the following: wireless LAN disconnection, loss of near-field wireless broadcast signal, change in Bluetooth signal, and change in motion state. The dynamic departure phase is determined based on dynamic fences, directional fences, offset fences, or short-term high-precision positioning corresponding to the associated location. The vehicle departure phase is determined based on in-vehicle Bluetooth connection, in-vehicle audio router, vehicle-machine connection, vehicle pairing device connection, or vehicle motion status. The fallback departure phase is determined based on system geofence departure events, distance thresholds, continuous moving away trends, or ordinary location results. Furthermore, the method allows the same reminder task to output reminder notifications at different triggering stages, and suppresses duplicate reminder notifications within the same triggering stage.

6. The method according to claim 5, characterized in that, The dynamic departure phase is determined based on a dynamic fence, and the dynamic fence is generated in the following ways: Obtain at least one of the following: the center location of the associated location, the boundary of the associated location, the user's current location, the user's orientation, the road direction, the building exit direction, or the historical departure direction; Based on the center location, boundary, user's current location, user's orientation, road direction, building exit direction, or historical departure direction, generate a fenced area offset relative to the associated location; When the electronic device is within a preset proximity range of the associated location and the target reminder task is in a triggerable state, the dynamic fence or short-term high-precision positioning judgment is activated. When it is detected that the electronic device crosses the dynamic fence, moves away from the associated location, or the distance from the associated location meets a preset departure threshold, the dynamic departure phase is determined as meeting the reminder output condition.

7. The method according to claim 5, characterized in that, The methods for determining the vehicle departure phase include: When at least one of the following vehicle evidences is detected: in-vehicle Bluetooth connection, in-vehicle audio routing switch, in-vehicle system connection, screen mirroring connection, or vehicle pairing device connection, vehicle connection evidence is generated. Determine whether the target reminder task is associated with a location departure trigger scenario, a vehicle-related trigger scenario, or travel-related semantics; When the confidence level of the vehicle connection evidence meets the vehicle trigger threshold, and the target reminder task is not in a completed state or the same vehicle departure phase is not triggered, the vehicle departure phase is determined to meet the reminder output condition. Specifically, when the electronic device enters the background, locks the screen, or switches application states, the collection of vehicle evidence does not stop immediately. Instead, it continues to collect vehicle connection evidence or audio routing evidence within a preset grace period window, and triggers the vehicle departure phase when vehicle evidence is detected within the preset grace period window.

8. The method according to claim 1, characterized in that, Record the triggering status of the target reminder task, and perform repetition control or stage escalation control on subsequent reminder notifications based on the triggering status, including: Generate a trigger cycle identifier for the target reminder task; Record the triggering phase that the target reminder task has been triggered under the triggering cycle identifier; When it is detected that a reminder notification has already been output under the same trigger period identifier and the same trigger phase, the repeated reminder notifications for that trigger phase are suppressed; When different triggering stages under the same triggering cycle identifier are detected to meet the reminder output conditions, the corresponding stage upgrade reminder notification is allowed to be output. When it is detected that the target reminder task has been completed by the user, marked as completed by the system, or all items in the list corresponding to the target reminder task have been completed, the subsequent triggering of the target reminder task shall be stopped or suppressed. Specifically, for shared reminder tasks or dispatched reminder tasks, the triggering state or completion state can be synchronized across multiple user devices, so that the completion operation on one user device suppresses the subsequent reminder triggering on another user device.

9. An intelligent reminder triggering system based on multi-source evidence fusion, characterized in that, include: The task acquisition module is used to acquire reminder tasks created by the user. The reminder task includes reminder content and triggering conditions associated with the reminder content. The semantic parsing module is used to perform semantic parsing on the reminder task to determine the reminder action, reminder object, and at least one triggering scenario corresponding to the reminder task; The candidate task filtering module is used to determine the target reminder task that is currently in a triggerable state from the candidate reminder task set based on the triggering scenario; The multi-source evidence acquisition module is used to acquire multi-source triggering evidence associated with the target reminder task. The multi-source triggering evidence includes at least two of the following: location evidence, wireless signal evidence, motion state evidence, vehicle connection evidence, time context evidence, and user operation state evidence. The evidence fusion module is used to determine the corresponding evidence confidence level according to the multi-source triggering evidence, and generate a fusion triggering result based on the preset evidence fusion rules; The triggering phase decision module is used to determine the triggering phase corresponding to the target reminder task based on the fusion triggering result. The notification output module is used to output a reminder notification corresponding to the triggering stage when the triggering stage meets the reminder output conditions of the target reminder task; The status control module is used to record the triggering status of the target reminder task and to perform repeat control or stage escalation control on subsequent reminder notifications based on the triggering status.

10. An electronic device or computer-readable storage medium, characterized in that, The electronic device includes a processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein the processor executes the computer program to implement the method as described in any one of claims 1 to 8; or, the computer-readable storage medium stores a computer program, which, when executed by the processor, implements the method as described in any one of claims 1 to 8.

Images