An intelligent reminding system for the travel of the elderly based on smart glasses

Abstract

The application relates to the technical field of intelligent wearable device data processing and control, and discloses an old person travel intelligent reminding system based on intelligent glasses, which comprises the following modules: a state quantization module extracts hardware features to construct a resource occupation state word containing a perception channel load bit; a mask configuration module splits reminding data into core and detail data and configures a demand mask; a concurrent routing module separates idle channels by bit operation, degrades output of the core prompt, and pushes the detail data into a background queue; an addressing control module dynamically updates a rendering template reading pointer based on suspension time and step number; a multi-modal verification module triggers wake-up verification through state word edge jump or field angle integral variation; and a safe rendering module realizes instantaneous state mask review and adaptive output by isolating peripheral interruption. The application avoids sudden stimulation interference during high-risk movement, offsets forgetting attenuation caused by task suspension, and eliminates a safety blind area in an output instant through bottom-layer atomic-level interception.

Description

Technical Field

[0001] This invention relates to the field of data processing and control technology for smart wearable devices, specifically a smart travel reminder system for the elderly based on smart glasses. Background Technology

[0002] Wearable devices such as smart glasses have penetrated into the daily health monitoring of the elderly population. Reminders for medication, follow-up visits, and moderate exercise constitute the core interactive services of the system. These devices have the technological advantages of first-person visual presentation and auditory transmission close to the ear. Near-eye optical display and bone conduction technology provide elderly people with intuitive and close-range information access methods for their cognitive functions that are gradually declining.

[0003] The current travel reminder logic of wearable devices mainly relies on the internal real-time clock module for unidirectional scheduling. After reaching the preset time node, the main control chip extracts the reminder text or graphic resources stored locally. The underlying driver encapsulates the above data into standard rendering instructions, which are then written to the micro-display and speaker registers through the internal bus to complete the information display. To avoid excessive interference, some conventional solutions will read the status characteristics of the basic sensors at the moment of triggering. When the device is found to be in a state of obvious physical movement, the operating system will transfer the current reminder task to the backend delay queue. Tasks in a suspended state rely on a fixed system timer to periodically wake up the peripheral hardware, detect whether the external environment has returned to stillness, and then, when the conditions are met, re-execute the pop-up window and voice broadcast with the originally set interface warning intensity.

[0004] However, due to the weaker balance control of the elderly, forcibly displaying highlighted graphics when going up or down stairs or crossing obstacles can instantly deprive them of their visual pathways. This sudden obstruction of vision directly amplifies the physical risk of falls and collisions. Simple delayed suspension operations abandon the tracking and management of task urgency. Given that human short-term memory declines irreversibly over time, a fixed retry prompt intensity is insufficient to rebuild cognitive expectations, leading to an increased probability that users will completely forget key schedules. Using timers to frequently wake up peripherals such as cameras to scan the environment continuously increases system noise and power consumption, and there is a processing time difference between software-level security judgments and actual bus writing to hardware. Therefore, this invention provides an intelligent travel reminder system for the elderly based on smart glasses to address the shortcomings of existing technologies. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an intelligent travel reminder system for the elderly based on smart glasses. It solves the problem that existing wearable reminder devices lack underlying scheduling intervention based on the user's real-time physical environment and the load status of the sensing channel when executing schedule reminders. Traditional devices usually use strong blocking prompts with fixed timing. When users are in high-risk activity scenarios such as going down stairs or running, or when there is severe external interference, sudden audiovisual rendering commands can easily cause users to jump or have their vision blocked, leading to personal safety hazards.

[0006] To achieve the above objectives, the present invention provides the following technical solution: The first aspect of this invention provides a smart travel reminder system for the elderly based on smart glasses, comprising: The state quantization module collects sensor characteristic parameters and constructs a user resource occupancy status word that includes visual, auditory, and motion load bits. The mask configuration module receives a trip task containing reminder data, splits the reminder data into core index data and detail load data, configures the required mask for each, and initializes the rendering template array. The concurrent routing module performs bitwise operations using the demand mask and the user resource occupancy status word, separates the available channels to send the core index data based on the operation result, and truncates and transfers the detailed load data to the asynchronous background queue. The addressing control module accumulates the suspension time and physical steps of the data in the asynchronous background queue as a spatiotemporal urgency parameter, calculates the offset index using the spatiotemporal urgency parameter, and continuously updates the read pointer of the rendering template array. The multimodal verification module listens for the falling edge transition of the motion load bit in the user resource occupancy status word or extracts the integral feature of the field of view change, thereby triggering wake-up and performing environment readiness verification. After the environment readiness verification is passed, the secure rendering module isolates external interrupts, completes transient mask verification, maps the updated read pointer, and executes adaptive prompts.

[0007] Furthermore, when processing the reminder data, the mask configuration module is configured to: acquire the infrared proximity pin level, attitude Euler angle, and illuminance value; perform cross-validation using Boolean logic equations to determine that the device is in an effectively worn state and initiate the corresponding data processing pipeline; for the in-body task, forcibly divide the original reminder data stream into core index data occupying a smaller storage space and detailed load data containing complete journey details according to the data frame header identifier; configure core requirement masks and detailed requirement masks based on binary sequences for the core index data and the detailed load data respectively, and allocate addressing areas in memory to generate a rendering template array sorted by security saliency in ascending order.

[0008] Preferably, when constructing the user resource occupancy status word, the state quantization module is configured as follows: within a set data acquisition time window, extract the vertical axis acceleration variance, the root mean square value of the audio sequence, and the instantaneous illuminance value as underlying hardware feature parameters; compare the extracted underlying hardware feature parameters with preset conventional motion variance thresholds, environmental noise thresholds, and environmental darkness thresholds, and discretize and map them into Boolean state variables; and concatenate the Boolean state variables to assemble the user resource occupancy status word containing the visual load bit, auditory load bit, and motion load bit, so as to output it to the concurrent routing module to represent the current high load occupancy state or idle state of each sensory channel.

[0009] Preferably, the state quantization module is also equipped with a hardware-level veto mechanism: the state quantization module simultaneously extracts the motion cycle frequency, the vertical axis acceleration variance, and the zero-crossing rate of the angular velocity corresponding to the head pitch motion, and combines them to generate a multi-dimensional feature vector characterizing the user's dynamic behavior; the multi-dimensional feature vector is input into a high-risk feature subspace defined by the motion variance limit value, the zero-crossing rate threshold, and the frequency threshold set to perform a spatial matching operation; when the matching determines that the multi-dimensional feature vector falls into the high-risk feature subspace, a hardware-level interception operation is triggered, forcibly skipping the independent environmental threshold comparison, and directly writing an instruction to the memory bus to overwrite the user resource occupancy status word to a fully blocked state, thereby cutting off the unsafe reminder event from the data processing front end.

[0010] Preferably, when performing bitwise operations and truncation transfers, the concurrent routing module is configured to: extract the detailed requirement mask and perform a bitwise AND operation with the real-time generated user resource occupancy status word; if the operation result determines that the channel set does not overlap, then the pass-through condition is met, the initialized read pointer is extracted and the complete output is performed; if the bitwise AND operation indicates a channel conflict, the core requirement mask and the bitwise inverted user resource occupancy status word are extracted and a composite bitwise operation is performed to generate a degraded security mask that removes environmental occupancy conflicts; based on the indicator bits of the degraded security mask, the available independent channel is allocated to send the core index data, and the detailed load data is synchronously retained and pushed into the asynchronous background queue of static access and updated to an asynchronous suspended state, so as to hand over to the addressing control module for subsequent scheduling.

[0011] Preferably, when calculating the offset index, the addressing control module is configured as follows: for the detailed load data transferred to the asynchronous background queue, a hardware timer and a step counting hardware interrupt are used to accumulate and generate a suspension time variable and a physical effective step count variable as parameters of spatiotemporal urgency; the suspension time variable and the physical effective step count variable are extracted, and multiplication and addition algebraic equation operations are performed in combination with the set time weight constant and step weight constant to construct an addressing offset calculation logic that directly maps to the evolution of the physical environment; an extreme value limiting function is used to perform out-of-bounds truncation protection on the result of the multiplication and addition algebraic equation operation after floating-point conversion, and a valid template index value is output to periodically overwrite the pointer variable address of the memory, thereby realizing the dynamic adaptive drift of its own rendering attribute intensity during the asynchronous data residence period.

[0012] Preferably, when the multimodal verification module is triggered to wake up, it is configured to: configure an edge-triggered interrupt service routine to continuously monitor the level state of the motion load bit in the user resource occupancy status word; when a falling edge is captured where the level state changes from high to low, load a preset stable anti-jitter window time parameter to perform state filtering; during the existence of the stable anti-jitter window, maintain polling and scanning of the level state of the motion load bit; if the level reverses, clear the interrupt flag bit; if the level remains low for the entire time window, determine that the physical environment has established a true steady state; in response to the establishment of a true steady state in the physical environment, send an initial wake-up interrupt signal to the internal bus to trigger the power-on of the peripheral radio frequency or optical hardware and perform environmental readiness verification of the target item.

[0013] Preferably, the multimodal verification module also has a secondary wake-up mechanism when the initial wake-up verification fails: when the system verification fails, causing the initial wake-up interrupt signal to terminate, and the user maintains a stationary posture so that the motion load bit remains in a low-level state for a long time, the yaw rate and pitch rate within the continuous sampling period are extracted; discrete integration is performed on the yaw rate and pitch rate data sequence to obtain the yaw rate increment and pitch rate increment; when it is determined that any angle increment exceeds the angle threshold calibrated by the human eye's foveal field of view, the field of view variation integral feature is generated; based on the field of view variation integral feature, a secondary wake-up interrupt signal is issued, and peripheral radio frequency components are preferentially allocated to broadcast detection frames to scan electronic tags in order of increasing power consumption, or the low-power image sensor is degraded to wake up and retrieve the reference plane mark pattern, and the environmental readiness verification pass mark is provided after successful matching.

[0014] Preferably, when the secure rendering module completes the transient mask verification, it is configured as follows: after obtaining the environment readiness verification pass flag, in the instruction cycle before preparing to import the rendering data frame into the display interface or digital-to-analog converter via the direct memory access controller, it masks external conventional hardware interrupts through kernel instructions and forces entry into the protected critical section; within the critical section, it bypasses the cache and directly reads the transient user resource occupancy status word re-collected in the underlying hardware register, and extracts the detailed demand mask to perform bitwise AND operation again for mask verification; if the operation result indicates that the transient channel availability status is true, it releases the interrupt mask and releases the processor resources, and loads the suspended detailed load data into the corresponding template data placeholder and converts it into underlying timing signals for presentation according to the read pointer updated by the addressing control module.

[0015] Preferably, the secure rendering module is further configured with a watchdog interception pipeline: maintaining an internal timed thread to periodically compare the suspension time variable of tasks in the asynchronous background queue with a preset limit security threshold; when it is determined that the suspension time variable exceeds the limit security threshold, the watchdog program cuts off the regular rendering pipeline pointing to the high-brightness screen display or the strong vibration peripheral, and reloads the detailed load data into the preset low-disturbance secure single-channel template for local output; synchronously extracts the process task identifier that caused the timeout, the instantaneous user resource occupancy status word, and the spatiotemporal stamp parameters to assemble an emergency data packet, and controls the underlying communication module to call the wireless communication protocol at the link layer to push the emergency data packet to the preset terminal to build a failover guarantee.

[0016] This invention provides a smart travel reminder system for the elderly based on smart glasses, which has the following beneficial effects: 1. This invention extracts sensor parameters from physical hardware in real time and quantifies them into status words containing multiple channel load bits. It then performs bitwise operations to compare these status words with a task requirement mask. When the system determines that the user is in a high-risk occupancy state such as motion or environmental interference, it actively intercepts the output of detailed load data and pushes it into the background queue. At the same time, it uses a degradation mask to match a safe, idle channel and outputs only the core prompt to establish the expectation. This pre-scheduling mechanism cuts off inappropriate audiovisual rendering instructions from the underlying data link, effectively avoiding startled reactions and visual obstruction caused by sudden strong reminders, and improving the interactive security of wearable devices in complex physical environments.

[0017] 2. This invention overcomes the passive forgetting defect caused by long-term background suspension of reminder tasks by constructing a hardware accumulation and pointer offset algorithm based on spatiotemporal urgency. The system uses an independent hardware timer and step counting interrupt to accumulate the suspension time variable and physical step count variable in real time without waking up the main processor core, and uses them as input parameters of the algebraic equation to dynamically calculate the offset index. Under the protection of the out-of-bounds amplitude function, this mechanism periodically updates the read pointer of the rendering template in memory, so that the reminder data in the background can establish a direct mapping with the evolution of the user's physical environment. This allows the rendering warning intensity of the suspended task to adaptively increase and drift over time, effectively offsetting the user forgetting decay caused by the delayed readiness of the environment.

[0018] 3. This invention significantly reduces system standby power consumption and eliminates high-risk race conditions by integrating an event-driven wake-up architecture with a transient mask verification mechanism before rendering. During task suspension, the system abandons high-frequency timed polling and instead listens for the falling edge of the motion load bit level or the integral increment event of the field of view to trigger hardware verification, avoiding battery consumption caused by continuous sensor power-on. More importantly, in the instruction cycle before confirming that the environment is ready and the bus is about to send rendering data, the system enters the critical section by isolating peripheral interrupts to perform atomic-level secondary reading of the status word and mask verification. This low-level interception mechanism prevents sudden environmental degradation that occurs during the time difference between confirming environmental safety and actual data writing, ensuring that the final delivery condition of the prompt information is absolutely safe, improving the task closure completion rate while protecting personal safety. Attached Figure Description

[0019] Figure 1 This is a system architecture diagram of the present invention; Figure 2 This is a flowchart of the method of the present invention; Figure 3 This is a flowchart illustrating the execution state verification and data decomposition process of the mask configuration module according to an embodiment of the present invention. Figure 4 A logical flowchart for constructing a user resource occupancy status word for a status quantization module in one embodiment of the present invention; Figure 5 This is a schematic diagram illustrating the principle structure of a concurrent routing module performing bitwise operations and suspending flow according to an embodiment of the present invention. Figure 6 This is a schematic diagram illustrating the calculation principle of dynamic updating of the template pointer by the addressing control module in one embodiment of the present invention. Figure 7 This is a flowchart illustrating the event-driven and wake-up verification process performed by the multimodal verification module according to an embodiment of the present invention. Figure 8This is a schematic diagram illustrating the timing control principle of closed-loop output and fallback in a secure rendering module according to an embodiment of the present invention. Figure 9 This is a graph comparing the experimental results of different reminder schemes of the present invention.

[0020] Among them, 101 is the mask configuration module; 102 is the state quantization module; 103 is the concurrent routing module; 104 is the addressing control module; 105 is the multimodal verification module; and 106 is the secure rendering module. Detailed Implementation

[0021] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0022] See attached document Figure 1 This invention provides a smart glasses-based intelligent travel reminder system for the elderly, running on a wearable hardware platform equipped with a microcontroller unit, a low-power image sensor, an ambient light sensor, a communication bus, and audiovisual output components. The system includes: The mask configuration module 101 is used to receive external task timestamps and corresponding reminder data, split the reminder data into core index data and detail load data, configure core requirement masks for core index data, configure detail requirement masks for detail load data, generate a rendering template array with security saliency in ascending order in the system memory, and initialize template index parameters.

[0023] The state quantization module 102 is used to collect sensor characteristic parameters obtained by the microcontroller unit; it integrates infrared proximity level, pitch and roll angle parameters, and ambient light intensity to determine the current on-site state of the device; and it constructs a user resource occupancy status word including visual load bit, auditory load bit, and motion load bit by extracting inertial measurement unit variance data, image motion amount, and audio sampling channel data.

[0024] The concurrent routing module 103, connection status quantization module 102, and mask configuration module 101 extract the demand mask of the task and perform bitwise logical verification with the real-time user resource occupancy status word to determine the availability status of channel resources. When the channel is completely idle, it directly triggers the complete data output. When the channel is in conflict, it calculates the degraded safety mask, outputs the core index data, and transfers the detailed load data to the background suspended area.

[0025] The addressing control module 104, for detailed load data entering the background suspended area, starts the spatiotemporal urgency recording unit, extracts the system real-time clock and step counting register data, accumulates and records the suspended time variable and the physical effective step count variable, calls the preset weight parameters to calculate the template index addressing formula and performs out-of-bounds processing, and periodically updates the read pointer of the rendering template array.

[0026] The multimodal verification module 105 is used to monitor the status level signal of the moving load position. When the falling edge of the level state is acquired and the state remains stable within the anti-shake window, the initial wake-up signal is triggered. When the moving load position remains stable and the initial wake-up verification is determined to be unsuccessful, the module enters a sleep monitoring state. The second wake-up signal is triggered by the integral yaw angle increment exceeding the limit. After responding to the wake-up signal, the low-power image sensor is woken up to identify the preset visual mark.

[0027] The security rendering module 106 responds to the verification success flag output by the multimodal verification module 105, enters the critical section to perform interrupt masking operation, reads the transient user resource occupancy status word through the underlying bus to perform mask verification calculation, releases interrupt masking after confirming that there is no environmental degradation or sudden change, loads detailed load data into the corresponding rendering template output according to the updated template index parameters, and initiates the network underlying communication protocol to encapsulate and send emergency data packets to the preset terminal for tasks that exceed the suspension time threshold.

[0028] See attached document Figure 2 This invention provides a method for intelligent travel reminders for the elderly based on smart glasses, comprising the following steps: S10, the device's on-state status flag is verified by the status quantization module 102, the task processing pipeline in the storage state is blocked, and the reminder data in the on-state state is split into core index data and detailed load data.

[0029] S20: Collect sensor characteristic parameters obtained by the microcontroller unit and construct a user resource occupancy status word including visual load bit, auditory load bit and motion load bit.

[0030] S30: Obtain the task requirement mask and user resource occupancy status word, perform bitwise operations, and send the detailed load data to the output pipeline by full bit verification, or separate the logic and operation of the status word after bit inversion into the available core channel and truncate the detailed data into the spatiotemporal background queue.

[0031] S40 extracts the timer and step counting hardware parameters to calculate the offset algebraic equation for the data in the background suspended state, and continuously updates the rendering template memory pointer by calculating the offset index through amplitude limiting correction.

[0032] S50 captures the physical static state level edge and field angle variation integral characteristics; after meeting the threshold condition, it wakes up the peripheral RF or optical hardware to verify the environment readiness.

[0033] S60 implements transient mask verification based on bus write duration isolation of peripheral interrupts, and maps drift pointer content to achieve adaptive prompting or relies on watchdog mechanism to transfer the underlying communication bus to upload intervention command packets.

[0034] The technical solution of the present invention will be described in detail below with reference to specific embodiments.

[0035] See attached document Figure 3 When the mask configuration module 101 receives an external task carrying a trigger timestamp, it triggers a device in-situ state verification to confirm whether the user is actually wearing the smart glasses, thus avoiding invalid use of computing resources. Specifically, the mask configuration module 101 acquires the underlying hardware signals and physical parameters collected by the state quantization module 102, including the pin levels of the infrared proximity sensor, the attitude Euler angles output by the inertial measurement unit, and the illuminance values ​​output by the ambient light sensor; and sets the in-situ state flag to... The infrared proximity sensor outputs a flag indicating that... The pitch angle calculated and extracted by the microcontroller based on the inertial measurement unit is denoted as... The roll angle is denoted as The illuminance value output by the ambient light sensor is recorded as .

[0036] To avoid false positives caused by the infrared proximity sensor outputting a high level due to physical obstruction when the device is in a clothing pocket or upside down on a table, this embodiment introduces cross-validation logic based on device posture and ambient brightness; the mask configuration module 101 executes the following Boolean logic equation to calculate the presence status flag bit. : ;

[0037] In the formula, For logical AND operation, For logical OR operation, For logical NOT operation; The abnormal wearing posture angle range of the device can be set, for example, it can be set to the range between 45 degrees and 60 degrees with an absolute value of the angle, to represent the flat or flipped state of the device; To determine the light intensity threshold in a dark room, it can be calibrated within the range of 0 to 10 lux, taking into account the specific hardware sensitivity.

[0038] According to the above equation, if And the device's orientation falls into interval, simultaneously Less than Then the calculation yields The mask configuration module 101 thus determines that the device is in a receiving state; in this state, the mask configuration module 101 forwards the received task data packet directly to the external associated terminal through the underlying communication bus, and terminates the subsequent processing pipeline for the task on the current device side to reduce system power consumption; conversely, if the calculation shows... If the device is in a valid wearing state, it is determined that the device is in the body state. Among them, for the matrix transformation operation of the inertial measurement unit to calculate Euler angles, those skilled in the art can use conventional attitude fusion algorithms such as Kalman filtering or complementary filtering, which are well known technologies in this field and will not be described in detail here.

[0039] For tasks determined to be in body posture, the mask configuration module 101 performs structured splitting on the original reminder data stream it carries, forcibly dividing it into core index data and detail load data according to the data frame header identifier; this splitting operation isolates the basic event trigger signal from the display of specific task content on the data link; the core index data is configured as a trigger signal that occupies a small storage space and does not contain specific task text, such as a monosyllabic prompt waveform file used to drive a bone conduction speaker, or a fixed duty cycle control command to drive a vibration motor; the detail load data is configured as a data block containing complete task details, such as an augmented reality graphics rendering command stream carrying absolute coordinate parameters, or complete text used for speech synthesis.

[0040] After data decomposition, the mask configuration module 101 parses the hardware output categories of the two types of data and configures the channel requirement mask in the memory respectively; specifically, the detail requirement mask corresponding to the detail payload data is set as follows: The core requirement mask corresponding to the core index data is The demand mask is constructed as a three-bit binary sequence containing visual demand bits, auditory demand bits, and motion demand bits. A bit set to 1 in this sequence indicates that the output of the corresponding data is strongly dependent on the sensory channel being in an idle state, while a bit set to 0 indicates that the physical occupancy state of the corresponding sensory channel does not interfere with the output of the data.

[0041] Meanwhile, the mask configuration module 101 allocates addressing regions within the storage space to generate a rendering template array. , where index variable , The preset maximum index value of the template; each element in the rendering template array is associated with the entry address of a different rendering function at the underlying level of the display driver or audio driver; according to the index variable The order of increasing values ​​corresponds to the visual contrast, flicker frequency, or audio gain in decibels output by the associated rendering functions; after memory allocation and entry address binding are completed, the mask configuration module 101 sets the template index parameters of the current task. Initialize the value to 0.

[0042] See attached document Figure 4 The state quantization module 102 extracts underlying hardware feature parameters through the microcontroller unit to evaluate the user's real-time physical motion intensity and environmental state. Within the set data acquisition time window, the state quantization module 102 reads the Z-axis acceleration sequence of the inertial measurement unit perpendicular to the horizontal plane, and the microcontroller unit performs mathematical statistics to obtain the variance value of the sequence. It is used to characterize the user's dynamic movement intensity and limb stability.

[0043] The state quantization module 102 calls the microelectromechanical system microphone to sample the ambient sound signal and calculates the root mean square value of the audio sequence within the corresponding time window. This reflects the current ambient noise level of the space; and simultaneously reads the instantaneous illuminance value output by the ambient light sensor. The time-domain calculations of acceleration variance and audio root mean square value can be implemented by conventional statistical algorithms in the field of digital signal processing, which are well-known technologies in this field and will not be elaborated here.

[0044] After obtaining the aforementioned basic physical characteristic parameters, the state quantization module 102 performs a quantization mapping operation, converting continuous sensor sample values ​​into discrete Boolean state variables, and then concatenating and assembling them into a user resource occupancy status word; this user resource occupancy status word is defined in memory as a binary sequence containing three data bits. In this sequence, For visual load bits, For auditory load, This is the motion load bit. A value of 1 in the sequence indicates that the sensory channel is in a high-load state, and a value of 0 indicates that the sensory channel is in an idle state.

[0045] The state quantization module 102 is configured with comparison logic to determine thresholds corresponding to each load bit; specifically, the conventional motion variance threshold is set as follows: If it can be calibrated as the empirical value of variance under the normal walking speed of adults on flat ground, when When the user is determined to be in a normal motion occupancy state, the state quantization module 102 will assign motion load bits. Set to 1; conversely, when hour, Set to 0; set the environmental noise threshold to... For example, the audio energy value can be calibrated to be in the range of 60 to 70 decibels. If the system audio output thread is in running state, and it is determined that the auditory channel is occupied, the auditory load bit will be switched off. Set to 1; otherwise, set to 0. Set the ambient darkness threshold to [value missing]. When the illuminance value When the environment is deemed too dark to hinder display interaction, the visual load is adjusted. Set to 1; otherwise, set to 0.

[0046] In addition to the conventional mapping mechanism mentioned above, to prevent the system from issuing reminder information when the user performs high-risk actions, which could trigger a startled reaction or obstruct the view, the state quantification module 102 is equipped with a higher-priority hardware-level veto mechanism.

[0047] The state quantization module 102 extracts the motion period frequency from the inertial measurement unit. Z-axis acceleration variance And the zero-crossing rate of the X-axis gyroscope angular velocity corresponding to head pitch movements. This leads to the generation of multidimensional feature vectors. Simultaneously, the system configures a high-risk feature subspace corresponding to high-risk actions within the memory. This subspace is defined by the set of high-risk motion variance limit, zero-crossing rate threshold and frequency threshold. The microcontroller performs spatial matching operation on the feature vectors extracted in real time.

[0048] When the matching result is determined Upon establishment, the state quantization module 102 triggers a hardware-level interception operation; this operation forcibly skips the aforementioned independent environmental threshold comparison steps and directly writes instructions to the memory bus, overwriting the user resource occupancy status word to a fully blocked state, i.e., performing an assignment:

[0049] ;

[0050] The overwrite operation generates a sequence of all 1s, forcing all sensory channels to be in a high-impedance state during the subsequent mask routing verification process. This mechanism directly cuts off the subsequent display or audio sound control flow from the data processing front end, building a physical safety barrier and thus avoiding unsafe warning events in extreme motion scenarios.

[0051] See attached document Figure 5 The concurrent routing module 103 is configured as the logical hub for system resource scheduling. It receives the channel demand mask constructed by the mask configuration module 101 and the user resource occupancy status word output by the status quantization module 102. The concurrent routing module 103 relies on the arithmetic logic unit of the microcontroller to perform binary bit operations and convert the environmental state conflict judgment into the underlying register instruction to achieve low-latency channel availability verification.

[0052] Specifically, the concurrent routing module 103 first executes the detail channel availability check algorithm, and the system extracts the detail demand mask associated with the detail load data. and real-time generated user resource occupancy status words ;in, The bit with an internal value set to 1 represents the idle sensor channel necessary for outputting detailed data. Bits with internal values ​​set to 1 indicate that the corresponding channel is currently in a high-load occupancy state; the microcontroller unit executes the following verification and determination equation: ;

[0053] In the formula, This is the bitwise AND operator; This is an equality comparison operator. If the above equation holds true, it indicates that the channel set required for the detailed load data does not overlap with the channel set currently occupied by the environment, meaning the detailed output channel meets the pass-through condition; in response to this determination, the microcontroller extracts the template index parameters of the initialization configuration. The address of the corresponding base rendering template is read from the self-rendering template array.

[0054] The concurrent routing module 103 then loads the detailed load data into the basic rendering template and directly imports it into the underlying pipeline such as the display driver or audio driver to perform the complete output; after the output is completed, the microcontroller erases the data cache resources corresponding to the task.

[0055] Conversely, if the above verification and judgment equation does not hold, it indicates... The bit requirement and When a conflict occurs in the occupancy of a position, such as visual requirement position 1 but the ambient darkness causes visual load position 1; for this type of task blockage caused by physical channel limitation, the concurrent routing module 103 triggers a derivative calculation process of the core channel security degradation mask.

[0056] In this process, the concurrent routing module 103 reads the core requirement mask for low cognitive load triggering. The microcontroller unit acquires real-time data. The sequence is then subjected to a bitwise inversion operation to generate a one's complement sequence. The bits set to 1 in the inverse sequence explicitly represent the idle and available channels in the current physical environment; subsequently, the concurrent routing module 103 executes the following composite bitwise operation equation to generate a degradation mask. : ; In the formula, This is a bitwise NOT operator; at the system's underlying level, this operation performs an intersection of the core requirement channel and the real-time security channel, thereby separating a degraded security output path that does not conflict with the occupied environment.

[0057] Based on the generated downgrade mask The concurrent routing module 103 implements downgraded interception and asynchronous suspension scheduling of conflicting tasks; the microcontroller unit verifies... The value: if the judgment This indicates the existence of a valid degraded output path. The concurrent routing module 103 then strips the core index data and routes it to... The hardware interface driver corresponding to setting bit 1.

[0058] When the audio / video channel is blocked but the haptic position is calibrated as available, the system only drives the vibration motor to execute a weak pulse output to construct the expected event, without triggering the original graphics or voice parsing stream; if determined This indicates that all demand channels are currently under severe restriction. The system cuts off the drive level of the corresponding peripherals and forces complete silence to ensure safe interaction.

[0059] After completing the downgraded output of the core index data or implementing silent interception, the concurrent routing module 103 switches to the asynchronous task suspension execution flow; the concurrent routing module 103 retains the detailed load data carrying complete information, does not destroy it, and pushes it from the active processing memory into the statically accessed asynchronous background queue cache area; the system then updates the status bit of the detailed load data to asynchronous suspension, hands over its master control to the underlying operating system background, and waits for subsequent modules to update its rendering parameter pointer and re-schedule based on time evolution or event triggering conditions.

[0060] See attached document Figure 6 For the detailed load data transferred to the asynchronous background queue by the concurrent routing module 103, the addressing control module 104 starts the background state update and memory pointer maintenance pipeline; the addressing control module 104 configures and activates the spatiotemporal urgency recording unit in the microcontroller unit, which calls a hardware timer independent of the main execution thread and registers a hardware interrupt service associated with the step count output pin of the inertial measurement unit with the interrupt controller.

[0061] In response to the activation state of the recording unit, the system allocates memory address to initialize the time variable and step count variable. The recording unit uses the clock pulses of the hardware timer to perform discrete counting, accumulates the time elapsed since the detailed load data entered the suspended state, and generates the time variable. Synchronously, the recording unit responds to the pulse event triggered by the step counting hardware interrupt, and accumulates to generate a variable of valid physical steps. This counting mechanism enables data residing in the background queue to establish a spatiotemporal relationship with the user's physical environment evolution without waking up the main processor kernel.

[0062] While recording the aforementioned spatiotemporal parameters in the background, the addressing control module 104 periodically performs template index addressing offset calculations via timer interrupts. This calculation logic is configured to dynamically upgrade the alert intensity of suspended data through environmental offset factors, thereby offsetting the attenuation caused by user forgetting over time. The addressing control module 104 extracts the real-time updated... and Execute the built-in algebraic equations to compute the template index at the current time step. The calculation formula is configured as follows: ; In the formula, The time weighting constant, This is the step weight constant. This is a floor function used to force the result of a floating-point multiplication and addition operation to be converted into an integer index value that conforms to the system memory addressing specification; This is an extreme value limiting function used to output the smaller of the two parameters within the parentheses; This is the maximum index boundary value of the template preset in the system memory. This boundary value is hard-coded to correspond to the starting address of the rendering template with the highest display contrast or the highest audio gain.

[0063] In the above formula, the weighting constant and The calibration value is associated with the urgency attribute of a specific reminder task; in practice, This can be calculated based on the set maximum tolerable suspension period. For example, if the highest alert level is triggered when the cumulative suspension time of a certain task reaches 600 seconds, then... Calibrate as the corresponding scaling factor Accordingly, The value can be derived and mapped based on the average human step frequency, and a linear response mechanism between the cumulative step count and the index offset can be established.

[0064] During the update of the underlying data structure, the extreme value limiting function A memory addressing out-of-bounds protection barrier has been constructed; when a continuous suspended state causes and When the polynomial accumulation result exceeds the preset rendering template array boundary, the clipping function forcibly truncates the index calculation result to be written and clamps it at the upper limit value. This protection mechanism prevents array overflow from occurring when the microcontroller performs address offset calculations at the instruction level, thus preventing system-wide resets or wild pointer read / write anomalies caused by this.

[0065] Addressing control module 104 relies on the underlying scheduling cycle to repeatedly call the algebraic equation to calculate valid ones. The value is obtained and the obtained index value is copied to the corresponding pointer variable address unit in the memory. Through periodic copying operations, the addressing control module 104 continuously and dynamically updates the template read pointer bound to the detailed load data, thereby realizing the dynamic adaptive drift of its own rendering attribute strength during the asynchronous data residence period.

[0066] See attached document Figure 7 The multimodal verification module 105 is configured to monitor the physical environment status during system background suspension to determine the appropriate time to wake up the underlying processing pipeline. The multimodal verification module 105 implements low-power wake-up based on an event-driven architecture. At the hardware connection level, the microcontroller allocates an external interrupt pin and establishes a level detection mapping with the motion load bit output by the state quantization module 102.

[0067] The multimodal verification module 105 initiates an edge-triggered interrupt service routine to monitor the level transitions of the motion load bit in real time. When it detects a change from 1 to 0, it indicates that the system has detected a falling edge signal, signifying that the user may have ended high-displacement motion and transitioned to a stable posture. To avoid false triggering of the system due to false falling edges caused by brief pauses or sensor electrical noise, the multimodal verification module 105 does not immediately wake up the system after receiving a falling edge. Instead, it loads a preset stable anti-shake window time parameter to perform state filtering. This time parameter can be set in conjunction with the inertial buffer period for the human body to return to a stationary state, for example, a fixed value in the range of 2 to 5 seconds.

[0068] During the stable anti-jitter window, the multimodal verification module 105 maintains a polling scan of the motion load bit level. If the scan detects that the level has changed from 0 back to 1, indicating an inversion, the system determines that the previously captured falling edge is a jitter interference signal. In response to this determination, the multimodal verification module 105 clears the flag in the interrupt service routine and restores the initial monitoring state. If the motion load bit level remains at 0 for the entire time window, the multimodal verification module 105 determines that the physical environment has formed a real and safe steady state, and then sends an initial wake-up interrupt signal to the microcontroller's internal bus to trigger the execution verification of the environment's physical readiness.

[0069] In response to the initial wake-up interrupt signal, the multimodal verification module 105 verifies the completeness of the external execution conditions. When the system verification fails, such as when the target reminder item is not within the current field of view, the initial wake-up process terminates. In this state, if the user maintains a low-variance posture such as sitting still, the level of the motion load bit will be anchored to 0 for a long time, causing the interrupt pin to be unable to capture a new falling edge signal, resulting in a wake-up deadlock of the suspended task. In order to resolve this logical deadlock and obtain a retry opportunity while avoiding the high power consumption caused by the high frequency of the timer pulling up the image sensor, the multimodal verification module 105 starts the secondary wake-up mechanism driven by the field of view incremental event.

[0070] Specifically, the multimodal verification module 105 continuously reads the yaw rate and pitch rate within the discrete sampling period from the inertial measurement unit, denoted as follows: and The microcontroller performs discrete integration on the angular velocity data sequence to obtain two predetermined time points. and The change in angle between them. Yaw angle increment. With pitch angle increment The calculation formula is as follows: ; In the formula, This is the integral time differential component. Absolute value calculation is used to isolate the rotation direction, extracting only the simple rotational displacement amplitude of the head; after calculation, the multimodal verification module 105 logically compares the angle increment with a pre-stored angle threshold; the system sets the yaw rotation angle threshold to... The pitch angle threshold is This threshold is combined with the calibration of the foveal field of view coverage boundary of the human eye, and can be configured, for example, within the range of 30 to 45 degrees.

[0071] When the multimodal verification module 105 compares and obtains... or When the system is established, it determines that although the user is in a static state, their head has made a rotational movement that causes a substantial change in the optical field of view. The system uses this incremental over-limit event as a trigger condition that the user may have refocused on the target object, and then issues a secondary wake-up interrupt signal.

[0072] Regardless of whether an initial or secondary wake-up interrupt signal is received, the multimodal verification module 105 executes a multi-level cross-validation strategy based on increasing power consumption for environmental physical readiness. The system prioritizes powering peripheral radio frequency components and broadcasts probe frames to scan for wireless response tags attached to target objects. If no corresponding radio frequency response is detected within a specified scanning period, the multimodal verification module 105 performs downgraded verification, waking up the low-power image sensor to acquire environmental frame images. The system calls the built-in optical tag matching algorithm to parse and retrieve preset reference plane tag patterns in the acquired frames. For baseband handshake parsing of radio frequency signals and contour matching of optical plane features, those skilled in the art can use standard radio protocol stacks or conventional computer vision algorithm libraries to implement such data processing methods, which are well-known technologies in the field and will not be elaborated here.

[0073] When the multimodal verification module 105 successfully matches the target feature signal through either the radio frequency link or the optical image, it determines that the environmental physical readiness verification has passed. The multimodal verification module 105 then sets the environmental verification success flag in the register and transfers it to the downstream security rendering module. If the multi-level verification method still returns a failure result after traversing all the steps, the system actively cuts off the power supply pins of the radio frequency circuit and the image sensor to maintain the sleep state, and continuously monitors the inertial parameter integration in anticipation of the next field of view increment event.

[0074] See attached document Figure 8 After obtaining the environment verification success flag output by the multimodal verification module 105, the secure rendering module 106 takes over the underlying control bus to execute the output pipeline. Considering that in embedded systems, there is a processing cycle delay between the sensor confirming the steady state of the environment and the rendering data being written to the peripheral register via the bus, if the user action or the environmental state deteriorates suddenly during this period, directly executing the rendering output will cause an interaction conflict. Therefore, the secure rendering module 106 is configured with a critical section transient state mask verification mechanism before the write deadline. In the instruction cycle before the microcontroller prepares to import the rendering data frame into the display interface or audio digital-to-analog converter via the direct memory access controller, the secure rendering module 106 triggers a global interrupt masking operation. The microcontroller masks external conventional hardware interrupts through kernel instructions, forcing the system into a protected critical section. During this period, the processor no longer responds to other peripheral pending interrupt requests to ensure the absolute atomicity of the current verification instruction sequence.

[0075] Within the critical section, the secure rendering module 106 bypasses the cache and directly reads the instantaneous user resource occupancy status word re-collected from the underlying hardware registers. And extract the detailed requirement mask for the current task. The microcontroller then executes the following masking kernel equation: ; In the formula, For bitwise AND operator, Boolean variable The determination result is used to characterize the instantaneous channel availability state; if the determination is calculated... If the result is false, it indicates that during the data transmission preparation period, the user's resource occupancy status word has changed, and the environment status no longer meets the output requirements. In response to this determination, the security rendering module 106 immediately terminates the data transmission, clears the transmission cache corresponding to the direct memory access controller, and cancels the current rendering instruction. The system then removes the global interrupt mask, resets the task status, and returns it to the asynchronous background queue to continue suspending.

[0076] If the calculation determines If true, it indicates that the physical environment remains stable; the secure rendering module 106 releases the global interrupt mask, freeing up processor resources and entering the adaptive output process; the secure rendering module 106 retrieves the current template index parameters, periodically updated by the addressing control module 104, from memory. Based on this, in the preset rendering template array The template storage area is addressed in the template index parameter; The suspension time and physical step count variables have been integrated, and the rendering template they point to is precisely adapted to the user's current forgetting tendency compensation needs.

[0077] After addressing is completed, the secure rendering module 106 loads the detailed payload data in the cache into the data placeholder of the specific rendering template. The display driver and audio driver generate underlying timing signals based on the loaded template parameters and send them to the screen display hardware and speaker hardware to complete the information presentation. After the rendering pipeline is completed, the microcontroller erases the residual data cache of the task in the memory and ends the task processing cycle.

[0078] For tasks that remain suspended due to unmet environmental conditions, the security rendering module 106 initiates a timeout watchdog interception and network-level intervention mechanism. The security rendering module 106 maintains an internal timed thread to periodically compare the discrete-time variables of the suspended tasks. With the preset limit safety threshold This extreme safety threshold Based on conventional understanding, the maximum permissible delay time can be set between 15 and 60 minutes, depending on the urgency of the specific trip; when the system determines... When the watchdog timer is activated, it indicates that the user is at serious risk of forgetting their schedule. The watchdog timer then takes over control of the task. It cuts off the regular rendering pipeline that is directed to a high-brightness screen display or a peripheral that is subject to strong vibration, terminates the call to the regular template array, and reloads the detailed payload data into the preset low-alarm safe single-channel template for local output, thus establishing a degradation warning baseline on the device side.

[0079] At the same time, the watchdog program extracts task-related parameters and assembles them into an emergency data packet conforming to a preset wireless transmission format. Emergency data package The structure encapsulates a process task identifier that triggers a timeout and a transient resource occupancy status word. And a spatiotemporal stamp parameter containing the accumulated time and steps; the watchdog program wakes up the underlying communication module of the wearable hardware platform, controlling the RF baseband chip to directly call the wireless communication protocol at the link layer to send emergency data packets. Pushed to bound external mobile devices or home gateway terminals; this fail-safe protocol builds a complete fallback mechanism with the help of external network intervention in environments where the front-end wearable devices are restricted.

[0080] This embodiment uses the example of reminding an elderly user to take antihypertensive medication every morning to illustrate the processing procedure of the present invention. After the user puts on the smart glasses, the system determines the user's physical presence using an infrared proximity sensor, posture angle, and ambient light intensity. When the system detects that the device is in an effective wearing state, it receives a medication schedule task from the bound mobile terminal. The task is triggered at 8:00 AM, and the reminder content is "Take 2 antihypertensive pills and confirm after taking the medication." The system breaks down this reminder content into core index data and detailed load data. The core index data is configured as a short bone conduction prompt, and the detailed load data is configured as a complete voice broadcast text and in-field text prompts.

[0081] At 8:00 AM, the user was walking from the bedroom downstairs to the kitchen. The state quantization module detected that the motion characteristics output by the inertial measurement unit were in a high displacement state, and the head pitch was changing frequently. The system set the motion load position to a high load state. Since the current environment is not suitable for direct full text or voice prompts, the concurrent routing module does not output detailed load data. Instead, it determines whether there is an available low-stimulation prompt channel based on the core requirement mask. If the auditory channel is not occupied, only a short bone conduction prompt sound is output. If a high-risk action state is detected at the same time, the system remains silent and pushes the detailed load data into the background suspended queue.

[0082] After the user arrives at the kitchen and stops walking, the multimodal verification module captures the falling edge of the motion load bit transitioning from a high-load state to an idle state, and confirms that the user has stopped within a stable window. Subsequently, the system prioritizes scanning the BLE tag or NFC tag on the medicine box. If the scan is successful, it determines that the target item is within the interactive range. If the RF verification fails, it wakes up the low-power image sensor to identify the QR code or ArUco mark on the surface of the medicine box. After successful identification, the secure rendering module rereads the instantaneous user resource occupancy status word before the formal output. After confirming that there are no new movements or environmental conflicts, it loads the detailed load data into the rendering template corresponding to the current template index for output. Since the task has accumulated a certain amount of time and steps during the background suspension, the template selected by the system is significantly higher than the initial template, but still within the safety threshold.

[0083] If the user does not approach the medicine box for an extended period, or fails to complete physical verification after multiple wake-ups, the system enters the watchdog fault tolerance process. The watchdog program does not use highlighted flashing or strong vibration templates, but instead uses a low-disturbance single-channel template to broadcast the minimum necessary prompts, such as "There is an incomplete medication reminder; please check in a safe location." Simultaneously, the system sends an intervention data packet to the caregiver's terminal via the linked mobile phone or home gateway. This data packet contains the task identifier, suspension duration, step count changes, and current resource usage status, allowing the caregiver to determine whether further intervention is needed.

[0084] The comparison object is set to three reminder schemes: Option A is a fixed-time direct reminder option, which outputs the complete reminder content immediately after the trigger time is reached; Option B is a motion state delayed reminder scheme, which means that the reminder is temporarily suspended when the user's movement is detected, and the complete reminder content is output directly after the user comes to a standstill; Scheme C is the scheme of this embodiment, which adopts the processing flow of splitting core index data and detail load data, determining resource occupancy status words, background suspension, physical verification, and security degradation output.

[0085] The test included scenarios such as going up and down stairs, walking indoors, standing in the kitchen, sitting quietly in the living room, low-light environment, and noisy TV environment. Each scenario executed 120 reminder tasks, including 60 high-risk movement trigger tasks and 60 normal state trigger tasks. The test statistics included the number of complete reminder outputs, task completion rate, number of error recovery outputs, average detail output delay, and number of nursing intervention triggers under high-risk states. The average test data are shown in Table 1 below. Total number of test tasks / times 120 120 120 High-risk sports trigger tasks / times 60 60 60 Number of times a complete alert is output under high-risk conditions / 58 12 3 Number of successful notifications in normal status / times 55 56 57 Number of times details were successfully restored after suspension / time 0 41 52 Number of successful verifications for the target item / times 0 0 49 Error recovery output count / time 17 9 2 Watchdog care intervention trigger count / time 0 0 7 Number of times the final closed loop of all tasks was completed / 88 96 112 Final closed-loop completion rate of the task / % 73.3 80.0 93.3 Average detail output delay / s 4.8 96.5 84.2 Average power consumption per task / mWh 5.6 4.9 5.2 .

[0086] See attached document Figure 9 According to the data in Table 1, the fixed-time direct reminder scheme does not consider the user's movement status at the trigger time. Therefore, the number of complete reminder outputs is relatively high in high-risk movement scenarios, which can easily cause visual obstruction or auditory interference when the user goes up or down stairs, turns around, or walks quickly. The movement delay reminder scheme can reduce some high-risk outputs, but it outputs details as soon as the user stops moving and lacks the verification of the target item or environment readiness. Therefore, there is still a certain proportion of error recovery outputs.

[0087] This embodiment uses a resource occupancy status word to pre-determine the alert channel and transfers complete details data to a background suspend queue. This reduces the number of complete alert outputs in high-risk situations. Through a stable window, radio frequency or optical physical verification, and transient mask verification mechanism, the probability of error recovery output can be reduced, making the alert content occur more concentratedly when the user is in a safe position and the target item is interactive.

[0088] The average delay for detailed output in this embodiment is slightly higher than that of the fixed-time direct reminder solution. However, this delay is a controllable delay caused by proactive safety waiting. Because the system has a watchdog-like intervention mechanism, tasks that fail to meet output conditions for an extended period will not remain indefinitely. Test results show that this embodiment achieves a good balance between safety, task completion rate, and anomaly mitigation capabilities, making it suitable for scenarios such as medication administration, follow-up appointments, blood pressure measurements, and reminders for elderly people to carry items when going out.

[0089] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A smart travel reminder system for the elderly based on smart glasses, characterized in that, include: The state quantization module collects sensor characteristic parameters and constructs a user resource occupancy status word that includes visual, auditory, and motion load bits. The mask configuration module receives a trip task containing reminder data, splits the reminder data into core index data and detail load data, configures the required mask for each, and initializes the rendering template array. The concurrent routing module performs bitwise operations using the demand mask and the user resource occupancy status word, separates the available channels to send the core index data based on the operation result, and truncates and transfers the detailed load data to the asynchronous background queue. The addressing control module accumulates the suspension time and physical steps of the data in the asynchronous background queue as a spatiotemporal urgency parameter, calculates the offset index using the spatiotemporal urgency parameter, and continuously updates the read pointer of the rendering template array. The multimodal verification module listens for the falling edge transition of the motion load bit in the user resource occupancy status word or extracts the integral feature of the field of view change, thereby triggering wake-up and performing environment readiness verification. After the environment readiness verification is passed, the secure rendering module isolates external interrupts, completes transient mask verification, maps the updated read pointer, and executes adaptive prompts.

2. The intelligent travel reminder system for the elderly based on smart glasses according to claim 1, characterized in that, The mask configuration module is specifically configured as follows when processing the reminder data: The infrared proximity pin level, attitude Euler angle, and illuminance value are acquired, and cross-validation is performed using Boolean logic equations to determine that the device is in an effective wearing state and to initiate the corresponding data processing pipeline. For in-process tasks, the original reminder data stream is forcibly divided into core index data, which occupies less storage space, and detailed load data containing complete process details, based on the data frame header identifier; Configure core requirement masks and detail requirement masks based on binary sequences for the core index data and the detail load data respectively, and allocate addressing regions in memory to generate a rendering template array with security saliency ascending order.

3. The intelligent travel reminder system for the elderly based on smart glasses according to claim 1, characterized in that, The state quantization module is specifically configured as follows when constructing the user resource occupancy status word: Within the set data acquisition time window, the vertical axis acceleration variance, the root mean square value of the audio sequence, and the instantaneous illuminance value are extracted as underlying hardware feature parameters. The extracted underlying hardware feature parameters are compared with preset conventional motion variance thresholds, environmental noise thresholds, and environmental darkness thresholds, and then discretized and mapped as Boolean state variables. The Boolean state variables are concatenated and assembled into a user resource occupancy status word containing visual load bits, auditory load bits, and motion load bits, and output to the concurrent routing module to represent the current high load occupancy state or idle state of each sensory channel.

4. The intelligent travel reminder system for the elderly based on smart glasses according to claim 3, characterized in that, The state quantization module is also equipped with a hardware-level veto mechanism: The state quantization module synchronously extracts the motion cycle frequency, the vertical axis acceleration variance, and the zero-crossing rate of the angular velocity corresponding to the head pitch motion, and combines them to generate a multi-dimensional feature vector characterizing the user's dynamic behavior. The multidimensional feature vector is input into a high-risk feature subspace defined by the set of motion variance limit, zero-crossing rate threshold and frequency threshold, and a spatial matching operation is performed. When the matching determines that the multi-dimensional feature vector falls into the high-risk feature subspace, a hardware-level interception operation is triggered, forcibly skipping the independent environmental threshold comparison and directly writing an instruction to the memory bus to overwrite the user resource occupancy status word to a fully blocked state, thereby completely cutting off the insecure alert event from the data processing front end.

5. The intelligent travel reminder system for the elderly based on smart glasses according to claim 1, characterized in that, The concurrent routing module is specifically configured as follows when performing bitwise operations and truncation transitions: The detailed requirement mask is extracted and a bitwise AND operation is performed with the real-time generated user resource occupancy status word. If the operation result determines that the channel set does not overlap, the pass-through condition is met and the initialized read pointer is extracted to perform complete output. If the bitwise AND operation indicates a channel conflict, the core requirement mask is extracted and the user resource occupancy status word after bitwise inversion is used to perform a composite bitwise operation to generate a degraded security mask that removes the environmental occupancy conflict. Based on the indicator bits of the downgraded security mask, core index data is distributed through an independent channel, while the detailed load data is simultaneously pushed into the asynchronous background queue of static access and updated to an asynchronous suspended state, so as to hand over the addressing control module for subsequent scheduling.

6. The intelligent travel reminder system for the elderly based on smart glasses according to claim 1, characterized in that, The addressing control module is specifically configured as follows when calculating the offset index: For the detailed load data transferred to the asynchronous background queue, a hardware timer and a step counting hardware interrupt are used to accumulate and generate a suspension time variable and a physical effective step count variable as parameters of spatiotemporal urgency. Extract the suspension time variable and the physical effective step number variable, and perform multiplication and addition algebraic equation operations in combination with the set time weight constant and step weight constant to construct the addressing offset calculation logic that directly maps to the evolution of the physical environment; The extremum limiting function is used to perform out-of-bounds truncation protection on the result of the multiplication-addition algebraic equation after floating-point conversion, and a valid template index value is output to periodically overwrite the memory pointer variable address, thereby realizing the dynamic adaptive drift of its own rendering attribute strength during the asynchronous data residence period.

7. The intelligent travel reminder system for the elderly based on smart glasses according to claim 1, characterized in that, The multimodal verification module is specifically configured as follows when wake-up is triggered: Configure an edge-triggered interrupt service routine to continuously monitor the level of the motion load bit in the user resource occupancy status word. When a falling edge is captured where the level changes from high to low, load the preset stable anti-jitter window time parameter to perform status filtering. During the duration of the stable anti-shake window, the polling scan of the bit level of the moving load is maintained. If the level is reversed, the interrupt flag is cleared. If the level remains low for the entire time window, it is determined that the physical environment has established a true steady state. In response to the establishment of a true steady state in the physical environment, an initial wake-up interrupt signal is sent to the internal bus to trigger the power-on of the peripheral radio frequency or optical hardware and perform environmental readiness verification of the target item.

8. The intelligent travel reminder system for the elderly based on smart glasses according to claim 7, characterized in that, The multimodal verification module also has a secondary wake-up mechanism when the initial wake-up verification fails: When the system verification fails, causing the initial wake-up interrupt signal to terminate, and the user maintains a stationary posture so that the motion load bit remains in a low level state for a long time, the yaw rate and pitch rate within the continuous sampling period are extracted. Discrete integral operations are performed on the yaw rate and pitch rate data sequences to obtain the yaw rate increment and pitch rate increment. When it is determined that any angle increment exceeds the angle threshold calibrated by the human eye's foveal field of view, the field of view angle variation integral feature is generated. Based on the integral characteristics of the field of view variation, a secondary wake-up interrupt signal is issued. The peripheral radio frequency components are preferentially allocated to broadcast detection frames to scan electronic tags in order of increasing power consumption, or the low-power image sensor is degraded to wake up and retrieve the reference plane mark pattern. After successful matching, the environmental readiness verification pass flag is provided.

9. The intelligent travel reminder system for the elderly based on smart glasses according to claim 1, characterized in that, The secure rendering module is specifically configured as follows when performing transient mask verification: After obtaining the environment readiness verification pass flag, in the instruction cycle before importing the rendered data frame into the display interface or digital-to-analog converter via the direct memory access controller, external conventional hardware interrupts are masked by kernel instructions, forcing entry into the protected critical section. Within the critical region, bypassing the cache, the instantaneous user resource occupancy status word re-collected from the underlying hardware register is directly read, and the detailed demand mask is extracted and bitwise AND operation is performed again to verify the mask; If the calculation result indicates that the instantaneous channel availability is true, the interrupt mask is removed to release processor resources. According to the read pointer updated by the addressing control module, the detailed load data in the suspended state is loaded into the corresponding template data placeholder and converted into underlying timing signals for presentation.

10. The intelligent travel reminder system for the elderly based on smart glasses according to claim 1, characterized in that, The secure rendering module is also equipped with a watchdog interception pipeline: Maintain an internal timer thread to periodically compare the suspension time variable of tasks in the asynchronous background queue with the preset limit safety threshold. When it is determined that the suspension time variable exceeds the limit safety threshold, the watchdog program cuts off the regular rendering pipeline pointing to the high-brightness screen display or the strong vibration peripheral, and reloads the detailed load data into the preset low-disturbance safe single-channel template for local output. The process timeout identifier, the instantaneous user resource occupancy status word, and the time stamp parameters are extracted and assembled into an emergency data packet. The underlying communication module is controlled to call the wireless communication protocol at the link layer to push the emergency data packet to a preset terminal to build a failover protection.

Images