Space positioning method and system based on intelligent wearable device

Abstract

The embodiment of the invention provides a spatial positioning method and system based on intelligent wearable equipment. The method comprises the following steps: acquiring satellite positioning signal intensity, plane position data, visual image data and inertial measurement data; judging whether the intelligent wearable device is in an indoor environment currently based on the satellite positioning signal intensity; in response to the judgment result that the environment is the indoor environment, executing the following steps: determining plane position data based on the satellite positioning signal, determining visual positioning data based on the visual image data, and generating visual inertial data based on fusion of the visual image data and the inertial measurement data; based on the satellite positioning signal intensity, distributing fusion weights of the plane position data, the visual positioning data and the visual inertial data, and obtaining a positioning result through weighted fusion and optimization; constructing an indoor scene structured knowledge graph; retrieving associated scene information from the structured knowledge graph based on a positioning result, and generating personalized service contents according to the associated scene information and user preferences; and outputting the personalized service content.

Description

Technical Field

[0001] This application relates to the field of positioning technology, and in particular to a spatial positioning method and system based on smart wearable devices. Background Technology

[0002] With the widespread adoption of smart wearable devices and the continuous growth in demand for indoor services, high-precision indoor spatial positioning technology has become one of the core infrastructures in many fields such as intelligent transportation, smart tourism, industrial warehousing, medical monitoring, and emergency rescue. Currently, the market demands increasingly higher indoor positioning performance, requiring not only sub-meter level or even higher accuracy in complex, obstructed environments, but also simultaneous compliance with diverse indicators such as high real-time performance, strong stability, and low power consumption. Especially in indoor environments where satellite navigation signals are severely attenuated or completely unavailable, reliable high-precision indoor positioning technology has irreplaceable application value.

[0003] Against this backdrop, a single technological approach struggles to simultaneously achieve accuracy, robustness, and universality, leading to multi-source information positioning solutions gradually becoming the mainstream direction of technological development. Among these, positioning methods incorporating wearable devices such as smart glasses have expanded the application boundaries of the technology in areas such as personnel navigation, equipment management, operational assistance, and user experience enhancement. However, despite the emergence of various multi-source fusion positioning solutions in the industry, existing technologies still possess several significant shortcomings.

[0004] Specifically, most smart glasses currently on the market integrate visual positioning modules that can achieve coarse-grained positioning in outdoor or simple, well-defined indoor environments. However, in complex indoor environments, their positioning accuracy and stability significantly decrease, making it difficult to consistently achieve sub-meter level positioning. Furthermore, existing solutions cannot achieve deep fusion of data from various sensors, leading to system drift and jumps when signals are missing or sensors are disturbed. They also generally suffer from weak scene adaptability, long initialization times, and high computational resource consumption.

[0005] Therefore, there is an urgent need for a positioning method that can deeply integrate multi-source information and maintain sub-meter accuracy and high real-time performance even in complex indoor environments. Summary of the Invention

[0006] The purpose of this application is to overcome the problems of insufficient positioning accuracy, poor environmental adaptability, and insufficient fusion depth in the indoor spatial positioning technology of existing smart wearable devices, and to provide a spatial positioning method and system based on smart wearable devices. In a first aspect, embodiments of this application provide a spatial positioning method based on a smart wearable device, including: Collect satellite positioning signal strength, planar position data, visual image data, and inertial measurement data; Determine whether a smart wearable device is currently in an indoor environment based on the strength of satellite positioning signals; In response to the judgment result being an indoor environment, the following steps are performed: Planar position data is determined based on satellite positioning signals, visual positioning data is determined based on visual image data, and visual inertial data is generated by fusing visual image data with inertial measurement data. Based on the satellite positioning signal strength, fusion weights are assigned to planar position data, visual positioning data, and visual-inertial data, and the positioning result is obtained through weighted fusion and optimization. Construct a structured knowledge graph of indoor scenes; retrieve related scene information from the structured knowledge graph based on the positioning results; generate personalized service content based on the related scene information and user preferences; and output the personalized service content.

[0007] In some embodiments, a structured knowledge graph includes: Location nodes are used to characterize the positions of selected fixed spatial features within an interior space; Resource nodes, associated with location nodes, are used to characterize the location of resources within an indoor space. Resources include facilities and services.

[0008] In some embodiments, when the positioning function is implemented, the weight of the positioning node exceeds a first threshold.

[0009] In some embodiments, constructing a structured knowledge graph of an indoor scene further includes real-time updating of the structured knowledge graph; real-time updating includes: Based on at least one of planar position data, visual image data, and inertial measurement data, determine changes in the actual indoor scene; based on the changes, perform add, delete, or modify operations on the positioning nodes, resource nodes, or the relationships between the two in the structured knowledge graph to maintain consistency between the structured knowledge graph and the actual indoor scene.

[0010] In some embodiments, assigning fusion weights includes: When the satellite positioning signal strength is greater than or equal to the second threshold, weights are allocated according to the first ratio, and planar position data, visual positioning data and visual-inertial data are fused. When the satellite positioning signal strength is less than the second threshold, the planar position data is shielded, and the weights are allocated according to the second ratio to fuse the visual positioning data and visual-inertial data.

[0011] In some embodiments, the step of obtaining the positioning result through weighted fusion and optimization specifically includes: Based on the assigned fusion weights, the least squares optimization algorithm is used to fuse planar position data, visual positioning data, and visual-inertial data to obtain preliminary positioning results. A sliding window optimization algorithm is used to smooth the continuous preliminary positioning results to obtain the final positioning result.

[0012] In some embodiments, before assigning fusion weights to the planar position data, visual positioning data, and visual-inertial data, the method further includes optimizing each data separately, specifically including: The Kalman filter algorithm is used to filter and optimize the planar position data. Based on ambient lighting conditions, the feature point extraction strategy for visual image data is adjusted to optimize visual positioning data. Visual inertial data is generated based on visual image data and inertial measurement data using the visual inertial odometry calculation method.

[0013] In some embodiments, personalized service content is output and presented in at least one of the following ways: The message is output via voice; or... Visual overlay display output is achieved through augmented reality.

[0014] Secondly, embodiments of this application provide a spatial positioning system based on a smart wearable device, used to implement any of the aforementioned spatial positioning methods based on a smart wearable device, including: The data acquisition module, integrated into the smart wearable device, includes a satellite positioning unit, a visual acquisition unit, and an inertial measurement unit; the visual acquisition unit includes a high-definition camera and an infrared camera; it is used to acquire satellite positioning signal strength, planar position data, visual image data, and inertial measurement data; The fusion positioning module is used to determine whether the smart wearable device is currently in an indoor environment based on the satellite positioning signal strength; in response to the determination result being an indoor environment, the following steps are executed: Based on satellite positioning signal strength, fusion weights are assigned to planar position data, visual positioning data generated from visual image data, and visual inertial data generated from the fusion of visual image data and inertial measurement data. The positioning result is obtained through weighted fusion and optimization. The scene association service module is used to construct a structured knowledge graph of indoor scenes; retrieve related scene information from the structured knowledge graph based on the positioning results; generate personalized service content based on the related scene information and user preferences; and output the personalized service content. The edge-side optimization module includes: a transmission unit for data transmission via Bluetooth and Wi-Fi; a power management unit for controlling the power supply status of each functional module in the smart wearable device through a dynamic sleep mechanism based on task load; and a thermal management unit for controlling the operating temperature of the smart wearable device by optimizing hardware layout and software scheduling strategies. The control unit is used to preprocess planar position data, visual image data, and inertial measurement data; and to control the operation of the data acquisition module, fusion positioning module, scene association service module, and edge optimization module. The power management unit provides power to the data acquisition module, the fusion positioning module, the scene association service module, the transmission unit, the thermal management unit, and the control unit.

[0015] Thirdly, embodiments of this application provide smart glasses that integrate a spatial positioning system based on a smart wearable device, as described above.

[0016] This application achieves sub-meter level indoor positioning accuracy by synergistically optimizing the fusion of planar position data, visual image data, and inertial measurement data. It also effectively reduces positioning interruption rates in typical scenarios, ensuring the continuity and stability of positioning output. Furthermore, it adapts and optimizes for the power consumption and heat dissipation limitations of smart wearable devices, improving battery life while maintaining high-performance positioning and controlling the operating temperature of core components to ensure reliability over long periods. The application provides efficient scene service response, enabling rapid integration from positioning to information services and significantly enhancing user experience. By relying on the device's own sensors for positioning, it eliminates the need for additional dedicated infrastructure deployment in indoor environments, thereby significantly reducing system deployment costs and subsequent maintenance complexity. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating a spatial positioning method based on a smart wearable device, according to an embodiment of this application. Detailed Implementation The following specific embodiments illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. Although the description of this application will be presented in conjunction with some embodiments, this does not mean that the features of this application are limited to this embodiment. On the contrary, the purpose of describing the application in conjunction with embodiments is to cover other options or modifications that may be derived based on the claims of this application. To provide a thorough understanding of this application, many specific details will be included in the following description. This application may also be implemented without using these details. Furthermore, to avoid confusion or obscuring the focus of this application, some specific details will be omitted in the description. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of this application can be combined with each other.

[0018] It should be noted that in this specification, similar reference numerals and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0019] In the description of this application, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0020] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0021] In the description of this application, it should be understood that "electrical connection" can be understood as physical contact and electrical conduction between components; it can also be understood as the form of connection between different components in a circuit structure through physical lines that can transmit electrical signals, such as copper foil or wires on a printed circuit board (PCB). "Coupled through..." can be understood as electrical conduction through indirect coupling. Indirect coupling can be understood as contactless coupling. Those skilled in the art will understand that coupling refers to the phenomenon where there is a close cooperation and mutual influence between the inputs and outputs of two or more circuit elements or electrical networks, and energy is transferred from one side to the other through interaction. To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be further described in detail below with reference to the accompanying drawings.

[0022] This invention provides a spatial positioning method based on a smart wearable device. Figure 1 This is a flowchart illustrating the spatial positioning method based on a smart wearable device provided in an embodiment of the present invention, as shown below. Figure 1 As shown, the method includes the following steps: Step S1: Collect satellite positioning signal strength, planar position data, visual image data, and inertial measurement data.

[0023] Step S2: Determine whether the smart wearable device is currently in an indoor environment based on the satellite positioning signal strength; in response to the determination result being an indoor environment, perform the following steps: Step S3: Determine planar position data based on satellite positioning signals, determine visual positioning data based on visual image data, and generate visual-inertial data by fusing visual image data and inertial measurement data; assign fusion weights to planar position data, visual positioning data, and visual-inertial data based on satellite positioning signal strength, and obtain positioning results through weighted fusion and optimization.

[0024] Step S4: Construct a structured knowledge graph of indoor scenes; retrieve related scene information from the structured knowledge graph based on the positioning results; generate personalized service content based on the related scene information and user preferences; output the personalized service content.

[0025] It should be noted that planar position data refers to positioning data provided by a Global Positioning System (GPS) receiver; visual image data refers to raw image data captured by a camera; inertial measurement data refers to angular velocity and acceleration data collected by an Inertial Measurement Unit (IMU); visual positioning data refers to absolute position and attitude data obtained through a Visual Positioning Service (VPS); and visual-inertial data refers to pose and map data optimized by a SLAM (Simultaneous Localization and Mapping) system after fusing visual image data and inertial measurement data.

[0026] In step S1, the smart wearable device collects satellite positioning signal strength and planar position data, visual image data, and inertial measurement data. The visual image data is used for visual positioning services and real-time positioning and mapping. The smart wearable device can be smart glasses, a smartwatch, a smart helmet, or other wearable devices. When moving, the smart wearable device acquires location-related data from multiple sources in real time to cope with complex environments and situations where satellite signals are weak or absent.

[0027] In one implementation, the smart wearable device first supports both the Global Positioning System (GPS) and the BeiDou Navigation Satellite System simultaneously via a built-in dual-mode satellite positioning module, collecting current satellite positioning signal strength and planar position data. This signal strength is measured by the signal-to-noise ratio (SNR) or signal quality index of the received signal, reflecting whether the current environment is suitable for satellite positioning. For example, signal strength may be higher indoors near a window, while it may significantly decrease deeper inside a building.

[0028] Furthermore, the smart wearable device collects visual image data through a visual acquisition unit. This unit includes a high-resolution camera and an infrared auxiliary camera. The infrared auxiliary camera can supplement the lack of visible light images in low-light environments, ensuring that clear image sequences can still be obtained in low-light scenarios such as underground parking lots, corridors, or at night. This visual image data is used for two purposes: firstly, for visual positioning services, which estimate location by matching image features with reference images in an existing map database; and secondly, for real-time positioning and map building, which builds local maps and synchronizes positioning in unknown or uncovered areas in real time.

[0029] Meanwhile, the smart wearable device collects inertial measurement data through an inertial measurement unit (IMU). This IMU integrates an accelerometer, gyroscope, and magnetometer, and can continuously record the device's linear acceleration, angular velocity, and directional changes. This data has high accuracy over a short period of time, making it suitable for compensating for positioning drift problems caused by satellite signal interruptions or loss of visual features.

[0030] It should be noted that after collecting the aforementioned multi-source data, the smart wearable device also preprocesses the collected data using a low-power computing chip. This computing chip employs a processor specifically designed for wearable devices, enabling efficient operation with limited power. The preprocessing process includes gross error removal from satellite planar position data, distortion correction and grayscale conversion of visual images, and zero-bias calibration and noise filtering of inertial measurement data, thereby providing clean and reliable input data for subsequent fusion. This preprocessing method can be completed on the device itself, avoiding the latency and increased power consumption caused by transmitting raw data to the cloud.

[0031] In step S2, the strength of the satellite positioning signal is monitored to determine whether the smart wearable device is currently in an indoor environment. Specifically, in open outdoor environments, the satellite positioning signal is stable and of high strength, sufficient for standard positioning needs; however, indoors, the signal strength decreases significantly due to signal obstruction and attenuation caused by building structures, typically increasing the positioning error to 5%. The distance is 10 meters, and signal drift is prone to occur, making it difficult to distinguish adjacent exhibits or achieve precise indoor positioning.

[0032] If the satellite positioning signal strength remains below a preset indoor judgment threshold, the device is determined to be in an indoor environment, triggering a subsequent indoor positioning process based on multi-source data fusion. This judgment mechanism provides a reliable basis for the system to adaptively switch positioning strategies between indoor and outdoor scenarios.

[0033] In step S3, the smart wearable device adaptively allocates weights to planar position data, visual positioning data, and visual-inertial data based on satellite positioning signal strength. The resulting positioning data is obtained through weighted fusion and optimization of the multi-source data. Specifically, the smart wearable device monitors satellite positioning signal strength in real time and uses this as the primary basis for environmental assessment, adaptively adjusting the weights of the three positioning sources. This adaptive allocation effectively addresses dynamically changing scenarios such as indoor / outdoor transitions and signal obstruction, avoiding the accumulation of positioning errors caused by a single dominant source.

[0034] In one implementation, the smart wearable device acquires the current satellite positioning signal strength. This strength is quantified by the number of received satellites, signal quality factor, or overall signal-to-noise ratio. Next, if the satellite positioning signal strength is greater than or equal to a second threshold, indicating good signal quality, a weight is allocated according to a first proportion. In this case, satellite positioning occupies a certain weight to ensure positioning accuracy indoors or near windows.

[0035] In one implementation, if the satellite positioning signal strength is less than a second threshold, for example, when the signal is almost unusable inside a building, satellite positioning is disabled and the visual positioning data and visual-inertial data are weighted according to a second ratio. In this case, the system relies entirely on the visual and inertial data from the device.

[0036] In one implementation, the second threshold is -100dBm.

[0037] In one implementation, when the satellite positioning signal strength is greater than or equal to -100dBm, the first ratio of fusion weights for planar position data, visual-inertial data, and visual positioning data is allocated as 20%, 35%, and 45%, respectively; when the satellite positioning signal strength is less than -100dBm, planar position data is masked, and the second ratio of fusion weights for visual-inertial data and visual positioning data is allocated as 40% and 60%, respectively.

[0038] In this embodiment, the smart wearable device obtains a preliminary positioning result by fusing weighted least squares multi-source data. Specifically, it first acquires weighted planar position data, visual positioning data, and visual-inertial data. Then, it performs least squares calculations on the weighted multi-source data to solve for the optimal position estimate, minimizing the sum of squared residuals. This fusion method comprehensively considers the reliability of each source to obtain the initial positioning result. A sliding window optimization is then applied to the initial positioning result. Specifically, a fixed-length historical position sequence is maintained using a sliding window optimization algorithm, and the position trajectory within the window is smoothed. This optimization effectively suppresses short-term fluctuations and improves positional stability in the user experience.

[0039] In this embodiment, before assigning fusion weights to the planar position data, visual positioning data, and visual-inertial data, each data point is optimized separately. Specifically, this includes: performing Kalman filtering on the planar position data to eliminate noise caused by multipath effects; adjusting the feature point extraction strategy of the visual image data based on ambient lighting conditions to optimize the visual positioning data; and generating visual-inertial data using visual-inertial odometry based on the visual image data and inertial measurement data. After these single-source optimizations, adaptive weight allocation and weighted fusion are performed, and finally, a sliding window optimization algorithm is used to smooth the fused positioning result. This progressive optimization approach can maintain sub-meter positioning accuracy in complex indoor environments.

[0040] In step S4, the smart wearable device retrieves related scene information from the structured knowledge graph based on the positioning results through a lightweight retrieval enhancement generative model, and generates personalized content based on user preferences and outputs it.

[0041] Specifically, after obtaining stable positioning results, the smart wearable device uses these results as the basis for querying, directly running a lightweight retrieval enhancement generative model on the device side to retrieve relevant scene information from a pre-stored structured knowledge graph. This device-side processing method avoids network dependence and ensures real-time response in indoor environments with no or weak network coverage.

[0042] In one implementation, a structured knowledge graph is pre-built and stored locally on the device. The knowledge graph includes location nodes, which represent the location of selected fixed spatial features within the interior space; and resource nodes, which are associated with the location nodes, and represent the location of resources within the interior space, including facilities and services.

[0043] Location nodes correspond to specific indoor locations, such as floors, load-bearing columns, and fixed walls; facilities include public amenities such as elevators, escalators, restrooms, and charging stations; services encompass promotional activities, navigation guidance, and emergency information. Nodes are linked by directed edges, for example, a spatial node may be connected to a nearby facility node or associated with a currently active service node. This graph structure facilitates efficient retrieval.

[0044] In one implementation, when the positioning function is required, the weight of the positioning node exceeds a first threshold. At this time, the weight of the positioning node is higher and the weight of the resource node is lower, resulting in more accurate positioning.

[0045] In one implementation, constructing a structured knowledge graph of an indoor scene further includes real-time updating of the structured knowledge graph; real-time updating includes: judging changes in the actual indoor scene based on at least one of planar position data, visual image data, and inertial measurement data; and based on the changes, performing add, delete, or modify operations on the positioning nodes, resource nodes, or the relationships between the two in the structured knowledge graph to maintain consistency between the structured knowledge graph and the actual indoor scene.

[0046] The "Add" operation involves creating a new resource node in the map when a new facility or service is detected within the area represented by a location node (e.g., "Rest Corner A"). The "Delete" operation involves deleting the corresponding resource node from the map and removing all its associated relationships when the system continuously detects the long-term disappearance of an existing resource through visual data, confirms this with maintenance personnel, or reaches a preset failure duration threshold. The "Modify" operation encompasses attribute modification and relationship reconstruction. Attribute modification involves updating the resource node's status attribute to "Under Maintenance" when a change in resource status is detected. Relationship reconstruction involves modifying the "Location" relationship between the resource node and the location node when it is confirmed that a resource has been permanently moved to another location, making it point to the new location node.

[0047] Furthermore, a lightweight retrieval-enhanced generative model is deployed on the edge of the smart wearable device. This model undergoes distillation and quantization, significantly reducing the number of parameters and making it suitable for the computing resources of wearable devices. The model first generates a location embedding vector based on the positioning result, then retrieves the nearest spatial node in the knowledge graph, expanding along the relationships to obtain information about surrounding facilities and services. For example, when the positioning result points to the atrium on the third floor of a shopping mall, the model retrieves that spatial node and further obtains related scene information such as nearby shops, restaurants, and events of the day.

[0048] Subsequently, smart wearable devices process relevant scene information based on user preferences to generate personalized content. User preferences can be pre-learned through device settings or historical behavior, such as preferences for food, shopping discounts, or accessible routes. The model filters, sorts, and reorganizes the retrieved information based on these preferences to generate natural language descriptions or structured guidance. For example, it prioritizes recommending nearby coffee shop promotions for coffee lovers and highlights accessible elevator locations for users with mobility impairments.

[0049] Finally, smart wearable devices output personalized content through voice prompts or augmented reality displays. Voice prompts are suitable for situations where both hands are occupied or for quick guidance, while augmented reality displays overlay arrows, labels, or virtual paths onto the smart glasses' field of view to enhance the immersive experience. This output method can be flexibly switched based on device type and user settings.

[0050] For example, when a user is standing in a museum exhibition hall, the location result is accurate to a certain exhibit. The lightweight model retrieves the exhibit's introduction, historical background, and related exhibition area information from the knowledge graph. Combined with the user's preference for art history, a concise audio explanation is generated. At the same time, the exhibit details and the next recommended path are highlighted in the augmented reality field of view.

[0051] This application embodiment also includes optimizations for smart wearable devices, specifically including: using Bluetooth and wireless LAN to transmit data, dynamic sleep mode to control power consumption, and hardware and software collaborative heat dissipation to optimize operating status.

[0052] Specifically, after completing location tracking and content generation, smart wearable devices need to maintain stable operation through efficient data transmission, power management, and heat dissipation control. This step is crucial for ensuring stable performance during prolonged use, directly impacting the continuity of user experience and the device's lifespan.

[0053] In one implementation, the smart wearable device integrates a dual-mode wireless transmission unit combining Bluetooth and Wi-Fi, enabling it to automatically select the transmission mode based on data type, transmission distance, and network environment. Bluetooth employs Bluetooth Low Energy technology, primarily used for connectivity with short-range devices such as smartphones and tablets, transmitting control commands, status information, and lightweight data. This module exhibits extremely low power consumption in standby mode and boasts rapid connection establishment, making it suitable for frequent short data packet transmissions. For instance, when a user adjusts the wearable device's positioning accuracy settings via a smartphone, control commands are rapidly transmitted via Bluetooth, with response latency typically within tens of milliseconds.

[0054] Furthermore, the wireless LAN supports dual-band operation, switching between the 2.4 GHz and 5 GHz bands for transmitting large amounts of data such as visual images, map updates, and knowledge graph synchronization. This module features adaptive channel selection, capable of detecting current channel occupancy and automatically selecting the channel with the least interference for transmission. In complex indoor electromagnetic environments, this adaptive mechanism effectively avoids conflicts with other wireless devices, maintaining a stable data transmission rate.

[0055] It should be noted that the transmission unit also features intelligent switching, dynamically selecting the transmission path based on data priority and network conditions. Emergency location data and safety-related information are preferentially transmitted through channels with better signal quality, while background data synchronization can be delayed until network conditions improve. This hierarchical transmission strategy ensures the real-time nature of critical data while avoiding the impact of network congestion on overall performance.

[0056] In this embodiment, the smart wearable device implements dynamic sleep mode based on the module's operating status through a power management unit.

[0057] Specifically, the power management unit is the core component for power consumption control in smart wearable devices. It can monitor the working status of each functional module in real time and dynamically adjust the power supply strategy according to usage requirements.

[0058] In one implementation, the power management unit maintains a module status table, recording the current operating mode, expected usage time, and power consumption level of each functional module. When a module has no data interaction or processing tasks within a preset time, the system automatically switches it to sleep mode. For example, when a user remains stationary indoors for a certain period of time, the sampling frequency of the inertial measurement unit automatically decreases, the satellite positioning unit enters intermittent operating mode, and the visual acquisition unit suspends image acquisition, maintaining only minimal environmental monitoring functions.

[0059] It's worth noting that the sleep control also features a fast wake-up mechanism. When user activity is detected or an external command is received, the relevant modules can resume normal operation within milliseconds. This rapid response capability ensures a smooth user experience and avoids functional delays caused by energy-saving measures.

[0060] For example, when a user is wearing smart glasses and working in an office building, sitting at their desk for an extended period, the system detects that their position is stable and there is no need for movement. It automatically switches the positioning-related modules to a low-power mode, maintaining only basic environmental awareness functions. When the user gets up to go to a meeting room, the accelerometer detects the change in motion and immediately activates the data acquisition module, preparing for the upcoming navigation service.

[0061] In this embodiment, the smart wearable device optimizes the hardware layout of the passive heat dissipation components and the temperature monitoring software scheduling strategy through a thermal management unit, thereby controlling heat dissipation.

[0062] Specifically, smart wearable devices generate heat during high-intensity computing and data transmission, especially when running lightweight retrieval enhancement generative models and multi-source data fusion, where the heat generation of various modules such as the control unit increases significantly. Effective heat dissipation control not only protects hardware components from thermal damage but also maintains stable device performance output, avoiding frequency reduction or functional limitations caused by overheating.

[0063] In one implementation, a passive heat dissipation component made of aluminum alloy is incorporated into the smart wearable device. Aluminum alloy has extremely high thermal conductivity, enabling it to evenly distribute heat across the device's outer surface, dissipating heat through natural convection and radiation. This passive heat dissipation method requires no additional electrical power and does not increase the device's power consumption.

[0064] In one implementation, the passive heat dissipation component is mounted in close contact with the control unit.

[0065] It's important to note that heat dissipation control is also closely related to user comfort. The temperature monitoring software not only monitors the internal temperature of the device but also the surface temperature of the areas in contact with the skin, ensuring that it does not cause discomfort to the user under any operating conditions. When the surface temperature exceeds the comfortable range, the system prioritizes reducing the workload of components near that area. If necessary, it activates a heat redistribution mechanism to redirect heat generated by the heat source away from the skin.

[0066] In one implementation, thermal control is further optimized using environmental sensing. The device acquires ambient environmental information through temperature and humidity sensors and dynamically adjusts its thermal strategy. In low-temperature environments, maintaining a certain device temperature helps battery performance and display quality, while in high-temperature and high-humidity environments, more aggressive thermal management is required. This environmentally adaptive thermal management maintains ideal device performance and user experience across various usage scenarios.

[0067] Based on the above embodiments, the optimized operation of smart wearable devices achieves a balance between device performance and energy consumption through dual-mode wireless transmission, dynamic sleep mode, and coordinated hardware and software heat dissipation. Dual-mode transmission ensures the reliability and efficiency of data communication, dynamic sleep mode extends the device's usage time, and heat dissipation control ensures long-term stable operation. This comprehensive optimization scheme is particularly suitable for indoor positioning applications that require long-term continuous operation, maintaining high-performance positioning while improving device battery life, and controlling the operating temperature of core components to ensure reliability for long-term use.

[0068] In one implementation, the three subunits can also coordinate their operations. For example, when the thermal management unit detects a high device temperature, it can notify the power management unit to appropriately reduce the workload of high-power modules, while simultaneously guiding the transmission unit to select a lower-power transmission method. This coordination maximizes the overall efficiency and stability of the device while ensuring core functionality.

[0069] This multi-source fusion spatial positioning method based on smart wearable devices can adapt to various complex indoor environments, providing accurate and reliable positioning services from small office buildings to large complexes, and from underground parking lots to high-rise buildings. By integrating planar position data, visual image data, and inertial measurement data for collaborative optimization, it achieves sub-meter level positioning accuracy indoors. Simultaneously, it effectively reduces positioning interruption rates in typical scenarios, ensuring the continuity and stability of positioning output. It is adapted and optimized to address the power consumption and heat dissipation limitations of smart wearable devices, maintaining high-performance positioning while improving device battery life and controlling the operating temperature of core components to ensure reliability over long periods. The method offers efficient scene service response, enabling rapid integration from positioning to information services, significantly improving user experience. Relying on the device's own sensors for positioning functionality eliminates the need for additional dedicated infrastructure deployment in indoor environments, thereby greatly reducing system deployment costs and subsequent maintenance complexity.

[0070] Secondly, embodiments of this application provide a spatial positioning system based on a smart wearable device, used to implement any of the aforementioned spatial positioning methods based on a smart wearable device, including: The data acquisition module, integrated into the smart wearable device, includes a satellite positioning unit, a visual acquisition unit, and an inertial measurement unit; the visual acquisition unit includes a high-definition camera and an infrared camera; it is used to acquire satellite positioning signal strength, planar position data, visual image data, and inertial measurement data; The fusion positioning module is used to determine whether the smart wearable device is currently in an indoor environment based on the satellite positioning signal strength. In response to the determination result being an indoor environment, the following steps are performed: based on the satellite positioning signal strength, fusion weights are assigned to the planar position data, the visual positioning data generated based on visual image data, and the visual inertial data generated by fusing visual image data and inertial measurement data, and the positioning result is obtained through weighted fusion and optimization. The scene association service module is used to construct a structured knowledge graph of indoor scenes; retrieve related scene information from the structured knowledge graph based on the positioning results; generate personalized service content based on the related scene information and user preferences; and output the personalized service content. The edge-side optimization module includes: a transmission unit for data transmission via Bluetooth and Wi-Fi; a power management unit for controlling the power supply status of each functional module in the smart wearable device through a dynamic sleep mechanism based on task load; and a thermal management unit for controlling the operating temperature of the smart wearable device by optimizing hardware layout and software scheduling strategies. The control unit is used to preprocess planar position data, visual image data, and inertial measurement data; and to control the operation of the data acquisition module, fusion positioning module, scene association service module, and edge optimization module. The power management unit provides power to the data acquisition module, the fusion positioning module, the scene association service module, the transmission unit, the thermal management unit, and the control unit.

[0071] In one implementation, the satellite positioning unit is connected to the control unit via an SPI interface, the visual acquisition unit is connected to the control unit via a MIPI interface, the inertial measurement unit is connected to the control unit via an I2C interface, the transmission unit is connected to the control unit via a PCIe interface, the power management unit supplies power to the control unit, and the control unit is connected to each data acquisition module, the fusion positioning module, the scene association service module, the transmission unit and the thermal management unit, and realizes data interaction through the transmission unit.

[0072] In one implementation, the control unit integrates a lightweighted Retrieval Augmentation (RAG) model for performing scene association services based on location results. This model reduces the number of parameters from approximately 1 billion to approximately 200 million through model pruning and quantization compression techniques, enabling efficient operation in resource-constrained embedded environments. Its workflow includes: receiving location results from the fusion positioning module; retrieving associated exhibit, facility, or scene information from a pre-built structured knowledge graph; and then generating corresponding personalized voice explanations or augmented reality display content based on user preference context. This effectively reduces system response latency and cloud dependency.

[0073] Thirdly, embodiments of this application provide smart glasses that integrate a spatial positioning system based on a smart wearable device, as described above.

[0074] Example 1 This embodiment uses a medium-sized smart museum with a building area of ​​approximately 8,000 square meters as the application scenario. The museum contains 3 themed exhibition halls and a total of 50 core exhibits. It needs to provide visitors with sub-meter real-time positioning and personalized tour guide services. The implementation process of this application is described in detail below: First, let's discuss hardware selection and connectivity: 1. Smart Glasses Main Body: Customized lightweight smart glasses, equipped with Android 12 wearable device system, supporting AR display (display resolution 1920×1080) and voice interaction; 2. Satellite positioning unit: Ublox NEO-7M module is selected, which supports GPS L1 band and Beidou B1 band, positioning accuracy ±2.5m (outdoor), signal strength detection range -159dBm~-40dBm, and is connected to the control unit through SPI interface; 3. Visual acquisition unit: The high-definition camera uses an OV5640 module (30fps, 1080P resolution), and the infrared auxiliary camera uses an IR-CUT module (supports acquisition in low light environments of 0-10 lux). Both are connected to the control unit via a MIPI interface. 4. Inertial Measurement Unit: The MPU6050 module is selected to collect acceleration (range ±2g) and angular velocity (range ±250° / s), and is connected to the control unit via an I2C interface; 5. Control Unit: The RK3399Pro chip (dual-core Cortex-A72 + quad-core Cortex-A53, integrated NPU, computing power 3.0 TOPS) based on the ARM architecture is selected as the data processing core, responsible for front-end data preprocessing and lightweight algorithm operation; 6. Transmission Unit: Bluetooth uses BK3288 (transmission rate ≥2Mbps), and wireless LAN uses AX200 (supports 2.4G / 5G dual-band), both connected to the control unit via PCIe interface to achieve bidirectional data transmission; 7. Power Management Unit: The LP3943 power management chip is used, which integrates a dynamic sleep unit and supplies power to each module through the power interface; 8. Thermal Management Unit: It adopts a collaborative design of passive heat dissipation components (aluminum alloy material) and temperature monitoring software. The passive heat dissipation components are installed in close contact with the control unit.

[0075] Electrical connection logic: The power management unit supplies power to the control unit, which is connected to various data acquisition modules, fusion positioning modules, scene association service modules, transmission units, and thermal management units.

[0076] Secondly, the specific implementation steps are mainly divided into three stages: The preliminary preparation stage includes the following steps: The system collects multi-view, multi-scale image sequences covering the entire museum, generates a dense 3D point cloud map using Structure from Motion (SfM) technology, and extracts ORB (Oriented Fast and Rotated BRIEF) feature points to construct a visual feature database, which is stored on the backend server (deploying structured knowledge graphs and recomputation tasks).

[0077] By integrating the layout of exhibition halls, the attributes of exhibits, the location of service facilities and their semantic information within the museum, a structured knowledge graph is constructed with location nodes and resource nodes as the core, and includes multiple types of relationships between nodes. At the same time, an efficient graph query interface is configured.

[0078] The Retrieval Augmentation (RAG) model was pruned and quantized to reduce its parameter count from approximately 1 billion to approximately 200 million, enabling lightweight deployment. The model retrieval module is deployed on the smart glasses, while the generation module is deployed on the backend server. Pre-configured voice wake-up words and corresponding command recognition models support real-time interaction.

[0079] The location service operation phase includes the following steps: The navigation service application in the smart glasses is launched, the control unit initializes the working parameters of each sensor module and transmission unit, and establishes a stable connection with the backend server with high bandwidth and low latency through the wireless local area network.

[0080] The satellite positioning unit collects planar position data and signal strength in real time; the visual acquisition unit collects environmental images through a high-definition camera and an infrared auxiliary camera, the latter of which is automatically activated in low-light conditions to ensure image quality; the inertial measurement unit continuously collects the device's acceleration and angular velocity data to track its motion attitude.

[0081] Kalman filtering is used to denoise and smooth the planar position data; an improved ORB-SLAM2 visual inertial odometry framework is adopted to fuse visual image data and inertial measurement data, construct a local map and estimate the device pose, effectively suppressing accumulated errors; illumination adaptive compensation is performed on the images used for visual positioning services to improve the stability and matching accuracy of feature point extraction and obtain visual positioning data.

[0082] The fusion weights of each data point are dynamically allocated based on the real-time satellite signal strength. The weighted least squares method is used for data fusion, and the continuous preliminary positioning results are smoothed and optimized through a sliding window optimization algorithm. The final output positioning result has an accuracy better than 0.5 meters.

[0083] The optimized positioning results are transmitted in real time to the AR display module of the smart glasses via Bluetooth for local augmented reality overlay; at the same time, the data is uploaded to the backend server via a wireless LAN link to drive scene association services.

[0084] The scenario-related service phase specifically includes the following steps: Based on the received location results, the backend server uses a lightweight RAG model retrieval module to efficiently retrieve exhibit information, facility status, or service content associated with the current spatial location from the structured knowledge graph, ensuring that the retrieval response time does not exceed 0.3 seconds.

[0085] The RAG model's generation module combines user presets or historical behavior analysis to obtain preference information, integrates, refines, and structures the search results, and generates personalized explanatory text or interactive prompts that match user preferences.

[0086] The generated personalized service content is presented to the user in at least one of the following ways: it is converted into an audio stream by a speech synthesis engine and broadcast through the smart glasses' speaker; or it is converted into augmented reality overlay information and rendered and positioned in real time in the user's field of vision through an AR display module.

[0087] It supports users to trigger follow-up questions on the current content or conduct secondary searches on related content through voice commands, enabling dynamic interaction; The power management unit dynamically adjusts the sleep and wake-up states of each functional module according to the system load to achieve energy efficiency optimization; The thermal management unit monitors the temperature of the control unit in real time and ensures that the equipment operates stably within a suitable temperature range by adjusting the operating frequency and other strategies.

[0088] In this embodiment, reliable electrical and data transmission links are established between system hardware modules through standard interfaces such as SPI, I2C, MIPI, and PCIe. Simultaneously, the system uses a pre-built visual point cloud map and structured knowledge graph as a static foundational data layer to support overall operation. Meanwhile, multi-source sensor data is dynamically processed by a fusion algorithm to continuously output high-precision positioning results. The RAG service module dynamically retrieves and generates scene-related information based on real-time positioning results. Power supply and thermal management adaptation units dynamically adjust their operating modes according to the actual load of the equipment and environmental conditions to optimize overall system performance.

[0089] Verified in real-world scenarios, this solution achieves the following comprehensive results: positioning error is controlled within 0.5m; no significant drift was observed during 2 hours of continuous operation; the positioning interruption rate is no more than 2%; the average response time for RAG retrieval is 0.2s; the accuracy rate of personalized explanations reaches 93%; the device can provide more than 8.5 hours of continuous service on a single charge; the operating temperature of the control unit is stably maintained between 38℃ and 42℃; no additional indoor positioning infrastructure is required, reducing system deployment costs by approximately 70% and subsequent maintenance workload by approximately 60%.

[0090] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A spatial positioning method based on a smart wearable device, characterized in that, include: Collect satellite positioning signal strength, planar position data, visual image data, and inertial measurement data; The smart wearable device is currently in an indoor environment based on the strength of the satellite positioning signal. In response to the judgment result being an indoor environment, the following steps are performed: The planar position data is determined based on satellite positioning signals, the visual positioning data is determined based on the visual image data, and the visual inertial data is generated by fusing the visual image data with the inertial measurement data. Based on the satellite positioning signal strength, fusion weights are assigned to the planar position data, visual positioning data, and visual-inertial data, and the positioning result is obtained through weighted fusion and optimization. Construct a structured knowledge graph for indoor scenes; Based on the location results, relevant scene information is retrieved from the structured knowledge graph, and personalized service content is generated according to the relevant scene information and user preferences. Output the personalized service content.

2. The spatial positioning method as described in claim 1, characterized in that, The structured knowledge graph includes: Location nodes are used to characterize the positions of selected fixed spatial features within an interior space; Resource nodes, associated with the location nodes, are used to characterize the location of resources within the indoor space, including facilities and services.

3. The spatial positioning method as described in claim 2, characterized in that, When the positioning function is implemented, the weight of the positioning node exceeds the first threshold.

4. The spatial positioning method as described in claim 3, characterized in that, The construction of the structured knowledge graph for indoor scenes also includes real-time updating of the structured knowledge graph; the real-time updating includes: Based on at least one of the planar position data, visual image data, and inertial measurement data, the changes in the actual indoor scene are determined; based on the changes, add, delete, or modify the positioning nodes, resource nodes, or the relationships between the two in the structured knowledge graph to maintain the consistency between the structured knowledge graph and the actual indoor scene.

5. The spatial positioning method as described in claim 1, characterized in that, The allocation of fusion weights includes: When the satellite positioning signal strength is greater than or equal to the second threshold, the weights are allocated according to the first ratio, and the planar position data, visual positioning data and visual inertial data are fused together. When the satellite positioning signal strength is less than the second threshold, the planar position data is masked, weights are allocated according to the second ratio, and the visual positioning data and visual inertial data are fused.

6. The spatial positioning method as described in claim 1, characterized in that, The steps for obtaining the positioning result through weighted fusion and optimization specifically include: Based on the assigned fusion weights, the least squares optimization algorithm is used to fuse the planar position data, visual positioning data, and visual-inertial data to obtain preliminary positioning results. The continuous preliminary positioning results are smoothed using a sliding window optimization algorithm to obtain the positioning result.

7. The spatial positioning method as described in claim 1, characterized in that, Before assigning fusion weights to the planar position data, visual positioning data, and visual-inertial data, the method further includes optimizing each data separately, specifically including: The planar position data is filtered and optimized using a Kalman filter algorithm. Based on ambient lighting conditions, the feature point extraction strategy of the visual image data is adjusted to optimize the visual positioning data; Visual inertial data is generated based on the visual image data and inertial measurement data using the visual inertial odometry method.

8. The spatial positioning method as described in claim 1, characterized in that, The personalized service content is presented in at least one of the following ways: The message is output via voice; or... Visual overlay display output is achieved through augmented reality.

9. A spatial positioning system based on a smart wearable device, used to implement the spatial positioning method based on a smart wearable device as described in any one of claims 1 to 8, characterized in that, include: The data acquisition module, integrated into the smart wearable device, includes a satellite positioning unit, a visual acquisition unit, and an inertial measurement unit; the visual acquisition unit includes a high-definition camera and an infrared camera; it is used to acquire satellite positioning signal strength, planar position data, visual image data, and inertial measurement data; The fusion positioning module is used to determine whether the smart wearable device is currently in an indoor environment based on the satellite positioning signal strength; in response to the determination result being an indoor environment, the following steps are performed: based on the satellite positioning signal strength, fusion weights are assigned to the planar position data, the visual positioning data generated based on the visual image data, and the visual inertial data generated based on the fusion of the visual image data and inertial measurement data, and the positioning result is obtained through weighted fusion and optimization; The scene association service module is used to build a structured knowledge graph of indoor scenes; Based on the location results, relevant scene information is retrieved from the structured knowledge graph, and personalized service content is generated according to the relevant scene information and user preferences. Output the personalized service content; The edge-side optimization module includes: a transmission unit for transmitting data via Bluetooth and a wireless local area network; a power management unit for controlling the power supply status of each functional module in the smart wearable device based on task load through a dynamic sleep mechanism; and a thermal management unit for controlling the operating temperature of the smart wearable device by optimizing hardware layout and software scheduling strategies. The control unit is used to preprocess the planar position data, visual image data, and inertial measurement data; and to control the operation of the data acquisition module, fusion positioning module, scene association service module, and edge optimization module. The power management unit provides power to the data acquisition module, the fusion positioning module, the scene association service module, the transmission unit, the thermal management unit, and the control unit.

10. A type of smart glasses, characterized in that, It integrates the spatial positioning system based on smart wearable devices as described in claim 9.

Images