A new desktop interaction system and method based on a globally draggable floating AI assistant

Abstract

This invention provides a novel desktop interaction system and method based on a globally draggable floating AI assistant, belonging to the fields of artificial intelligence, smart terminals, and human-computer interaction technology. This invention achieves lightweight intelligent interaction across all scenarios through hardware-layer touch event interception, application-layer dual-mode rendering, and direct connection via the TLS 1.3 protocol. Screenless devices can achieve equivalent interaction through lights, vibrations, and projections. For screenless devices, this system achieves equivalent interaction through light color, vibration frequency, and sound prompts, ensuring that screenless devices can also fully implement security alerts. This invention provides users with a low-latency, standardized, and interactive upper-layer intelligent interaction entry point, specifically applicable to desktop interaction scenarios on mainstream mobile terminal operating systems such as Android, iOS, and HarmonyOS.

Description

Technical Field

[0001] This invention belongs to the fields of artificial intelligence, smart terminals and human-computer interaction technology, and relates to a novel desktop interaction system and method based on a globally draggable floating AI assistant. Background Technology

[0002] With the deep integration of artificial intelligence and mobile terminal technology, intelligent assistants have become an important entry point for user interaction with devices. However, existing intelligent assistant technologies on mobile terminals still have the following technical shortcomings in terms of interaction efficiency, system integration, security linkage, and personalized expansion: 1) Deep interaction entry points and low operation efficiency: Existing smart assistants mostly rely on clicking application icons, requiring a three-level operation chain of "icon click - page jump - function trigger". For users with weaker operation skills, such as the elderly and children, the operation path is long and the success rate is low, failing to provide a quick global interaction entry point at the system level; 2) Low UI rendering priority, unable to remain globally persistent: Existing AI assistants' interactive interfaces are typically implemented based on application-level Activity components, whose rendering priority is the same as ordinary applications, making it impossible to achieve global floating across application scenarios. When the user switches to other applications, the assistant interface is covered or exited, failing to provide continuous and uninterrupted intelligent services; 3) High data synchronization latency and lack of standardization: Data communication between multiple modules often adopts the method of inter-application interface calls, which lacks standardized data formats. This results in data synchronization latency that is usually greater than 500ms, making it impossible to achieve real-time aggregation and unified display of multi-dimensional information such as security status, to-do tasks, and family status. 4) Insufficient linkage between security alerts and interface interaction: In existing technologies, the linkage between security alerts and interface interaction is mostly a passive response mode of "message push - pop-up prompt", which fails to achieve real-time synchronization of status codes at the system level and active linkage rendering at the UI level. The response latency is high, and the status display lacks a unified technical standard. 5) Poor persistence and low interactive flexibility of floating controls: Conventional floating controls are developed based on in-application Service components. After exiting the application, the Service process is easily reclaimed by the system. They lack technical solutions for background persistence and system-level window registration, making it impossible to achieve global persistence across applications. At the same time, traditional floating controls have a fixed form and do not support customization of brand logos or import of vector files, failing to meet the dual needs of partner manufacturers for brand visualization and users for personalization.

[0003] In summary, there is an urgent need for a desktop interaction system that can achieve global persistence, low-latency linkage, convenient interaction, and ecological expansion capabilities. Summary of the Invention

[0004] In view of this, the purpose of the present invention is to provide a novel desktop interaction system and method based on a globally draggable floating AI assistant, which can overcome the above-mentioned technical problems existing in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A novel desktop interaction system based on a globally draggable floating AI assistant, comprising: The global floating control module calls the mobile terminal WindowManager system-level service to create a floating control, sets the window type to TYPE_APPLICATION_OVERLAY or UIWindowLevelAlert, makes the display priority of the floating control higher than all applications and system interfaces, binds the OnTouchListener touch listening interface to the floating control to enable free dragging, and achieves automatic edge snapping through the screen boundary detection algorithm. The module for personalized shape and brand logo adaptation includes shape editing and vector file parsing interfaces, which support basic geometric shape splicing, import of brand logo vector files and parsing of bitmap files. The shape of the floating control can be edited in the range of 8~32dp, and all shapes are optimized with rounded corners to prevent accidental touch. The voice interaction module includes a local lightweight automatic speech recognition model and a hardware-layer long-press event interception unit. It is used to trigger local voice acquisition when the floating control is long-pressed for 300ms. After being parsed by the local lightweight automatic speech recognition model, cross-module instruction execution is achieved through the event bus. The background persistent module listens to the application lifecycle through application lifecycle callbacks and is used to start a background keep-alive process when the application is moved to the background, so that the floating control remains displayed and responsive after the application is exited. The secondary security architecture interface module establishes a long TCP connection with the secondary security architecture via the TLS 1.3 encryption protocol to enable timed synchronization of security status codes and real-time push of abnormal statuses.

[0006] Furthermore, the global floating control module is also used to: develop dual-form custom view layouts, including in-application linear layouts and out-of-application frame layouts, and to achieve automatic switching between the two layouts through application lifecycle awareness, wherein the configured brand logo and personalized form remain unchanged during the switching process.

[0007] Furthermore, the global floating control module also includes a personalized image customization unit, which supports the import of image resources in multiple formats and realizes rendering through a custom drawing interface. The image resources are decoupled from the core interactive logic.

[0008] Furthermore, the interactive system also includes a state-driven UI refresh engine, which adjusts the color, transparency, size, blinking frequency, and animation style of the floating control in real time according to the security status. The state priority is set by integer assignment, and the color, animation, and badge of the floating control are adjusted in real time according to the security status, reminder events, and to-do tasks. A high-risk status triggers a sound or vibration reminder, and the brand logo and personalized form remain unchanged during the UI refresh process.

[0009] Furthermore, the voice interaction module also includes offline parsing of the collected voice data using a local lightweight automatic speech recognition model. If the parsing fails, the data is uploaded to the cloud-based automatic speech recognition server via the HTTP / 2 protocol. The parsing result is processed into a preset command type by an intent recognition algorithm, and the parsing latency is less than 200ms.

[0010] Furthermore, it also includes a standardized data parsing module, which defines a unified data format. It collects data from the secondary security architecture, family interaction module, to-do task module, and navigation module. After data deduplication and format conversion, the data is stored in the memory cache, with a data reading latency of less than 10ms.

[0011] This invention also provides a novel desktop interaction method based on a globally draggable floating AI assistant, the method comprising the following steps: S1. Call the mobile terminal WindowManager system-level service to create and register a global floating control, set the highest display priority, bind the touch listening interface to enable free dragging, and use the screen boundary detection algorithm to achieve automatic edge snapping. S2. Load shape and brand trademark resources. Through shape editing and vector file parsing interfaces, load the shape resources configured by users or partners. After format verification, size adaptation and anti-accidental touch rounded corner optimization, complete the rendering through the custom drawing interface. S3. Bind a long press event interception unit to the floating control. After triggering, collect local voice at a preset sampling rate, parse it into a preset instruction type by a local lightweight automatic speech recognition model, and realize cross-module instruction execution through the event bus. S4. Monitor the application lifecycle. When the application is moved to the background, start a background keep-alive process and switch the floating control from its full form within the application to its simplified form outside the application. S5. Establish a long TCP connection with a secondary security architecture through the TLS 1.3 encryption protocol, periodically synchronize security status codes and push abnormal statuses in real time, and store the synchronized data in the memory cache after standardized parsing; S6. Develop a state-driven UI refresh engine that detects memory cache data at a preset frequency, dynamically adjusts the color, animation, and badge of the floating control according to the security status, and triggers sound or vibration alerts in high-risk states.

[0012] Furthermore, the global floating control supports personalized image customization. It obtains the image resource path through a file selector, and after format verification and size adaptation, it is stored in a local specified directory and rendered through a custom drawing interface, so that the image and form rendering are independent and conflict-free.

[0013] Furthermore, the method enables unified execution of both voice and manual dual-mode commands. Both modes are converted into standardized event objects, which are sent through the event bus. Each module subscribes to the event object and calls the core interface to execute the command. The execution results are then synchronized to the local cache, the system notification center, and the floating control.

[0014] The beneficial effects of this invention are as follows: 1) Significantly improved global interaction efficiency: Based on the WindowManager system-level service, global registration and persistence of floating controls are achieved, drag response latency is less than 10ms, the automatic edge-fitting algorithm accuracy reaches 1dp, and cross-application interaction efficiency is improved by more than 80%. 2) The convenience of voice interaction is greatly improved: by long-pressing the floating control for 300ms to trigger voice commands, combined with the local lightweight ASR model, the parsing latency is less than 200ms, reducing the high-frequency operation steps from the traditional 3 steps to 1 step, and increasing the success rate of operation for the elderly and children by more than 90%. 3) Strong branding and personalization expansion capabilities: It supports direct import of brand trademark vector files in SVG / AI and other formats, as well as customization of personalized forms and images in multiple formats. The form rendering response latency is less than 100ms, and the brand adaptation development cycle of cooperative manufacturers is shortened by 90%, which has extremely strong ecological scalability. 4) Low latency of security linkage: Based on the TLS 1.3 protocol and a two-level security architecture, a direct connection channel is established, with a security status synchronization latency of less than 50ms and a control command issuance and receipt latency of less than 200ms. Real-time linkage protection of home security is achieved without zero image or audio acquisition. 5) Accurate and stable state rendering: Based on the priority sorting UI refresh engine, the state detection and rendering latency is less than 100ms, the high-risk state reminder reach rate is 100%, the system operation stability is greater than 99.9%, and the brand logo and personalized form remain unchanged during the state refresh process.

[0015] Other advantages, objectives, and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination, or may be learned from practice of the invention. The objectives and other advantages of the invention can be realized and obtained through the following description. Attached Figure Description

[0016] To make the objectives, technical solutions, and advantages of the present invention clearer, the preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings, wherein: Figure 1 This is a block diagram of the core modules of the system of the present invention. Detailed Implementation

[0017] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0018] The accompanying drawings are for illustrative purposes only and are schematic diagrams, not actual pictures. They should not be construed as limiting the invention. To better illustrate the embodiments of the invention, some parts in the drawings may be omitted, enlarged, or reduced, and do not represent the actual product dimensions. It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0019] In the accompanying drawings of the embodiments of the present invention, the same or similar reference numerals correspond to the same or similar components. In the description of the present invention, it should be understood that if terms such as "upper," "lower," "left," "right," "front," and "rear" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing the present invention 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, the terms used to describe positional relationships in the drawings are only for illustrative purposes and should not be construed as limiting the present invention. For those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances. Example 1: System Architecture

[0020] This embodiment provides a novel desktop interaction system based on a globally draggable floating AI assistant. The system is deployed on mobile terminals (such as smartphones and tablets). Figure 1 The following is a block diagram of the core modules of the system of this invention, including the following core modules: Global Floating Control Module: As the core interaction carrier of the system, this module creates and manages floating controls by calling the WindowManager system-level service provided by the mobile terminal operating system. Specifically, in the Android system, views are added using the WindowManager.addView() method, and the LayoutParams type is set to TYPE_APPLICATION_OVERLAY; in the iOS system, UIWindow is used, and windowLevel is set to UIWindowLevelAlert. This makes the floating controls have a higher display priority than all ordinary applications, allowing them to be globally persistent. This module binds an OnTouchListener to the controls, listens for touch events in real time, and refreshes the control coordinates in real time (with an accuracy of 1dp) using the WindowManager.updateViewLayout() method during the ACTION_MOVE phase, enabling smooth dragging. At the same time, a built-in screen boundary detection algorithm automatically calibrates the coordinates to the edge position when the control is dragged to the edge of the screen (such as within 16dp-24dp from the boundary), ensuring that it does not obstruct the main interface content.

[0021] Personalized Shape and Brand Logo Adaptation Module: This module provides highly customizable interface rendering capabilities. On one hand, it allows users to stitch together basic geometric shapes through a shape editing interface to create personalized control shapes. On the other hand, it provides a brand logo adaptation interface for partner manufacturers, supporting direct import of vector files in formats such as SVG (Scalable Vector Graphics) or AI (Adobe Illustrator) (resolution no less than 1080P). The system parses the vector files using parsers such as SAXParser and performs size adaptation using the Bitmap.createScaledBitmap() method (editable size range of 8~32dp, supporting stepless scaling in 0.5dp increments). Finally, rendering is completed through a custom Drawable drawing interface. To avoid accidental touches, all shapes are optimized with rounded corners during rendering, with corner radii set between 2dp and 8dp.

[0022] Voice Interaction Module: This module implements hardware-level voice-activated wake-up. The module binds an OnLongClickListener to the floating control and sets the long-press trigger threshold to 300ms. Upon triggering, local voice acquisition is initiated via the AudioRecord component, with parameters set to a sampling rate of 16kHz, a sampling bit depth of 16bit, and mono to ensure lightweight and clear acquisition. The acquired voice data is first sent to a local lightweight ASR model (based on TF-Lite, model file less than 50MB) for offline parsing, with a parsing latency of less than 200ms. If local parsing fails, the voice data is uploaded to a cloud-based ASR server via HTTP / 2 for secondary parsing. The parsed text is processed by an intent recognition algorithm to generate standardized event objects, which are then published via the EventBus event bus to enable cross-module execution of commands without opening any application interface.

[0023] Background Persistent Module: To ensure the global availability of the floating control, this module monitors the entire application lifecycle using `Application.ActivityLifecycleCallbacks`. When the application's main screen is detected as being moved to the background (i.e., `onPause()` is triggered), the module automatically starts a foreground service and sets its startup mode to `START_STICKY` to prevent the process from being reclaimed by the system. Simultaneously, the module calls the `View.setVisibility()` method to switch the floating control from its fully functional "complete form" (e.g., Linear Layout, 60dp-160dp) to a "simplified form" (e.g., Frame Layout, 40dp-60dp) that retains only the core icon and brand logo, adapting to background operation scenarios and ensuring it always appears on top.

[0024] Level 2 Security Architecture Interface Module: This module is responsible for secure, low-latency data interaction with the "AI Intelligent Independent Level 2 Security Monitoring System and Device Based on Device Status Data" (hereinafter referred to as "Level 2 Security Architecture") submitted by the inventor on the same day. The module establishes a long TCP connection with the Level 2 Security Architecture via a Socket client based on the TLS 1.3 protocol, uses the ECDHE algorithm for key exchange, and the AES-256-GCM algorithm for data encryption. After the connection is established, the module receives and synchronizes standardized security status codes (binary stream format, including device ID, status code, and timestamp) at a frequency of 1Hz, achieving a latency of less than 50ms. When the Level 2 Security Architecture detects a high-risk abnormal state, it triggers a real-time push through an emergency push flag, and the floating AI responds immediately upon receiving the push. For screenless devices (such as elderly walking aids, home appliances, etc.), this system uses this module to map status information to light colors (green, yellow, red), vibration frequencies (short vibrations, long vibrations), or sound prompts (different tones), achieving equivalent interaction.

[0025] Example 2: Dual-morphological adaptive and personalized rendering This embodiment details the dual-shape adaptive switching and personalized rendering process of the floating control. Users or partner manufacturers configure the shape and appearance through a customized page within the app. Users can select basic geometric shapes for splicing or directly import local brand logo vector files (SVG / AI) and image images (PNG / JPG / WEBP). After the system calls the file selector to obtain the file path, it performs format verification and size adaptation, and finally saves the resource file to the specified local directory (e.g., / data / data / package name / files / shape / ), and writes the resource ID to SharedPreferences.

[0026] When the application starts, the system calls the WindowManager service in Application.onCreate(), loads the corresponding appearance and avatar resources based on the configuration information in SharedPreferences, and creates a fully-formed floating control, which is displayed in the lower right corner of the screen by default. At the same time, the ActivityLifecycleCallbacks listener is started.

[0027] When the user interacts with the application, the floating control is presented in its full form, including an image display area, a function shortcut area, and a status display area. When the user presses the Home button or switches to another application, the onPause() callback is triggered, and the background resident module immediately switches the floating control to a simplified form, retaining only the core brand logo and image icon, which remains permanently displayed on the desktop. Throughout the entire switching process, the configured brand logo and personalized form remain unchanged, and the core interaction logic and form rendering are completely decoupled, with the switching achieved solely through resource ID mapping.

[0028] Example 3: Voice / Manual Dual-Mode Collaborative Control and Status-Driven UI Refresh This embodiment illustrates the unified execution logic of dual-mode commands and the dynamic UI refresh mechanism. Command execution: Whether it's a voice command triggered by long-pressing the floating control or a command triggered by clicking the floating control... Manual commands generated by shortcut buttons on the device (such as security alerts, family contact, navigation, and to-do lists) are converted into standardized Event objects (containing command type, parameters, and execution target) by the system. These event objects are then published to the event bus using the `EventBus.getDefault().post()` method. Various functional modules in the system (such as the reminder module and navigation module) subscribe to the corresponding events using the `@Subscribe` annotation and call their respective core execution interfaces (such as `createRemind()` and `startNavi()`) upon receiving an event. The execution results are broadcast to the entire system via `BroadcastReceiver` and synchronized to the local cache, the system notification center, and the memory cache of the floating control, achieving real-time feedback of the operation results.

[0029] UI Refresh: The system has developed a state-driven UI refresh engine. This engine uses a Handler to create a 100ms timed task to continuously monitor state data (safety status, number of pending tasks, etc.) in the LruCache memory cache. Internally, the engine defines state priorities using integer assignment: high risk = 5, medium risk = 4, navigation = 3, low risk = 2, pending = 1. When a state change is detected, the engine triggers the corresponding UI rendering based on the highest priority state. Color rendering: Set the background color using view.setBackgroundColor(), such as red for a high-risk status; Animation rendering: Blinking or pulse effects are achieved using ValueAnimator. For example, a blinking effect with a frequency of 1.2Hz is triggered in high-risk situations; Subtitle rendering: A numerical subtitle is drawn in the upper right corner of the control using Canvas.drawText() to display the number of tasks to be done; Sound / Vibration Alerts: When a high-risk situation is detected, a local warning sound is played via MediaPlayer, and the audio priority is set via AudioManager to ensure that vibration alerts can still be triggered in silent mode. During this process, the brand logo and personalized form of the floating control remain unchanged as the base layer, with only the upper-level visual effects such as color and animation being dynamically updated, ensuring the consistency of brand visualization and meeting users' personalized needs.

[0030] Example 4: Low-latency linkage with a two-level security architecture This embodiment details the linkage process with the secondary security architecture. Upon startup of the suspended AI, its secondary security architecture interface module immediately initiates TLS 1.3 communication via a Socket client. A handshake is initiated to establish a persistent TCP connection with the secondary security architecture. After successful device authentication, both parties maintain the connection and send heartbeat packets at 30-second intervals. Under normal conditions, the secondary security architecture pushes standardized binary status codes to the suspended AI at a frequency of 1Hz. Upon receiving these codes, the suspended AI's data receiving thread parses them using the standardized data parsing module and updates its memory cache.

[0031] When the secondary security architecture detects an anomaly (such as device failure or security risk), it sets an emergency push flag in the push data packet. The Floating AI's Socket listening thread identifies this flag in real time, immediately receives and processes the abnormal status data, with a synchronization latency of less than 50ms. After detecting a change in security status in the memory cache, the status refresh engine triggers dynamic UI rendering according to preset priorities, achieving second-level early warning for high-risk states. Users can initiate control commands (such as deceleration or stopping) through the Floating AI. Commands are standardized and encoded in JSON format, then sent to the secondary security architecture via an established TLS 1.3 channel. After executing the command, the secondary security architecture immediately returns an execution result receipt. Upon receiving the receipt, the Floating AI updates the UI and logs the results; the entire process of command delivery and receipt processing has a latency of less than 200ms. Example 5: Application Scenarios

[0032] This system can be deployed independently or serve as the central interactive hub of the "Red Bean Leaf Ecosystem," achieving deep technical synergy with the secondary security architecture and AI education ecosystem. In home smart interaction scenarios: it serves as a global floating entry point for family care, smart home control, and home security alerts. Through linkage with the secondary security architecture, the floating control can display the device's security status in real time (via color and animation changes), and users can issue control commands with a single click, achieving real-time protection of home security. In mobile terminal quick operation scenarios: it provides a simplified interaction method for users such as the elderly and children. By long-pressing the floating control, users can quickly complete high-frequency operations such as creating voice reminders, starting navigation, querying to-do lists, and contacting family members, effectively lowering the interaction threshold. In business cooperation and brand adaptation scenarios: it provides customized brand logo forms for partner manufacturers. Manufacturers can import their brand logos as vector files, and the system automatically adapts them into floating control forms that conform to interaction specifications, achieving a seamless combination of brand exposure and smart interaction.

[0033] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A novel desktop interaction system based on a globally draggable floating AI assistant, characterized in that, The system includes: The global floating control module calls the mobile terminal WindowManager system-level service to create a floating control, sets the window type to TYPE_APPLICATION_OVERLAY or UIWindowLevelAlert, makes the display priority of the floating control higher than all applications and system interfaces, binds the OnTouchListener touch listening interface to the floating control to enable free dragging, and achieves automatic edge snapping through the screen boundary detection algorithm. The module for personalized shape and brand logo adaptation includes shape editing and vector file parsing interfaces, which support basic geometric shape splicing, import of brand logo vector files and parsing of bitmap files. The shape of the floating control can be edited in the range of 8~32dp, and all shapes are optimized with rounded corners to prevent accidental touch. The voice interaction module includes a local lightweight automatic speech recognition model and a hardware-layer long-press event interception unit. It is used to trigger local voice acquisition when the floating control is long-pressed for 300ms. After being parsed by the local lightweight automatic speech recognition model, cross-module instruction execution is achieved through the event bus. The background persistent module listens to the application lifecycle through application lifecycle callbacks and is used to start a background keep-alive process when the application is moved to the background, so that the floating control remains displayed and responsive after the application is exited. The secondary security architecture interface module establishes a long TCP connection with the secondary security architecture via the TLS 1.3 encryption protocol to enable timed synchronization of security status codes and real-time push of abnormal statuses.

2. A novel desktop interaction system based on a globally draggable floating AI assistant as described in claim 1, characterized in that, The global floating control module is also used to: develop dual-form custom view layouts, including in-application linear layouts and out-of-application frame layouts, and achieve automatic switching between the two layouts through application lifecycle awareness, wherein the configured brand logo and personalized form remain unchanged during the switching process.

3. A novel desktop interaction system based on a globally draggable floating AI assistant as described in claim 1, characterized in that, The global floating control module also includes a personalized image customization unit, which supports the import of image resources in multiple formats and implements rendering through a custom drawing interface. The image resources are decoupled from the core interactive logic.

4. A novel desktop interaction system based on a globally draggable floating AI assistant as described in claim 1, characterized in that, The interactive system also includes a state-driven UI refresh engine, which adjusts the color, transparency, size, blinking frequency, and animation style of the floating control in real time according to the security status. The state priority is set by integer assignment, and the color, animation, and badge of the floating control are adjusted in real time according to the security status, reminder events, and to-do tasks. A sound or vibration reminder is triggered in a high-risk state, and the brand logo and personalized form remain unchanged during the UI refresh process.

5. A novel desktop interaction system based on a globally draggable floating AI assistant as described in claim 1, characterized in that, The voice interaction module also includes offline parsing of the collected voice data using a local lightweight automatic speech recognition model. If the parsing fails, the data is uploaded to the cloud-based automatic speech recognition server via the HTTP / 2 protocol. The parsing result is processed into a preset command type by an intent recognition algorithm, and the parsing latency is less than 200ms.

6. A novel desktop interaction system based on a globally draggable floating AI assistant as described in claim 1, characterized in that, It also includes a standardized data parsing module, which defines a unified data format. Data is collected from the secondary security architecture, family interaction module, to-do task module and navigation module. After data deduplication and format conversion, it is stored in memory cache with a data reading latency of less than 10ms.

7. A novel desktop interaction method based on a globally draggable floating AI assistant, characterized in that, The method includes the following steps: S1. Call the mobile terminal WindowManager system-level service to create and register a global floating control, set the highest display priority, bind the touch listening interface to enable free dragging, and use the screen boundary detection algorithm to achieve automatic edge snapping. S2. Load shape and brand trademark resources. Through shape editing and vector file parsing interfaces, load the shape resources configured by users or partners. After format verification, size adaptation and anti-accidental touch rounded corner optimization, complete the rendering through the custom drawing interface. S3. Bind a long press event interception unit to the floating control. After triggering, collect local voice at a preset sampling rate, parse it into a preset instruction type by a local lightweight automatic speech recognition model, and realize cross-module instruction execution through the event bus. S4. Monitor the application lifecycle. When the application is moved to the background, start a background keep-alive process and switch the floating control from its full form within the application to its simplified form outside the application. S5. Establish a long TCP connection with a secondary security architecture through the TLS 1.3 encryption protocol, periodically synchronize security status codes and push abnormal statuses in real time, and store the synchronized data in the memory cache after standardized parsing; S6. Develop a state-driven UI refresh engine that detects memory cache data at a preset frequency, dynamically adjusts the color, animation, and badge of the floating control according to the security status, and triggers sound or vibration alerts in high-risk states.

8. A novel desktop interaction method based on a globally draggable floating AI assistant according to claim 7, characterized in that, The global floating control supports personalized image customization. It obtains the image resource path through a file selector, and after format verification and size adaptation, it is stored in a local specified directory. It is then rendered through a custom drawing interface, making the image and form rendering independent and conflict-free.

9. A novel desktop interaction method based on a globally draggable floating AI assistant according to claim 7, characterized in that, The method enables unified execution of both voice and manual commands. Both modes are converted into standardized event objects, which are sent through the event bus. Each module subscribes to the event object and calls the core interface to execute the command. The execution results are then synchronized to the local cache, system notification center, and floating control.

Images