Intelligent desktop management system and dynamic resource allocation method

Abstract

The invention discloses an intelligent desktop management system and a dynamic resource allocation method. The intelligent desktop management system comprises a main control module; a multi-mode sensing module; a resource execution module; a man-machine interaction module; the dynamic resource allocation method comprises the following steps: S1, sensing and fusing multi-source data; s2, performing multi-level identification on a user state and an intention; s3, dynamic resource mapping based on a strategy library; and S4, performing cooperative control and instruction execution. According to the intelligent desktop management system, user states, behaviors and environment data are collected in real time through the multi-mode sensing module, a dynamic resource allocation method is combined, resource requirements are accurately mapped to cooperative control of units such as a displayer, a lighting unit and a power source, the energy allocation unit supporting the USBPD protocol can dynamically adjust power output, and the intelligent desktop management system is high in practicability. And the energy utilization efficiency is maximized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent office, and particularly relates to an intelligent desktop management system and a dynamic resource allocation method. BACKGROUND

[0002] The intelligent desktop management technology develops rapidly with the growth of intelligent office demand. In the prior art, the desktop management system mainly realizes part of automatic functions through sensors and control units, such as detecting the presence state of a user based on an infrared sensor or adjusting the height of a display and the brightness of a light through manual input. Some smart home systems introduce single modal perception such as a camera capturing user actions or a simple resource allocation mechanism such as a timing power switch, and combine with a basic human-computer interaction interface such as a touch screen or a mobile phone application to realize user control. These systems have been applied to some extent in office environments and can improve basic work efficiency, but are usually limited to single resource management or static configuration and are difficult to adapt to dynamic needs in complex scenarios.

[0003] However, the prior art has significant defects. First, single modal perception such as a camera or an infrared sensor has limited accuracy, and a camera may cause privacy problems, limiting its application in privacy-sensitive scenarios. Second, resource allocation is mostly static or preset mode, lacking dynamic adaptation to real-time user behavior and environmental changes, resulting in low resource utilization efficiency. In addition, the existing systems lack personalized learning function and cannot optimize resource allocation according to user habits, resulting in poor user experience. Therefore, there is an urgent need for an intelligent desktop management system with multi-modal perception, dynamic resource allocation and personalized learning support. SUMMARY

[0004] The present application relates to the technical field of intelligent office, and particularly relates to an intelligent desktop management system and a dynamic resource allocation method.

[0005] To achieve the above-mentioned purpose, the present application adopts the following technical solutions: An intelligent desktop management system and a dynamic resource allocation method, comprising a main control module, a multi-modal perception module, a resource execution module, and a human-computer interaction module. The multi-modal perception module is in communication connection with the main control module and is used for collecting user and environmental data, and comprises a non-visual biological perception unit, a posture and behavior perception unit, and an article perception unit. The resource execution module is in communication connection with the main control module and is used for executing adjustment operations on desktop electronic and physical resources, and comprises at least one of a display support unit, a lighting adjustment unit, an energy distribution unit, and a physical configuration unit. The human-computer interaction module is in communication connection with the main control module. The main control module is configured to execute a dynamic resource allocation method.

[0006] Preferably, the non-visual biological perception unit is a millimeter wave radar sensor; the posture and behavior perception unit is a distributed pressure sensor array and / or a 3D-ToF sensor; the article perception unit is an RFID reader / writer or a strain gauge sensor.

[0007] Preferably, the energy distribution unit supports the USBPD power distribution protocol.

[0008] Preferably, the master module further comprises a personalized learning unit for optimizing the rule parameters in the resource allocation strategy library according to the manual operation records of the user through the human-computer interaction module.

[0009] A dynamic resource allocation method, executed by a master module such as an intelligent desktop management system, comprising the following steps: S1. Multi-source data perception and fusion step: acquiring and fusing heterogeneous sensor data streams from multi-modal perception modules to generate comprehensive perception data; S2. Multi-level identification of user state and intention step: based on the comprehensive perception data, sequentially performing presence state identification, behavior pattern identification, and external demand identification; S3. Dynamic resource mapping based on the strategy library step: taking the identified state, pattern, and demand as input, querying the resource allocation strategy library, and mapping to obtain the corresponding multi-resource collaborative allocation scheme; S4. Collaborative control and instruction execution step: compiling the resource allocation scheme into control instructions and sending them to the corresponding units in the resource execution module to drive their execution.

[0010] Preferably, the multi-level identification of user state and intention step comprises: (1) Based on millimeter wave radar signals, determine whether the user is in a presence, temporary leave, or long-term leave state; (2) Based on posture signals, identify user behavior patterns through a lightweight convolutional neural network model, including typing, reading and writing, video conferencing, or resting; (3) Based on article identification signals and ambient light signals, identify the user's peripheral device demand and lighting demand.

[0011] Preferably, the mapping rules in the resource allocation strategy library are IF-THEN rules, where the conditions in the IF part are the logical combinations of multi-modal perception information, and the actions in the THEN part are collaborative control instructions for multiple units in the resource execution module.

[0012] Preferably, after the collaborative control and instruction execution step, it further comprises: Personalized learning and strategy optimization step: record the user's manual adjustment instructions issued through the human-computer interaction module, and constitute a training sample with the system state before adjustment, and fine-tune and optimize the rule parameters in the resource allocation strategy library.

[0013] A computing device comprising a memory and a processor, the memory having stored thereon a computer program which, when executed by the processor, implements the dynamic resource allocation method.

[0014] A computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the dynamic resource allocation method.

[0015] The present application has the following beneficial effects: 1. The intelligent desktop management system of the present application acquires user state, behavior and environment data in real time through a multi-modal perception module, and accurately maps resource demand to the collaborative control of display, lighting and power units in combination with the dynamic resource allocation method, so that the energy distribution unit supporting the USBPD protocol can dynamically adjust power output, maximizing energy utilization efficiency.

[0016] 2. The intelligent desktop management system of the present application has a built-in personalized learning unit, which records user manual adjustment instructions and system state, and optimizes IF-THEN rules in the resource allocation strategy library using the gradient descent method, so that the optimized rules can dynamically adjust resource allocation according to user habits.

[0017] 3. The present application uses a non-visual biological perception unit and a 3D-ToF sensor to avoid using a camera in the intelligent desktop management system, protect user privacy, and achieve high-precision user state and behavior recognition, while the millimeter wave radar can accurately detect the presence of the user, and the 3D-ToF and pressure sensor array support behavior pattern recognition. Compared with traditional systems that rely on visual perception, the present application is more suitable for privacy-sensitive scenarios, provides more accurate perception data, and improves the accuracy of resource allocation. BRIEF DESCRIPTION OF DRAWINGS

[0018] Fig. 1 is a system block diagram of the intelligent desktop management system; Fig. 2 is a flowchart of the dynamic resource allocation method. DETAILED DESCRIPTION

[0019] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all.

[0020] Reference Figs. 1-2The application discloses an intelligent desktop management system and a dynamic resource allocation method, and relates to the technical field of intelligent desktop management.

[0021] In the embodiment, the master control module serves as the core of the system and plays a key role in overall coordination of operation of the modules. The built-in resource allocation strategy library of the master control module provides decision basis and rule support for subsequent dynamic resource allocation, and ensures rationality and standardization of resource allocation. The multi-modal perception module is in communication connection with the master control module. The core role of the multi-modal perception module is to comprehensively and real-timely collect user and environment data, and to provide a data basis for system perception of user state and environment condition. The non-visual biological perception unit is responsible for collection of non-visual biological data of the user, avoids privacy problems caused by visual perception, and acquires basic existence information of the user. The posture and behavior perception unit focuses on collection of posture features and behavior action data of the user, and helps the system to judge a current behavior mode of the user. The article perception unit is used for collection of related information of various articles on the desktop, so that the system can master article resource conditions on the desktop. The resource execution module is in communication connection with the master control module. The core function of the resource execution module is to convert resource adjustment instructions generated by the master control module into actual operations, and to execute adjustment of electronic and physical resources on the desktop. The display support unit is used for adjustment of a support structure or display related parameters of a display device on the desktop, so as to adapt to viewing or use requirements of the user. The lighting adjustment unit is used for adjustment of a lighting state of a desktop area, so as to create a suitable light environment for the user. The energy distribution unit is used for distribution and regulation of energy supply of electronic devices on the desktop, so as to realize efficient use of energy. The physical configuration unit is used for adjustment of a physical structure of the desktop, so as to satisfy individualized requirements of the user on the physical form of the desktop. The module contains at least one of the above units, and can ensure that the system has basic desktop resource adjustment capability. The man-machine interaction module is in communication connection with the master control module. The role of the man-machine interaction module is to build an interaction bridge between the user and the system, to facilitate the user to input operation instructions to the system or to acquire feedback information of the system, and to improve controllability of the system by the user and use experience of the system. The master control module is configured to execute the dynamic resource allocation method. The master control module can generate accurate control instructions for the resource execution module based on fusion data collected by the multi-modal perception module and in combination with rules in the resource allocation strategy library. Finally, the master control module realizes dynamic and intelligent allocation and adjustment of electronic and physical resources on the desktop through the resource execution module, and thus achieves the overall effect of efficient use of desktop resources and adaptation to requirements of the user.

[0022] In the present application, the non-visual biological perception unit is a millimeter wave radar sensor; the posture and behavior perception unit is a distributed pressure sensor array and / or a 3D-ToF sensor; and the article perception unit is an RFID reader / writer or a strain gauge sensor.

[0023] In the present embodiment, the non-visual biological perception unit specifically adopts a millimeter wave radar sensor, which has the core function of accurately capturing the non-visual biological signals and spatial position information of the user without relying on visual collection, can effectively detect whether the user is present, in a temporary absence state or a long-term absence state, and is not affected by environmental factors such as light brightness, obstacle blocking, etc., thereby providing stable and reliable basic data support for subsequent user state recognition; the posture and behavior perception unit selects a distributed pressure sensor array and / or a 3D-ToF sensor, wherein the distributed pressure sensor array can recognize posture information such as the sitting posture and hand operation position of the user, and behavior characteristics such as typing operation and writing action, by sensing the pressure distribution change of different areas of the table top, and the 3D-ToF sensor can obtain three-dimensional depth data by emitting and receiving infrared light, and capture the body movement trajectory of the user in real time, and then accurately recognize behavior patterns such as reading and writing, video conference or rest, and both can realize high-precision, multi-dimensional perception of the user's posture and behavior, thereby providing detailed data for the system to judge the current behavior needs of the user; the article perception unit adopts an RFID reader / writer or a strain gauge sensor, the RFID reader / writer can quickly identify the type, ownership and use state of the article by reading the RFID tag information attached to the article on the table top, and the strain gauge sensor can obtain information such as the weight, placement position and whether the article is moved by sensing its own deformation, and either of the two sensors can enable the system to clearly grasp the distribution and state of the articles on the table top, thereby providing accurate article data support for the subsequent resource execution module and ensuring that the resource allocation is more in line with the actual article use needs of the table top.

[0024] In the present application, the energy distribution unit supports the USBPD power distribution protocol.

[0025] In this embodiment, the energy distribution unit that undertakes the energy supply regulation function of the desktop electronic device has the core characteristics of high adaptability, dynamic regulation and high energy efficiency because it supports the USBPD power distribution protocol. The USBPD power distribution protocol is an advanced power transmission standard commonly used in the industry, which can realize wide-range power output adjustment of 5V / 0.5A to 20V / 5A and is compatible with multiple interface types such as USB-A and USB-C, providing a general solution for the power supply needs of diversified desktop electronic devices. Specifically, after supporting this protocol, the energy distribution unit can dynamically adjust the power distribution scheme in response to the device usage state feedback from the multi-modal sensing module under the unified scheduling of the host module. For example, when the user is working on the go and the laptop battery is low, the energy distribution unit can prioritize high-power output to the laptop for fast charging through the USBPD protocol. When only low-power peripherals such as wireless mice and keyboards need to be powered, the energy distribution unit automatically switches to low-power output to avoid energy waste in the fixed power supply mode. At the same time, the universality of the USBPD power distribution protocol can significantly reduce the number of desktop-specific adapters, reduce cable clutter, and improve the cleanliness and convenience of the desktop space. In addition, the protocol has built-in safety mechanisms such as over-voltage protection, over-current protection, short-circuit protection, and over-temperature protection. The energy distribution unit can monitor the voltage, current, and temperature state of the power supply loop in real time, and can quickly disconnect the faulty circuit or adjust the output parameters if an anomaly is detected, effectively protecting the desktop electronic device from damage and ensuring user safety. In terms of overall system operation, the support of the energy distribution unit for the USBPD power distribution protocol and the dynamic resource allocation method of the host module form a deep synergy, enabling the desktop energy supply to change from "passive fixed" to "active on-demand distribution", which not only meets the differentiated energy needs of users in different scenarios such as video conferencing, document processing, and device charging, but also significantly improves energy efficiency, reduces unnecessary energy consumption, and further highlights the advantages of the intelligent desktop management system in resource optimization and user experience improvement.

[0026] In the present application, the host module further includes a personalized learning unit for optimizing the rule parameters in the resource allocation strategy library based on the user's manual operation records through the human-computer interaction module.

[0027] In the main control module of the intelligent desktop management system in this embodiment, the newly added personalized learning unit serves as the core component for realizing adaptive optimization. Its core role is to build a closed-loop mechanism from "user feedback to policy iteration". Specifically, this unit continuously collects and stores various manual operation records initiated by users through the human-computer interaction module. These records include both direct adjustment instructions of the user on the resource execution module and system state information at the time of operation. Based on these records, the personalized learning unit accurately locates the inadequacy of rule parameters in the resource allocation strategy library by analyzing the differences between "user manual adjustment" and "automatic configuration before adjustment". For example, if most users tend to manually adjust the lighting brightness from the default 500 lux in the policy library to 700 lux in "reading mode", the unit will identify that the threshold parameter in the "reading mode and lighting brightness" rule needs to be optimized. Subsequently, the personalized learning unit will make targeted fine-tuning of these rule parameters, making the rules of the resource allocation strategy library more in line with users' actual usage habits and preferences. The direct effect of this optimization mechanism is that the system can automatically output resource allocation schemes that better meet users' expectations in similar scenarios without users repeating the same manual adjustments, significantly reducing user operation costs and improving the system's personalized service capabilities. At the same time, as users use the system for a longer period of time, the operation records accumulated by the personalized learning unit become increasingly rich, and the rule parameters of the resource allocation strategy library are continuously iteratively optimized, continuously improving the resource allocation accuracy of the intelligent desktop management system, and ultimately realizing the intelligent experience of "system evolving with user habits", further enhancing the adaptability of the system to user needs.

[0028] The embodiment also provides a dynamic resource allocation method, which is executed by a main control module such as an intelligent desktop management system, and includes the following steps: S1. Multi-source data perception and fusion step: acquiring and fusing heterogeneous sensor data streams from multi-modal perception modules to generate comprehensive perception data; S2. Multi-level identification of user state and intention step: based on the comprehensive perception data, sequentially performing presence state identification, behavior pattern identification, and external demand identification; S3. Dynamic resource mapping based on policy library step: taking the identified state, pattern, and demand as input, querying the resource allocation strategy library, and mapping to obtain the corresponding multi-resource collaborative allocation scheme; S4. Collaborative control and instruction execution step: compiling the resource allocation scheme into control instructions and sending them to the corresponding units in the resource execution module to drive their execution.

[0029] In this embodiment, the intelligent allocation of desktop resources is achieved through four closely connected steps. The multi-source data perception and fusion step is the basic link of the method, and its core role is to integrate various heterogeneous sensor data streams from multi-modal perception modules. Through data cleaning, spatio-temporal alignment and feature fusion, the dispersed and heterogeneous raw data is converted into unified and comprehensive comprehensive perception data, solving the problem of one-sidedness of single sensor data and providing complete and reliable data support for subsequent user state and intent recognition. The multi-level recognition step S2 of user state and intent is based on the comprehensive perception data generated by S1, and plays a key role through the progressive recognition logic of "from shallow to deep and from coarse to fine". First, the presence state is recognized through non-visual biological perception data, then the user behavior pattern is recognized in combination with posture and behavior perception data, and finally the external demand of the user is recognized by fusing the object perception data and environmental data. This multi-level recognition mechanism greatly improves the understanding accuracy of user state and real intent, and avoids misjudgment caused by single-dimensional recognition. The dynamic resource mapping step S3 based on the strategy library serves as a bridge connecting user intent and resource allocation, which takes the user state, behavior pattern and external demand recognized by S2 as input conditions, accurately queries the IF-THEN rules in the resource allocation strategy library, and maps to generate a collaborative allocation scheme covering display, lighting, energy, physical configuration and other dimensions through rule matching and logical reasoning, ensuring that resource allocation not only meets the current core needs of users, but also realizes the linkage optimization of multiple resources, avoiding the limitations of single resource adjustment. The collaborative control and instruction execution step S4 is the final landing link of the method, which compiles the multi-resource collaborative allocation scheme obtained by S3 into specific control instructions recognizable by each unit of the resource execution module, and accurately sends the instructions to the corresponding execution unit to drive it to perform the corresponding operation, thereby converting the abstract allocation scheme into actual desktop resource adjustment effect, and finally realizing the dynamic, intelligent and collaborative allocation of desktop resources according to user state and intent, improving the convenience, comfort and resource utilization efficiency of desktop use.

[0030] In the present application, the multi-level recognition step of user state and intent includes: (1) determining whether the user is in the presence, temporary leave or long-term leave state based on the millimeter wave radar signal; (2) identifying the user behavior pattern through a lightweight convolutional neural network model based on the posture signal, the behavior pattern including typing operation, reading and writing, video conference or rest; (3) identifying the user's peripheral device demand and lighting demand based on the object identification signal and the ambient light signal.

[0031] In this embodiment, the multi-level recognition of user state and intention is achieved through the progressive logic of "basic state judgment to behavior pattern analysis to specific demand mining", which realizes the accurate perception of user demand. The first step is to judge the user state based on the millimeter wave radar signal, which can detect human micro-movement, vital signs and other non-visual information, and is not affected by environmental factors such as light brightness, obstacle blocking, etc. It can accurately distinguish whether the user is present, temporarily away or permanently away. This judgment provides a basic premise for subsequent resource allocation and avoids the privacy risks of visual perception. The second step is to identify user behavior patterns based on posture signals with the help of a lightweight convolutional neural network model. Due to its small number of parameters and fast inference speed, the lightweight convolutional neural network model can adapt to the computing power requirements of the intelligent desktop management system master module, realize real-time processing, accurately identify typing operations, reading and writing, video conferencing or resting behavior patterns, and let the system clearly understand the user's current core behavior scene, providing key evidence for resource allocation direction. The third step is to identify the user's peripheral device demand and lighting demand based on the object identification signal and the ambient light signal. The object identification signal can help the system determine whether the user has connected or placed a peripheral device that needs to be powered / enabled, and then identify the peripheral device demand. The ambient light signal can intuitively reflect whether the current desktop light is suitable for user behavior, thereby accurately identifying the lighting demand. This step allows the system to go from "understanding behavior" to "meeting specific needs", ensuring that the subsequent resource allocation scheme can directly address the user's actual use pain points. Overall, this multi-level recognition step achieves accurate and real-time recognition results through phased and multi-dimensional perception and analysis, and realizes the depth of understanding from "knowing that" to "knowing why", providing comprehensive and accurate input information for the dynamic resource mapping step based on the strategy library, and laying a reasonable foundation for resource collaborative allocation.

[0032] In this application, the mapping rules in the resource allocation strategy library are IF-THEN rules, where the conditions of the IF part are the logical combination of multi-modal perception information, and the actions of the THEN part are the collaborative control instructions for multiple units in the resource execution module.

[0033] In the resource allocation strategy library of the intelligent desktop management system in this embodiment, the mapping rule for realizing the accurate association between user demand and resource configuration specifically adopts an IF-THEN rule. Due to the characteristics of clear logic and efficient matching, this rule structure can provide explicit and implementable decision basis for the main control module to execute dynamic resource allocation. The IF part serves as the triggering condition of the rule, and its core composition is the logical combination of multi-modal perception information. These multi-modal perception information covers various types of key data collected by the multi-modal perception module. Through logical relationships such as "and / or / not", these scattered perception information is integrated to accurately lock the specific scene where the user is currently located, avoid misjudgment caused by triggering rules with single perception information, and ensure that the rule triggering condition is highly consistent with the user's real use scene. The THEN part serves as the execution action of the rule, and its content is the cooperative control instruction for multiple units in the resource execution module, rather than the independent instruction for a single unit. For example, for the IF condition of the above video conference scene, the THEN part can generate the cooperative instruction "the lighting adjustment unit adjusts the brightness to 500 lux and the color temperature to 4500K, the display support unit adjusts the display angle to 15° of the elevation angle, and the energy distribution unit allocates 10W of power supply power for the external camera". This multi-unit cooperative control can avoid the experience fragmentation caused by single resource adjustment and realize the linkage adaptation of multi-dimensional resources such as display, lighting, and energy. Overall, the IF-THEN rule adopts the mode of "multi-condition logical combination triggering and multi-resource cooperative control response", which guarantees the scene pertinence of resource allocation and ensures the integrity and coordination of resource configuration. When the main control module executes the dynamic resource mapping step based on the strategy library, it can quickly query the resource allocation scheme matching the current scene, provide accurate and feasible operation basis for the subsequent cooperative control and instruction execution step, and finally improve the accuracy of resource allocation and the coherence of user experience of the intelligent desktop management system.

[0034] In the present application, after the cooperative control and instruction execution step, it further includes: Personalized learning and strategy optimization step: record the manual adjustment instruction issued by the user through the human-computer interaction module, constitute a training sample with the manual adjustment instruction and the system state before adjustment, and fine-tune and optimize the rule parameters in the resource allocation strategy library.

[0035] In this embodiment, after the cooperative control and instruction execution step completes the resource allocation operation, the newly added personalized learning and strategy optimization step serves as the closed-loop optimization link of the dynamic resource allocation method, and undertakes the core role of continuously adapting the system to the user's habits and improving the accuracy of resource allocation. This step first records the manual adjustment instructions actively issued by the user through the human-computer interaction module in real time. These instructions cover the user's autonomous correction operations on various resources on the desktop. These manual instructions directly reflect the user's supplementary needs and personalized preferences for the system's automatic configuration in the current scenario. Subsequently, this step integrates the recorded manual adjustment instructions and the system state before adjustment to form a complete training sample. This "scene state and user feedback" sample structure can accurately locate the deviation between the system's automatic configuration and the user's actual needs. Finally, based on these training samples, the personalized learning and strategy optimization step fine-tunes and optimizes the rule parameters in the resource allocation strategy library, avoiding the need for manual adjustment by the user in the same or similar scenarios in the future. From the overall effect, this step forms a complete closed loop of "execution, feedback, and iteration" through the logic of "recording feedback, building samples, and optimizing rules." This not only gradually eliminates the deviation between the system's automatic configuration and the user's habits, significantly reduces the frequency of repeated manual operations by the user, and improves the system's personalized service capabilities, but also, as the user's usage time increases and training samples accumulate, the rule parameters in the resource allocation strategy library are continuously iterated and upgraded, making the resource allocation of the intelligent desktop management system more and more in line with the user's usage preferences, ultimately achieving the intelligent experience of "the system evolving with the user's habits," and further enhancing the practicality and adaptability of the dynamic resource allocation method.

[0036] The embodiment also provides a computing device, including a memory and a processor, and the memory stores a computer program. When the computer program is executed by the processor, the dynamic resource allocation method is realized.

[0037] In this embodiment, the computing device provided as a hardware implementation carrier of the dynamic resource allocation method builds the core hardware architecture of the method through the memory and processor contained therein. The core role of the memory is to stably store the computer program for implementing the dynamic resource allocation method, ensuring that the program will not be lost or damaged due to power failure, restart, etc. during device operation, providing a reliable data storage foundation for the continuous calling and execution of the dynamic resource allocation method, and avoiding problems such as method failure to start or interrupted operation due to program loss. The processor, as the operation core of the computing device, undertakes the key responsibility of calling, parsing and executing the computer program in the memory. When the processor executes the computer program, it can rely on its computing power to strictly follow the step logic of the dynamic resource allocation method, i.e. sequentially complete the acquisition and integration of heterogeneous sensor data streams in the multi-source data perception and fusion step, the progressive judgment of presence state, behavior pattern and external demand in the multi-level identification step of user state and intent, the matching generation of resource allocation scheme in the dynamic resource mapping step based on the strategy library, and the compilation and sending of control instructions in the collaborative control and instruction execution step. If there is a personalized learning and strategy optimization step, it will also complete the manual adjustment instruction record, training sample construction and strategy library rule parameter optimization, converting abstract program code into specific hardware operation and system control instructions, ensuring that each step of the method can be efficiently and accurately advanced, and will not be delayed or deviated due to insufficient computing power. The computer program, as the core bridge connecting the hardware of the computing device and the logic of the dynamic resource allocation method, has the characteristics of "stored in the memory and executed in the processor", which enables the dynamic resource allocation method to be implemented in actual operation relying on the hardware entity of the computing device, such as driving the computing device and the multi-modal perception module of the intelligent desktop management system to establish data communication through the processor executing the program, real-time acquisition of heterogeneous data streams such as millimeter wave radar signal, pressure sensor data, RFID identification information and completion of fusion, then querying the IF-THEN rules in the resource allocation strategy library to generate a multi-resource collaborative allocation scheme, and finally sending control instructions to the corresponding unit of the resource execution module to drive it to perform specific adjustment operations. Overall, the computing device provides a stable and efficient hardware operating environment for the dynamic resource allocation method through the collaborative work of the memory, processor and computer program, not only enabling the dynamic resource allocation method to fall from the logical concept to the actual operation system function, avoiding the method remaining at the theoretical level, but also ensuring the real-time and accuracy of method execution relying on the computing performance of hardware, and thus ensuring that the intelligent desktop management system can successfully complete the dynamic and intelligent allocation of desktop resources with the computing device as the main control module of the hardware carrier, ultimately realizing the core effects of improving desktop resource utilization efficiency, adapting to user personalized needs and optimizing user experience.

[0038] The embodiment further provides a computer readable storage medium, which stores a computer program. The computer program is executed by a processor to implement the dynamic resource allocation method.

[0039] In this embodiment, the computing device provided as a hardware implementation carrier of the dynamic resource allocation method builds the core hardware architecture of the method through the memory and processor contained therein. The core role of the memory is to stably store the computer program for implementing the dynamic resource allocation method, ensuring that the program will not be lost or damaged due to power failure, restart, etc. during device operation, providing a reliable data storage foundation for the continuous calling and execution of the dynamic resource allocation method, and avoiding problems such as method failure to start or interrupted operation due to program loss. The processor, as the operation core of the computing device, undertakes the key responsibility of calling, parsing and executing the computer program in the memory. When the processor executes the computer program, it can rely on its computing power to strictly follow the step logic of the dynamic resource allocation method, i.e. sequentially complete the acquisition and integration of heterogeneous sensor data streams in the multi-source data perception and fusion step, the progressive judgment of presence state, behavior pattern and external demand in the multi-level identification step of user state and intent, the matching generation of resource allocation scheme in the dynamic resource mapping step based on the strategy library, and the compilation and sending of control instructions in the collaborative control and instruction execution step. If there is a personalized learning and strategy optimization step, it will also complete the manual adjustment instruction record, training sample construction and strategy library rule parameter optimization, converting abstract program code into specific hardware operation and system control instructions, ensuring that each step of the method can be efficiently and accurately advanced, and will not be delayed or deviated due to insufficient computing power. The computer program, as the core bridge connecting the hardware of the computing device and the logic of the dynamic resource allocation method, with the characteristics of "stored in the memory and executed in the processor", enables the dynamic resource allocation method to be implemented in actual operation relying on the hardware entity of the computing device, for example, by executing the program through the processor, the computing device and the multi-modal perception module of the intelligent desktop management system can establish data communication, real-time acquire heterogeneous data streams such as millimeter wave radar signal, pressure sensor data, RFID identification information and complete fusion, then identify the user's presence state and analyze the behavior pattern based on the integrated perception data, subsequently query the IF-THEN rules in the resource allocation strategy library to generate a multi-resource collaborative allocation scheme, and finally send control instructions to the corresponding unit of the resource execution module to drive it to perform specific adjustment operations. Overall, the computing device provides a stable and efficient hardware operating environment for the dynamic resource allocation method through the collaborative work of the memory, processor and computer program, not only enabling the dynamic resource allocation method to fall from the logical concept to the actual operation system function, avoiding the method remaining at the theoretical level, but also ensuring the real-time and accuracy of method execution by relying on the computing performance of the hardware, and thus ensuring that the intelligent desktop management system can successfully complete the dynamic and intelligent allocation of desktop resources with the computing device as the main control module hardware carrier, ultimately realizing the core effects of improving desktop resource utilization efficiency, adapting to user personalized needs and optimizing user experience.

[0040] The above merely provides the preferred embodiment of the present application, and the protection scope of the present application is not limited thereto, and any person skilled in the art, according to the technical scheme and the inventive concept of the present application, can make equivalent replacements or changes within the technical scope disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. An intelligent desktop management system, characterized by, Comprise: a master module, which is built-in with a resource allocation strategy library; a multi-modal perception module, which is in communication connection with the master module, for collecting user and environment data, including a non-visual biological perception unit, a posture and behavior perception unit, and an article perception unit; a resource execution module, which is in communication connection with the master module, for executing adjustment operations on desktop electronic and physical resources, including at least one of a display support unit, a lighting adjustment unit, an energy distribution unit, and a physical configuration unit; a human-computer interaction module, which is in communication connection with the master module; the master module is configured to execute a dynamic resource allocation method.

2. The system according to claim 1, wherein: the non-visual biological perception unit is a millimeter wave radar sensor; the posture and behavior perception unit is a distributed pressure sensor array and / or a 3D-ToF sensor; the article perception unit is an RFID reader / writer or a strain gauge sensor.

3. The system of claim 1, wherein, the energy distribution unit supports the USB PD power distribution protocol.

4. The system of claim 1, wherein, the master module further comprises a personalized learning unit for optimizing the rule parameters in the resource allocation strategy library according to the manual operation records of the user through the human-computer interaction module.

5. A method of dynamic resource allocation, characterized by, executed by the master module of the intelligent desktop management system according to any one of claims 1-4, comprising the following steps: S1. Multi-source data perception and fusion step: acquiring and fusing heterogeneous sensor data streams from the multi-modal perception module to generate comprehensive perception data; S2. Multi-level identification of user state and intention step: based on the comprehensive perception data, sequentially performing in-place state identification, behavior pattern identification, and external demand identification; S3. Dynamic resource mapping step based on the strategy library: taking the identified state, pattern, and demand as input, querying the resource allocation strategy library, and mapping to obtain the corresponding multi-resource collaborative allocation scheme; S4. Collaborative control and instruction execution step: compiling the resource allocation scheme into control instructions and sending them to the corresponding units in the resource execution module to drive their execution.

6. The dynamic resource allocation method of claim 5, wherein, The multi-level identification of user state and intention step comprises: (1) determining whether the user is in the in-place, temporary leave, or long-term leave state based on millimeter wave radar signals; (2) identifying user behavior patterns, including typing, reading and writing, video conferencing, or resting, through a lightweight convolutional neural network model based on posture signals; (3) identifying user peripheral device requirements and lighting requirements based on article identification signals and ambient light signals.

7. The dynamic resource allocation method of claim 5, wherein, The mapping rules in the resource allocation strategy library are IF-THEN rules, where the conditions in the IF part are the logical combinations of multi-modal perception information, and the actions in the THEN part are collaborative control instructions for multiple units in the resource execution module.

8. The dynamic resource allocation method of claim 5, wherein, After the collaborative control and instruction execution step, further comprising: a personalized learning and strategy optimization step: recording the manual adjustment instructions issued by the user through the human-computer interaction module, forming training samples with the system state before adjustment, and fine-tuning and optimizing the rule parameters in the resource allocation strategy library.

9. A computing device comprising a memory and a processor, said memory having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the dynamic resource allocation method as claimed in any one of claims 5 to 8.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the dynamic resource allocation method as claimed in any one of claims 5 to 8.

Images