AIoT-based building low-carbon edge computing intelligent device

Abstract

The utility model discloses a kind of building low-carbon edge computing intelligent devices based on AIoT, comprising: core processing module, including ARM multicore processor and AI acceleration chip, and the computing power of AI acceleration chip is not less than 3TOPS;Data acquisition module, including at least two-way RS485 interface, at least one-way RS232 interface, at least four-way analog input interface and gigabit ethernet port;Data acquisition module is connected core processing module by wiring;Multi-modal communication module, integrated with 4G / 5G and LoRa multimode communication;Multi-modal communication module is connected core processing module by PClePCIe interface;Power management module, by DC12V power input, simultaneously support PoE power supply can be through PoE power supply and DC 12V dual power input;Heat dissipation structure, through aluminum fin carries out fanless passive cooling. By integrating the AI processor of higher energy efficiency, multifunctional sensor interface and efficient energy management architecture, the collection of internal environment data of building, local execution AI optimization algorithm and real-time distribution equipment control instruction can be realized.

Description

Technical Field

[0001] This utility model relates to the field of building intelligence technology, specifically to a building low-carbon edge computing intelligent device based on AIoT. Background Technology

[0002] Existing building automation systems mostly employ centralized cloud processing architectures, which suffer from high data transmission latency, high bandwidth consumption, and insufficient privacy and security. Meanwhile, traditional edge computing devices generally have limitations in computing power, single interfaces, and poor scalability, making it difficult to meet the real-time acquisition, analysis, and AI inference needs of multi-source heterogeneous sensor data in building scenarios. Therefore, there is an urgent need for a high-performance, modular AIoT edge computing hardware device to support the localized deployment and real-time response of building low-carbon optimization control algorithms. Utility Model Content

[0003] The purpose of this utility model is to provide an AIoT-based low-carbon edge computing intelligent device for buildings, which integrates a high-performance AI chip and multiple sensor interfaces to solve the hardware bottleneck problems of real-time data acquisition, AI inference and energy-saving control in building scenarios, thereby improving the response speed and efficiency of low-carbon building management.

[0004] The technical solution of this utility model is as follows: A smart building low-carbon edge computing device based on AIoT, characterized in that it includes: The core processing module includes an ARM multi-core processor and an AI acceleration chip, and the computing power of the AI ​​acceleration chip is no less than 3 TOPS. The data acquisition module includes at least two RS485 interfaces, at least one RS232 interface, at least four analog input interfaces, and a gigabit Ethernet port; the data acquisition module is connected to the core processing module via a ribbon cable. The multi-modal communication module integrates 4G / 5G and LoRa multi-mode communication; the multi-modal communication module is connected to the core processing module via a PCIe interface. The power management module supports both DC 12V power input and PoE power supply, and can be powered by both PoE and DC 12V dual power inputs. The heat dissipation structure uses aluminum heat sinks for fanless passive cooling.

[0005] Through the aforementioned structure, the device integrates a highly energy-efficient AI processor, multi-functional sensor interfaces, and a high-efficiency energy management architecture. This enables it to collect data on the building's internal environment, execute AI optimization algorithms locally, and distribute equipment control commands in real time. The device is suitable for deployment in a building's computer room or equipment room, providing hardware support for low-carbon operation of the building's air conditioning and lighting systems. Its design emphasizes high response speed, modularity for future expansion, and reduced long-term operation and maintenance costs.

[0006] Furthermore, the data acquisition module also includes a USB 3.0 interface and a GPIO expansion interface for analog input. The data acquisition module connects external temperature and humidity sensors, current sensors, and voltage sensors through these interfaces.

[0007] Through the aforementioned structure and leveraging its multi-interface compatibility, this device can be directly and efficiently integrated with HVAC sensors and electrical monitoring devices, enabling collaborative operation with localized artificial intelligence algorithms. This ensures that data from different monitoring points can be uniformly collected, processed, and analyzed in depth using AI algorithms to optimize building environment control and energy management strategies.

[0008] Furthermore, the heat dissipation structure has wavy aluminum heat dissipation fins inside its outer shell, and these fins are in direct contact with the ARM multi-core processor and AI acceleration chip of the core processing module via thermal paste.

[0009] Furthermore, the multimodal communication module also integrates a Wi-Fi 6 wireless communication unit for short-range interaction with the local control system.

[0010] Furthermore, the power management module typically consumes ≤20W and operates in a temperature range of -20℃ to 80℃.

[0011] Furthermore, it also includes a storage module with a built-in eMMC memory of no less than 16GB for caching sensor data and pre-built AI models.

[0012] Through the above structure, the storage module incorporates lightweight AI models (such as lightweight time-series prediction models and closed-loop control algorithms) to achieve millisecond-level closed-loop control response. The AI ​​inference algorithm is deployed directly on the edge computing hardware device, allowing for real-time analysis and decision-making on-site without uploading data to the cloud.

[0013] Furthermore, the AI ​​acceleration chip supports localized inference using TensorFlow Lite and PyTorch Mobile frameworks.

[0014] The above structure employs a CPU+AI coprocessor heterogeneous computing architecture, allocating tasks to dedicated computing units on demand, thus reducing overall device energy consumption. It supports local real-time inference and optimization control of complex AI algorithms, meeting the requirements of building low-carbon control for computational efficiency and real-time performance.

[0015] Furthermore, the device measures 220mm × 150mm × 50mm, and the outer casing has standard guide rail slots and wall-mounting holes on its sidewalls.

[0016] Furthermore, the core processing module has at least one reserved space for expanding other communication modules, storage modules, or computing power modules. The data acquisition module also includes an expansion interface: at least one M.2 slot for connecting external computing power modules or sensor adapter boards.

[0017] The above structure provides an M.2 slot and a USB 3.0 interface, supporting the connection of external sensors or the upgrading of computing modules.

[0018] Furthermore, the power management module employs dynamic voltage and frequency regulation technology.

[0019] Compared with existing technologies, the beneficial effects of this utility model are: Energy efficiency improvement: Reduces cloud dependence and lowers communication energy consumption through localized AI inference; Real-time optimization: Supports millisecond-level issuance of control commands for building air conditioning systems; Flexible deployment: The device measures 220mm×150mm×50mm, supports wall mounting or rack mounting, supports external connection of multiple types of sensors, and is adaptable to both new and old building renovation scenarios. Attached Figure Description

[0020] Figure 1 This is a hardware structure block diagram of the device in this application. Detailed Implementation

[0021] It should be noted that relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0022] The features and performance of this utility model will be further described in detail below with reference to the embodiments.

[0023] Please see Figure 1 A smart building low-carbon edge computing device based on AIoT, comprising: The core processing module includes an ARM multi-core processor and an AI acceleration chip, with the AI ​​acceleration chip having a computing power of no less than 3 TOPS. The AI ​​acceleration chip supports TensorFlow Lite and PyTorch Mobile frameworks for localized inference. The core processing module has at least one reserved M.2 slot for expanding other communication modules, storage modules, or computing power modules.

[0024] The data acquisition module includes at least two RS485 interfaces, at least one RS232 interface, at least four analog input interfaces, and a Gigabit Ethernet port. The data acquisition module connects to the core processing module via a ribbon cable. The module also includes a USB 3.0 interface, and the analog input interfaces are GPIO expansion interfaces. Temperature and humidity sensors, current sensors, and voltage sensors can be externally connected to the data acquisition module via the RS485 interface, RS232 interface, Gigabit Ethernet port, USB 3.0 interface, or GPIO expansion interface. The data acquisition module also includes an expansion interface: at least one M.2 slot for connecting an external computing module or sensor adapter board.

[0025] The multimodal communication module integrates 4G / 5G and LoRa multimodal communication; the multimodal communication module is connected to the core processing module through a PCIe / PCIe interface; the multimodal communication module also integrates a Wi-Fi 6 wireless communication unit for short-range interaction with the local control system.

[0026] The storage module has a built-in eMMC memory of no less than 16GB for caching sensor data and pre-installed AI models; the eMMC is pre-installed with a Linux operating system and deploys sensor drivers, AI model libraries and MQTT communication protocol stack.

[0027] The power management module accepts a DC 12V power input and also supports PoE power supply, allowing for dual power input via both PoE and DC 12V. The typical power consumption of the device using this module is ≤20W, and its operating temperature range is -20℃ to 80℃. The power management module employs dynamic voltage and frequency regulation technology.

[0028] The heat dissipation structure utilizes aluminum heat sinks for fanless passive cooling. The interior of the heat dissipation structure features wave-shaped aluminum heat sink fins, which directly contact the ARM multi-core processor and AI acceleration chip of the core processing module via thermal paste.

[0029] The device measures 220mm × 150mm × 50mm, and the side wall of the outer casing is equipped with standard guide rail slots and wall mounting holes.

[0030] By employing edge computing hardware, the localized deployment of building low-carbon optimization control algorithms has been achieved. This device, through its built-in AI model, supports direct on-site data processing and decision-making, eliminating the intermediate step of transmitting data to the cloud for processing. This significantly reduces data transmission latency and network bandwidth consumption, enabling rapid issuance of low-carbon optimization control commands. Simultaneously, data processing and storage locally avoids the transmission of sensitive information over public networks, effectively reducing the risk of data leakage and enhancing privacy protection. Therefore, this device not only solves the problems of high data transmission latency and high bandwidth consumption but also significantly improves data security and privacy.

[0031] The embodiments described above merely illustrate specific implementation methods of this application, and while the descriptions are detailed and specific, they should not be construed as limiting the scope of protection of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the technical solution of this application, and these modifications and improvements all fall within the scope of protection of this application.

Claims

1. A smart building low-carbon edge computing device based on AIoT, characterized in that, include: The core processing module includes an ARM multi-core processor and an AI acceleration chip, and the computing power of the AI ​​acceleration chip is no less than 3 TOPS. The data acquisition module includes at least two RS485 interfaces, at least one RS232 interface, at least four analog input interfaces, and a gigabit Ethernet port; the data acquisition module is connected to the core processing module via a ribbon cable. The multi-modal communication module integrates 4G / 5G and LoRa multi-mode communication; the multi-modal communication module is connected to the core processing module via a PCIe interface. The power management module supports both DC 12V power input and PoE power supply, and can be powered by both PoE and DC 12V dual power inputs. The heat dissipation structure uses aluminum heat sinks for fanless passive cooling.

2. The building low-carbon edge computing intelligent device based on AIoT according to claim 1, characterized in that, The data acquisition module also includes a USB 3.0 interface and a GPIO expansion interface. The analog input interface is a GPIO expansion interface. The data acquisition module can connect to external temperature and humidity sensors, current sensors, and voltage sensors through the USB 3.0 interface and the GPIO expansion interface.

3. The building low-carbon edge computing intelligent device based on AIoT according to claim 1, characterized in that, The heat dissipation structure has a wavy aluminum heat dissipation fin inside its outer shell. The aluminum heat dissipation fin is in direct contact with the ARM multi-core processor and AI acceleration chip of the core processing module through thermal paste.

4. The AIoT-based low-carbon edge computing intelligent device for buildings according to claim 1, characterized in that, The multimodal communication module also integrates a Wi-Fi 6 wireless communication unit for short-range interaction with the local control system.

5. The AIoT-based low-carbon edge computing intelligent device for buildings according to claim 1, characterized in that, The power management module typically consumes ≤20W during operation and operates in a temperature range of -20℃ to 80℃.

6. The building low-carbon edge computing intelligent device based on AIoT according to claim 1, characterized in that, It also includes a storage module with at least 16GB of built-in eMMC memory for caching sensor data and pre-built AI models.

7. The AIoT-based low-carbon edge computing intelligent device for buildings according to claim 1, characterized in that, The AI ​​acceleration chip supports localized inference using TensorFlow Lite and PyTorch Mobile frameworks.

8. The building low-carbon edge computing intelligent device based on AIoT according to claim 1, characterized in that, The device measures 220mm × 150mm × 50mm, and its outer casing has standard guide rail slots and wall-mounting holes on its side walls.

9. A smart building low-carbon edge computing device based on AIoT according to claim 1, characterized in that, The core processing module has at least one reserved slot for expanding other communication modules, storage modules, or computing power modules. The data acquisition module also includes an expansion interface: at least one M.2 slot for connecting external computing power modules or sensor adapter boards.

10. A smart building low-carbon edge computing device based on AIoT according to claim 1, characterized in that, The power management module employs dynamic voltage and frequency regulation technology.

Images