Intelligent energy-saving control method and system for LED of Internet of Things based on artificial synaptic device

Abstract

The invention discloses an intelligent energy-saving control method and system for an LED of the Internet of Things based on an artificial synaptic device. The system comprises an infrared sensing unit, an LED lighting unit, an artificial synapse learning unit, a local control and decision unit, an LED driving and power adjusting unit and an Internet of Things communication module. The infrared sensing unit is used for identifying whether pedestrians or moving objects pass or not in the LED lighting area. The artificial synapse learning unit is composed of a deep ultraviolet LED light source and a gallium oxide-based deep ultraviolet artificial synapse device. The gallium oxide-based deep ultraviolet artificial synapse device is constructed by a Se-doped Ga2O3 low-dimensional structure, and has ultraviolet light sensing and neural-like learning and memory functions. By constructing a closed-loop self-learning control mechanism with an artificial synaptic device as a core, distribution information of pedestrian or moving object flow in each time period of each day is obtained, self-adaptive switching of LED illumination between efficient learning and low-power-consumption operation is achieved, and energy consumption overflow is effectively reduced.

Description

Technical Field

[0001] This invention belongs to the fields of semiconductor optoelectronic devices, neuromorphic devices, and Internet of Things (IoT) intelligent control technology, specifically an IoT-based LED intelligent energy-saving control method and system based on artificial synaptic devices. Background Technology

[0002] With the rapid development of IoT, smart building, and intelligent lighting technologies, LED lighting systems have been widely applied in public spaces, industrial plants, and smart buildings. How to further reduce energy consumption while meeting actual lighting needs has become a critical issue that intelligent lighting systems urgently need to address. However, existing LED lighting energy-saving control schemes mostly rely on preset time strategies, simple sensor triggers, or centralized digital algorithms, lacking the ability to continuously learn from environmental changes, human activity characteristics, and long-term usage patterns. This can easily lead to high-power lighting even in situations with few people or low usage frequency, resulting in lighting power redundancy and energy overflow.

[0003] In typical IoT lighting systems, front-end sensing units usually only have single functions such as light intensity detection or human presence detection. The collected data needs to be uploaded to a local controller or cloud platform via a communication module for centralized processing before control decisions are made. This system architecture, which separates "sensing-transmission-computation-control," not only increases communication and computing energy consumption but also makes it difficult to achieve real-time response and low-power operation in application scenarios with a large number of nodes or complex environments, thus hindering the further promotion of smart lighting systems.

[0004] Therefore, there is an urgent need for an intelligent device capable of simultaneously performing environmental perception, information accumulation, and decision support at the system's front end. This would enable the lighting system to learn autonomously and dynamically adjust based on the actual usage conditions within the area, thereby minimizing energy consumption while ensuring lighting comfort and safety. Meanwhile, traditional silicon-based semiconductor devices are gradually approaching their physical limits as feature sizes continue to shrink, facing bottlenecks in improving power consumption, parallel information processing capabilities, and intelligence levels. Integrating perception and computing functions at the device level to construct neuromorphic devices with brain-like learning and memory capabilities has become a crucial technological direction for overcoming these limitations.

[0005] Among numerous candidate materials, gallium oxide (Ga2O3), as an ultrawide bandgap semiconductor material with a bandgap of approximately 4.8 eV, naturally possesses solar-blind deep ultraviolet spectral selectivity and excellent resistance to visible light interference. When gallium oxide is constructed into low-dimensional nanowires and other structures, its high specific surface area and strong spatial confinement effect are beneficial for enhancing the capture, release, and storage behavior of photogenerated carriers, providing a good physical basis for constructing artificial synaptic devices with time-dependent responses to ultraviolet light stimulation. However, during the growth of low-dimensional gallium oxide nanostructures, defect states such as oxygen vacancies are inevitably introduced into the material's interior and surface. These defect states are important physical sources of photogenerated carrier capture and slow release, and under deep ultraviolet light stimulation, they can induce significant sustained photoconductivity and time-delay response effects, which are key foundations for realizing functions such as short-term memory and synaptic weight evolution in artificial synapses. However, due to the difficulty in precisely controlling the generation location, energy level distribution, and density of oxygen vacancies during traditional growth processes, the defect state distribution of different devices often exhibits significant dispersion. This results in problems such as obvious dark current fluctuations, significant differences in synaptic response amplitude and time constant, limited upper and lower limits of device performance, and insufficient long-term operational stability. These issues severely restrict the engineering application of artificial synaptic devices in IoT lighting and other application scenarios that require long-term stability and autonomous learning control.

[0006] Existing research primarily focuses on modulating synaptic behavior through device structure design, applied electric fields, or post-processing techniques. However, methods for in-situ control of defect states during material growth remain limited, making it difficult to achieve large-scale fabrication while simultaneously maintaining deep-ultraviolet selectivity, low power consumption, and stable synaptic plasticity. In particular, during plasma-enhanced chemical vapor deposition (PECVD) growth, mature and engineerable technical solutions are still lacking for synergistically controlling the distribution of defect states and carrier dynamics in gallium oxide nanostructures through doping to endow devices with stable, spontaneous learning and memory capabilities at the material level.

[0007] Based on this, this invention proposes a gallium oxide-based deep ultraviolet artificial synapse device and its fabrication method that incorporates selenium (Se) doping during PECVD growth. By synergistically regulating the oxygen vacancy concentration and carrier trapping-release process through selenium-related defect states, the device exhibits stable and tunable synaptic weight evolution characteristics under deep ultraviolet light stimulation, and achieves a controllable transition from short-term memory to long-term memory. Thus, at the device level, it possesses the ability to accumulate and characterize environmental ultraviolet information over time.

[0008] Furthermore, this artificial synaptic device is integrated into the front end of the IoT LED lighting system as a learning unit. It continuously senses and learns ultraviolet environmental information related to lighting usage within the area, such as the frequency of human activity, dwell time, and historical usage intensity. Through temporal correlation learning of this information, the artificial synaptic device autonomously establishes a mapping relationship between regional pedestrian traffic and lighting power demand. The learning results are then fed back to the local control unit, enabling adaptive switching between high-power and low-power energy-saving modes for LED lighting. This effectively reduces overall system energy consumption while meeting actual lighting needs, providing a new technical path for building low-power, high-reliability intelligent lighting and neuromorphic IoT systems. Summary of the Invention

[0009] The purpose of this invention is to provide an IoT LED intelligent energy-saving control system and method based on gallium oxide artificial synaptic devices. By introducing selenium (Se) doping during the material growth stage, the defect states and carrier dynamics in the gallium oxide low-dimensional nanostructure are modulated, enabling the device to exhibit stable and tunable synaptic weight evolution characteristics under deep ultraviolet light stimulation, and possessing the ability to controllably transition from short-term memory to long-term memory, thereby achieving front-end autonomous learning and closed-loop energy control. This invention uses artificial synaptic devices as the front-end learning unit of the LED lighting system, sensing signals from the illuminated area in real time and continuously accumulating information on pedestrian or moving object traffic distribution within the learning area. This forms a mapping relationship between LED output power and pedestrian or moving object traffic, enabling adaptive switching between low-power and high-power modes without the need for complex digital algorithms, significantly reducing overall system energy consumption and improving energy efficiency.

[0010] To achieve the above technical objectives, the present invention adopts the following technical solution:

[0011] I. Device Fabrication Method

[0012] This invention provides a gallium oxide-based deep ultraviolet artificial synapse device. By introducing selenium (Se) doping during the material growth stage, the defect state distribution and carrier dynamics in the low-dimensional gallium oxide nanostructure are controlled, enabling the device to exhibit stable and tunable synaptic weight evolution characteristics under deep ultraviolet light stimulation, and realizing the controllable transition from short-term memory to long-term memory, providing a physical basis for front-end autonomous learning.

[0013] Furthermore, the method for fabricating the gallium oxide-based deep ultraviolet artificial synapse device provided by the present invention specifically includes the following steps:

[0014] (1) Substrate pretreatment: Select a c-plane sapphire substrate, and ultrasonically clean it in acetone, anhydrous ethanol and deionized water for 5 to 10 minutes in sequence. After drying with high-purity nitrogen, dry it at 100 to 120 °C for 10 to 15 minutes to obtain a clean and uniform growth surface.

[0015] (2) Preparation of metal catalyst layer: A high-purity gold (Au, 99.999%) thin film with a thickness of about 5~20 nm is thermally evaporated on the substrate surface, followed by high-temperature annealing to form nanoparticles (diameter 20~50 nm), which provide uniform catalytic sites for the nucleation and directional growth of gallium oxide nanowires;

[0016] (3) Growth of Se-doped gallium oxide nanowire thin film: High-purity gallium metal and selenium doping source are placed in a quartz boat of PECVD system in proportion, oxygen (1~5 sccm) and inert gas (10~100 sccm) are introduced, the substrate temperature is 800~1000 ℃, the radio frequency power is 50~200 W, and the growth time is 30~120 minutes to achieve vertical directional growth of gallium oxide nanowires and uniform doping of selenium. By controlling the amount of selenium doping and the growth conditions, the oxygen vacancy and related defect state density can be precisely adjusted, thereby controlling the capture, release and postsynaptic current response of photogenerated carriers.

[0017] (4) Metal electrode fabrication: Ti / Au (metal electrode layer thickness 50~200nm) is magnetron sputtered on the nanowire thin film, and the electrode pattern is defined by photolithography or mask technology to realize the test electrode structure and electrical interface of the device. Then, low temperature annealing (200~300 ℃, 10~30 minutes) can be performed to improve the contact between the metal and the nanowire and improve the electrical performance of the device.

[0018] (5) Device packaging and testing: The prepared device is packaged to prevent the influence of humidity and contamination, and performance tests such as dark current, photoresponse, and postsynaptic current are performed to verify the artificial synaptic behavior and device consistency.

[0019] The gallium oxide-based solar-blind artificial synapse device obtained by the above preparation method possesses stable and tunable synaptic weight evolution characteristics. It can achieve a controllable transition from short-term memory to long-term memory under deep ultraviolet light stimulation, making it suitable for constructing the front-end learning and sensing unit of an IoT LED intelligent energy-saving control system. This invention achieves highly efficient energy-saving control of the lighting area through the device's front-end self-learning and closed-loop power regulation capabilities, balancing low power consumption, high reliability, and environmental adaptability. It is suitable for smart buildings, public spaces, industrial lighting, and neuromorphic IoT applications.

[0020] II. System Composition and Working Principle

[0021] Based on the aforementioned artificial synaptic device, the present invention further provides an IoT LED intelligent energy-saving control system, comprising:

[0022] Infrared sensing unit: used to identify whether a pedestrian or moving object is passing through the LED lighting area;

[0023] LED lighting unit: used to illuminate the target area;

[0024] Artificial Synaptic Learning Unit: Composed of a deep ultraviolet LED light source and gallium oxide-based deep ultraviolet artificial synaptic devices. It is used for deep ultraviolet light emission, perception, and neural-like learning and memory continuation, establishing a mapping relationship between LED output power and pedestrian or moving object traffic flow.

[0025] Local control and decision-making unit: Generates LED power adjustment commands based on artificial synaptic learning results, enabling high-power rapid learning and low-power steady-state mode switching;

[0026] LED driver and power regulation unit: used to adjust the LED lighting power according to control commands;

[0027] IoT communication module: used to upload system operating status and learning parameters to enable multi-node collaboration or remote management.

[0028] System Working Principle: The system uses an infrared sensing unit to detect pedestrian traffic or object movement within a target area in real time, feeding the signal back to an artificial synaptic learning unit. This unit continuously learns the temporal evolution of the ultraviolet light signal, autonomously establishing a mapping relationship between LED lighting power and pedestrian or moving object traffic in the target area. Based on this learning, the system forms a closed-loop self-learning energy-saving control mechanism. This mechanism can acquire pedestrian or object traffic distribution information at different times of the day and adaptively switch between a high-power rapid learning mode and a low-power steady-state operation mode. This mechanism adaptively switches between high-power and low-power modes based on changes in the number of people in the area, lighting needs, and historical usage patterns, without requiring complex digital algorithms. This effectively reduces overall system energy consumption and ensures long-term stable operation.

[0029] This invention provides an IoT-based intelligent energy-saving control method and system for LEDs using artificial synaptic devices. It offers the following advantages:

[0030] 1. This invention effectively modulates the defect state distribution and carrier dynamics within gallium oxide nanostructures by introducing selenium doping during the growth process. This artificial synaptic device exhibits stable and tunable conductance changes under deep ultraviolet light stimulation, simulating the weight evolution and memory transition characteristics of biological synapses. This simulation-based computational method, grounded in materials physics, enables the device to directly process environmental information at the front end. Furthermore, the device possesses excellent solar-blindness characteristics, effectively resisting interference from ambient visible light and ensuring the accuracy of signal sensing.

[0031] 2. This invention constructs a closed-loop adaptive control system based on artificial synaptic devices, without relying on complex digital logic algorithms or high-performance processors. The system utilizes the characteristic that the device's conductivity cumulatively changes with external stimuli to directly establish a mapping relationship between lighting power and ambient pedestrian density, achieving automatic switching between high-power rapid response and low-power steady-state operation modes. This hardware-level self-learning mechanism significantly reduces the system's computational energy consumption and response latency, solving the problem of strong dependence on preset rules in traditional control methods.

[0032] 3. This invention supports multi-node IoT collaborative control and is suitable for large-scale smart lighting scenarios. By sharing synaptic states through communication modules, adjacent nodes can collaboratively adjust lighting strategies, forming a dynamic lighting distribution that moves with the target. This distributed collaborative mechanism effectively reduces overall energy consumption while ensuring the brightness requirements of the actual lighting area, improving the reliability and environmental adaptability of the system in smart buildings and public spaces. Attached Figure Description

[0033] Figure 1 This is a flowchart illustrating the fabrication process of the gallium oxide-based deep ultraviolet artificial synapse device obtained by the method of this invention.

[0034] Figure 2 This is a schematic diagram of the structure of the gallium oxide-based deep ultraviolet artificial synapse device prepared by the method of the present invention.

[0035] Figure 3 The graph shows the PPF performance of the gallium oxide-based deep ultraviolet artificial synapse device prepared by the method of the present invention under a 10V bias voltage and 254nm illumination with a light intensity of 400μW / cm2.

[0036] Figure 4 The graph shows the learning and memory performance of the gallium oxide-based deep ultraviolet artificial synapse device prepared by the method of the present invention under a 10V bias voltage and 254nm light illumination with an intensity of 400μW / cm2.

[0037] Figure 5 This is a flowchart illustrating the IoT LED intelligent energy-saving control system described in this invention.

[0038] Figure 6 This is a schematic diagram of the structure of the IoT LED intelligent energy-saving control system described in this invention.

[0039] The components include: 1. Infrared sensing unit; 2. LED lighting unit; 3. Artificial synaptic learning unit; 4. Local control and decision-making unit; 5. LED driving and power regulation unit; and 6. Internet of Things communication module. Detailed Implementation

[0040] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0041] Please see the appendix Figure 1 To be continued Figure 6 This invention provides an IoT-based LED intelligent energy-saving control system and device fabrication method based on artificial synaptic devices.

[0042] Example 1: Fabrication of a selenium-doped gallium oxide-based deep ultraviolet artificial synapse device. First, the fabrication process of the selenium-doped gallium oxide-based deep ultraviolet artificial synapse device, which is the core component of the system, is described. The specific steps are as follows:

[0043] Step 1: Substrate Pretreatment. A sapphire substrate was selected as the growth substrate. The sapphire substrate was ultrasonically cleaned in acetone, anhydrous ethanol, and deionized water in sequence. Each cleaning session lasted 10 minutes. After cleaning, the substrate was dried with high-purity nitrogen gas for later use.

[0044] Step 2: Preparation of the metal catalyst layer. A high-purity gold film with a thickness of 3 nm to 6 nm was deposited on the surface of the pretreated sapphire substrate using a thermal evaporation process. Subsequently, the substrate with the deposited gold film was placed in a tube furnace and subjected to heating annealing under a nitrogen atmosphere at a temperature of 500°C to 600°C for 30 minutes. During this process, the gold film rearranged due to surface tension, forming discretely distributed metal nanoparticles, which serve as catalytic sites for the subsequent nucleation and directional growth of gallium oxide nanostructures.

[0045] Step 3: Growth of selenium-doped gallium oxide nanostructures. A plasma-enhanced chemical vapor deposition (PECVD) system was used. High-purity metallic gallium (Ga) and selenium powder (Se) were placed in quartz boats and positioned upstream of the reaction chamber. A mixture of oxygen and argon gas at a flow rate ratio of 20:80 was introduced into the reaction chamber. Growth was carried out at a reaction temperature of 850℃ to 950℃, a radio frequency power of 150W, and a growth time of 60 minutes. Gallium oxide nucleated and grew at the metal catalyst particles, forming a low-dimensional nanostructure film. During the growth process, selenium vapor was transported into the gallium oxide lattice by the carrier gas. The distribution of oxygen vacancies and carrier trapping behavior were controlled by selenium-related defect states, thereby achieving excellent carrier storage and release characteristics while maintaining the solar-blind deep ultraviolet spectral selectivity of gallium oxide.

[0046] Step 4: Electrode Fabrication. A 20nm / 80nm thick titanium / gold (Ti / Au) composite metal film is deposited on the surface of the grown selenium-doped gallium oxide nanostructure film using photolithography and magnetron sputtering. After a lift-off process, source and drain structures are formed, resulting in a gallium selenium-doped gallium oxide-based deep ultraviolet artificial synapse device.

[0047] See attached document Figure 2 The fabricated artificial synaptic device includes a substrate, a selenium-doped gallium oxide nanostructure functional layer grown on the substrate, and source / drain electrodes. (See attached diagram.) Figure 3 and attached Figure 4 The device's photoelectric performance was tested: under 10V bias and deep ultraviolet light pulse stimulation at a wavelength of 254nm, the device's conductivity reversibly changed with the intensity of the light stimulation, the irradiation time, and the number of repetitions. At a light intensity of 400μW / cm², the device's response current stabilized above 300nA and exhibited significant two-pulse facilitated (PPF) characteristics. Repeated ultraviolet stimulation caused the device's conductivity to gradually accumulate; after the stimulation was removed, the conductivity slowly decreased over time. This non-volatile change in conductivity based on light stimulation constitutes the physical basis for the system's continuous perception and learning of environmental information.

[0048] Example 2: Internet of Things LED Intelligent Energy-Saving Control System, refer to Appendix Figure 5 The IoT LED intelligent energy-saving control system described in this embodiment of the invention includes: an infrared sensing unit 1, an LED lighting unit 2, an artificial synaptic learning unit 3, a local control and decision-making unit 4, an LED driving and power adjustment unit 5, and an IoT communication module 6.

[0049] The specific connections and functions of each module are as follows: Infrared sensing unit 1 is used to monitor the presence of pedestrians or moving objects within the illuminated area in real time and output a trigger signal. LED lighting unit 2 is used to illuminate the target area. Artificial synapse learning unit 3 consists of a deep ultraviolet LED light source and the aforementioned selenium-doped gallium oxide-based deep ultraviolet artificial synapse device. The two are packaged or arranged using optical coupling, so that the light emitted by the deep ultraviolet LED light source illuminates the photosensitive area of ​​the artificial synapse device. This unit is used to establish a physical mapping relationship between LED output power and pedestrian or moving object flow. Local control and decision unit 4 is electrically connected to artificial synapse learning unit 3 and is used to read the conductivity value of the artificial synapse device and generate LED power adjustment commands based on the value. LED driving and power adjustment unit 5 is connected to local control and decision unit 4 and LED lighting unit 2 respectively and is used to execute adjustment commands. Internet of Things communication module 6 is used to upload system operating status and realize information interaction between multiple nodes.

[0050] See attached document Figure 6 The specific operation and control methods of this system are as follows:

[0051] When the infrared sensing unit 1 detects a pedestrian or moving object in the illuminated area, it triggers the activation of the deep ultraviolet LED light source in the artificial synapse learning unit 3, projecting a light pulse signal onto the gallium oxide-based deep ultraviolet artificial synapse device. Under the stimulation of the light signal, the internal carrier concentration of the artificial synapse device changes, resulting in a change in its conductivity.

[0052] The local control and decision-making unit 4 reads the current conductivity value of the artificial synaptic device (corresponding to the learning weight) in real time and compares it with the preset threshold range to perform hierarchical control of the LED lighting unit 2:

[0053] When there are few pedestrians or moving objects in the area, the deep ultraviolet LED light source has a low trigger frequency and the cumulative conductivity of the artificial synaptic device is small and below the first threshold. The local control and decision unit 4 controls the LED lighting unit 2 to be in a low-power energy-saving mode.

[0054] When the flow of pedestrians or moving objects in the area increases, the triggering frequency rises, causing the conductivity of the artificial synapse device to accumulate and increase. When the conductivity is between the first threshold and the second threshold, the system controls the LED lighting unit 2 to maintain normal lighting power.

[0055] When the accumulated conductivity value exceeds the second threshold, the area is determined to be densely populated, and the system controls LED lighting unit 2 to switch to high-power enhanced lighting mode.

[0056] When a pedestrian or moving object leaves the area, the infrared sensing unit 1 stops triggering, and the deep ultraviolet LED light source turns off. The conductivity of the artificial synaptic device decays naturally over time due to the recombination of internal charge carriers (forgetting characteristic). After detecting the decrease in conductivity, the local control and decision unit 4 gradually reduces the output power of the LED lighting unit 2 until it returns to a low-power steady-state mode.

[0057] This system is suitable for scenarios such as smart building corridors. In actual deployment, several lighting nodes can be arranged along the top or side walls of the corridor. Each lighting node can operate independently or form a collaborative network through the IoT communication module 6. In collaborative mode, adjacent lighting nodes share synaptic weight information. When people move along the corridor, the artificial synaptic learning units 3 of adjacent nodes are activated sequentially, making the LED lighting brightness of the entire corridor present a "following" lighting effect that gradually changes with the direction of pedestrian flow. While ensuring lighting needs, inactive areas automatically maintain low-power operation.

[0058] In this way, the system utilizes the physical memory characteristics (conductivity accumulation and decay) of selenium-doped gallium oxide material to perform integral calculations on the history of environmental interactions. It can achieve adaptive lighting control based on actual usage frequency at the edge without the need for pre-setting complex digital statistical algorithms or relying on cloud big data training.

[0059] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An IoT-based intelligent energy-saving control method for LEDs using artificial synaptic devices, characterized in that, Includes the following steps: S1. The infrared sensor unit monitors in real time whether there are pedestrians or moving objects in the illuminated area and outputs a trigger signal. S2. When a pedestrian or moving object is detected in the illuminated area, the deep ultraviolet LED light source in the artificial synapse learning unit is turned on, and a light pulse signal is projected onto the selenium-doped gallium oxide-based deep ultraviolet artificial synapse device. The light signal is used to stimulate and change the conductivity value of the selenium-doped gallium oxide-based deep ultraviolet artificial synapse device. S3. The local control and decision unit reads the current conductivity value of the selenium-doped gallium oxide-based deep ultraviolet artificial synapse device in real time. The local control and decision unit generates an LED power adjustment command based on the current conductivity value and drives the LED lighting unit to execute the corresponding power mode through the LED driver and power adjustment unit.

2. The IoT LED intelligent energy-saving control method based on artificial synaptic devices according to claim 1, characterized in that, The step of the local control and decision-making unit generating an LED power adjustment command based on the current conductivity value specifically includes: The current conductivity value is compared with a preset threshold range, which includes a first threshold and a second threshold, wherein the first threshold is less than the second threshold. When the current conductivity value is below the first threshold, it is determined that the pedestrian flow in the lighting area is low, and an instruction is generated to control the LED lighting unit to enter a low-power energy-saving mode. When the current conductivity value is between the first threshold and the second threshold, it is determined that the lighting area is in normal use, and an instruction is generated to control the LED lighting unit to maintain normal lighting power. When the current conductivity value is higher than the second threshold, it is determined that the pedestrian flow in the lighting area is dense, and an instruction is generated to control the LED lighting unit to switch to a high-power enhanced lighting mode.

3. The IoT LED intelligent energy-saving control method based on artificial synaptic devices according to claim 2, characterized in that, It also includes the following steps: When the infrared sensing unit does not detect a pedestrian or moving object in the illuminated area, it stops triggering the deep ultraviolet LED light source; the conductivity of the selenium-doped gallium oxide-based deep ultraviolet artificial synapse device decays naturally over time; when the local control and decision unit detects a decrease in the current conductivity, it controls the LED driver and power adjustment unit to gradually reduce the output power of the LED lighting unit until it returns to the low-power energy-saving mode.

4. The IoT LED intelligent energy-saving control method based on artificial synaptic devices according to claim 1, characterized in that, The IoT-based LED intelligent energy-saving control method is applied to a collaborative network composed of multiple lighting nodes. The method further includes: sharing synaptic weight information between adjacent lighting nodes through an IoT communication module; when a pedestrian or moving object moves along the area, the artificial synaptic learning units of adjacent lighting nodes are activated in sequence, so that the LED lighting units of adjacent lighting nodes present a brightness distribution that gradually changes with the direction of pedestrian flow.

5. The IoT-based LED intelligent energy-saving control method based on artificial synaptic devices according to claim 1, characterized in that, The selenium-doped gallium oxide-based deep ultraviolet artificial synapse device is prepared by a method including the following steps: S1. Select a sapphire substrate for cleaning and drying pretreatment; S2. A gold film with a thickness of 3 nm to 6 nm is deposited on the surface of the pretreated sapphire substrate using a thermal evaporation process, and then subjected to heating annealing in a nitrogen atmosphere to form discretely distributed metal nanoparticles. S3. Using the metal nanoparticles as catalytic sites, a selenium-doped gallium oxide nanostructure thin film is grown in the reaction chamber of a plasma-enhanced chemical vapor deposition system. S4. Source and drain structures are prepared on the surface of the selenium-doped gallium oxide nanostructure thin film.

6. The IoT LED intelligent energy-saving control method based on artificial synaptic devices according to claim 5, characterized in that, The annealing temperature for the heat annealing treatment is 500°C to 600°C, and the annealing time is 30 minutes. The specific steps for growing selenium-doped gallium oxide nanostructure thin films are as follows: placing metallic gallium and selenium powder in a quartz boat respectively, introducing a mixture of oxygen and argon gas with a flow rate ratio of 20:80 into the reaction chamber, and growing for 60 minutes at a reaction temperature of 850°C to 950°C and a radio frequency power of 150W.

7. The IoT LED intelligent energy-saving control method based on artificial synaptic devices according to claim 5, characterized in that, The specific steps for preparing the source and drain structures are as follows: a titanium / gold composite metal film with a thickness of 20nm / 80nm is deposited on the surface of the selenium-doped gallium oxide nanostructure film using photolithography and magnetron sputtering processes, and the source and drain structures are formed by a lift-off process.

8. The IoT LED intelligent energy-saving control method based on artificial synaptic devices according to claim 1, characterized in that, The deep ultraviolet LED light source in the artificial synaptic learning unit is arranged with the selenium-doped gallium oxide-based deep ultraviolet artificial synaptic device by optical coupling, and the emission wavelength of the deep ultraviolet LED light source is 254nm.

9. The IoT LED intelligent energy-saving control method based on artificial synaptic devices according to claim 1, characterized in that, The selenium-doped gallium oxide-based deep ultraviolet artificial synapse device utilizes the defect states introduced by selenium doping to regulate oxygen vacancy distribution and carrier trapping behavior. Under deep ultraviolet light pulse stimulation, it exhibits a double-pulse facilitating characteristic, achieving reversible accumulation and decay of conductivity.

10. The IoT LED intelligent energy-saving control system based on artificial synaptic devices according to any one of claims 1-9, characterized in that, include: Infrared sensing unit is used to monitor the presence of pedestrians or moving objects in the illuminated area in real time and output trigger signals; LED lighting units are used to illuminate the target area; An artificial synaptic learning unit includes a deep ultraviolet LED light source and a gallium selenide-doped gallium oxide-based deep ultraviolet artificial synaptic device. The deep ultraviolet LED light source is used to project light pulse signals onto the gallium selenide-doped gallium oxide-based deep ultraviolet artificial synaptic device according to the trigger signal. A local control and decision unit, electrically connected to the artificial synapse learning unit, is used to read the conductivity value of the selenium-doped gallium oxide-based deep ultraviolet artificial synapse device and generate a power adjustment command; an LED driver and power adjustment unit, connected to the local control and decision unit and the LED lighting unit respectively, is used to respond to the power adjustment command and adjust the power of the LED lighting unit. The IoT communication module is used to upload system operating status and enable information exchange between multiple nodes.

Images