Sensor experiment system and method based on artificial intelligence

Abstract

The invention provides a sensor experiment system and method based on artificial intelligence, and belongs to the technical field of experiment teaching equipment. The intelligent system comprises a display module and a decision module; the operation system comprises a tracking trolley and a plurality of independent sensor modules, the top of the tracking trolley is provided with an adapter plate, and the sensor modules are connected with the tracking trolley through the adapter plate; the display module is used for acquiring an experiment command and displaying an experiment scheme, so that an experimenter can perform an experiment according to the displayed experiment scheme; and the decision module is used for receiving an experiment command and generating an experiment scheme based on the tracking trolley and the plurality of independent sensor modules, so that an experimenter can operate an experiment. Through modular hardware design and a multi-sensor fusion algorithm, the problems that a traditional experiment system is poor in flexibility and low in intelligent degree are solved, and the teaching efficiency and scientific research innovation space are remarkably improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of experimental teaching equipment, and in particular relates to a sensor experiment system and method based on artificial intelligence. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] To cultivate students' ability to solve complex engineering problems and foster innovative thinking, sensor-related experiments are essential for achieving these goals. However, sensor experimentation is limited. Existing sensor experiments involve students experimenting with individual sensors in a lab box, passively following fixed steps. This lacks in-depth exploration of sensor design principles, complex system integration, and real-world scenarios.

[0004] With the rapid development of artificial intelligence, existing sensor-based experimental systems face multiple technical bottlenecks. First, data processing in existing sensor-based experimental systems relies on manual labor, which is time-consuming, labor-intensive, and slow. Second, existing experimental systems suffer from limited experimental equipment, single-function equipment, rigid scenarios, and fixed experimental procedures, which restrict students' ability to independently design experiments. Summary of the Invention

[0005] To overcome the shortcomings of the above-mentioned existing technologies, the present invention proposes an artificial intelligence-based sensor experiment system and method. Through modular hardware design and multi-sensor fusion algorithms, it solves the problems of poor flexibility and low intelligence of traditional experimental systems, significantly improving teaching efficiency and scientific research innovation space.

[0006] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions: In the first aspect, a sensor experiment system based on artificial intelligence is disclosed, including an intelligent system and an operating system; The intelligent system includes a display module and a decision module; the operating system includes a tracking car and several independent sensor modules. An adapter plate is installed on the top of the tracking car, and the sensor module is connected to the tracking car through the adapter plate; The display module is used to obtain experimental commands and display experimental plans, so that the experimenter can conduct experiments according to the displayed experimental plans; The decision-making module is used to receive experimental commands and use a data parsing unit to convert the experimental commands into experimental parameter vectors. The data parsing unit includes constructing a semantic parsing framework based on an attention mechanism; using a solution generation unit based on reinforcement learning to construct an experimental solution; using a verification and error correction unit to perform logical verification on the generated experimental solution, and generating an experimental solution based on a tracking car and several independent sensor modules for experimenters to operate the experiment.

[0007] In a second aspect, a sensor experiment method based on artificial intelligence is disclosed, comprising: Step S1: Initialize the experimental system. The intelligent system obtains the user's experimental command through the display module or voice processing module. Step S2: The decision module of the intelligent system generates an experimental plan according to the experimental command through a pre-trained artificial intelligence mapping model; The experimental plan includes obstacle avoidance and tracking experiments, intelligent fire extinguishing experiments, intelligent transportation experiments and intelligent search and rescue experiments.

[0008] Compared with the prior art, the present invention has the following beneficial effects: The present invention recommends experimental plans through AI algorithms, dynamically optimizes the experimental process, and makes the experimental process more reasonable; by integrating multiple types of sensors, such as temperature, humidity, images, sound, etc., multimodal data fusion realizes cross-modal analysis; the present invention connects the intelligent system to the artificial intelligence big model, improves teaching effect, optimizes experimental design, enhances data analysis capabilities, and provides students with a personalized learning experience; the sensor module of the present invention modularizes the learned experiments, and combines the software layer to automatically generate sensor-algorithm-actuator combination solutions according to the experimental objectives, associates sensor characteristics, experimental scenarios and algorithm libraries, provides intelligent recommendations, supports flexible combinations of sensors, algorithms, and actuators, and meets the needs of different scenarios.

[0009] Advantages of additional aspects of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The accompanying drawings, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0011] Figure 1 A schematic diagram of the system structure of an artificial intelligence-based sensor experiment system provided in Example 1 of the present invention; Figure 2 A schematic diagram of the intelligent system structure of the artificial intelligence-based sensor experiment system provided in Example 1 of the present invention; Figure 3A schematic diagram of the structure of a tracking car in an artificial intelligence-based sensor experimental system provided in Example 1 of the present invention; Figure 4 Schematic diagram of the tracking car adapter plate of the artificial intelligence-based sensor experiment system provided in Example 1 of the present invention; Figure 5 This is a structural diagram of the tracking vehicle system provided in Example 1 of the present invention; Figure 6 A schematic diagram of a tracking circuit for a tracking vehicle according to the first embodiment of the present invention; Figure 7 This is a flow chart of the tracking circuit of the tracking car provided in the first embodiment of the present invention; Figure 8 A schematic diagram of a driving circuit for a tracking vehicle according to a first embodiment of the present invention; Figure 9 This is a schematic diagram of the Bluetooth module circuit of the tracking car provided in Example 1 of the present invention; Figure 10 This is a flow chart of the automatic recharging function of the tracking car provided in Example 1 of the present invention; Figure 11 This is a flow chart of the obstacle avoidance function of the tracking car provided in Example 1 of the present invention; Figure 12 A circuit diagram of a vibration sensor module provided in Example 1 of the present invention; Figure 13 A schematic diagram of a temperature and humidity sensor module circuit according to the first embodiment of the present invention; Figure 14 A circuit diagram of a sound sensor module provided in Example 1 of the present invention; Figure 15 This is a circuit diagram of a digital temperature sensor module provided in Example 1 of the present invention; Figure 16 This is a circuit diagram of an ultrasonic sensor module provided in Example 1 of the present invention; Figure 17 A schematic diagram of the infrared sensor module circuit provided in Example 1 of the present invention; Figure 18 A schematic diagram of a flame sensor module circuit according to the first embodiment of the present invention; Figure 19 A flow chart of the artificial intelligence-based sensor experiment method provided in Example 2 of the present invention; Figure 20 This is a schematic diagram of the structure of the intelligent fire extinguishing experiment tracking vehicle system provided in the second embodiment of the present invention; Figure 21 A schematic diagram of the fire extinguishing process of the intelligent fire extinguishing experiment provided in the second embodiment of the present invention; Figure 22A schematic diagram of the operation flow of the fire extinguishing unit of the intelligent fire extinguishing experiment provided in the second embodiment of the present invention; Figure 23 This is a schematic diagram of the structure of the tracking vehicle system for intelligent handling experiments provided in Example 2 of the present invention; Figure 24 A schematic diagram of the transport process of the intelligent transport experiment provided in the second embodiment of the present invention; Figure 25 This is a schematic diagram of the structure of the intelligent search and rescue experiment tracking vehicle system provided in Example 2 of the present invention; Figure 26 This is a schematic diagram of the search and rescue process of the intelligent search and rescue experiment provided in Example 2 of the present invention.

[0012] In the figure: 1. Tracking module; 2. Obstacle avoidance module; 3. Adapter board; 4. STM32 main control board; 5. Driver module. DETAILED DESCRIPTION

[0013] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise indicated, all technical and scientific terms used herein have the same meanings as those generally understood by those of ordinary skill in the art to which the present invention belongs. It should be noted that the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit the exemplary embodiments according to the present invention. In the absence of conflict, the features in the embodiments of the present invention and the embodiments may be combined with each other.

[0014] Example 1 In one or more embodiments, a sensor experiment system based on artificial intelligence is disclosed, such as Figure 1 As shown, including intelligent system and operating system; Intelligent systems, such as Figure 2 As shown in the figure, it includes display module, voice processing module and decision module; the operating system includes tracking car and sensor module. The display module, voice processing module and decision module constitute the intelligent system, such as Figure 3-Figure 4As shown, an adapter plate is installed on the top of the tracking car, through which the sensor module is connected to the tracking car. The adapter plate includes several sensor module interfaces. The display module and voice processing module are used to obtain experimental commands, which are transmitted to the decision-making module. Based on the experimental commands, the decision-making module uses a pre-trained artificial intelligence mapping model to generate and optimize the experimental plan for the experimenter to operate the experiment, thereby achieving dynamic experimental control and process scheduling. The display module includes a touch screen for user interaction. The user enters the experimental command by touch, and the system displays the corresponding plan on the display module. The voice processing module performs voice input and feature extraction, and the extracted signal is also input to the decision-making module. The experimental command entered by the user by touch and the signal extracted from the voice signal are both input to the decision-making module as the text sequence of the input command. The decision-making module uses the artificial intelligence mapping model to generate experimental plan data based on the experimental command data for the experimenter to operate the experiment. The decision-making module includes a data parsing unit, a plan generation unit, a verification and error correction unit, and a data storage and management unit. The data parsing unit includes a semantic parsing framework based on the attention mechanism to convert the experimental command into an experimental parameter vector.

[0015] Specifically, the input command is a text sequence , generate word vector matrix through BERT pre-training model ,in, represents the set of real numbers, n is the length of the input text sequence, d is the vector dimension, W=BERT (S) In this embodiment, the last hidden state of the BERT model Finally, a self-attention layer is added to specifically handle long-distance dependencies in experimental commands. The attention mechanism Attn is introduced to calculate feature weights, highlighting keywords such as experiment type and parameters. The self-attention expression is:

[0016] Where, is the weight matrix, b a is the bias term, 、 is the input text sequence.

[0017] The eigenvector is expressed as:

[0018] Where, is the self-attention weight. The feature vector is then mapped to the experimental parameter vector F through linear transformation.

[0019] The solution generation unit uses a strategy based on reinforcement learning to build experimental solution generation. According to the key information extracted by the data analysis unit, it matches the corresponding template and combines the built-in algorithm model to generate detailed and operational experimental solution data. Assume that the template library is , each template corresponds to a feature vector , match the optimal template by cosine similarity: , matching template , where arg max is a mathematical formula used to find the input parameter that can make the function reach the maximum value. Then optimize the experimental process and define the sequence of experimental steps ,in, For the experimental steps 1 to 1, construct the objective function to minimize the experimental cost ,in, is the resource consumption of step t, is the risk factor, For time cost, , , The weight parameter is used to generate a preliminary experimental plan for the experimental steps through the reinforcement learning model. Furthermore, the Q-learning algorithm is introduced to dynamically optimize the preliminary experimental plan and output the adjusted execution plan. The state s is the current experimental progress, the action a is the next step selection, and the reward function R combines the experimental efficiency and accuracy to adjust the experimental plan:

[0020] Where η is the learning rate, is the discount factor, r is the immediate reward, is the expected long-term cumulative reward for selecting action a in state s, For the next state Select the next action The expected long-term cumulative rewards that can be obtained, For the best value.

[0021] The verification and error correction unit is used to perform logical verification on the generated experimental plan data and check the rationality of the experimental steps. The probability graph model is used to quantify the risk of parameter conflict and the dynamic threshold is combined to achieve adaptive error correction. First, the parameter rationality is verified: define the safety constraint set , the verification function is , where || is an indicator function, which is 1 when the condition is met and 0 otherwise. is a parameter within the safe set. When the parameters conflict, the posterior probability of the corrected parameters is calculated based on Bayesian inference:

[0022] Where, is the posterior probability of the corrected parameter, P' is the corrected parameter, P(Error|P') is the error likelihood, and P(P') is the parameter prior distribution. Then, based on the historical experimental data based on the exponential moving average Adaptively adjust the verification threshold, the expression is:

[0023] Where, is the verification threshold, β is the smoothing coefficient, is the previous verification threshold, is the optimal threshold for the current experiment.

[0024] In this embodiment, the safety constraint set defines multiple constraints, including experimental step timing constraints (which regulate the order of experimental operations) and equipment capacity constraints (which set limits based on the performance parameters of experimental equipment). Based on the safety constraint set, the verification and error correction unit performs full-process verification and correction on the experimental plan data output by the plan generation unit.

[0025] During the initial verification phase, the experimental plan data output by the plan generation unit is received and the steps and other information in the plan are matched with the conditions in the safety constraint set. Logical conflict detection checks for timing inconsistencies between experimental steps. Furthermore, an automatic error correction mechanism addresses issues discovered during verification based on built-in error correction rules. Finally, feedback optimization is performed, feeding the corrected experimental plan data back to the plan generation unit. The data from this verification and error correction process is recorded and used to update the dynamic thresholds of the safety constraint set, optimize the template library and algorithm model of the plan generation unit, and achieve continuous optimization of the system.

[0026] The data storage and management unit includes constructing a knowledge graph of experimental data and realizing efficient retrieval and association analysis through graph embedding algorithms. The experimental data is represented as a triple G=(Entity,Relation,Attribute) , Entity The key object in the experimental plan, Relation For the connection between experimental items, Attribute For the characteristics of the experimental objects themselves, a node embedding vector is generated through a graph convolutional network (GCN):

[0027] Where, N(u) For nodes v The neighbor set of c uv is the normalization coefficient, l is the number of network layers, is the activation function, u is the neighbor node, For the The weight matrix of the layer graph convolution, For the The embedding vector of the neighbor node u at the layer.

[0028] This embodiment uses graph embedding to calculate the semantic similarity between experimental commands and historical cases, retrieving the most relevant historical experimental plans. When the knowledge graph detects a highly similar subgraph structure between a new experimental plan and a historical failure case, a risk warning is triggered. The similarity threshold for triggering the warning is pre-set. When there are insufficient cases in the target domain, mature operational procedures from other domains can be migrated through the domain associations of the knowledge graph.

[0029] The data parsing unit in the decision-making module utilizes pre-trained language models such as BERT combined with an attention mechanism to accurately capture key parameters, operational steps, and sequential logic in experimental requirements. This effectively avoids the reliance on manual interpretation of experimental requirements in traditional large-scale model-based experimental design, which can easily lead to information omissions or misunderstandings due to subjective factors. The solution generation module integrates template matching and reinforcement learning. Template matching ensures that the generated solution has a basic experimental process framework and that the components used in the generated solution are all available in the equipment. Reinforcement learning dynamically adjusts the solution parameters and step sequence based on real-time experimental feedback. Existing large-scale model-based solution generation methods are mostly based on fixed templates and lack adaptability to new experimental scenarios and requirements. The verification and error correction unit constructs a dynamic safety constraint set that not only verifies individual parameters but also comprehensively verifies parameter combinations and experimental step sequence, ensuring that the experimental solution is more compatible with the experimental system. Conventional large-scale model-based solutions lack effective verification mechanisms, making it impossible to evaluate the rationality and safety of the solution. The data and storage unit utilizes knowledge graph technology to store experimental data in a structured format, clearly displaying the complex connections between entities, relationships, and attributes. However, existing data storage technologies can only perform simple storage and cannot effectively organize and manage data.

[0030] The various units in the decision-making module work together to comprehensively ensure the accuracy of experimental plans, from the precise analysis of experimental requirements to the rationality of plan generation, the comprehensiveness of verification and error correction, and the effective utilization of historical experience through data storage and management. Compared with traditional plans and ordinary large models, this system can more accurately understand experimental intent and generate error-free experimental plans that meet actual needs, greatly increasing the probability of experimental success. During the experiment, users can simply provide simple instructions such as "I want to do *** experiment" or "I want to study *** sensor." The system will generate an error-free experimental plan based on the keywords in the instructions, which meets the hardware requirements of the equipment. After the experiment, the experimental plan is analyzed and optimized. Because different hardware requires different connections and code between components, a correct experimental plan can greatly improve the success rate of the experiment.

[0031] The tracking car also includes a control system and a power module. The control system includes a tracking module, an obstacle avoidance module, a communication module, a drive module, and an automatic recharge module. When executing the existing modules, the tracking car uses the existing control algorithm. In this embodiment, Figure 5 As shown in the figure, the control system of the tracking car adopts the SMT32 microcontroller.

[0032] Specifically, such as Figure 6 As shown, the tracking module includes an infrared tracking sensor and adopts a four-way infrared tracking sensor, specifically four infrared transmitting / receiving sensors. Each infrared transmitting / receiving sensor can transmit and receive at the same time. Four are placed in a row in front of the front of the car to realize the tracking function. The detection distance is 0 to 2~3cm. The circuit consists of four infrared receivers U7U10 and two LM393 comparators U11 and U12. Pin 1 of U7U10 is connected to GND, pin 3 is connected to 3.3V power supply, pin 2 is connected to the in-phase input of U11 and U12 (pin 3 or 5) after the respective pull-down resistors, and pin 4 is connected to the inverting input of the comparator (pin 2 or 6) through a current-limiting resistor; the output of the comparator (pin 1 or 7) is connected to the negative pole of LED1~LED4 respectively, and the positive pole is connected to the power supply, and the current is controlled by the current-limiting resistor; when the infrared signal reflection is received, the comparator output level changes, and the corresponding LED is lit to realize the detection indication of the reflection status. Each channel works independently and is suitable for scenarios such as path recognition. The working process of the tracking module is as follows Figure 7As shown, the infrared transmitter generates an infrared beam that illuminates a track on the ground. The infrared receiver receives the infrared light reflected from the ground. As the vehicle moves, black lines or other high-contrast tracks on the ground absorb and reflect the infrared light more effectively. The infrared receiver in the tracking module detects this reflection, while areas outside the track have stronger reflections. When the infrared receiver receives the reflected infrared light, the voltage signal it generates changes accordingly, based on the different reflectivity of black and white colors. This change in voltage signal is related to the vehicle's current position relative to the track. Based on the voltage signal generated by the infrared receiver, the tracking module can determine the vehicle's position relative to the track. For example, if the left sensor detects a stronger reflection and the right sensor detects a weaker reflection, it indicates that the right sensor has detected a black line. The system may determine that the vehicle has deviated to the left of the track and adjust the vehicle's wheel speed so that the right wheel speed is slower than the left wheel speed, for example, by slowing down the right wheel and accelerating the left wheel. This allows the vehicle to adjust its direction to the left and return to the track, and vice versa. This information is provided to the vehicle's control system, which adjusts the vehicle's direction to keep it on the planned path and, if any other unusual circumstances arise, stops the vehicle for inspection. The tracking module plays a crucial role in the vehicle's tracking and obstacle avoidance, enabling it to navigate autonomously in complex environments and follow a predetermined trajectory. Therefore, in the implementation of the intelligent tracking and obstacle avoidance vehicle, the drive module ensures the vehicle's stable and flexible movement according to the system's design requirements, executing obstacle avoidance and tracking strategies when necessary.

[0033] The obstacle avoidance module includes an infrared sensor. An infrared sensor is installed on each side of the front of the car, which is equivalent to an infrared electronic switch. It outputs a low level when an obstacle is detected and a high level at other times.

[0034] Specifically, the working principle of the obstacle avoidance module is as follows: when there is an obstacle ahead, the infrared signal emitted by the infrared tube of the infrared sensor is received back by the infrared receiving tube, amplified by the integrated chip, and outputs a low level after comparison, lighting up the LED light-emitting tube on the module, and outputting a low-level signal at the same time. When the control system receives the signal from the obstacle avoidance module, it will adjust the movement of the car by adjusting the state of the motor according to the preset obstacle avoidance strategy, including adjusting the direction of the car, slowing down or stopping, to avoid obstacles. When no obstacle is detected, continue to go straight; when the left probe of the car detects an obstacle, it will retreat for 500 ms and then turn right for 500 ms; the same applies when the right probe detects an obstacle. If there are other special circumstances, the car will stop and check. The obstacle avoidance principle process is as follows: Figure 11As shown above. In summary, the obstacle avoidance module enables the car to detect and avoid obstacles ahead, ensuring it does not collide or become obstructed while following a vehicle. The obstacle avoidance module uses sensors to monitor the surrounding environment. When an obstacle is detected, it triggers a corresponding control strategy, causing the car to adjust its direction or speed to avoid the obstacle and improve safety.

[0035] The communication module uses BT04-A Bluetooth communication module for communication, such as Figure 9 As shown, this module uses the BT04-E chip (U13). Pin 1 connects to GND and pin 12 to VCC, providing power supply support. TX (pin 22) and RX (pin 21) connect to the transmit and receive pins of the external communication interface, respectively. P31 to P35 and P10 to P14 are multi-function I / O ports, with P13 connected to an external key (KEY) and P34 to an LED indicator (LED7) for status display. Pin 15 (NRST) is the reset pin, triggered by a low level. LED7 is connected to GND via a current-limiting resistor (1kΩ) R23 to indicate the Bluetooth module's operating status. Because this module utilizes the Bluetooth 3.0 transmission protocol, transmission distance and stability are significantly improved. The module has six pins, of which TXD and RXD are serial communication pins. TXD is used for data transmission and is connected to the receive pin (RXD) of the microcontroller's serial port 2. Similarly, RXD controls data reception and is connected to the transmit pin (TXD) of the main control chip's serial port 2, forming a complete data transmission path.

[0036] Driver modules, such as Figure 8 As shown, the RZ7899 chip drives a DC reduction motor. The RZ7899 quad-high-current H-bridge driver integrates a motor chip. The RZ7899 motor driver chip can control two DC motors, achieving forward and reverse motion. It can provide a bidirectional drive current of up to 600 mA and is equipped with a freewheeling circuit and overcurrent protection circuit to effectively ensure circuit safety. Suitable for low-power DC motors, it simplifies motor connection and control, is easy to operate, and is simple to debug. This module converts the PWM waves sent by the control system into motor motion. By adjusting the motors, the smart car's speed and direction are adjusted to execute corresponding tracking and obstacle avoidance strategies.

[0037] Automatic recharge module, including power monitoring unit, Hall sensor, signal receiving unit, such as Figure 10As shown, when the power monitoring unit of the car detects that the power is low, the recharging process is triggered, and the signal sending unit of the charging base communicates with the signal receiving unit of the automatic recharging module of the car to realize the automatic recharging function. Specifically, the signal sending unit of the charging base includes two infrared transmitters and magnetic strips or magnetic nails laid around the charging base. The car detects the magnetic field strength through the Hall sensor and navigates along the magnetic track; the signal receiving unit of the car includes two infrared receiving heads and a camera. The two infrared transmitters of the charging base transmit infrared modulated signals to the infrared receiving head of the car for receiving signals, guiding the car to approach the charging base and further move to the vicinity of the central axis of the charging base. The camera captures and detects the precise position of the two infrared receiving heads, guiding the car to accurately dock with the charging base.

[0038] The docking process for the tracking robot and the charging base is a two-step process. The main idea is to utilize the cooperation of an infrared transmitter, a receiver, and a camera. The robot first uses the infrared transmitter to approach the charging base and gradually moves to the center axis of the charging base. Finally, the camera is used to precisely dock the robot with the charging base. The charging base is designed with V-shaped or funnel-shaped metal rails. The robot triggers a fine-tuning program via a collision switch to gradually correct its position. The charging contacts have built-in electromagnets (such as 12V coils), which are powered after docking to enhance connection stability. A communication-first, power-on strategy is adopted, with the charging relay controlled by optocoupler isolation. A PTC resettable fuse is added to the charging base output to prevent short circuits. It is important to ensure that each step is supported by the corresponding module, such as power monitoring, sensor detection, navigation adjustment, and charging detection. Step 1: Dock using infrared signals. When any infrared receiver receives an infrared signal, the robot turns so that both infrared receivers can receive the signal from the infrared transmitter. If both receivers can receive signals from transmitters a and b, the robot has exited transmission zone 3, and camera docking begins. If receiving heads A and B can only receive a signal from either transmitter a or transmitter b, the robot is in Transmitter Area 1 or 2. The robot then turns to face Transmitter Area 3, confirming it is in Transmitter Area 2. Step 2: Dock using camera information. After the robot completes infrared signal docking, it will be in Transmitter Area 3 facing the charging dock. At this point, there will still be a few centimeters of error between the robot and the camera. This requires docking with the camera. The camera captures the charging dock and extracts the light spots formed by the two transmitters in the image. Based on the areas of the two light spots, the robot can be determined to be on the left or right side of the charging dock.

[0039] Sensor modules include vibration sensor module, sound sensor module, temperature and humidity sensor module, digital temperature sensor module, ultrasonic sensor module, infrared sensor module, flame sensor module, color sensor module, 24L01 wireless module, etc. According to different experimental requirements, they can also include fire extinguishing unit, color recognition module, grasping device, etc. Each sensor module can be built with the functions required by the experiment according to the course requirements.

[0040] Vibration sensor module circuit as shown Figure 12 As shown, the circuit consists of a vibration sensor (U15) and an LM393 comparator (U14). The vibration sensor is connected to VCC via R25 (a 1kΩ resistor), and the sensor output signal is connected to the LM393's inverting input (pin 3). The LM393's non-inverting input (pin 4) is grounded, and its output (pin 1) is connected to the system's input port via diode D9 for subsequent processing. The circuit also includes C17 and C19 (0.1μF capacitors) for filtering and signal stabilization. The LM393's VCC is connected to the positive power supply (pin 8), and GND is connected to ground (pin 5) to ensure proper operation. The operating principle of a vibration sensor relies primarily on the interaction between a mass and an inertial space within its internal structure. When the sensor senses vibration, the mass moves relative to a fixed part. This relative motion is converted into an electrical signal by a transducer element (such as a piezoelectric material). The core of a piezoelectric vibration sensor is the ability of certain crystalline materials to generate an electrical charge when subjected to mechanical stress. When these materials are compressed or stretched, positive and negative charges form on their surfaces, generating a voltage output.

[0041] The vibration sensor module can be used in various vibration trigger detection, theft alarm, unmanned vehicles, electronic building blocks and other related designs.

[0042] Sound sensor modules such as Figure 14As shown in the figure, this circuit consists of a microphone sensor (MIC), an LED indicator (LED8), an amplifier (LM386, U16), and related external components. The microphone output signal is filtered by R2, C22, and C21 before being input to pin 2 (IN1-) of the LM386. Pin 3 (IN1+) is also connected to ground for signal amplification. The LM386 output (pin 5, VOUT) is coupled via capacitor C24 to LED8, which indicates the audio signal strength. The negative terminal of LED8 is connected to ground. The circuit also includes a gain adjustment potentiometer RP1, which filters and stabilizes the signal via C23 and C25. D2 is a protection diode, and RP1 is used to adjust the output gain. J4 is a 3-pin connector for connecting external devices, providing VCC, GND, and a signal output. This typically includes a microphone (such as a condenser electret microphone), which converts sound waves into corresponding voltage changes. When sound waves strike the diaphragm inside the microphone, the diaphragm vibrates, which in turn changes the capacitance of the connected capacitor, generating small voltage fluctuations. These voltage signals are then amplified and processed, ultimately outputting analog or digital signals to a host controller or other circuits for further analysis and response. This module can be used in intelligent lighting control systems, audio-visual interactive devices, noise alarm systems, and specific sound recognition.

[0043] Temperature and humidity sensor modules, such as Figure 13 As shown, this circuit consists of a DHT11 temperature and humidity sensor U17 and connector J5. Pin 1 (VDD) of the DHT11 is connected to the power supply VCC via resistor R30 (1kΩ). Pin 4 (GND) is connected to ground. Pin 2 (DATA) is connected to VCC via R30 and is used to transmit temperature and humidity data. Pin 3 (NC) is a non-connected pin and is unused. J5 is a 3-pin connector for connecting external modules. Pin 1 is connected to VCC, pin 3 to GND, and pin 2 is the data transmission line, transmitting temperature and humidity signals to the system. This module is an electronic device that integrates temperature and humidity measurement functions. It can be used to monitor temperature and humidity changes in the environment and convert this information into easy-to-process data signals. This type of module is widely used in various fields, such as meteorological observation, industrial automation, agriculture, and household appliances.

[0044] Digital temperature sensor modules such as Figure 15As shown, the DS18B20 digital temperature sensor (U18) is connected to power via VCC (pin 3) and GND (pin 1). The data line DQ (pin 2) is connected to VCC via a 1kΩ resistor R31. This data line is used to transmit temperature data to the control system. This module converts temperature into an easily processable digital signal. Unlike analog temperature sensors, digital temperature sensors output data in a discrete, binary format, making them ideal for direct interfacing with microcontrollers or computer systems without the need for an additional analog-to-digital converter (ADC). The DS18B20 is a key example.

[0045] Ultrasonic sensor modules such as Figure 16 As shown, pin 1 (VCC) of the HC-SR04 ultrasonic sensor (U19) is connected to the positive power supply, and pin 4 (GND) is connected to ground. Pin 2 (TRIG) is the trigger pin, used to control the transmission of the ultrasonic signal. Pin 3 (ECHO) receives the echo signal and outputs the reflection time for distance calculation. This module is a sensor device that uses ultrasonic waves (sound waves with a frequency greater than 20kHz) for distance measurement and object detection. This type of module is widely used in robotic navigation, smart homes, automation equipment, and other fields.

[0046] Infrared sensor modules such as Figure 17 As shown, this circuit includes an infrared transmitter (U22) and a receiver (U23). The transmitter is connected to the power supply via resistors R36 and R37, while the receiver is connected to an LM358 amplifier (U24.1). The input signal is regulated via RP3, and the output signal is regulated via RP4 to drive an LED (LED11). The negative terminal of LED11 is connected to ground via a current-limiting resistor R38, indicating the infrared signal detection status. This sensor's detection range can be adjusted using a potentiometer, and it features low interference, easy assembly, and convenient use. It is widely used in applications such as robot obstacle avoidance, obstacle avoidance vehicles, assembly line counting, and black and white line tracking.

[0047] Flame sensor module, such as Figure 18As shown in the figure, the flame sensor circuit consists of an infrared emitting diode (U21), a receiving head (U22), an LM393 comparator (U20), and related components. U21 is connected to VCC via a current-limiting resistor (R32) and continuously transmits infrared signals. U22 receives the infrared signal reflected by the flame and connects its output to the inverting input (pin 2) of the LM393 via R33. The non-inverting input is connected to the reference voltage formed by the voltage divider R34 and R35. The output (pin 1) of the LM393 controls the lighting status of LED 10. The positive terminal of LED 10 is connected to VCC, and the negative terminal is connected to the output terminal through a current-limiting resistor in series, indicating whether a flame is detected. Capacitors C26 and C27 provide signal filtering, power supply stability, and enhanced anti-interference capabilities. This module is a flame sensor commonly used to detect flames or infrared light of a specific wavelength (760nm-1100nm). It has a detection angle of approximately 60° and is particularly sensitive to the flame spectrum. The sensitivity is adjustable (adjusted by the blue potentiometer in the figure). The flame detection range depends on the sensitivity and flame intensity and is generally within 1 meter. It has strong applicability and is suitable for industrial automation, security monitoring, fire warning, robotics and other fields.

[0048] The operating principle of a color sensor module is based on the optical properties of color. It typically illuminates an object using its own light source, typically red, green, and blue LEDs. Objects of different colors reflect light of varying wavelengths, and the sensor's photodetectors, such as photodiodes or phototransistors, receive this reflected light. Furthermore, different colors of light reflect off the surface in varying proportions, resulting in varying light intensities received by the detectors. After photoelectric conversion, the sensor's internal processing circuitry analyzes the electrical signals to determine the object's color.

[0049] The 24L01 wireless module, including the NRF24L01, is a new single-chip RF transceiver device operating in the 2.4GHz-2.5GHz ISM band. It features a built-in frequency synthesizer, power amplifier, crystal oscillator, modulator, and other functional modules, incorporating enhanced ShockBurst technology. Its output power and communication channel are programmable. The NRF24L01 boasts low power consumption, consuming only 9mA when transmitting at -6dBm power and 12.3mA when receiving. It offers multiple low-power operating modes, including 160mA at 100mW. While its data transmission range is longer than Wi-Fi, its transmission throughput is lower than that of Wi-Fi (due to power-down and idle modes), making energy-saving designs more convenient.

[0050] Optionally, the experimental system can be designed to conduct intelligent fire-fighting experiments, intelligent handling experiments, and intelligent search and rescue experiments based on the sensors used in teaching and combined with practical applications. The intelligent fire-fighting experiment involves adding a fire-fighting unit to the tracking vehicle. The fire-fighting unit is relatively simple in design, consisting primarily of an infrared flame detector, a relay, and a micro water pump. The flame detector utilizes a flame-light-sensitive element. The intelligent handling experiment involves adding a color recognition module and a gripping device to the tracking vehicle, placing the vehicle in a designated area for testing. The intelligent search and rescue experiment involves adding a human perception module and a sound module to the tracking vehicle, placing the vehicle in a designated area for testing.

[0051] Example 2 In one or more embodiments, a sensor experiment method based on artificial intelligence is disclosed, using the above-mentioned sensor experiment system based on artificial intelligence, such as Figure 19 Shown, including: Step S1: Initialize the experimental system. The intelligent system obtains the user's experimental command through the display module or voice processing module. Step S2: The decision module of the intelligent system generates an experimental plan based on the experimental command through the AI ​​algorithm; the user can confirm the experimental plan. If confirmed, the next step is carried out; if not, the experimental command is obtained again through the display module or the voice processing module; The experimental scheme includes intelligent fire extinguishing experiment, intelligent transportation experiment and intelligent search and rescue experiment; Furthermore, by displaying the experimental steps according to the experimental plan and executing the experiment according to the steps, this system implements trajectory estimation based on a wheel odometer on the STM32 platform. Timed interrupts periodically read the speed information of the left and right motors, and combined with the preset wheel radius and wheel spacing, the vehicle's travel distance and steering angle are calculated in real time. The system maintains a coordinate state variable, and each iteration incrementally updates the position based on speed and direction. All trajectory points are sequentially stored in a memory buffer as two-dimensional coordinates. After the experiment concludes, the trajectory is transmitted to the intelligent system via the Bluetooth serial port, and a report is generated.

[0052] The experimental steps of the intelligent fire extinguishing experiment include: the intelligent fire extinguishing experiment requires the installation of a fire extinguishing unit on the basis of the obstacle avoidance tracking car. The fire extinguishing unit is relatively simple in design and mainly consists of an infrared flame sensor, a relay, and a micro water pump. The flame sensor is made of a flame light sensitive element. The system structure of the intelligent fire extinguishing experiment tracking car is as follows: Figure 20 The infrared flame sensor is installed at position 1 on the adapter board, while the relay and micro water pump can be installed at positions 2 and 3 respectively.

[0053] Intelligent fire extinguishing process Figure 21As shown, the flame intensity is detected and the controller transmits the sensed signal to the intelligent vehicle. The flame detector, which can distinguish electronic signals, can be used to detect the location of the fire source or sense fire sources within the wavelength range of 760 to 1100 nm. The detector board interface can be directly connected to the microcontroller's IO port. When the flame detector detects a flame and reaches the threshold set by the potentiometer, the green indicator light above illuminates and the DO digital switch output is low; otherwise, the indicator light remains off. The system uses a water pump to spray fires. An optoelectronic isolator is combined with a transistor to control a relay interrupt, and an intermediate relay controls the water pump's on / off. Within a specified area, when the infrared flame detector detects the location of a flame, the intelligent fire truck starts spraying water, completing the firefighting operation.

[0054] like Figure 22 As shown, to improve fire extinguishing accuracy, the car needs to be pointed directly at the fire source; otherwise, the fire may be difficult to extinguish. In this case, the smart car needs to determine the strength of the fire on both sides. If the left side is stronger, the car turns 45 degrees to the left, and vice versa. The smart car retests after each turn to ensure accurate fire extinguishing. After aiming at the fire source and waiting for 0.5 seconds, it activates the water pump to spray water to extinguish the fire. After waiting for 5 seconds, it reads the data to confirm whether the fire is extinguished. If so, it turns off the water pump and continues to search for other fire sources.

[0055] Step S4-1: Build a simulated building structure consisting of multiple rooms connected by corridors. Set up possible fire sources in each room (use candles or small flame simulators to simulate fire sources). Lay tracking path markers on the corridors and room floors to guide the car. Also, place some common obstacle props (such as table and chair models) in the room to simulate the complexities of a real environment.

[0056] Step S4-2: At the starting point of the experimental site, place the intelligent tracking and obstacle avoidance vehicle on the tracking path, ensuring that it can accurately identify the path markings. Turn on the vehicle's power switch, activate the control system, and the vehicle begins to drive along the path according to the preset tracking mode.

[0057] Step S4-3: As the car drives, its infrared tracking sensor continuously detects path markings on the ground, guiding it along the corridor. When the car approaches a room entrance, the tracking path guides it inside. Before entering the room, the car's ultrasonic obstacle avoidance sensor detects obstacles at the entrance. If an obstacle is blocking the way, the car automatically adjusts its driving direction based on its obstacle avoidance strategy and seeks a passable path into the room.

[0058] Step S4-4: The vehicle enters the room, and the flame sensor module begins operating, detecting flame signals in the room in real time. The flame sensor scans the room in all directions at a constant frequency, searching for the characteristic signal of a fire source. When the flame sensor detects a fire source signal, it immediately transmits the signal to the microcontroller control system.

[0059] Step S4-5: After receiving the fire source signal, the microcontroller processes and analyzes the signal to determine the location and intensity of the fire. Based on the fire source's location information, the control system adjusts the vehicle's direction and speed to move it toward the fire source and activates the fire extinguishing device to prepare for fire extinguishing operations. As the vehicle approaches the fire source, the flame sensor continuously monitors changes in the fire source signal and continuously updates the fire source's location information, allowing the vehicle to accurately locate the fire source. When the vehicle reaches a certain distance from the fire source, the control system triggers the fire extinguishing device to extinguish the fire. The fire extinguishing device sprays fire extinguishing agent or releases small fire extinguishing bombs to extinguish the fire.

[0060] Step S4-6: During the fire extinguishing process, the flame sensor will continue to monitor the flame signal of the fire source to determine whether the fire source has been completely extinguished. If the fire source still exists, the robot will adjust the spray direction and force of the fire extinguishing device based on the feedback information from the flame sensor and continue the fire extinguishing operation until the fire source is completely extinguished. After the robot successfully extinguishes a fire source, it will return to the entrance of the room according to the preset program, re-traverse the tracking path, and enter the next room to search for the fire source. During the driving process, the robot's tracking sensor and obstacle avoidance sensor will work together to ensure that the robot can smoothly shuttle between rooms while avoiding obstacles.

[0061] Step S4-7: Repeat the process of fire source search, location and fire extinguishing in steps S4-4 to S4-6. The trolley checks and extinguishes each room in turn until the fire sources in all rooms are extinguished.

[0062] After the experiment, the vehicle transmits its path and the location of the fire source via Bluetooth to the intelligent system, which then generates an experiment report. The goal of this experiment is to allow students to install a flame sensor module on the tracking vehicle and learn how to use it to detect fire sources and extinguish them using a fire extinguishing device. This experiment will help students understand the working principles, data processing, and applications of flame sensors, while also practicing how to design and implement a fire extinguishing control system.

[0063] The experimental steps of the intelligent handling experiment include: the intelligent handling vehicle control system is based on the obstacle avoidance and tracking unmanned vehicle control, and a color recognition module and a grasping device are added. Figure 23As shown in the figure, the color recognition module is installed at position 1 on the adapter plate, and the grabbing device is installed at position 5.

[0064] Color Recognition Module: The color recognition module uses the TCS230 photoelectric sensor. The TCS230 is a color photoelectric sensor. It consists of a color photoelectric converter, a silicon photodiode, and a current frequency generator, forming a CMOS integrated circuit. It also incorporates red, green, and blue filters and is digitally compatible. The TCS230 has a fast response time, outputs a digital signal, is easy to use, and has strong anti-interference capabilities.

[0065] The grabbing device consists of four EMAXES08A simulated servos and a grab hook. The servos rotate 180°, so using two servos in combination can achieve two degrees of freedom within 180°. This meets the requirements of the car design in this article. The grab hook has a mass of 4g, which can ensure the required rotation accuracy.

[0066] The intelligent transport vehicle uses color recognition to distinguish the colors of objects and uses a servo plus mechanical structure to grab objects. Figure 24 As shown in the figure, the designed control program can achieve the following functions: (1) realize automatic tracking of the intelligent car; (2) automatically identify the color of the object and grab it.

[0067] Step S5-1: Environment Setup: Build an environment that simulates a warehouse or production workshop. Lay tracking path markers (such as black lines or specific patterns) on the floor. Set up multiple storage points and target placement points along the path. The storage points contain objects of different colors (such as small balls or blocks in red, blue, and green, of appropriate size and weight for easy grasping by the robotic arm). The target placement points are used to store objects after handling.

[0068] Step S5-2, Experimental Preparation: Check whether all modules of the intelligent tracking and obstacle avoidance robot are working properly, including the color sensor module, grasping device, tracking sensor module, ultrasonic sensor module, drive module, power module, etc., to ensure the sensitivity and accuracy of the sensors, as well as the movement flexibility and grasping stability of the robotic arm.

[0069] Step S5-3: Comprehensive Test: At the starting point of the experimental site, place the intelligent tracking and obstacle avoidance vehicle on the tracking path to ensure that it can accurately identify the path markings. Turn on the vehicle's power switch, activate the control system, and the vehicle will begin to drive along the path according to the preset tracking mode.

[0070] As the car moves, the tracking sensor continuously detects path markings on the ground, guiding it along the pre-set path. When the car approaches a storage point, the color sensor module activates, detecting the color of objects within the storage point in real time. The color sensor scans the storage point at a predetermined frequency (e.g., once per second) and transmits the detected color information to the microcontroller control system.

[0071] After receiving the color information, the microcontroller compares it with the preset target object color. If the detected object's color matches the target color, the control system immediately issues a command to the robotic arm to prepare for the grasping operation. If the color does not match, the robot continues along the tracking path to search for the next target object.

[0072] Once the robot locates the target object, the robotic arm begins its movement. First, the arm's telescopic mechanism extends, moving the arm over the target object. Then, the lifting mechanism descends, allowing the gripping mechanism to approach the object's surface. Next, the gripping mechanism (such as the gripper) automatically adjusts the gripping force and angle based on the object's shape and size, ensuring a stable grasp. During the grasping process, the robotic arm's sensors provide real-time feedback on the gripping status, such as gripping force and object position. Based on this information, the control system fine-tunes the robotic arm's movements to ensure accurate and reliable grasping.

[0073] After a successful grasp, the robotic arm's lifting mechanism rises, lifting the object to a desired height (to avoid collision with the ground or other objects). The telescopic mechanism then retracts, moving the object onto a small shelf or pallet carried by the robot and placing it in the designated location. The entire grasping and placement process is precisely controlled by a control system to ensure the object does not fall or become damaged.

[0074] After the robot has grasped and placed the object, it continues along the tracking path toward the target placement point. During this process, ultrasonic obstacle avoidance sensors monitor the surrounding environment in real time. When an obstacle is detected, the robot automatically adjusts its direction and speed based on its obstacle avoidance algorithm to avoid the obstacle, ensuring safety and stability during the handling process.

[0075] When the robot arrives near the target placement point, the tracking path guides the robot to the precise location. The robotic arm moves again, grabs the object from the shelf or pallet, and then places it on the target placement point according to the preset placement position and posture. After placement, the robotic arm returns to its initial position, and the robot prepares for the next handling task.

[0076] After the experiment, the robot transmits its path, the location of the objects it carried, and the location where it placed them via Bluetooth to the intelligent system, which then generates an experiment report. The goal of this experiment is to allow students to learn how to use sensors to identify object colors and accurately grasp them, using a tracking robot combined with a color sensor and a gripper. This experiment not only helps students master the use of color sensors but also deepens their understanding of robotic arm control and the integrated operation of sensors and actuators.

[0077] The experimental steps of the intelligent search and rescue experiment include: like Figure 25 As shown, place the infrared sensor module and sound sensor module at positions 1 and 5 on the adapter board. The infrared sensor module can use the HC-SR501 human infrared sensor module and the HC-SR501LHI778 probe, which are highly sensitive and reliable, with a detection range of up to 7 meters, fully meeting the requirements. Additionally, a sound sensor with a built-in sound-sensitive resistor is installed to detect sound. The LM386 sound sensor module can amplify audio signals 200 times, accurately detecting surrounding sounds and assisting the human sensor module in precisely locating trapped individuals.

[0078] This module is designed to cope with the situation where trapped people use sound sources as distress signals under poor lighting conditions, and to increase the adaptability of the car. Figure 26 The figure shows the process of intelligent search and rescue experiment.

[0079] Step S6-1, Environment Construction: Build an environment that simulates a disaster scene, such as a ruins scene. Use bricks, wooden boards, and other simple obstacles and shelters. Lay tracking path markings (such as black lines or specific patterns) on the ground. Hide simulated trapped people (using mannequins or real people) at different locations along the path. At the same time, set up some interference sources (such as other heat-generating objects, noise sources, etc.) to test the anti-interference ability of the robot.

[0080] Step S6-2, sensor calibration; check whether each module of the intelligent tracking and obstacle avoidance car is working properly, including the infrared sensor module, sound sensor module, tracking sensor module, ultrasonic sensor module, drive module, power module, etc., to ensure the sensitivity and accuracy of the sensors and the normal operation of the motors.

[0081] Program and debug the robot's control system so that it can perform tracking, obstacle avoidance, target detection, and marking functions based on sensor feedback. Set the thresholds for the infrared sensor module and the sound sensor to accurately identify target signals. Adjust the parameters of the tracking sensor and ultrasonic obstacle avoidance sensor to ensure that the robot can stably follow the preset path and avoid obstacles.

[0082] Step S6-3: Comprehensive Test: At the starting point of the experimental site, place the intelligent tracking and obstacle avoidance vehicle on the tracking path to ensure that it can accurately identify the path markings. Turn on the vehicle's power switch, activate the control system, and the vehicle will begin to drive along the path according to the preset tracking mode.

[0083] As the car drives, its tracking sensors continuously detect path markings on the ground, guiding it along the pre-set path. Meanwhile, its ultrasonic obstacle avoidance sensors monitor the surrounding environment in real time. When an obstacle is detected, the car automatically adjusts its direction and speed based on its obstacle avoidance algorithm to avoid it, ensuring the safety and continuity of the search and rescue process.

[0084] The human infrared sensor module and sound sensor are also active, detecting target signals in the surrounding environment in real time. The human infrared sensor scans the surrounding area at a certain frequency (for example, once per second) to detect infrared signals similar to those emitted by human bodies. The sound sensor continuously monitors the surrounding sounds, detecting possible cries for help or other unusual sounds.

[0085] When the human infrared sensor module detects an infrared signal or the sound sensor captures a sound signal, the car immediately pauses its tracking. The control system processes and analyzes the sensor feedback signal, using existing algorithms to determine the target's location and approximate direction. For example, the infrared sensor module estimates the relative distance and angle between the target and the car based on the intensity and direction of the infrared signal; the sound sensor determines the location of the sound source based on the intensity and direction of the sound.

[0086] Based on the target's location, the car adjusts its direction and moves toward it. Meanwhile, its sensors continuously monitor the target, updating its location information to more accurately locate it. As it approaches the target, its ultrasonic obstacle avoidance sensors more carefully monitor its surroundings to prevent collisions with obstacles.

[0087] When the vehicle approaches a target within a certain distance (e.g., 0.5 meters) and identifies it as a target object (e.g., a simulated trapped person), a warning light or buzzer on the vehicle activates, sending a clear signal indicating that the target has been found. Simultaneously, the control system records the target's location (e.g., through a built-in positioning system or relative distance and direction from the starting point) and sends a search and rescue report to a pre-set receiving device (e.g., the experimenter's phone or computer). The report includes information such as the time, location, and status of the target. The experimenter then determines whether the experiment is complete based on the search and rescue report.

[0088] After completing the marking and reporting of a target, the car will return to the tracking path according to the preset program and continue to drive along the path to search for other possible targets. If the car's sensors do not detect any target signals during the search and rescue process, the car will continue to drive along the tracking path until it completes the search and rescue mission of the entire experimental site or receives a stop command. After the experiment is over, the car will transmit the path it has run during the experiment and the location of the rescued people to the intelligent system via the Bluetooth module, and the intelligent system will then generate an experimental report. The purpose of this experiment is to allow students to use infrared sensor modules and sound sensors to identify signals from trapped people, and learn how to locate targets and conduct search and rescue through sensors in simulated disaster scenes. Through this experiment, students can become familiar with the application of human perception and sound detection sensors, and improve their multi-sensor data fusion and practical problem-solving capabilities.

[0089] The description of each embodiment in the above embodiments has different emphases. For parts not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0090] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A sensor experiment system based on artificial intelligence, characterized in that: Including intelligent systems and operating systems; The intelligent system includes a display module and a decision module; the operating system includes a tracking car and several independent sensor modules. An adapter plate is installed on the top of the tracking car, and the sensor module is connected to the tracking car through the adapter plate; The display module is used to obtain experimental commands and display experimental plans, so that the experimenter can conduct experiments according to the displayed experimental plans; The decision-making module is used to receive experimental commands and use a data parsing unit to convert the experimental commands into experimental parameter vectors. The data parsing unit includes constructing a semantic parsing framework based on an attention mechanism; using a solution generation unit based on reinforcement learning to construct an experimental solution; using a verification and error correction unit to perform logical verification on the generated experimental solution, and generating an experimental solution based on a tracking car and several independent sensor modules for experimenters to operate the experiment.

2. The sensor experiment system based on artificial intelligence according to claim 1, characterized in that: The experimental commands include text experimental commands and voice experimental commands; The display module includes a touch screen display for inputting textual experimental commands; The intelligent system also includes a voice processing module for performing voice input and voice processing to obtain voice experiment commands.

3. The sensor experiment system based on artificial intelligence according to claim 1, characterized in that: The tracking car includes a control system, which includes a tracking module and an obstacle avoidance module; The tracking module includes an infrared tracking sensor, which includes four infrared transmitting / receiving sensors. The four sensors are placed in a row in front of the front of the car to realize the tracking function. The sensor generates an infrared beam, illuminates the track on the ground, and receives the infrared light reflected from the ground at the same time, so that the car can drive on the predetermined path according to the track. The obstacle avoidance module includes an infrared sensor, one installed on each side of the front of the car. It outputs a low level when an obstacle is detected and a high level at normal times. The movement of the car is adjusted based on the level signal to avoid the obstacle.

4. The sensor experiment system based on artificial intelligence according to claim 1, characterized in that: Each independent sensor module is a combination of one or more of a vibration sensor module, a sound sensor module, a temperature and humidity sensor module, a digital temperature sensor module, an ultrasonic module, an infrared sensor module, a flame sensor module, a color sensor module, and a 24L01 wireless module.

5. The sensor experiment system based on artificial intelligence according to claim 1, characterized in that: The solution generation unit matches the corresponding template according to the key information extracted by the data analysis unit, and constructs the experimental cost minimization objective function combined with the reinforcement learning model to generate preliminary solution data. At the same time, the Q-learning algorithm is used to dynamically optimize the preliminary experimental solution and output the adjusted execution solution to obtain the adjusted experimental solution: In the formula, s is the state; a is the action; is the expected long-term cumulative reward that can be obtained by selecting action a in state s; For the next state Select the next action The expected long-term cumulative rewards that can be obtained; η is the learning rate; is the discount factor; r is the immediate reward, For the best value.

6. The sensor experiment system based on artificial intelligence according to claim 1, characterized in that: The decision module also includes a data storage and management unit for constructing a knowledge graph of experimental data and implementing retrieval and association analysis through a graph embedding algorithm; Specifically, the semantic similarity between experimental commands and historical cases is calculated through graph embedding, and the most relevant historical experimental plans are retrieved; when the knowledge graph detects that the new experimental plan has a highly similar subgraph structure with historical failure cases, a risk warning is triggered.

7. A sensor experiment method based on artificial intelligence, characterized in that: include: Step S1: Initialize the experimental system. The intelligent system obtains the user's experimental command through the display module or voice processing module. Step S2: The decision module of the intelligent system generates an experimental plan according to the experimental command through a pre-trained artificial intelligence mapping model; The experimental plan includes intelligent fire extinguishing experiment, intelligent transportation experiment and intelligent search and rescue experiment.

8. The sensor experiment method based on artificial intelligence according to claim 7, characterized in that: The intelligent fire extinguishing experiment includes: The experiment was conducted using a fire-fighting tracking vehicle in a constructed building scene. The fire-fighting tracking vehicle includes a tracking module and a fire-fighting unit. The fire-fighting unit includes a flame sensor module, a relay, and a water pump. The flame sensor module detects whether there is a flame and the intensity of the flame. When the flame reaches the threshold, the relay controls the water pump switch, and the water pump sprays water to extinguish the fire.

9. The sensor experiment method based on artificial intelligence according to claim 7, characterized in that: The intelligent handling experiment includes: In the simulated transport scene, a tracking trolley was used to conduct experiments; The tracking transport vehicle includes a color sensor module, a gripping device, a tracking module, an ultrasonic sensor module, a driving module and a power supply module; The car travels along the tracking path based on the tracking module, and the color sensor module detects the color of the object. If the detected color of the object matches the target color, the grasping device grasps and places the object, and the ultrasonic sensor module monitors the surrounding environment in real time to avoid obstacles.

10. The sensor experiment method based on artificial intelligence according to claim 7, characterized in that: The intelligent search and rescue experiment includes: The experiment was conducted using a firefighting tracking vehicle at a simulated disaster scene. The fire extinguishing tracking vehicle includes an infrared sensor module, a sound sensor module, a tracking module, an ultrasonic sensor module, a driving module and a power supply module; The car travels along the tracking path based on the tracking module, and the ultrasonic sensor module monitors the surrounding environment in real time to avoid obstacles, the infrared sensor module detects infrared signals of human bodies in the surrounding area, and the sound sensor module detects surrounding sounds; When the infrared sensor module detects an infrared signal or the sound sensor detects a sound signal, the car stops tracking and determines the target position, and then moves closer to the target according to the target position; When the car approaches the target to a certain distance and is confirmed to be the target object, the car sends a marking signal and records the location information, generating a search and rescue report so that the experimenter can determine whether the experiment is completed based on the search and rescue report.

Images