Intelligent car task planning teaching method and system

Abstract

According to the intelligent dolly task planning teaching method and system provided by the invention, through an optimal strategy generated by artificial intelligence, a clear quantitative benchmark is provided for students, a learning target is improved from function realization to strategy optimization, and the students can not only manually control the dolly through programming, but also compare with a model recommendation result, so that the teaching efficiency is improved. The optimization capability and principle of deep learning can be intuitively understood, and the learning interest of students can be greatly stimulated.

Description

Technical Field

[0001] The present invention relates to the technical field of artificial intelligence teaching equipment, and in particular to a teaching method and system for task planning of an intelligent vehicle. Background Art

[0002] In the field of project-based learning that combines robotics and artificial intelligence, especially in task-oriented competition scenarios (such as simulated rescue, smart logistics, etc.), learners usually need to control autonomous mobile robots (such as smart cars) through programming to complete a series of preset tasks, such as material transportation, path planning and navigation.

[0003] Currently, the technical solutions of the teaching and evaluation systems used in this field usually focus on controlling and verifying the basic functions of robots, such as controlling the robot to travel along a preset trajectory, identifying specific visual beacons, or executing a fixed sequence of actions.

[0004] However, existing teaching and assessment systems suffer from the following technical flaws: They typically only determine whether a task is completed but fail to provide an optimal strategy as a reference. Consequently, even if learners successfully complete a task, they are unable to quantify the difference between their strategy and the theoretically optimal solution. This significantly limits their ability to explore and validate advanced optimization algorithms.

[0005] Traditional feedback mechanisms often focus solely on end-results, such as total time or final score. This feedback model fails to provide in-depth, quantitative analysis and diagnosis of learners' specific decisions during task execution. Consequently, the feedback provided is relatively coarse, lacking sufficient diagnostic and guidance, making it difficult to effectively support personalized instruction and iterative improvement.

[0006] Therefore, this field urgently needs a technical solution that can generate optimal strategy benchmarks and provide multi-dimensional, process-based diagnostic feedback. Summary of the Invention

[0007] In order to overcome the shortcomings of the above existing technologies, a method for teaching task planning of an intelligent vehicle is provided, which is characterized by comprising the following steps: S1. Site digitization: Convert a mission scenario containing multiple key locations and physical paths into a weighted digital graph data structure. The site digitization steps specifically include: S1.1 Node definition: Based on the multiple key locations in the scenario, define multiple nodes, and assign each node a unique node ID and a node type used to characterize its function; S1.2 Path Definition and Weight Calculation: Find and connect the physical paths between any two nodes in the mission scenario to determine the path connection relationship between the nodes, and calculate a weight representing the travel cost for each path connection relationship; S2 Strategy Generation: Based on the digital graph data structure generated in the site digitization step in step S1, combined with the task rules, an artificial intelligence optimization algorithm is used to deduce and generate a task execution strategy including an optimal action sequence and an optimal travel path; wherein, the strategy generation step further includes: S2.1 Environment Modeling: Converting the digital graph structure and task rules into a reinforcement learning environment, wherein the environment comprises at least a state space, an action space, and a reward function; S2.2 Model Training: Through multiple rounds of simulation in the reinforcement learning environment, a deep Q-network model capable of predicting action values ​​based on state is trained; S2.3 Strategy Extraction: Utilizing the trained deep Q-network model, starting from the initial state, the optimal task execution strategy is generated by iteratively selecting the action with the highest action value in each state; S3. User execution and data recording step: Provides a programming interface that allows users to independently write control programs to drive the smart car to perform tasks in the task scenario, and records the user task process data including timestamps, trajectories, and key actions generated by the smart car under the user's control in real time; S4. Comparative analysis and feedback step: Perform a multi-dimensional comparative analysis of the user task process data recorded in S3 and the optimal task execution strategy generated in S2 to generate a teaching feedback report containing at least one key efficiency indicator difference and a diagnostic optimization suggestion.

[0008] Furthermore, the state space described in step S2.1 is used to describe the current position, load status, and status of multiple task nodes in the task scenario of the smart car; the action space includes at least moving actions and actions required by task rules; the reward function is used to give positive rewards to actions that complete task goals, and give negative rewards to actions that consume time or resources.

[0009] Furthermore, step S2.2 also includes: storing the interactive experience including state, action and reward generated in the simulation in an experience replay pool, and randomly sampling a batch of interactive experiences from the experience replay pool for updating the weight parameters of the deep Q network model.

[0010] Furthermore, the multiple rounds of simulation and deduction in step S2.2 adopt a greedy strategy; the iterative selection in step S2.3 is a greedy selection.

[0011] Furthermore, the key efficiency indicators in S4 include at least one of path efficiency, task sequence efficiency, and average scoring time. The diagnostic optimization suggestions in S4 include: analyzing the strategy type of the user control program based on the differences in the key efficiency indicators and providing improved algorithm suggestions for the strategy type; the teaching feedback report also includes a visual comparison diagram of the trajectory of the intelligent car under user control and the optimal travel path.

[0012] Furthermore, the path efficiency is determined by calculating the ratio of the total travel distance of the intelligent car under user control to the distance of the optimal travel path; the task sequence efficiency is determined by comparing the difference between the order in which the robot visits the task nodes under user control and the node access order in the optimal action sequence.

[0013] An intelligent car task planning teaching system, characterized by including A map digitization module is used to convert the physical layout and constraints of a mission scenario containing multiple key locations and physical paths into a digital graph structure containing nodes and weighted paths; A strategy generation module is used to generate a task execution strategy including an optimal action sequence and an optimal travel path based on the digital graph structure and task rules through an artificial intelligence optimization algorithm; The data recording module is used to record the user task process data generated when the user autonomously controls the smart car through the programming interface; a comparison and diagnosis module, configured to compare and analyze the user task process data with the task execution strategy to generate a teaching feedback report including at least one key efficiency indicator difference and diagnostic optimization suggestions; Programming development module, used for users to perform graphical or code programming to compile and generate task instructions; The network module is used for Wi-Fi communication and task instruction transmission with the smart car.

[0014] Furthermore, the system also includes an intelligent car, which adopts a dual MCU structure. The dual MCU structure includes: an upper circuit board and a lower circuit board. The upper circuit board is equipped with a first master MCU, which is responsible for communicating with the network module and receiving and executing instructions; the lower circuit board is equipped with a second master MCU, which receives internal instructions from the first master MCU through the UART protocol, connects the USART TX pin of the first master MCU to the USART RX pin of the second master MCU; and connects the USART RX pin of the first master MCU to the USART TX pin of the second master MCU to achieve control of the underlying hardware. The lower circuit board is provided with a battery to power the first and second master MCUs. Furthermore, the upper circuit board is provided with a gateway node module and a sensor interface, and the sensor interface is connected using a separate coupled modular magnetic POGO PIN connector. Corresponding to the sensor provided with a magnetic POGO PIN connector interface, the first main control MCU is electrically connected to the gateway node module and the sensor interface; the lower circuit board is provided with a dual-channel infrared tracking module, a power management module, a DC motor, a wheel and motor drive module, an RFID identification module, an ultrasonic sensor module, and a camera module. The dual-channel infrared tracking module is arranged at the bottom of the lower circuit board, and the ultrasonic sensor module is arranged at the front end of the lower circuit board. The second main control MCU is electrically connected to each module of the underlying hardware.

[0015] Furthermore, the gateway node module includes a WiFi module, a display module, and a Zigbee module.

[0016] Beneficial effects of the invention: The invention provides a teaching method and system for intelligent car task planning. It provides students with a clear quantitative benchmark through the optimal strategy generated by artificial intelligence, and elevates the learning goal from functional realization to strategy optimization. Students can not only manually control the car through programming, but also intuitively understand the optimization capabilities and principles of deep learning by comparing with the model recommendation results, which greatly stimulates students' interest in learning. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] For ease of explanation, the present invention is described in detail with reference to the following preferred embodiments and the accompanying drawings.

[0018] Figure 1 This is a flow chart of a method for teaching task planning of an intelligent vehicle according to the present invention; Figure 2 This is a functional block diagram of an intelligent vehicle task planning teaching system according to the present invention; Figure 3 This is a circuit diagram of the first main control MCU of an intelligent car task planning teaching system of the present invention; Figure 4 This is a circuit diagram of the second main control MCU of an intelligent car task planning teaching system of the present invention; Figure 5 This is a circuit diagram of a power management module for an intelligent vehicle task planning teaching system according to the present invention; Figure 6 This is a circuit diagram of a motor drive module of an intelligent vehicle task planning teaching system according to the present invention; Figure 7 This is a schematic diagram of the intelligent car results of an intelligent car task planning teaching system of the present invention; Figure 8 This is a bottom view of an intelligent vehicle in the intelligent vehicle task planning teaching system of the present invention.

[0019] In the figure, 1-upper circuit board, 2-lower circuit board, 3-sensor interface, 3.1-magnet, 4-gateway node module, 5-ultrasonic sensor module, 7-infrared tracking module, 8-RFID identification module. DETAILED DESCRIPTION

[0020] To make the implementation objectives, technical solutions, and features of the present invention more clear, the following will be combined with the accompanying drawings of the embodiments of the present invention to clearly and completely describe the technical solutions implemented in the present invention. Obviously, the described embodiments are only some examples of the present invention, not all implementation cases. Generally, the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in different configurations.

[0021] Therefore, the detailed description of the embodiments provided in the invention application is not intended to limit the scope of protection claimed in this application, but merely represents specific implementation cases of this application. Based on the implementation cases in this application, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection claimed in this invention application.

[0022] The invention is described in detail below with reference to the following examples and accompanying drawings. This example is based on the "Rapid Rescue Challenge" scenario, which features a "Rapid Rescue" mission. To enhance post-flood rescue awareness, participants are required to control a "smart rescue vehicle" to simulate post-disaster rescue operations and conduct rescue operations. The competition map consists of the following elements: a high-frequency card (hidden below the map), start / finish areas, supply stations, residential areas, roads / buildings, and a tracking line. In line with the "Rapid Rescue" theme, participants must complete the rescue mission and reach the finish line as quickly as possible. The faster the time, the higher the score.

[0023] like Figure 1-2 The method for teaching intelligent vehicle task planning includes the following steps: S1. Site Digitization: Convert the mission scenario, which includes multiple key locations and physical paths (tracing lines), into a weighted digital graph data structure (Graph). The site digitization steps specifically include: S1.1 Node Definition: Based on the multiple key locations in the scenario, define multiple nodes and assign each node a unique node ID and a node type that characterizes its function; identify all key locations on the physical map that are marked by high-frequency RFID cards. According to the rules of the "Rapid Rescue Challenge", these locations include the start / end area, supply stations, residential areas, and all road intersections. Each location where a high-frequency card is located will become a "node" in the graph structure. S1.2 Path Definition and Weight Calculation: Identify and connect the physical paths between any two nodes in the mission scenario to determine the path connection between the nodes and calculate a weight representing the travel cost for each path connection. Path Relationship Determination: Determine whether a directly traversable path exists between any two nodes. A path corresponds to a black trace line connecting two high-frequency card nodes on the map. The vehicle traverses all trace lines autonomously or under guidance. Based on the sequence of consecutively read RFID card IDs, the system can automatically infer the adjacency relationship between nodes. For example, if the vehicle departs from node '0005' and the next ID read is node '0009', the system records a path from '0005' to '0009'. Path Weight Calculation: Assign a numerical "weight" or "cost" to each identified path. The weight can represent distance or travel time: this travel time is used as the weight of the directed path connecting the two nodes. For example, the weight of going from node '0005' to node '0009' is 3.2 seconds; traversing all defined nodes and paths and building an adjacency list, an entry in the adjacency list may be as follows, indicating that starting from node '0001', you can reach '0002' and '0004': { "0001":[ { "target_node": "0002", "weight": 2.5}, / / It takes 2.5 seconds to reach 0002 { "target_node": "0004", "weight": 3.0} / / It takes 3.0 seconds to reach 0004 ], ...}.

[0024] S2 Strategy Generation: Based on the digital graph data structure generated in the site digitization step in step S1, combined with the task rules, an artificial intelligence optimization algorithm is used to deduce and generate a task execution strategy including an optimal action sequence and an optimal travel path; wherein, the strategy generation step further includes: S2.1 Environment modeling: convert the digital graph structure and task rules into a reinforcement learning environment, which at least includes a state space, an action space and a reward function; the state space in step S2.1 is used to describe the current position, load status and status of multiple task nodes in the task scenario of the smart car: the current position of the car: represented by the node ID where it is located; the load status of the car: an integer representing the current amount of materials on board; the status of the material station: a list or dictionary that records the current remaining amount of materials at each material station node.

[0025] Cell state: A list or dictionary that records the number of supplies required for each cell node. The action space includes at least movement actions and actions required by the task rules. Move to (target node): Move from the current node to an adjacent node along a defined path. Load supplies: This action is executed when the cart is at a "supply station" type node.

[0026] Unloading Supplies: This action is performed when the vehicle is located at a "cell" type node. The reward function is used to assign positive rewards to actions that achieve the mission objectives and negative rewards to actions that consume time or resources. Primary Positive Reward: +10 points are awarded for each successful unloading of supplies into a cell; Final Positive Reward: A large reward, such as +200 points, is given for successfully reaching the destination to encourage the vehicle to complete the entire mission; Continuous Negative Reward (Cost): -1 point is awarded for each "Move" action performed. This encourages the AI ​​to find shorter paths to reduce time consumption; Penalty: Negative rewards, such as -5 points, are given for performing invalid actions (such as loading at an empty supply station or unloading into a cell where demand has already been met), allowing the AI ​​to learn to avoid invalid operations.

[0027] S2.2 Model training: Through multiple rounds of simulation and deduction in the reinforcement learning environment, a deep Q network model capable of predicting the value of actions based on states is trained; step S2.2 also includes: storing the interaction experience containing states, actions, and rewards generated in the simulation and deduction in an experience replay pool, and randomly sampling a batch of interaction experiences from the experience replay pool for updating the weight parameters of the deep Q network model.

[0028] a) Initialize the Q-Network: Create a deep neural network (Q-Network) that will receive the “state” as input and output the value (Q-value) of each optional “action”.

[0029] b) Start a simulation episode: In the simulation environment, place the car at the starting point and start a complete mission simulation.

[0030] c) Trial-and-Error Exploration: At each simulation step, the AI ​​selects an action based on the current Q-network prediction and a certain amount of randomness (ε-greedy strategy). This ensures that the AI ​​both leverages existing experience and explores unknown possibilities.

[0031] d) Observation and Storing Experience: After executing an action, the AI ​​observes the environment entering a new state and receives a reward from the reward function. This process (current state, action, reward, new state) is stored as valuable experience in the "experience replay pool."

[0032] e) Learning and Reflection: The AI ​​randomly draws a batch of past experiences from its "experience replay pool" to train the Q network. By comparing the predicted action value at the time with the action value that appears more accurate in hindsight, the network continuously adjusts its parameters to make its predictions increasingly accurate.

[0033] f) Iterate: Repeat steps b) through e) over thousands or even tens of thousands of simulation rounds. In this process, the AI's Q network gradually evolves from a "novice" to an "expert," and its judgment of the long-term total reward of any action in any state becomes increasingly accurate.

[0034] S2.3 Strategy extraction: Using the trained deep Q network model, starting from the initial state, generate the optimal task execution strategy by iteratively selecting the action with the highest action value in each state; the multiple rounds of simulation and deduction described in step S2.2 adopt a greedy strategy; the iterative selection described in step S2.3 is a greedy selection.

[0035] Generate optimal action sequence a) Reset environment: Restore the simulation environment to its initial state, with the car at the starting point.

[0036] b) Greedy decision-making: In the current state, the state information is input into the trained Q network, and the network will give the value scores (Q values) of all optional actions.

[0037] c) Select the best action: Instead of random exploration, firmly choose the action with the highest Q value and record it.

[0038] d) Update state and loop: Execute the optimal action in the simulation environment, enter the next state, and repeat steps b) and c).

[0039] e) Develop a final strategy: Chain together the optimal actions recorded throughout the entire process, for example: [Go to Supply Station A -> Load 5 items -> Go to Community C -> Unload 5 items -> Go to Supply Station B -> ...], to form the final "optimal action sequence." Based on all the "movement" actions in this sequence, an "optimal path" can be mapped.

[0040] S3. User execution and data recording step: Provides a programming interface that allows users to independently write control programs to drive the smart car to perform tasks in the task scenario, and records the user task process data including timestamps, trajectories, and key actions generated by the smart car under the user's control in real time; S4. Comparative analysis and feedback step: Perform a multi-dimensional comparative analysis of the user task process data recorded in S3 and the optimal task execution strategy generated in S2 to generate a teaching feedback report containing at least one key efficiency indicator difference and a diagnostic optimization suggestion.

[0041] Furthermore, the key efficiency indicators in S4 include at least one of path efficiency, task sequence efficiency, and average score time. The diagnostic optimization suggestions in S4 include: analyzing the strategy type of the user control program based on the differences in the key efficiency indicators, and providing improved algorithm suggestions for the strategy type; the teaching feedback report also includes a visual comparison chart of the trajectory of the smart car under user control and the optimal travel path. Visual trajectory comparison chart: The student's and AI's travel routes are simultaneously presented on a map; KPI differentiation panel: clearly lists the scores and comparisons of various efficiency indicators; Diagnostic comments and optimization suggestions: Combined with the built-in expert rule library, it points out the essential differences between student strategies (such as "greedy algorithm") and optimal strategies.

[0042] Furthermore, the path efficiency is determined by calculating the ratio of the total travel distance of the intelligent car under user control to the distance of the optimal travel path; the task sequence efficiency is determined by comparing the difference between the order in which the robot visits the task nodes under user control and the node access order in the optimal action sequence.

[0043] like Figure 3-8 The embodiment of this specification also includes an intelligent vehicle task planning teaching system, including A map digitization module is used to convert the physical layout and constraints of a mission scenario containing multiple key locations and physical paths into a digital graph structure containing nodes and weighted paths; A strategy generation module is used to generate a task execution strategy including an optimal action sequence and an optimal travel path based on the digital graph structure and task rules through an artificial intelligence optimization algorithm; The data recording module is used to record the user task process data generated when the user autonomously controls the smart car through the programming interface; a comparison and diagnosis module, configured to compare and analyze the user task process data with the task execution strategy to generate a teaching feedback report including at least one key efficiency indicator difference and diagnostic optimization suggestions; The programming development module is used for users to perform graphical or code programming to compile and generate task instructions. Students use graphical code tools to drag and drop code function modules or directly write function codes. After compilation, the code is written into the first main control MCU of the car through the network module.

[0044] The network module is used for Wi-Fi communication and task instruction transmission with the smart car.

[0045] Furthermore, the system also includes an intelligent car, and each module of the system runs on a computer in the same network environment as the intelligent car.

[0046] The smart car adopts a dual MCU structure, which includes: an upper circuit board 1 and a lower circuit board 2. The upper circuit board 1 is equipped with a first master MCU, which adopts STM32F103RET6 and is responsible for communicating with the network module, receiving and executing instructions; the lower circuit board 2 is equipped with a second master MCU, which adopts AT32F403RCT6 and receives internal instructions from the first master MCU through the UART protocol, connects the USART TX pin of the STM32F103RET6 to the USART RX pin of the AT32F403RCT6; and connects the USART RX pin of the STM32F103RET6 to the USART TX pin of the AT32F403RCT6 to realize control of the underlying hardware. A battery is provided on the lower circuit board to power the first and second master MCUs. Furthermore, the upper circuit board 1 is provided with a gateway node module 4 and a sensor interface 3. The sensor interface 3 is connected by a modular magnetic POGO PIN connector with separate coupling. The corresponding sensor is provided with a magnetic POGO PIN connector interface. Magnets 3.1 are provided on both sides of the connector. The sensor is also provided with magnets 3.1 of opposite polarity for adsorption on the connector and connected through POGO pins. The sensor supports sound sensors, gas sensors, human sensors, light sensors, smoke modules, condensation sensors, temperature and humidity sensors, flame sensors, etc. to meet different task requirements. The first main control MCU and the gateway node module 4 include a WiFi module, which is ESP8266-12F, a display module, and a Zigbee module. It supports ZigBee network number configuration, and the display module and sensor interface 3 are electrically connected, displaying configuration information and enabling touch control. The lower circuit board 2 is equipped with a dual-channel infrared tracking module 7, a power management module IP5108, a DC motor, a wheel and motor driver module TB6612, an RFID identification module, an ultrasonic sensor module 5, and a camera module. The dual-channel infrared tracking module 7 is located at the bottom of the lower circuit board 2, and the ultrasonic sensor module 5 is located at the front of the lower circuit board 2. The first and second master MCUs are electrically connected to the output of the power management module. Each of these modules is connected to the first and second master MCUs via I2C interfaces.

Claims

1. A teaching method for intelligent car task planning, characterized by: The following steps are involved: S1. Site digitization: Convert a mission scenario containing multiple key locations and physical paths into a weighted digital graph data structure. The site digitization steps specifically include: S1.1 Node definition: Based on the multiple key locations in the scenario, define multiple nodes, and assign each node a unique node ID and a node type used to characterize its function; S1.2 Path Definition and Weight Calculation: Find and connect the physical paths between any two nodes in the mission scenario to determine the path connection relationship between the nodes, and calculate a weight representing the travel cost for each path connection relationship; S2 Strategy Generation: Based on the digital graph data structure generated in the site digitization step in step S1, combined with the task rules, an artificial intelligence optimization algorithm is used to deduce and generate a task execution strategy including an optimal action sequence and an optimal travel path; wherein, the strategy generation step further includes: S2.1 Environment Modeling: Converting the digital graph structure and task rules into a reinforcement learning environment, wherein the environment comprises at least a state space, an action space, and a reward function; S2.2 Model Training: Through multiple rounds of simulation in the reinforcement learning environment, a deep Q-network model capable of predicting action values ​​based on state is trained; S2.3 Strategy Extraction: Utilizing the trained deep Q-network model, starting from the initial state, the optimal task execution strategy is generated by iteratively selecting the action with the highest action value in each state; S3. User execution and data recording step: Provides a programming interface that allows users to independently write control programs to drive the smart car to perform tasks in the task scenario, and records the user task process data including timestamps, trajectories, and key actions generated by the smart car under the user's control in real time; S4. Comparative analysis and feedback step: Perform a multi-dimensional comparative analysis of the user task process data recorded in S3 and the optimal task execution strategy generated in S2 to generate a teaching feedback report containing at least one key efficiency indicator difference and a diagnostic optimization suggestion.

2. The intelligent vehicle task planning teaching method according to claim 1 is characterized in that: The state space described in step S2.1 is used to describe the current position, load status, and status of multiple task nodes in the task scenario of the smart car; the action space includes at least moving actions and actions required by task rules; the reward function is used to assign positive rewards to actions that complete task goals, and assign negative rewards to actions that consume time or resources.

3. The intelligent vehicle task planning teaching method according to claim 1 is characterized in that: The step S2.2 also includes: storing the interactive experience including state, action and reward generated in the simulation in an experience replay pool, and randomly sampling a batch of interactive experiences from the experience replay pool for updating the weight parameters of the deep Q network model.

4. The intelligent vehicle task planning teaching method according to claim 1, characterized in that: The multiple rounds of simulation and deduction in step S2.2 adopt a greedy strategy; the iterative selection in step S2.3 is a greedy selection.

5. The intelligent vehicle task planning teaching method according to claim 1 is characterized in that: The key efficiency indicators in S4 include at least one of path efficiency, task sequence efficiency, and average scoring time. The diagnostic optimization suggestions in S4 include: analyzing the strategy type of the user control program based on the differences in the key efficiency indicators and providing improved algorithm suggestions for the strategy type; the teaching feedback report also includes a visual comparison diagram of the trajectory of the intelligent car under user control and the optimal travel path.

6. The intelligent vehicle task planning teaching method according to claim 5 is characterized in that: The path efficiency is determined by calculating the ratio of the total distance traveled by the intelligent car under user control to the distance of the optimal travel path; the task sequence efficiency is determined by comparing the difference between the order in which the robot visits task nodes under user control and the node access order in the optimal action sequence.

7. An intelligent car task planning teaching system, characterized by: Includes modules that run on your computer: A map digitization module is used to convert the physical layout and constraints of a mission scenario containing multiple key locations and physical paths into a digital graph structure containing nodes and weighted paths; A strategy generation module is used to generate a task execution strategy including an optimal action sequence and an optimal travel path based on the digital graph structure and task rules through an artificial intelligence optimization algorithm; The data recording module is used to record the user task process data generated when the user autonomously controls the smart car through the programming interface; a comparison and diagnosis module, configured to compare and analyze the user task process data with the task execution strategy to generate a teaching feedback report including at least one key efficiency indicator difference and diagnostic optimization suggestions; Programming development module, used for users to perform graphical or code programming to compile and generate task instructions; The network module is used for Wi-Fi communication and task instruction transmission with the smart car.

8. The intelligent vehicle task planning teaching system according to claim 7, characterized in that: The system also includes an intelligent car, which adopts a dual MCU structure. The dual MCU structure includes: an upper circuit board and a lower circuit board. The upper circuit board is equipped with a first master MCU, which is responsible for communicating with the network module and receiving and executing instructions; the lower circuit board is equipped with a second master MCU, which receives internal instructions from the first master MCU through the UART protocol, connects the USART TX pin of the first master MCU to the USART RX pin of the second master MCU; connects the USART RX pin of the first master MCU to the USART TX pin of the second master MCU to achieve control of the underlying hardware; and a battery is provided on the lower circuit board to power the first and second master MCUs.

9. The intelligent vehicle task planning teaching system according to claim 7, characterized in that: The upper circuit board is provided with a gateway node module and a sensor interface. The sensor interface is connected using a separate coupled modular magnetic POGO PIN connector. Corresponding to the sensor provided with a magnetic POGO PIN connector interface, the first main control MCU is electrically connected to the gateway node module and the sensor interface; the lower circuit board is provided with a dual-channel infrared tracking module, a power management module, a DC motor, a wheel and motor drive module, an RFID identification module, an ultrasonic sensor module, and a camera module. The dual-channel infrared tracking module is arranged at the bottom of the front end of the lower circuit board, and the ultrasonic sensor module is arranged at the front end of the lower circuit board. The second main control MCU is electrically connected to each module of the underlying hardware.

10. The intelligent vehicle task planning teaching system according to claim 9, characterized in that: The gateway node module includes a WiFi module, a display module, and a ZigBee module.

Images