Flight control system and method for an educational drone
The flight control system, which integrates multi-sensor fusion and edge processing, solves the problems of obstacle avoidance and teaching adaptability of educational drones in complex scenarios. It realizes a closed loop of multi-drone collaboration and programming learning, provides dynamic adjustment and visual evaluation functions, and improves the teaching effectiveness of STEM education.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- LANZHOU XINXING LONGYUAN INFORMATION TECHNOLOGY CO LTD
- Filing Date
- 2026-03-18
- Publication Date
- 2026-07-14
AI Technical Summary
Existing educational drones have low obstacle avoidance success rates in complex scenarios, insufficient adaptability to teaching, a disconnect between programming learning and flight practice, rigid multi-drone collaboration, and an inability to dynamically adjust.
The flight control system employs a combination of multi-sensor fusion and edge processing, integrating RGB-D cameras, ultrasonic sensors, infrared sensors, barometers, and IMU inertial measurement units. It performs obstacle avoidance through a lightweight multimodal fusion model and a reinforcement learning decision model, supports multi-aircraft collaborative communication and dynamic formation, and generates visualized teaching evaluation reports.
It improves the success rate of obstacle avoidance in complex scenarios, achieves compatibility and adaptation between flight control and children's programming tools, opens up a closed loop between programming learning and flight practice, has dynamic adjustment capabilities, and provides quantitative teaching evaluation tools.
Smart Images
Figure CN122387124A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of unmanned aerial vehicle (UAV) control technology, specifically to a flight control system and method for an educational UAV designed for youth teaching scenarios, combining teaching adaptability with dynamic collaborative capabilities. Background Technology
[0002] With the strengthening of science and technology education, the focus has shifted from early niche skills training to comprehensive literacy development that integrates knowledge from multiple disciplines, emphasizing a model that prioritizes science, technology, engineering, and mathematics (STEM).
[0003] Educational drones, as a core carrier of STEM education, need to balance flight performance with teaching adaptability, but existing technologies have the following key pain points:
[0004] Existing educational drones mostly rely on single visual obstacle avoidance or preset flight paths, resulting in low obstacle avoidance success rates and a high risk of collisions in complex scenarios such as classrooms and playgrounds.
[0005] Insufficient adaptability to teaching; the flight control logic is disconnected from children's programming education. The existing system only supports fixed command control and cannot be deeply compatible with children's programming tools. Furthermore, it does not adapt to the cognitive characteristics of teenagers, resulting in a disconnect between programming learning and flight practice during the teaching process.
[0006] Multi-drone collaboration is rigid and lacks dynamic adjustment capabilities; when multiple educational drones fly together, they only support preset formations and cannot adapt to changes in real-time scenarios. Summary of the Invention
[0007] In view of the technical deficiencies mentioned in the background art, the purpose of this invention is to provide a flight control system and method for educational drones, aiming to at least partially solve one of the technical problems in the related art.
[0008] To achieve the above objectives, in a first aspect, embodiments of the present invention provide a flight control system for an educational unmanned aerial vehicle (UAV), the system comprising:
[0009] The drone terminal hardware module is used to collect environmental data, flight attitude data, and to achieve physical safety protection;
[0010] The programming adaptation module is used to achieve compatibility between flight control and children's programming tools. It converts graphical / text programming instructions received from users into standardized control instructions.
[0011] Edge processing module, used for:
[0012] The collected data is processed through multimodal fusion to output obstacle information and flight status assessment results;
[0013] Based on the deployed reinforcement learning decision model, combined with the standardized control commands, obstacle information and flight status evaluation results, the flight control parameters are output to drive the UAV to perform corresponding actions.
[0014] If a collision risk is detected, an obstacle avoidance command will be automatically triggered.
[0015] Multi-drone collaborative communication module, used for real-time data synchronization and dynamic formation control among multiple drones;
[0016] The teaching data feedback module is used to collect flight process data and generate visual teaching evaluation reports.
[0017] As one specific implementation of this application, the UAV terminal hardware module includes a multi-sensor fusion unit and a safety execution unit;
[0018] The multi-sensor fusion unit integrates an RGB-D camera, an ultrasonic sensor, an infrared sensor, a barometer, and an IMU (Inertial Measurement Unit), wherein:
[0019] RGB-D cameras are used to identify static obstacles and scene environments;
[0020] Ultrasonic sensors are used to detect dynamic obstacles at close range;
[0021] Infrared sensors are used to compensate for visual blind spots in strong light / backlight scenes;
[0022] The barometer and the IMU (Inertial Measurement Unit) work together to collect flight attitude data;
[0023] The safety actuator includes an electromagnetic braking assembly, a detachable anti-collision buffer cover, and a low-voltage automatic return-to-home trigger module.
[0024] As one specific implementation of this application, the edge processing module performs multimodal fusion processing on the collected data through a deployed lightweight multimodal fusion model;
[0025] The lightweight multimodal fusion model is based on an improvement of MobileNet-YOLO and is used for cleaning, feature extraction, and obstacle category and coordinate identification of multi-sensor data.
[0026] The state space of the reinforcement learning decision model includes flight attitude, obstacle distance, and user commands; the action space includes speed, turning angle, and altitude; and the reward function is configured with safe flight weights, command execution accuracy weights, and cooperative stability weights, which are used to dynamically adjust flight parameters to adapt to the environment and user operations.
[0027] As one specific implementation of this application, the programming adaptation module includes a standardized programming interface and a cognitive adaptation unit;
[0028] The standardized programming interface is compatible with Scratch graphical programming and Python text programming, encapsulating complex flight control logic into visual building block instructions;
[0029] The cognitive adaptation unit is divided into low stage and high stage according to age; the low stage only opens the core control commands, while the high stage opens the advanced parameter configuration permissions, including obstacle avoidance sensitivity, formation speed and sensor sampling frequency.
[0030] As one specific implementation of this application, the multi-machine collaborative communication module adopts a dual-mode communication architecture of Bluetooth and UWB positioning;
[0031] The Bluetooth module is used to transmit control commands and non-real-time data;
[0032] The UWB positioning module is used to achieve centimeter-level positioning for multiple devices;
[0033] The multi-machine collaborative communication module has a built-in distributed dynamic collaborative algorithm, which is based on an improved Raft protocol and supports real-time data synchronization and dynamic formation adjustment of multi-machine nodes.
[0034] As a specific implementation of this application, the flight process data collected by the teaching data feedback module includes the success rate of programming instruction execution, the number of obstacle avoidance attempts, the formation coordination accuracy, the flight trajectory deviation, and the validity of sensor data;
[0035] The teaching data feedback module generates a visual report through edge computing. The report includes data statistics charts, annotations of students' operational weaknesses, and teaching optimization suggestions, and supports export in PDF format.
[0036] As one specific implementation of this application, the electromagnetic braking component of the safety execution unit automatically triggers braking deceleration when a collision risk is detected;
[0037] When the battery voltage is lower than the preset voltage, the low-voltage automatic return-to-home trigger module automatically generates a return-to-home command and sends it to the edge processing module, prioritizing the execution of the return-to-home operation.
[0038] Secondly, embodiments of the present invention also provide a flight control method for an educational unmanned aerial vehicle (UAV), applied to the flight control system of an educational UAV as described in the first aspect, the method comprising:
[0039] Collect environmental data, flight attitude data, and implement physical safety protection;
[0040] Based on the graphical / text programming instructions received from the user, they are converted into standardized control instructions;
[0041] The collected data is processed through multimodal fusion to output obstacle information and flight status assessment results;
[0042] Based on the deployed reinforcement learning decision model, combined with the standardized control commands, obstacle information and flight status evaluation results, the flight control parameters are output to drive the UAV to perform corresponding actions.
[0043] If a collision risk is detected, an obstacle avoidance command will be automatically triggered.
[0044] Real-time synchronization of position, attitude, and environmental data of each UAV; adjustment of formation structure through distributed dynamic collaborative algorithms.
[0045] Real-time data collection during flight and generation of visual teaching evaluation reports.
[0046] As one specific implementation of this application, a lightweight multimodal fusion model is deployed to perform multimodal fusion processing on the collected data;
[0047] The lightweight multimodal fusion model is based on an improvement of MobileNet-YOLO and is used for cleaning, feature extraction, and obstacle category and coordinate identification of multi-sensor data.
[0048] The state space of the reinforcement learning decision model includes flight attitude, obstacle distance, and user commands; the action space includes speed, turning angle, and altitude; and the reward function is configured with safe flight weights, command execution accuracy weights, and cooperative stability weights, which are used to dynamically adjust flight parameters to adapt to the environment and user operations.
[0049] As a specific implementation of this application, the method of adjusting the formation structure through a distributed dynamic cooperative algorithm, which is based on an improved Raft protocol, specifically includes:
[0050] Node initialization and role assignment;
[0051] Real-time data synchronization;
[0052] Real-time assessment of formation status;
[0053] Dynamically adjust decision generation;
[0054] Formation adjustment execution;
[0055] Adjusted verification and reorganization.
[0056] The technical solution provided by this invention combines multi-sensor fusion with edge processing for obstacle avoidance, improving the success rate of obstacle avoidance in complex scenarios; it supports compatibility and adaptation between flight control and children's programming tools, and classifies them by age to create a closed loop between programming learning and flight practice, improving teaching adaptability; at the same time, through a dual-mode communication architecture and a distributed dynamic collaborative algorithm, it has dynamic adjustment capabilities, effectively solving the problem of existing formation rigidity; and it generates a visualized teaching evaluation report, providing a quantitative evaluation tool for STEM education and realizing innovation in teaching auxiliary functions. Attached Figure Description
[0057] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below.
[0058] Figure 1 This is a schematic block diagram of a flight control system for an educational drone provided in an embodiment of the present invention;
[0059] Figure 2 This is a flowchart of a flight control method for an educational drone provided in an embodiment of the present invention. Detailed Implementation
[0060] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0061] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.
[0062] Throughout this specification, references to "an embodiment," "an example," or "an example" mean that a particular feature, structure, or characteristic described in connection with that embodiment or example is included in at least one embodiment of the invention. Therefore, the phrases "in an embodiment," "in an embodiment," "an example," or "an example" appearing in various places throughout the specification do not necessarily refer to the same embodiment or example. Furthermore, specific features, structures, or characteristics can be combined in one or more embodiments or examples in any suitable combination and / or sub-combination.
[0063] It should be noted that, unless otherwise stated, the technical terms used in this embodiment have the common meaning as understood in the relevant technical field.
[0064] Please refer to Figure 1 The present invention provides a flight control system for an educational unmanned aerial vehicle (UAV), the system comprising:
[0065] The drone terminal hardware module is used to collect environmental data, flight attitude data, and to achieve physical safety protection;
[0066] The programming adaptation module is used to achieve compatibility between flight control and children's programming tools. It converts graphical / text programming instructions received from users into standardized control instructions.
[0067] Edge processing module, used for:
[0068] The collected data is processed through multimodal fusion to output obstacle information and flight status assessment results;
[0069] Based on the deployed reinforcement learning decision model, combined with the standardized control commands, obstacle information and flight status evaluation results, the flight control parameters are output to drive the UAV to perform corresponding actions.
[0070] If a collision risk is detected, an obstacle avoidance command will be automatically triggered.
[0071] Multi-drone collaborative communication module, used for real-time data synchronization and dynamic formation control among multiple drones;
[0072] The teaching data feedback module is used to collect flight process data and generate visual teaching evaluation reports.
[0073] The UAV terminal hardware module includes a multi-sensor fusion unit and a safety execution unit;
[0074] The multi-sensor fusion unit integrates an RGB-D camera, an ultrasonic sensor, an infrared sensor, a barometer, and an IMU (Inertial Measurement Unit), wherein:
[0075] RGB-D cameras are used to identify static obstacles and scene environments;
[0076] Ultrasonic sensors are used to detect dynamic obstacles at close range;
[0077] Infrared sensors are used to compensate for visual blind spots in strong light / backlight scenes;
[0078] The barometer and the IMU (Inertial Measurement Unit) work together to collect flight attitude data;
[0079] The safety actuator includes an electromagnetic braking assembly, a detachable anti-collision buffer cover, and a low-voltage automatic return-to-home trigger module.
[0080] Furthermore, the electromagnetic braking component of the safety execution unit automatically triggers braking and deceleration when a collision risk is detected;
[0081] When the battery voltage is lower than the preset voltage, the low-voltage automatic return-to-home trigger module automatically generates a return-to-home command and sends it to the edge processing module, prioritizing the execution of the return-to-home operation.
[0082] The edge processing module performs multimodal fusion processing on the collected data through a deployed lightweight multimodal fusion model;
[0083] The lightweight multimodal fusion model is based on an improvement of MobileNet-YOLO and is used for cleaning, feature extraction, and obstacle category and coordinate identification of multi-sensor data.
[0084] The state space of the reinforcement learning decision model includes flight attitude, obstacle distance, and user commands; the action space includes speed, turning angle, and altitude; and the reward function is configured with safe flight weights, command execution accuracy weights, and cooperative stability weights, which are used to dynamically adjust flight parameters to adapt to the environment and user operations.
[0085] It should be noted that the training process of the lightweight multimodal fusion model includes:
[0086] S1: Collect a large number of multi-sensor data samples from educational scenarios, including classrooms, playgrounds, roads, and laboratories, and label obstacle categories, coordinates, and ambient light intensity levels;
[0087] S2: The sample set was expanded using Mosaic data augmentation, random cropping, and brightness adjustment methods;
[0088] S3: Trained using the PyTorch framework, employing the Adam optimizer, with dynamic adjustment of the learning rate, until the model loss value is ≤0.01.
[0089] Visual sensors identify static obstacles, such as tables, chairs, and walls; ultrasonic sensors detect nearby dynamic obstacles, such as moving people; infrared sensors compensate for blind spots in bright / backlit scenes; barometers and IMU work together to correct flight attitude.
[0090] The embedded reinforcement learning decision model dynamically adjusts flight parameters, such as speed, turning angle, and altitude, based on the teenager's operating instructions and real-time environmental data. This reduces the difficulty of operation while automatically correcting operational deviations and avoiding collisions caused by misoperation.
[0091] Furthermore, by employing edge processing, a balance is struck between real-time performance and power consumption; this overcomes the shortcomings of current cloud inference, such as high communication power consumption, shortened battery life, and dependence on the network environment.
[0092] In this embodiment, the programming adaptation module includes a standardized programming interface and a cognitive adaptation unit;
[0093] The standardized programming interface is compatible with Scratch graphical programming and Python text programming, encapsulating complex flight control logic into visual building block instructions.
[0094] The cognitive adaptation unit is divided into low stage and high stage according to age; the low stage only opens the core control commands, while the high stage opens the advanced parameter configuration permissions, including obstacle avoidance sensitivity, formation speed and sensor sampling frequency.
[0095] Furthermore, the multi-machine collaborative communication module adopts a dual-mode communication architecture of Bluetooth and UWB positioning;
[0096] The Bluetooth module is used to transmit control commands and non-real-time data;
[0097] The UWB positioning module is used to achieve centimeter-level positioning for multiple devices;
[0098] The multi-machine collaborative communication module has a built-in distributed dynamic collaborative algorithm, which is based on an improved Raft protocol and supports real-time data synchronization and dynamic formation adjustment of multi-machine nodes.
[0099] By deploying a distributed dynamic collaborative algorithm, each drone acts as a node, sharing its position, attitude, and environmental data in real time, and dynamically adjusting the formation structure through a consensus protocol; when encountering obstacles, the formation is automatically split, avoided, and then reassembled to solve the problem of rigid preset formations.
[0100] When applied, the flight process data collected by the teaching data feedback module includes the success rate of programming command execution, number of obstacle avoidance attempts, formation coordination accuracy, flight trajectory deviation, and sensor data validity.
[0101] The teaching data feedback module generates visual reports through edge computing. These reports include data statistics charts, annotations of students' operational weaknesses, and teaching optimization suggestions, and can be exported in PDF format. By generating visual teaching data reports, the module assists teachers in evaluating students' learning outcomes.
[0102] The above solution improves obstacle avoidance success rates in complex scenarios by combining multi-sensor fusion and edge processing; it supports compatibility and adaptation between flight control and children's programming tools, and categorizes users by age to create a closed loop between programming learning and flight practice, thus enhancing teaching adaptability; it also enables dynamic adjustment capabilities through a dual-mode communication architecture and distributed dynamic collaborative algorithms, effectively solving the problem of rigid formations; and it generates visualized teaching evaluation reports, providing quantitative evaluation tools for STEM education and achieving innovation in teaching support functions.
[0103] Reference Figure 2 Based on the same inventive concept, this invention also provides a flight control method for an educational drone, applied to the flight control system of the educational drone described above, the method comprising:
[0104] S101 collects environmental data, flight attitude data, and provides physical safety protection;
[0105] S102, based on the graphical / text programming instructions received from the user, converts them into standardized control instructions;
[0106] S103 performs multimodal fusion processing on the collected data and outputs obstacle information and flight status assessment results;
[0107] S104, Based on the deployed reinforcement learning decision model, combined with the standardized control commands, obstacle information and flight status evaluation results, the flight control parameters are output to drive the UAV to perform corresponding actions;
[0108] S105: If a collision risk is detected, an obstacle avoidance command will be automatically triggered.
[0109] S106 synchronizes the position, attitude and environmental data of each UAV in real time, and adjusts the formation structure through a distributed dynamic cooperative algorithm.
[0110] S107 collects flight process data in real time and generates a visual teaching evaluation report.
[0111] Furthermore, the collected data is processed through multimodal fusion using a lightweight multimodal fusion model.
[0112] The lightweight multimodal fusion model is based on an improvement of MobileNet-YOLO and is used for cleaning, feature extraction, and obstacle category and coordinate identification of multi-sensor data.
[0113] The state space of the reinforcement learning decision model includes flight attitude, obstacle distance, and user commands; the action space includes speed, turning angle, and altitude; and the reward function is configured with safe flight weights, command execution accuracy weights, and cooperative stability weights, which are used to dynamically adjust flight parameters to adapt to the environment and user operations.
[0114] Furthermore, the adjustment of the formation structure through a distributed dynamic cooperative algorithm, which is based on an improved Raft protocol, specifically includes:
[0115] Node initialization and role assignment; When multiple drones are networked, one drone is randomly selected as a temporary master node, and the rest are slave nodes. The master node issues initial formation parameters, including target formation, spacing and flight altitude, etc. All nodes synchronously load basic configurations such as UWB positioning sampling frequency and Bluetooth data synchronization period.
[0116] Real-time data synchronization, including:
[0117] Each UAV collects its own real-time position coordinates through the UWB module, attitude data through the IMU, and environmental obstacle data, such as category and distance, through the multi-sensor fusion unit.
[0118] The slave node synchronizes the above data to the master node via Bluetooth according to the synchronization cycle. The master node then summarizes the data and broadcasts the global data to all nodes, including the positions / attitudes of all drones and the global obstacle distribution.
[0119] Real-time formation status assessment, including:
[0120] The master node evaluates the formation status based on global data:
[0121] Verify the positional deviation and attitude consistency of each node;
[0122] Identify whether there are obstacles intruding into the formation's safe zone or whether node offset exceeds a threshold;
[0123] Dynamically adjust decision generation, including:
[0124] If there are no abnormalities, the master node issues a command to maintain formation, and each node fine-tunes its position / attitude to maintain formation.
[0125] If an obstacle / node offset is detected, the master node calculates the optimal adjustment strategy based on the improved Raft algorithm, including splitting the formation into multiple sub-formations, planning obstacle avoidance paths, and adjusting node spacing.
[0126] The decision is synchronized to all nodes. Once confirmed, the adjustment is triggered. Nodes can raise objections. If more than 50% of the nodes object, the calculation is recalculated.
[0127] Formation adjustments will be implemented, including:
[0128] After receiving the adjustment command, each node independently executes flight actions, such as acceleration / deceleration, turning, and altitude gain / loss, based on the obstacle avoidance parameters output by its own edge processing module. During the process, it feeds back the execution status to the master node at a set period.
[0129] Post-adjustment verification and reorganization include:
[0130] After completing obstacle avoidance / offset correction, the master node reassesses the formation status and confirms that all nodes have returned to the preset safe range.
[0131] Once the formation reorganization command is issued, each node works together to restore the initial formation or a new formation adapted to the scenario. After synchronization is complete, the normal data synchronization rhythm is restored.
[0132] It should be noted that for a more detailed description of the workflow of the method embodiments, please refer to the aforementioned system embodiments section, which will not be repeated here.
[0133] The entire solution improves obstacle avoidance success rates in complex scenarios by combining multi-sensor fusion and edge processing; it supports compatibility and adaptation between flight control and children's programming tools, and categorizes users by age to create a closed loop between programming learning and flight practice, thus enhancing teaching adaptability; it also enables dynamic adjustment through a dual-mode communication architecture and distributed dynamic collaborative algorithms, effectively solving the problem of rigid formations; and it generates visualized teaching evaluation reports, providing quantitative evaluation tools for STEM education and achieving innovation in teaching support functions.
[0134] In the embodiments provided in this application, it should be understood that the disclosed methods and systems can also be implemented in other ways. The system embodiments described above are merely illustrative. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the figures. For example, two consecutive blocks may actually be executed substantially in parallel, or they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, can be implemented using a dedicated hardware-based system that performs the specified functions or actions, or using a combination of dedicated hardware and computer instructions.
[0135] Furthermore, the functional modules in the various embodiments of this invention can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part. When using each module, information is collected and stored only with the full authorization of the relevant user or organization and in compliance with relevant laws and regulations, and the security and privacy of the data are protected. Unauthorized access is strictly prohibited. Data processing will be carried out within the scope stipulated by law and will not exceed the authorized purpose and scope. At the same time, the authorizing party has the right to access, correct, delete, restrict processing, refuse, etc., of its personal data, and strictly comply with applicable laws and regulations and conduct compliance reviews.
[0136] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. It should be noted that, in this document, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. The above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention can have various modifications and variations. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.
[0137] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.
Claims
1. A flight control system for an educational unmanned aerial vehicle (UAV), characterized in that, The system includes: The drone terminal hardware module is used to collect environmental data, flight attitude data, and to achieve physical safety protection; The programming adaptation module is used to achieve compatibility between flight control and children's programming tools. It converts graphical / text programming instructions received from users into standardized control instructions. Edge processing module, used for: The collected data is processed through multimodal fusion to output obstacle information and flight status assessment results; Based on the deployed reinforcement learning decision model, combined with the standardized control commands, obstacle information and flight status evaluation results, the flight control parameters are output to drive the UAV to perform corresponding actions. If a collision risk is detected, an obstacle avoidance command will be automatically triggered. Multi-drone collaborative communication module, used for real-time data synchronization and dynamic formation control among multiple drones; The teaching data feedback module is used to collect flight process data and generate visual teaching evaluation reports.
2. The flight control system for an educational unmanned aerial vehicle as described in claim 1, characterized in that, The UAV terminal hardware module includes a multi-sensor fusion unit and a safety execution unit; The multi-sensor fusion unit integrates an RGB-D camera, an ultrasonic sensor, an infrared sensor, a barometer, and an IMU (Inertial Measurement Unit), wherein: RGB-D cameras are used to identify static obstacles and scene environments; Ultrasonic sensors are used to detect dynamic obstacles at close range; Infrared sensors are used to compensate for visual blind spots in strong light / backlight scenes; The barometer and the IMU (Inertial Measurement Unit) work together to collect flight attitude data; The safety actuator includes an electromagnetic braking assembly, a detachable anti-collision buffer cover, and a low-voltage automatic return-to-home trigger module.
3. The flight control system for an educational unmanned aerial vehicle as described in claim 2, characterized in that, The edge processing module performs multimodal fusion processing on the collected data through a deployed lightweight multimodal fusion model; The lightweight multimodal fusion model is based on an improvement of MobileNet-YOLO and is used for cleaning, feature extraction, and obstacle category and coordinate identification of multi-sensor data. The state space of the reinforcement learning decision model includes flight attitude, obstacle distance, and user commands; the action space includes speed, turning angle, and altitude; and the reward function is configured with safe flight weights, command execution accuracy weights, and cooperative stability weights, which are used to dynamically adjust flight parameters to adapt to the environment and user operations.
4. The flight control system for an educational unmanned aerial vehicle as described in claim 1, characterized in that, The programming adaptation module includes a standardized programming interface and a cognitive adaptation unit; The standardized programming interface is compatible with Scratch graphical programming and Python text programming, encapsulating complex flight control logic into visual building block instructions. The cognitive adaptation unit is divided into low stage and high stage according to age; the low stage only opens the core control commands, while the high stage opens the advanced parameter configuration permissions, including obstacle avoidance sensitivity, formation speed and sensor sampling frequency.
5. The flight control system for an educational unmanned aerial vehicle as described in claim 1, characterized in that, The multi-machine collaborative communication module adopts a dual-mode communication architecture of Bluetooth and UWB positioning; The Bluetooth module is used to transmit control commands and non-real-time data; The UWB positioning module is used to achieve centimeter-level positioning for multiple devices; The multi-machine collaborative communication module has a built-in distributed dynamic collaborative algorithm, which is based on an improved Raft protocol and supports real-time data synchronization and dynamic formation adjustment of multi-machine nodes.
6. The flight control system of an educational unmanned aerial vehicle as described in any one of claims 2-5, characterized in that, The flight process data collected by the teaching data feedback module includes the success rate of programming command execution, number of obstacle avoidance attempts, formation coordination accuracy, flight trajectory deviation, and sensor data validity. The teaching data feedback module generates a visual report through edge computing. The report includes data statistics charts, annotations of students' operational weaknesses, and teaching optimization suggestions, and supports export in PDF format.
7. The flight control system for an educational unmanned aerial vehicle as described in claim 6, characterized in that, The electromagnetic braking component of the safety execution unit automatically triggers braking and deceleration when a collision risk is detected; When the battery voltage is lower than the preset voltage, the low-voltage automatic return-to-home trigger module automatically generates a return-to-home command and sends it to the edge processing module, prioritizing the execution of the return-to-home operation.
8. A flight control method for an educational unmanned aerial vehicle (UAV), characterized in that, The method, applied to the flight control system of an educational unmanned aerial vehicle as described in claim 1, comprises: Collect environmental data, flight attitude data, and implement physical safety protection; Based on the graphical / text programming instructions received from the user, they are converted into standardized control instructions; The collected data is processed through multimodal fusion to output obstacle information and flight status assessment results; Based on the deployed reinforcement learning decision model, combined with the standardized control commands, obstacle information and flight status evaluation results, the flight control parameters are output to drive the UAV to perform corresponding actions. If a collision risk is detected, an obstacle avoidance command will be automatically triggered. Real-time synchronization of position, attitude, and environmental data of each UAV; adjustment of formation structure through distributed dynamic collaborative algorithms. Real-time data collection during flight and generation of visual teaching evaluation reports.
9. The method as described in claim 8, characterized in that, The collected data is processed through multimodal fusion using a lightweight multimodal fusion model. The lightweight multimodal fusion model is based on an improvement of MobileNet-YOLO and is used for cleaning, feature extraction, and obstacle category and coordinate identification of multi-sensor data. The state space of the reinforcement learning decision model includes flight attitude, obstacle distance, and user commands; the action space includes speed, turning angle, and altitude; and the reward function is configured with safe flight weights, command execution accuracy weights, and cooperative stability weights, which are used to dynamically adjust flight parameters to adapt to the environment and user operations.
10. The method as described in claim 8, characterized in that, The process of adjusting the formation structure through a distributed dynamic cooperative algorithm, which is based on an improved version of the Raft protocol, specifically includes: Node initialization and role assignment; Real-time data synchronization; Real-time assessment of formation status; Dynamically adjust decision generation; Formation adjustment execution; Adjusted verification and reorganization.