Visual navigation AGV transportation system

Abstract

The invention relates to the technical field of visual navigation, and discloses a visual navigation AGV (Automatic Guided Vehicle) transportation system, which comprises a visual guidance module, a driving module, a path planning and control module and an obstacle avoidance and safety module, the visual guidance module is responsible for acquiring real-time image or video data of an environment through a camera and transmitting the real-time image or video data to the path planning and control module, and comprises a 3D deep high-definition camera, an image acquisition card, an image processor and a display screen; the driving module is responsible for driving the AGV to move according to the control instruction, so that the AGV moves along a predetermined path, receiving speed and direction instructions from the path planning and control module, and executing the instructions; the path planning and control module is responsible for receiving environment data from the visual guidance module and calculating an optimal driving path in combination with a built-in algorithm; the obstacle avoidance and safety module is used for detecting and preventing the AGV from colliding with an obstacle in the driving process. The material carrying device has the advantage of meeting the material carrying requirement under the complex and changeable environment.

Description

Technical Field

[0001] The present invention relates to the field of vision navigation technology, in particular to a vision-guided guided vehicle (AGV) transportation system. Background Art

[0002] With the rapid development of intelligent manufacturing and automated logistics, the demand for automatic guided vehicles (AGVs) is growing. As the core equipment in intelligent warehousing and logistics transportation, AGVs undertake tasks such as material handling, transportation, and sorting, greatly improving production efficiency and reducing labor costs. Traditional AGVs mostly use technologies such as magnetic strips, lasers, or RFID for navigation. Although these methods have been widely used, they still have certain limitations. The magnetic strip navigation system requires laying magnetic strips or tracks on the ground, resulting in high installation costs and long construction periods; laser navigation requires high-precision laser scanners, which have high equipment costs and are more sensitive to changes in environmental light, reflection, dust, etc., limiting its application in complex environments; RFID technology relies on the layout of ground tags and also faces the complexity of installation and maintenance. Most importantly, traditional technologies are difficult to achieve flexible and efficient path planning and real-time adaptation in a dynamically changing environment, especially when the environment changes significantly or obstacles appear frequently, the navigation accuracy and stability are easily affected. Therefore, it is necessary to design a low-cost, highly flexible, and environmentally adaptable visual navigation AGV transportation system. Summary of the invention

[0003] 1. Technical issues to be solved

[0004] In view of the deficiencies in the prior art, the present invention provides a vision-guided AGV transportation system, which has the advantages of meeting the material handling needs in complex and changeable environments and solves the problems in the above-mentioned background technology.

[0005] (II) Technical solution

[0006] In order to achieve the above-mentioned purpose of meeting the material handling needs in complex and changing environments, the present invention provides the following technical solutions: a visually guided AGV transportation system, including a visual guidance module, a driving module, a path planning and control module, and an obstacle avoidance and safety module;

[0007] The visual guidance module is responsible for collecting real-time images or video data of the environment through the camera and transmitting it to the path planning and control module, including a 3D depth high-definition camera, an image acquisition card, an image processor and a display screen. The 3D depth high-definition camera is responsible for collecting environmental images in real time; the image acquisition card transmits the collected image data to the image processor; the image processor uses image recognition and processing technology to analyze lane lines, markers, and obstacles in the image;

[0008] The driving module is responsible for driving the movement of the AGV according to the control instructions, making the AGV move along the predetermined path, receiving the speed and direction instructions from the path planning and control module, and executing the instructions;

[0009] The path planning and control module is responsible for receiving environmental data from the visual guidance module and calculating the optimal driving path in combination with the built-in algorithm;

[0010] The obstacle avoidance and safety module is used to detect and avoid collisions between the AGV and obstacles during driving.

[0011] Preferably, the visual guidance module further includes using a deep learning algorithm to pre-process, extract features and classify images, and combining SLAM technology to autonomously locate the vehicle and construct a map;

[0012] Preprocessing: including image denoising, normalization, and color adjustment;

[0013] Feature extraction: Automatically extract high-level features from images, including edge features, texture features, and geometric features;

[0014] Classification and recognition: Use convolutional neural networks to classify areas in the image and identify vehicles, pedestrians, obstacles, and road signs.

[0015] Preferably, the driving module further includes a motor drive, combined with an encoder to control the speed and feedback position, and controls the forward and reverse rotation and speed of the motor to operate the vehicle forward, backward, left turn, and right turn.

[0016] Preferably, the path planning and control module further includes:

[0017] Path planning: Based on the environmental information recognized by the visual system and the preset target point, the path planning algorithm is used to generate the optimal driving path. The environmental information recognized by the visual system is usually input into the system through sensors to form a digital representation of the environment. In the grid map, each cell represents the state of an environmental area. The preset target point is the location that the AGV needs to reach, recorded as , where is the position of the target point in the coordinate system. The path planning algorithm is used to generate the optimal path from the starting point to the target point;

[0018] Motion control: According to the path planning results, the vehicle's speed and direction are adjusted by controlling the drive system, the vehicle status is monitored in real time, and abnormal situations are handled promptly.

[0019] Preferably, the obstacle avoidance module further includes immediately starting an obstacle avoidance algorithm, calculating a safe obstacle avoidance path, and controlling the vehicle to perform obstacle avoidance actions when the visual system detects an obstacle ahead;

[0020] The visual system obtains information about the surrounding environment in real time through sensors and generates a map or environmental model. The system processes the sensor data in real time, identifies obstacles in front or around, and confirms the location, shape and size of the obstacles, as well as the distance between the obstacles and the AGV.

[0021] The system uses the obstacle position information detected by the visual system to analyze the relationship between the obstacle and the current position of the AGV and evaluate whether there is a collision risk. If the obstacle is on the AGV's driving path and conflicts with the vehicle's current speed and motion state, it is determined to be a dangerous situation. The system also evaluates the type and size of the obstacle, taking into account the different response strategies for dynamic and static obstacles. Based on the preset safety distance, the system evaluates whether obstacle avoidance is needed. If the distance of the obstacle is less than the safety distance, the obstacle avoidance mechanism is triggered to determine whether there is a collision risk. If there is a risk, the obstacle avoidance algorithm is started.

[0022] Preferably, starting the obstacle avoidance algorithm includes selecting an obstacle avoidance algorithm based on the position of the obstacle and the current position of the AGV, using a mechanical model, treating the obstacle as a repulsive source through a virtual force field, and the AGV treating the target as a suction source, calculating a path to avoid the obstacle, performing real-time scanning and evaluation around the obstacle, calculating a safe path, and selecting the optimal path, and generating a new target position or trajectory point based on the output of the obstacle avoidance algorithm.

[0023] Preferably, the safety model further includes an emergency stop button, an ultrasonic sensor and a collision sensor;

[0024] The emergency stop button is used to immediately stop all actions of the AGV and cut off the power supply or control signal when encountering sudden danger or abnormal situation;

[0025] Ultrasonic sensors are used to measure the distance between the AGV and obstacles, detect collision risks, and trigger obstacle avoidance or emergency stop mechanisms;

[0026] The collision sensor determines whether a collision has occurred by sensing the contact between the AGV and the object. When the collision sensor senses a physical collision, the system will immediately activate the protection mechanism, including automatic shutdown or reverse adjustment of the path.

[0027] (III) Beneficial effects

[0028] Compared with the prior art, the present invention provides a visual navigation AGV transportation system, which has the following

[0029] Beneficial effects:

[0030] 1. All modules work closely together, combining the environmental data provided by the visual guidance module, the path planning and control module generates the optimal path, the drive module performs precise actions, and the obstacle avoidance and safety module responds to environmental and dynamic changes in real time to ensure that AGV can operate stably in complex and dynamic environments.

[0031] 2. Through the precise coordination of path planning, motion control and obstacle avoidance modules, the system can perform tasks efficiently and safely, reduce path deviation and collision risks, and ensure efficient task completion through dynamic adjustment.

[0032] 3. By utilizing the image processing and analysis capabilities of the visual system, combined with intelligent obstacle avoidance and dynamic path adjustment algorithms, the system can not only handle static obstacles, but also respond to dynamic changes in real time, improving the autonomous navigation capabilities of AGV in automated warehousing, manufacturing and logistics.

[0033] 4. AGV can respond and adjust quickly in various complex environments to cope with emergencies or environmental changes, while maintaining high transportation efficiency and extremely high safety, maximizing the working reliability of AGV. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is a schematic diagram of the structure of the present invention. DETAILED DESCRIPTION

[0035] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0036] Embodiment 1, the present invention provides a technical solution: a visually guided AGV transportation system, comprising a visual guidance module, a driving module, a path planning and control module and an obstacle avoidance and safety module;

[0037] The visual guidance module is responsible for collecting real-time images or video data of the environment through the camera and transmitting it to the path planning and control module, which includes a 3D depth high-definition camera, an image acquisition card, an image processor and a display screen. The 3D depth high-definition camera is responsible for collecting environmental images in real time; the image acquisition card transmits the collected image data to the image processor; the image processor uses image recognition and processing technology to analyze lane lines, markers, and obstacles in the image.

[0038] Deep learning algorithms are used to preprocess images, extract features, and classify them, and combined with SLAM technology, vehicles can be autonomously positioned and maps can be constructed.

[0039] Visual SLAM obtains environmental images through cameras, and uses algorithms to infer the position of AGVs and build environmental maps. The SLAM system extracts significant feature points in the image and matches them between different image frames. Based on the matched feature points, the SLAM algorithm estimates the motion trajectory of the AGV, combines the position information of the AGV at different times with the observed feature points, and builds a dense or sparse map of the environment. This process will gradually accumulate data and continuously improve the accuracy of the map.

[0040] Traditional SLAM methods use manually designed features, while deep learning can extract richer and more robust features through models such as CNN, making the SLAM system more stable in complex and dynamic environments. For example, deep neural networks can help extract more reliable feature points in low-texture environments or scenes with uneven lighting; by combining deep learning models with SLAM, the accuracy of map construction and positioning can be enhanced. For example, convolutional neural networks are used to extract features from each frame of an image, and the extracted deep feature points are input into the SLAM algorithm for matching, thereby improving positioning accuracy; by using deep learning optimization algorithms, errors in SLAM can be adjusted in real time to reduce positioning errors caused by dynamic environments or sensor noise.

[0041] The drive module is responsible for driving the movement of the AGV according to the control instructions, making the AGV move along the predetermined path, receiving speed and direction instructions from the path planning and control module, and executing the instructions.

[0042] It adopts motor drive, combined with encoder to control speed and feedback position, and controls the forward and reverse rotation and speed of the motor to operate the vehicle forward, backward, left turn and right turn.

[0043] Select the appropriate motor type, determine the power and response characteristics of the motor according to the load and motion requirements of the AGV, use the motor driver to control the speed and direction of the motor, control the speed of the motor by adjusting the voltage, frequency or duty cycle, and control the direction of the motor by adjusting the current direction. Encoder feedback is used to monitor the actual speed and position of the motor in real time to ensure that the motor's action is consistent with the control instructions and reduce errors caused by friction or load changes. Adjust the input of the motor driver through feedback to achieve precise control. Combined with the PID control algorithm, the speed and position of the motor can be closed-loop controlled. By monitoring the encoder feedback value in real time, comparing it with the target speed or position, adjust the input signal of the motor to maintain the stability and accuracy of the system. Adjust the speed of the motor according to the target speed. The speed information provided by the encoder is compared with the target speed, and the PWM signal of the motor is adjusted to maintain a constant speed. According to the difference between the target position and the current position, the motion state of the motor is adjusted to ensure that the AGV reaches the predetermined position accurately.

[0044] Forward and reverse control: The vehicle can move forward or backward by controlling the forward or reverse rotation of the motor. Usually, the vehicle is equipped with two or four drive motors, and forward or backward movement can be achieved by adjusting the speed and direction of the motor.

[0045] Left and right turn control: Change the direction of the vehicle by controlling the speed or direction of the left and right motors. For example:

[0046] Turn left: The left motor keeps the same or lower speed, and the right motor increases speed or reverses.

[0047] Turn right: The right motor maintains the same or lower speed, and the left motor increases speed or reverses.

[0048] The path planning and control module is responsible for receiving environmental data from the visual guidance module and calculating the optimal driving path based on the built-in algorithm.

[0049] Path planning: Based on the environmental information recognized by the visual system and the preset target point, the path planning algorithm is used to generate the optimal driving path. The environmental information recognized by the visual system is usually input into the system through sensors to form a digital representation of the environment. In the grid map, each cell represents the state of an environmental area. The preset target point is the location that the AGV needs to reach, denoted as T = (x T ,y T ), where T = (x T ,y T ) is the position of the target point in the coordinate system. Through the path planning algorithm, a path from the starting point P0 = (x0, y0) to the target point T = (x T ,y T The purpose of path planning is to find a feasible path from the starting point to the target point in the map and minimize a certain cost function.

[0050] Motion control: According to the path planning results, the vehicle's speed and direction are adjusted by controlling the drive system, the vehicle status is monitored in real time, and abnormal situations are handled promptly.

[0051] The motion control system calculates and adjusts the speed and steering angle of the AGV in real time based on the path planning results, and ensures that the vehicle accurately travels along the optimal path by precisely controlling the drive system. The system uses a variety of sensors for real-time status monitoring, and continuously feeds back the current position, speed, acceleration and posture of the AGV to ensure that the vehicle always remains on the predetermined trajectory. When abnormal situations are detected, such as sudden appearance of obstacles, sensor failure, low battery power, etc., the control system can respond in time and initiate emergency parking, automatic obstacle avoidance, path replanning or redundant control mechanisms to ensure the safe and efficient operation of the AGV. By adopting advanced control algorithms such as PID control and model predictive control (MPC), as well as sensor fusion and data processing technology, the system can achieve high-precision path tracking, dynamic environment adaptation and rapid processing of abnormal situations, thereby improving the overall operating efficiency, safety and reliability of the AGV, reducing the impact of system failures and ensuring the smooth completion of tasks.

[0052] The obstacle avoidance and safety module is used to detect and avoid collisions between AGV and obstacles during driving.

[0053] When the vision system detects an obstacle ahead, it immediately starts the obstacle avoidance algorithm, calculates a safe obstacle avoidance path, and controls the vehicle to perform obstacle avoidance actions.

[0054] The visual system obtains information about the surrounding environment in real time through sensors and generates a map or environmental model. The system processes the sensor data in real time, identifies obstacles in front or around, and confirms the location, shape and size of the obstacles, as well as the distance between the obstacles and the AGV.

[0055] The obstacle position information detected by the visual system, the system analyzes the relationship between the obstacle and the current position of the AGV, and evaluates whether there is a collision risk. If the obstacle is located on the AGV's driving path and conflicts with the vehicle's current speed and motion state, it is determined to be a dangerous situation; evaluate the type and size of the obstacle, taking into account the different response strategies for dynamic obstacles and static obstacles; based on the preset safety distance, evaluate whether obstacle avoidance is needed. If the distance of the obstacle is less than the safety distance, the obstacle avoidance mechanism is triggered to determine whether there is a collision risk. If there is a risk, the obstacle avoidance algorithm is started. The system not only relies on the visual sensor to detect the current position of the obstacle, but also combines the AGV's motion trajectory prediction with the obstacle's position change to dynamically analyze whether the obstacle may enter the AGV's path. By combining real-time motion information with environmental models, the system can predict the future position of the obstacle to avoid the risk of collision caused by dynamic obstacle changes. The visual system not only determines the location of the obstacle, but also analyzes the type and size of the obstacle. Depending on the different types of obstacles, the system will choose different response strategies:

[0056] Static obstacles: Usually avoided by replanning the path or emergency stopping.

[0057] Dynamic obstacles: Track their movement trajectory in real time, predict their future positions through intelligent obstacle avoidance algorithms, and adjust the path or speed to avoid collisions.

[0058] Based on the current speed, acceleration, steering angle and position of the obstacle of the AGV, the system makes a collision prediction to determine whether the AGV will collide with the obstacle. This prediction is based on the vehicle dynamics model and the motion state of the obstacle, and uses a physical model to calculate the shortest time to collision (TTC) and the collision point between the two. If a collision risk is predicted, the system immediately starts the obstacle avoidance algorithm. The preset safety distance is adaptively adjusted based on factors such as the speed, acceleration, and steering radius of the AGV. In dangerous situations, if the obstacle enters the preset safety distance, the system will trigger the obstacle avoidance mechanism. This mechanism will consider the current speed and acceleration of the AGV to calculate the most suitable obstacle avoidance path or parking distance to avoid other safety hazards caused by too hasty obstacle avoidance actions. When executing the obstacle avoidance command, the system will make dynamic adjustments based on the real-time status of the vehicle. Avoid new collision risks caused by excessive or too fast obstacle avoidance operations. At the same time, the system adjusts the obstacle avoidance strategy by real-time monitoring of the AGV's motion trajectory and feedback to ensure that the obstacle avoidance process is safe, smooth and efficient. Through accurate obstacle detection, real-time collision prediction and dynamic obstacle avoidance decision-making, the system significantly reduces the risk of AGV collision. Whether it is a static or dynamic obstacle, the AGV can take timely and accurate actions to avoid accidents and improve the safety of the operating environment. The system can not only avoid obstacles in a static environment through path replanning, but also respond to changes in dynamic obstacles in real time, optimize the obstacle avoidance path, and ensure that the AGV can operate efficiently in complex environments. The ability to avoid dynamic obstacles ensures that the AGV can still perform tasks quickly and accurately in crowded or dynamically changing environments.

[0059] Starting the obstacle avoidance algorithm includes selecting the obstacle avoidance algorithm according to the obstacle position and the current position of the AGV, using the mechanical model, treating the obstacle as a repulsive source through a virtual force field, and the AGV treating the target as a suction source, calculating the path to avoid the obstacle, performing real-time scanning and evaluation around the obstacle, calculating the safe path, and selecting the optimal path, and generating a new target position or trajectory point based on the output of the obstacle avoidance algorithm.

[0060] After starting the obstacle avoidance algorithm, the system selects a suitable obstacle avoidance strategy based on the position of the obstacle and the current position of the AGV, such as the obstacle avoidance algorithm based on the virtual force field, the artificial potential field method, or the A* path planning. Using the mechanical model, the obstacle is regarded as a repulsive force source, and the AGV regards the target as a suction source. By calculating the force of the virtual force field, the current position and motion state of the AGV are evaluated in real time to generate a safe path to avoid obstacles. The system scans and evaluates the area around the obstacle in real time, combines sensor data, dynamically analyzes the type, size, and motion state of the obstacle, and calculates the safest and shortest obstacle avoidance path. Based on the output of the obstacle avoidance algorithm, a new target position or trajectory point is generated, and the speed, direction, and motion strategy of the AGV are adjusted to ensure that no new collision risks are generated during the obstacle avoidance process, while avoiding instability caused by drastic changes. The system will also optimize the path planning in real time through continuous feedback and updates to ensure that the AGV completes its tasks safely and efficiently in a complex and dynamic environment.

[0061] Example 2 is as follows:

[0062] Scenario description: In an automated warehouse, a vision-guided vehicle is required to move goods from area A to area B.

[0063] System configuration: The vision-guided vehicle is equipped with two high-definition cameras (installed on the front and rear of the vehicle body respectively), an image processor, and a complete vehicle structure and drive system.

[0064] Operation process:

[0065] 1. Start the vehicle, the vision system starts working, and collects environmental images in real time.

[0066] 2. The image processor analyzes the image and identifies lane lines, obstacles and target point locations.

[0067] 3. Based on the recognition results, the path planning algorithm generates the optimal driving path.

[0068] 4. The control system controls the vehicle to travel along the planned path based on the path planning results.

[0069] 5. During driving, the visual system continuously monitors the environment and immediately activates the obstacle avoidance algorithm once an obstacle is detected.

[0070] 6. The vehicle arrives safely at Area B, completing the handling mission.

[0071] It should be noted that, in this article, relational terms such as first and second, etc. 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. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0072] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A visually guided AGV transportation system, characterized in that: It includes visual guidance module, driving module, path planning and control module and obstacle avoidance and safety module; The visual guidance module is responsible for collecting real-time images or video data of the environment through the camera and transmitting it to the path planning and control module, including a 3D depth high-definition camera, an image acquisition card, an image processor and a display screen. The 3D depth high-definition camera is responsible for collecting environmental images in real time; the image acquisition card transmits the collected image data to the image processor; the image processor uses image recognition and processing technology to analyze lane lines, markers, and obstacles in the image; The driving module is responsible for driving the movement of the AGV according to the control instructions, making the AGV move along the predetermined path, receiving the speed and direction instructions from the path planning and control module, and executing the instructions; The path planning and control module is responsible for receiving environmental data from the visual guidance module and calculating the optimal driving path in combination with the built-in algorithm; The obstacle avoidance and safety module is used to detect and avoid collisions between the AGV and obstacles during driving.

2. The visually guided AGV transportation system according to claim 1, characterized in that: The visual guidance module further includes using a deep learning algorithm to pre-process, extract features and classify images, and combining SLAM technology to autonomously locate the vehicle and build a map; Preprocessing: including image denoising, normalization, and color adjustment; Feature extraction: Automatically extract high-level features from images, including edge features, texture features, and geometric features; Classification and recognition: Use convolutional neural networks to classify areas in the image and identify vehicles, pedestrians, obstacles, and road signs.

3. The visually guided AGV transportation system according to claim 1, characterized in that: The driving module further includes a motor drive, combined with an encoder to control speed and feedback position, and controls the forward and reverse rotation and speed of the motor to operate the vehicle forward, backward, left turn, and right turn.

4. The visually guided AGV transportation system according to claim 1, characterized in that: The path planning and control module further includes: Path planning: Based on the environmental information recognized by the visual system and the preset target point, the path planning algorithm is used to generate the optimal driving path. The environmental information recognized by the visual system is usually input into the system through sensors to form a digital representation of the environment. In the grid map, each cell represents the state of an environmental area. The preset target point is the location that the AGV needs to reach, denoted as T = (x T ,y T ), where T = (x T ,y T ) is the position of the target point in the coordinate system. Through the path planning algorithm, a path from the starting point P0 = (x0, y0) to the target point T = (x T ,y T )’s optimal path; Motion control: According to the path planning results, the vehicle's speed and direction are adjusted by controlling the drive system, the vehicle status is monitored in real time, and abnormal situations are handled promptly.

5. The visually guided AGV transportation system according to claim 1, characterized in that: The obstacle avoidance module further includes immediately starting an obstacle avoidance algorithm, calculating a safe obstacle avoidance path, and controlling the vehicle to perform obstacle avoidance actions when the visual system detects an obstacle ahead; The visual system obtains information about the surrounding environment in real time through sensors and generates a map or environmental model. The system processes the sensor data in real time, identifies obstacles in front or around, and confirms the location, shape and size of the obstacles, as well as the distance between the obstacles and the AGV. The obstacle position information detected by the vision system, the system analyzes the relationship between the obstacle and the current position of the AGV, and evaluates whether there is a collision risk. If the obstacle is located in the AGV's driving path and conflicts with the vehicle's current speed and motion state, it is judged as a dangerous situation; evaluate the type and size of the obstacle, taking into account the different response strategies for dynamic obstacles and static obstacles; Based on the preset safety distance, evaluate whether obstacle avoidance is needed. If the distance of the obstacle is less than the safety distance, the obstacle avoidance mechanism is triggered to determine whether there is a risk of collision. If there is a risk, the obstacle avoidance algorithm is started.

6. The visually guided AGV transportation system according to claim 5, characterized in that: The starting obstacle avoidance algorithm includes selecting an obstacle avoidance algorithm according to the obstacle position and the current position of the AGV, using a mechanical model, treating the obstacle as a repulsive source through a virtual force field, and the AGV treating the target as a suction source, calculating a path to avoid the obstacle, performing real-time scanning and evaluation around the obstacle, calculating a safe path, and selecting the optimal path, and generating a new target position or trajectory point according to the output of the obstacle avoidance algorithm.

7. The visually guided AGV transportation system according to claim 1, characterized in that: The safety model further includes being equipped with an emergency stop button, an ultrasonic sensor and a collision sensor; The emergency stop button is used to immediately stop all actions of the AGV and cut off the power supply or control signal when encountering sudden danger or abnormal situation; Ultrasonic sensors are used to measure the distance between the AGV and obstacles, detect collision risks, and trigger obstacle avoidance or emergency stop mechanisms; The collision sensor determines whether a collision has occurred by sensing the contact between the AGV and the object. When the collision sensor senses a physical collision, the system will immediately activate the protection mechanism, including automatic shutdown or reverse adjustment of the path.

Images