Automatic following and autonomous navigation system for fruit and vegetable collection and transportation

Abstract

Disclosed is an automatic following and autonomous navigation system for fruit and vegetable collection and transportation, falling within the technical field of following navigation. The system includes a control center, and the control center is communicatively connected to an environment perception module, a target recognition and positioning module, a relative heading deviation detection module, a deviation adjustment decision-making module, a power and steering execution module, an autonomous navigation planning module and a human—machine interaction module. Through autonomous navigation and obstacle avoidance functions, a follower vehicle can automatically and accurately track a leader vehicle or autonomously plan its path, accomplishing complex transportation tasks without human intervention. Simultaneously, the system can rapidly calculate an optimal path based on target and current locations, minimizing transportation time and enhancing operational efficiency. The system can obtain the position and attitude information of the follower vehicle in real time, ensuring stable driving in complex environments. FIG. 1

Description

TECHNICAL FIELD The present invention relates to the technical eld of following navigation, and specically to an . BACKGROUND With the advancement of technology and the acceleration of agricultural modernization, agricultural production methods are gradually transitioning towards intelligent and automated approaches. In the traditional process of fruit and vegetable collection and transportation, a signicant amount of manual labor is required for handling and following, which not only results in high labor intensity but also low efciency. However, with the increase in consumer demand for high-quality fruits and vegetables and the expansion of agricultural production scale, the need for efcient and intelligent fruit and vegetable collection and transportation systems is growing. By realizing the automatic following and autonomous navigation of follower vehicles for fruit and vegetable collection and transportation, agricultural production efciency can be signicantly improved and labor costs can be reduced, conforming to the trend of agricultural modernization. During trajectory tracking and automatic following, the heading deviation of the follower vehicle relative to the leader vehicle may be affected due to insufcient width of the perception eld of view, which can easily result in the follower vehicle deviating from the intended direction and losing track of the target, making it difcult to regain its position. Therefore, the problem to be solved lies in dynamically adjusting the relative heading deviation between the follower vehicle and the leader vehicle while achieving continuous deviation detection, thereby improving the system's control precision. For this purpose, an is provided. SUMMARY An objective of the present invention is to provide an , for solving the problems raised in the above background. To solve the above technical problems, the present invention employs the following technical solutions. The includes a control center, and the control center is communicatively connected to an environment perception module, a target recognition and positioning module, a relative heading deviation detection module, a deviation adjustment decisionmaking module, a power and steering execution module, an autonomous navigation planning module and a human-machine interaction module, in which each module is connected by electrical signals. The environment perception module is congured to collect environmental information around a follower vehicle for fruit and vegetable collection and transportation containing position, velocity, direction, surrounding obstacles and road boundary information of a leader vehicle, thereby providing realtime and accurate perception data, and laying the foundation for subsequent decision-making and control. The target recognition and positioning module is congured to identify and locate a coordinate position and attitude information of the leader vehicle in space based on the data obtained by the environment perception module, distinguish the leader vehicle from other interfering objects, and continuously lock the leader vehicle, thereby ensuring that the follower vehicle's target remains unambiguous, guaranteeing that the follower vehicle always tracks the correct target, and improving the accuracy of following. The relative heading deviation detection module is congured to compare the position and attitude data of the follower vehicle and the leader vehicle, and analyze a heading deviation of the follower vehicle relative to the leader vehicle to achieve continuous deviation detection, thereby providing a precise basis for direction adjustment of the follower vehicle. The deviation adjustment decisionmaking module is congured to determine whether the current following situation is normal or not based on analysis results of the relative heading deviation detection module, issue early warning to a detected deviation trend, and update a direction adjustment strategy for the follower vehicle. The power and steering execution module is congured to issue a control instruction based on the strategy of the deviation adjustment decision-making module, and control steering, acceleration and braking actions of the follower vehicle to realize the precise control of the follower vehicle, thereby improving the control precision and response speed of the system. The autonomous navigation planning module is congured to plan an optimal path from a current position to a target position by relying on autonomous navigation technology while considering comprehensive obstacle factors when either no leader vehicle is present or its signal is lost, and guide the follower vehicle forward accordingly, thereby causing the follower vehicle to reach a destination safely and efciently, and ensuring continuous operation even in complex environments. The human-machine interaction module is congured to provide a visual operation interface that displays a running status of the system, including the positions of the follower vehicle and the leader vehicle, the relative heading deviation, and whether there is abnormal information, thereby allowing operators to perform necessary parameter settings and manual intervention control operations, and facilitating exible adjustments to the system under special circumstances. In a further improvement of the present invention, a process of collecting the environmental information within the environment perception module includes the steps of: deploying various types of sensors such as cameras and laser radars on the follower vehicle for fruit and vegetable collection and transportation to capture surrounding environmental information, such as relative position, velocity and direction information of the leader vehicle, distance, shape and size information of surrounding obstacles, and outline and position information of road boundaries; collecting and processing images through the cameras, identifying lanes and obstacles ahead, sending light beams through laser radars, synchronously receiving reected light, measuring a distance from the obstacles using a time difference, and drawing a point cloud map to determine the shape, size and distance of the obstacles; transmitting the relevant data of the collected surrounding environment information to an electronic control unit of the vehicle for preprocessing such as denoising, ltering, data format conversion to improve the accuracy and usability of data, and calibrating and synchronizing the data collected by different sensors to keep the data consistent in time and space; and fusing the data from different sensors to obtain comprehensive and accurate environmental information, and extracting the position, velocity, acceleration of the obstacles and geometric features of the road boundaries from the fused data to obtain a feature dataset. In a thher improvement of the present invention, a process of distinguishing the leader vehicle from other interfering objects within the target recognition and positioning module includes the steps of: iterating through each feature data in the feature dataset obtained by the environment perception module, and extracting the feature data including the size, shape and color of the leader vehicle using image processing technology to facilitate the recognition and positioning; matching the extracted feature data of the leader vehicle with a pre-trained leader vehicle feature model through a machine learning algorithm, screening out candidate targets that match the leader vehicle features from matching results, comparing the features of these candidate targets with the obstacle features provided by the environment perception module, and eliminating interference objects that deviate from the leader vehicle features; calculating the coordinate position of the leader vehicle in space using the image processing technology of image registration based on the selected leader vehicle features, and estimating the attitude information of an orientation and a tilt angle of the leader vehicle by combining the shape and size information of the leader vehicle; and associating the coordinate position, orientation and tilt angle information of the leader vehicle with point cloud data obtained from the sensors to improve the accuracy of positioning, continuously tracking and locking the leader vehicle using a tracking algorithm (Kalman ltering) to ensure that the follower vehicle can always accurately follow the leader vehicle, and dynamically updating the position and attitude information of the leader vehicle based on the real-time data provided by the environment perception module to adapt to environmental changes. In a further improvement of the present invention, a calculation formula of the coordinate position of the leader vehicle in space is denoted as: A 2 A =(, )=( / MT» ( , +?» where is a coordinate position of the leader vehicle in space, and represent coordinate positions of the leader vehicle on a horizontal plane, is a horizontal distance between the leader vehicle and the follower vehicle, A is a vertical distance between the leader vehicle and the follower vehicle, is a tilt angle of the leader vehicle relative to the horizontal plane, and is a maximum value in the vertical direction for limiting a range of ; a calculation formula of the orientation of the leader vehicle is denoted as: = ZOL) where is an orientation of the leader vehicle, 2 is an arctangent function considering quadrants for calculating an orientation angle of the leader vehicle, A is the vertical distance between the leader vehicle and the follower vehicle, and is the horizontal distance between the leader vehicle and the follower vehicle; and a calculation formula of the tilt angle of the leader vehicle is denoted as: = (All), where is a tilt angle of the leader vehicle, Ah is a height change of the leader vehicle, is the horizontal distance between the leader vehicle and the follower vehicle, and is an arctangent function for calculating the tilt angle of the leader vehicle. In a further improvement of the present invention, an implementation process of continuous deviation detection within the relative heading deviation detection module includes the steps of: obtaining the position and attitude data of the follower vehicle and the leader vehicle through the environment perception module, preprocessing the obtained position and attitude information such as denoising, ltering and data alignment to improve the accuracy and reliability of the data, and extracting feature information for calculating the relative heading deviation from the preprocessed data, with the position data being global positioning system (GPS) coordinates, and the attitude information including velocity vectors (velocity magnitude and direction) and an angle of orientation (vehicle orientation); comparing the heading deviation of the follower vehicle (based on the angle of orientation) with that of the leader vehicle, calculating a difference between the two, resulting in a relative heading deviation, and denoting a calculation formula of the relative heading deviation as: A = (| |, 360o I I), where A is a relative heading deviation, is a heading angle of the leader vehicle, and is a heading angle of the follower vehicle; and setting a timer to trigger a data update and calculation process every 500 milliseconds, updating the position and attitude data of the follower vehicle and the leader vehicle regularly, and repeating a heading angle comparison process for the continuous deviation detection. In a further improvement of the present invention, a process of deviation trend warning within the deviation adjustment decision-making module includes the steps of: setting a normal range and an early warning threshold of the relative heading deviation, with the normal range being i 50 and the early warning threshold being i 10°, and receiving a realtime analysis result from the relative heading deviation detection module including a current value and historical data of the relative heading deviation; determining whether a following situation of the current follower vehicle relative to the leader vehicle is within the normal range based on the received relative heading deviation data, the current following situation being normal and continued to be monitored if it is within the normal range, and a deviation early warning and processing ow being entered if it exceeds the normal range; determining a deviation degree based on an absolute value of the relative heading deviation, calculating a deviation early warning coefcient by combining the relative heading deviation data at multiple consecutive time points, and analyzing the deviation trend of the relative heading deviation; combining analysis results of the deviation trend and the calculated deviation early warning coefcient, and triggering an early warning mechanism to notify a driver through sound and light if the deviation early warning coefcient approaches or exceeds the early warning threshold and the deviation trend continues or deteriorates; and updating the direction adjustment strategy for the follower vehicle to correct the deviation and keep the follower vehicle on a correct heading based on the deviation early warning coefcient and the analysis results of the deviation trend, determining the direction (left, right or hold) to be adjusted by the follower vehicle based on the relative heading deviation and the deviation trend, calculating an angle or velocity variable quantity to be adjusted by combining the deviation early warning coefcient and the preset adjustment strategy, and sending an adjustment instruction to the power and steering execution module of the follower vehicle to execute the direction adjustment operation. In a further improvement of the present invention, a calculation formula of the deviation early warning coefcient is denoted as: 1 _ A | A 1+ (A_0) >< 1 >< A where is a deviation early warning coefcient, A is a current value of the relative heading deviation, 0 is a preset deviation early warning threshold, is a standard deviation for adjusting the sensitivity of the deviation early warning coefcient, A is a relative heading deviation data at an th consecutive time point, is a quantity of consecutive time points, A is a maximum relative heading deviation within the continuous time points, A is a minimum relative heading deviation within the continuous time points, and a value range of is limited between 0 and 1, with O indicating no deviation and 1 indicating a serious deviation. In a irther improvement of the present invention,an implementation process of controlling the follower vehicle within the power and steering execution module includes the steps of: receiving the adjustment strategy updated by the deviation adjustment decisionmaking module through a communication interface within the power and steering execution module, extracting the control instruction in the adjustment strategy, and analyzing the instruction; extracting specic control parameters of steering angle, acceleration or braking force based on analysis results of the control instruction, and determining types of actions to be executed including steering, acceleration or braking based on the analysis results; controlling a steering mechanism based on the steering angle parameters to perform the steering if the control instruction is a steering action during steering action execution,and monitoring the steering angle and steering velocity in real time through the sensors to ensure the accuracy and stability of the steering action during the steering; controlling an output power of engine based on the acceleration parameters to achieve the acceleration if the control instruction is an acceleration action during acceleration action execution, and monitoring the vehicle speed and acceleration in real time through the sensors to ensure the stability and safety of acceleration action in the process of acceleration; and controlling a braking system to generate corresponding braking force based on the braking force parameters for deceleration or stopping if the control instruction is a braking action during braking action execution, monitoring the vehicle speed and deceleration in real time through the sensors to ensure the timeliness and reliability of the braking action during the braking, and simultaneously or sequentially executing the steering,acceleration and braking actions to realize precise control of the follower vehicle; and monitoring the actual state of the vehicles in real time through the sensors during the execution of the steering, acceleration and braking actions, comparing the actual state with an expected target, evaluating execution effects, feeding back execution results (actual steering angle, vehicle speed, acceleration) to the deviation adjustment decision-making module through the communication interface, and further adjusting the control strategy based on feedback within the deviation adjustment decisionmaking module to improve the control precision and response speed of the system. In a irther improvement of the present invention, a process of guiding the follower vehicle forward using the autonomous navigation planning within the autonomous navigation planning module includes the steps of: processing the surrounding environment information collected by the follower vehicle,converting it into a raster map for representation, and intuitively reecting the obstacles and feasible areas in the environment using the raster map; estimating the current position of the follower vehicle using a visual simultaneous localization and mapping (SLAM) algorithm based on the environmental information and the known map, and obtaining the attitude information of the follower vehicle in real time using an inertial measurement unit (IMU); detecting the obstacles in the environment in real time based on the sensor data and map information, formulating obstacle avoidance strategies such as bypassing, stopping and waiting based on positions, shapes and sizes of the obstacles, introducing an obstacle cost function into the path planning,and evaluating inuence degrees of obstacles in a path to avoid the obstacles and optimize the path; searching for a shortest path to avoid obstacles and follow trafc rules in the map using a search algorithm (Astar algorithm) based on the target position and current position, and performing dynamic adjustment and optimization according to a real-time state of the follower vehicle and changes of the surrounding environment based on a global path planning, ensuring that the follower vehicle can operate smoothly; and generating the control instruction according to path tracking results and the realtime state of the follower vehicle, and controlling the steering, acceleration and braking actions of the follower vehicle to achieve the precise control. In a further improvement of the present invention, an expression of the obstacle cost function is denoted as: 2 =1 _ ( > > , = < >< (1, _), where is an obstacle cost inction for evaluating the inuence degrees of obstacles in the path, is a distance from the follower vehicle to a th obstacle, is a standard deviation of obstacle distance for adjusting the sensitivity of the cost inction to the distance, is a total number of obstacles in the path, is a constant for adjusting a curvature of an exponential mction and affecting the inuence degree of the obstacle distance on the cost mction, is a maximum angular deviation allowed in the path, is an angular deviation between the follower vehicle and the th obstacle, and a value range of is limited between 0 and 1, with 0 indicating that the path completely avoids the obstacles, and l indicating that the path passes directly through the obstacles. By employing the above technical solutions, the present invention has achieved the following technical advancements compared to the related art. 1. In the provided by the present invention, through autonomous navigation and obstacle avoidance functions, a follower vehicle can automatically and accurately track a leader vehicle or autonomously plan its path, accomplishing complex transportation tasks without human intervention. Simultaneously, the system can rapidly calculate an optimal path based on target and current locations, minimizing transportation time and enhancing operational efciency. Moreover, the system, which has high-precision positioning and attitude estimation capabilities, can obtain the position and attitude information of the follower vehicle in real time, ensuring stable driving in complex environments. 2. In the provided by the present invention, the follower vehicle can drive smoothly along the planned path through precise path planning and control instruction generation, thereby preventing hazardous situations caused by improper operations such as sharp turns or sudden acceleration. Additionally, through the realtime l ] monitoring and analysis of surrounding environmental information, potential hazardous obstacles can be detected and avoided in time, thereby preventing collision accidents. BRIEF DESCRIPTION OF THE DRAWINGS To more clearly illustrate the technical solution in the embodiment of the present application or in the related art, a brief description of the drawings required to be used in the embodiment is presented below. Obviously, the drawings described below are only some embodiments recorded in the present invention, and for those ordinary skilled in the art, other drawings may be obtained based on these drawings. FIG. 1 is a module composition diagram of the present invention; FIG. 2 is an implementation owchart of continuous deviation detection of the present invention; FIG. 3 is a owchart of deviation trend early warning of the present invention; and FIG. 4 is a owchart showing a process of guiding a follower vehicle forward using an autonomous navigation planning of the present invention. DETAILED DESCRIPTION In order to make the purposes, technical solutions and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention are irther described clearly and completely below in combination with the accompanying drawings. Obviously, the embodiments described are only some, rather than all embodiments of the present invention. Based on the embodiment of the present invention, all other embodiments obtained by those ordinary skilled in the art without creative efforts fall within the scope of protection of the present invention. Embodiment ]: referring to FIGS. 1-2, the present invention provides an , including a control center. The control center is communicatively connected to an environment perception module, a target recognition and positioning module, a relative heading deviation detection module, a deviation adjustment decision-making module, a power and steering execution module, an autonomous navigation planning module and a human-machine interaction module, in which each module is connected by electrical signals. The environment perception module is congured to collect environmental information around a follower vehicle for fruit and vegetable collection and transportation containing position, velocity, direction, surrounding obstacles and road boundary information of a leader vehicle, thereby providing real-time and accurate perception data, and laying the foundation for subsequent decisionmaking and control. Various types of sensors such as cameras and laser radars are deployed on a follower vehicle for fruit and vegetable collection and transportation to capture surrounding environmental information, such as relative position, velocity and direction information of a leader vehicle, distance, shape and size information of surrounding obstacles, and outline and position information of road boundaries. Images are collected and processed through the cameras, and lanes and obstacles ahead are identied. Light beams are sent through the laser radars and reected light is received synchronously. A distance from the obstacles is measured using a time difference, and a point cloud map is drawn to determine the shape, size and distance of the obstacles. The relevant data of the collected surrounding environment information are transmitted to an electronic control unit of the vehicle for preprocessing, and the preprocessing includes denoising, ltering, data format conversion to improve the accuracy and usability of data. The data collected by different sensors are calibrated and synchronized to keep the data consistent in time and space. The data from different sensors are fused to obtain comprehensive and accurate environmental information, and the position, velocity, acceleration of the obstacles and geometric features of the road boundaries are extracted from the ised data to obtain a feature dataset. The target recognition and positioning module is congured to identify and locate a coordinate position and attitude information of the leader vehicle in space based on the data obtained by the environment perception module, distinguish the leader vehicle from other interfering objects, and continuously lock the leader vehicle, thereby ensuring that the follower vehicle's target remains unambiguous, guaranteeing that the follower vehicle always tracks the correct target, and improving the accuracy of following. Each feature data in the feature dataset obtained by the environment perception module is iterated through, l3 and the feature data including the size, shape and color of the leader vehicle are extracted using image processing technology to facilitate the recognition and positioning. Through a machine learning algorithm, the extracted feature data of the leader vehicle are matched with a pretrained leader vehicle feature model, and candidate targets that match the leader vehicle features are screened out from matching results. The features of these candidate targets are compared with the obstacle features provided by the environment perception module, and interference objects that deviate from the leader vehicle features are eliminated, such as trees, buildings, and other vehicles, in which the pretraining of the leader vehicle feature model involves the following processes: collecting a large volume of image data of the leader vehicles from multiple angles, lighting conditions, and backgrounds, and ensuring the dataset encompasses diverse types of leader vehicles to enhance model generalization capability; annotating the collected image data with bounding boxes and class labels for the leader vehicles; conducting normalization, cropping, rotation, and other operations on the image data to improve the training efciency and performance of the model; feeding the annotated image data into a convolutional neural network (CNN) architecture for training, and implementing backpropagation algorithm to iteratively update model weights and biases during training; utilizing a validation set to evaluate model performance, and adjusting the architecture, parameters, or data preprocessing methods of the model based on validation feedback to optimize model capabilities; assessing the optimized model on a test set to verify its reliable recognition performance in real-world scenarios; and deploying the trained model into the automatic following system for fruit and vegetable collection and transportation, and integrating it into the environment perception module for real-time recognition and positioning of the leader vehicles. Based on the selected leader vehicle features, the coordinate position of the leader vehicle in space is calculated using the image processing technology of image registration, and the attitude information of an orientation and a tilt angle of the leader vehicle is estimated by combining the shape and size information of the leader vehicle. Moreover, the coordinate position, orientation and tilt angle information of the leader vehicle are associated with point cloud data obtained from the sensors to improve the accuracy of positioning, and the leader vehicle is continuously tracked and locked using a tracking algorithm (Kalman ltering) to ensure that the follower vehicle can always accurately follow the leader vehicle. According to the real-time data provided by the environment perception module, the position and attitude information of the leader vehicle is updated dynamically to adapt to environmental changes. In a irther description, a calculation formula of the coordinate position of the leader vehicle in space is denoted as: A 2 A =(, >=(J aw), ( w)), where is a coordinate position of the leader vehicle in space, and represent coordinate positions of the leader vehicle on a horizontal plane, is a horizontal distance between the leader vehicle and the follower vehicle, A is a vertical distance between the leader vehicle and the follower vehicle, is a tilt angle of the leader vehicle relative to the horizontal plane, and is a maximum value in the vertical direction for limiting a range of ; a calculation formula of the orientation of the leader vehicle is denoted as: = ze), where is an orientation of the leader vehicle, 2 is an arctangent inction considering quadrants for calculating an orientation angle of the leader vehicle, A is the vertical distance between the leader vehicle and the follower vehicle, and is the horizontal distance between the leader vehicle and the follower vehicle; and a calculation formula of the tilt angle of the leader vehicle is denoted as: = (Al) where is a tilt angle of the leader vehicle, Ah is a height change of the leader vehicle, is the horizontal distance between the leader vehicle and the follower vehicle, and is an arctangent function for calculating the tilt angle of the leader vehicle. The relative heading deviation detection module is congured to compare the position and attitude data of the follower vehicle and the leader vehicle, and analyze a heading deviation of the follower vehicle relative to the leader vehicle to achieve continuous deviation detection, thereby providing a precise basis for direction adjustment of the follower vehicle. The position and attitude data of the follower vehicle and the leader vehicle are obtained through the environment perception module. The obtained position and attitude information including denoising, ltering and data alignment are preprocessed to improve the accuracy and reliability of the data, and feature information is extracted for calculating the relative heading deviation from the preprocessed data, with the position data being GPS coordinates, and the attitude information including velocity vectors (velocity magnitude and direction) and an angle of orientation (vehicle orientation). The heading deviation of the follower vehicle (based on the angle of orientation) is compared with that of the leader vehicle to calculate a difference between the two, resulting in a relative heading deviation. A calculation formula of the relative heading deviation is denoted as: A = (I I, 3600 I I), where A is a relative heading deviation, is a heading angle of the leader vehicle, and is a heading angle of the follower vehicle. A timer is set to trigger a data update and calculation process every 500 milliseconds, the position and attitude data of the follower vehicle and the leader vehicle are updated regularly, and a heading angle comparison process is repeated for the continuous deviation detection. The deviation adjustment decision-making module is congured to determine whether the current following situation is normal or not based on analysis results of the relative heading deviation detection module, issue early warning to a detected deviation trend, and update a direction adjustment strategy for the follower vehicle. The power and steering execution module is congured to issue a control instruction based on the strategy of the deviation adjustment decisionmaking module, and control steering, acceleration and braking actions of the follower vehicle to realize the precise control of the follower vehicle, thereby improving the control precision and response speed of the system. The autonomous navigation planning module is congured to plan an optimal path from a current position to a target position by relying on autonomous navigation technology while considering comprehensive obstacle factors when either no leader vehicle is present or its signal is lost, and guide the follower vehicle forward accordingly, thereby causing the follower vehicle to reach a destination safely and efciently, and ensuring continuous operation even in complex environments. The humanmachine interaction module is congured to provide a visual operation interface that displays a running status of the system, including the positions of the follower vehicle and the leader vehicle, the relative heading deviation, and whether there is abnormal information, thereby allowing operators to perform necessary parameter settings and manual intervention control operations, and facilitating exible adjustments to the system under special circumstances. Embodiment 2: referring to FIGS. 34, the present invention provides a technical solution based on Embodiment l. Preferably, a process of deviation trend warning within the deviation adjustment decisionmaking module includes the steps that: A normal range and an early warning threshold of the relative heading deviation are set, with the normal range being i 5° and the early warning threshold being i 10°, and a real-time analysis result is received from the relative heading deviation detection module including a current value and historical data of the relative heading deviation. Based on the received relative heading deviation data, it is determined whether a following situation of the current follower vehicle relative to the leader vehicle is within the normal range. If it is within the normal range, the current following situation is normal and monitoring shall continue; and if it exceeds the normal range, a deviation early warning and processing ow shall be initiated. A deviation degree is determined based on an absolute value of the relative heading deviation, and a deviation early warning coefcient is calculated by combining the relative heading deviation data at multiple consecutive time points to analyze the deviation trend of the relative heading deviation. Analysis results of the deviation trend and the calculated deviation early warning coefcient are combined, and if the deviation early warning coefcient approaches or exceeds the early warning threshold and the deviation trend continues or deteriorates, an early warning mechanism is triggered to notify a driver through sound and light. Based on the deviation early warning coefcient and the analysis results of the deviation trend, the direction adjustment strategy for the l7 follower vehicle is updated to correct the deviation and keep the follower vehicle on a correct heading, and the direction (left, right or hold) to be adjusted by the follower vehicle is determined based on the relative heading deviation and the deviation trend. An angle or velocity variable quantity to be adjusted is calculated by combining the deviation early warning coefcient and the preset adjustment strategy, and an adjustment instruction is sent to the power and steering execution module of the follower vehicle to execute the direction adjustment operation. In a further description, a calculation formula of the deviation early warning coefcient is denoted as: = Ê >< (1, Æ) >< F, where is a deviation early warning coefcient, A is a current value of the relative heading deviation, 0 is a preset deviation early warning threshold, is a standard deviation for adjusting the sensitivity of the deviation early warning coefcient, A is a relative heading deviation data at an th consecutive time point, is a quantity of consecutive time points, A is a maximum relative heading deviation within the continuous time points, A is a minimum relative heading deviation within the continuous time points, and a value range of is limited between 0 and 1, with O indicating no deviation and 1 indicating a serious deviation. When the relative heading deviation A approaches or exceeds the early warning threshold 0, the value of tends to 1, indicating an immediate early warning is required. Conversely, when an absolute value of A remains small and exhibits minimal change across consecutive time points, the value of stays low, signifying a lower degree of deviation. An implementation process of controlling the follower vehicle within the power and steering execution module includes the steps that: The adjustment strategy updated by the deviation adjustment decisionmaking module is received through a communication interface within the power and steering execution module, and the control instruction in the adjustment strategy is extracted for instruction parse. Based on analysis results of the control instruction, specic control parameters of l8 steering angle, acceleration or braking force are extracted, and types of actions to be executed including steering, acceleration or braking are determined based on the analysis results. For steering action execution, a steering mechanism is controlled based on the steering angle parameters to perform the steering if the control instruction is a steering action. During the steering, the steering angle and steering velocity are monitored in real time through the sensors to ensure the accuracy and stability of the steering action. For acceleration action execution, an output power of engine is controlled based on the acceleration parameters to achieve the acceleration if the control instruction is an acceleration action. In the process of acceleration, the vehicle speed and acceleration are monitored in real time through the sensors to ensure the stability and safety of acceleration action. For braking action execution, a braking system is controlled to generate corresponding braking force based on the braking force parameters for deceleration or stopping if the control instruction is a braking action. During the braking, the vehicle speed and deceleration are monitored in real time through the sensors to ensure the timeliness and reliability of the braking action. The steering, acceleration and braking actions are executed either simultaneously or sequentially to realize precise control of the follower vehicle. During the execution of the steering, acceleration and braking actions, the actual state of the vehicles is monitored in real time through the sensors, the actual state is compared with an expected target to evaluate execution effects, and execution results (actual steering angle, vehicle speed, acceleration) are fed back to the deviation adjustment decisionmaking module through the communication interface. Based on feedback, the control strategy is further adjusted within the deviation adjustment decisionmaking module, thereby improving the control precision and response speed of the system. A process of guiding the follower vehicle forward using the autonomous navigation planning within the autonomous navigation planning module includes the steps that: The surrounding environment information collected by the follower vehicle is processed, and converted into a raster map for representation, and the obstacles and feasible areas in the environment are intuitively reected using the raster map. Based on the environmental information and the known map, the current position of the follower l9 vehicle is estimated using a visual SLAM algorithm, and the attitude information of the follower vehicle is obtained in real time using an IMU. Based on the sensor data and map information, the obstacles in the environment is detected in real time, and obstacle avoidance strategies such as bypassing, stopping and waiting are formulath based on positions, shapes and sizes of the obstacles. Moreover, an obstacle cost function is introduced into the path planning, and inuence degrees of obstacles in a path are evaluated to avoid the obstacles and optimize the path. Based on the target position and current position, a shortest path to avoid obstacles and follow trafc rules is searched in the map using a search algorithm (A-star algorithm). Based on a global path planning, dynamic adjustment and optimization is performed according to a realtime state of the follower vehicle and changes of the surrounding environment, thereby ensuring that the follower vehicle can operate smoothly. According to path tracking results and the realtime state of the follower vehicle, the control instruction is generated to control the steering, acceleration and braking actions of the follower vehicle, thereby achieving the precise control. In a further description, an expression of the obstacle cost function is denoted as: 2 H > = < >< (1, _), where is an obstacle cost function for evaluating the inuence degrees of obstacles in the path, is a distance from the follower vehicle to a th obstacle, is a standard deviation of obstacle distance for adjusting the sensitivity of the cost function to the distance, is a total number of obstacles in the path, is a constant for adjusting a curvature of an exponential function and affecting the inuence degree of the obstacle distance on the cost function, is a maximum angular deviation allowed in the path, is an angular deviation between the follower vehicle and the th obstacle, and a value range of is limited between 0 and 1, with 0 indicating that the path completely avoids the obstacles, and 1 indicating that the path passes directly through the obstacles. When the obstacle distance is small, an exponential term amplies, causing the value of to decrease, thereby indicating path adjustments to avoid these obstacles. Simultaneously, as the angular deviation approaches , a minimum function will reduce the value of , necessitating path adjustments to minimize angular deviation. By adjusting the values of and , the sensitivity of the cost mction to the obstacle distance and the angular deviation to adapt to different driving environments and requirements. The above mentioned are only the specic embodiments, but the scope of protection of the present application is not limited thereto, and within the scope of the technology disclosed in the present application, any person skilled in the art may readily think of changes or substitutions, and these changes or substitutions shall be covered by the scope or protection of the present application. Therefore, the scope of protection shall be determined by the scope of protection of the claims. CLAIMS 1. Automatic tracking and autonomous navigation system for collecting and transporting of fruits and vegetables, consisting of a control center, where the control center is communicatively connected to an environmental monitoring module, a module for target recognition and positioning, a module for detecting a relative course deviation, a decision-making module for adjusting the course deviation, a power and control module, autonomous module navigation planning and a human-machine interaction module, with each module connected is by electrical signals; the environmental sensing module, configured to collect environmental information around a follow-up vehicle for collecting and transporting fruit and vegetables; the target recognition and positioning module, configured to provide a coordinate position and to identify and locate position information of a leader vehicle in space, to distinguish the leader vehicle from other disturbing objects and the leader vehicle continuously lockable; the relative course deviation detection module, configured to detect the position and to compare the position data of the follow vehicle and the follow vehicle and a to analyze the course deviation of the follow vehicle relative to the follow vehicle in order to to achieve continuous deviation detection, involving a continuous implementation process deviation detection within the relative course deviation detection module includes the steps by: obtaining position and status data of the follow vehicle and the follow vehicle via the environmental perception module, preprocessing the obtained position and position information through denoising, filtering and data alignment, and the extracting feature information for calculating relative course deviation from the preprocessed data, where the position data are GPS coordinates (Global Positioning System) and the attitude information velocity vectors and an orientation angle includes; comparing the course deviation of the following vehicle with that of the leading one vehicle, calculating a difference between the two, resulting in a relative course deviation, and specifying a calculation formula for the relative course deviation as: A = (| |, 360° | |) , where A is a relative course deviation is, a course angle of the follow vehicle and a course angle of the follow-up vehicle; and setting a timer to update data every 500 milliseconds and to start the calculation process, which includes the position and attitude data of the following vehicle and the follow vehicle is updated regularly and a course angle comparison process is performed repeated for continuous deviation detection; the deviation adjustment decision module, configured to determine whether the current tracking situation is normal or not, to provide an early warning for a detected deviation trend and a price adjustment strategy for the to update follower vehicle, which includes a process of warning for deviation trend within The deviation adjustment decision-making module consists of the steps of: setting a normal range and a threshold for early warning of the relative course deviation, where the normal range is i 50 and the early warning threshold i 10°, and receiving a real-time analysis result of the relative course deviation detection module which is a current contains value and historical data of the relative price deviation; determine whether a following situation of the current following vehicle in relation to the leading vehicle is within normal range based on the data received from the relative course deviation, where the current tracking situation is normal and further monitored if it is within normal range, and an early warning and processing flow is entered if it exceeds the normal range; determining a degree of deviation based on an absolute value of the relative course deviation, calculating an early warning coefficient for deviation by combining the relative course deviation data on multiple consecutive points in time, and analyzing the deviation trend of the relative course deviation; combining the analysis results of the deviation trend and the calculated deviation warning coefficient, and activating a warning mechanism to warn a driver by means of sound and light if the deviation warning coefficient approaches or exceeds the warning threshold and the deviation trend continues or worsens; updating the strategy for adjusting the direction of the follow-up vehicle to to correct the deviation and keep the following vehicle on a correct course based on the early warning coefficient for the deviation and the analysis results of the deviation trend, determining the direction to be taken by the following vehicle adjusted based on the relative course deviation and deviation trend, calculating of an angle or speed variable that needs to be adjusted by the early warning coefficient for the deviation and the preset adjustment strategy to combine and send an adjustment instruction to the module for the perform power and control of the follower vehicle to perform the direction adjustment to feed, and a calculation formula of the early warning coefficient for deviation as indicate as follows: 1 _ A | A 1+ (10 ) >< 1 >< A _ <, 1' > F; where is a deviation warning coefficient, A is a current value of the relative course deviation, 0 a preset deviation warning threshold, a standard deviation for adjusting the sensitivity of the deviation warning coefficient, A a measure of the relative course deviation on a consecutive point in time, n is a number of consecutive time points, A is a maximum relative course deviation is within consecutive time points, A a minimum relative course deviation is within the consecutive time points, and a value range is between 0 and 1, where 0 indicates no deviation and 1 indicates a serious deviation indicates; the power and control output module, which is configured to perform a to provide control instruction based on the strategy of the decision-making module for adjusting the deviation, and the steering, acceleration and braking of the to arrange follow vehicle to realize the precise control of the follow vehicle; the autonomous navigation planning module, configured to plan an optimal path from a current position to a target position, taking into account extensive obstacle factors when there is no following vehicle present or its signal is lost gone, and to lead the following vehicle forward accordingly; and the human-machine interaction module, configured to provide a visual control interface to provide a running status of the system. 2. Automatic tracking and autonomous navigation system for collecting and transporting fruits and vegetables according to claim 1, in which a process of the collecting the environmental information within the environmental perception module the steps includes: the deployment of various types of sensors, including cameras and laser radars, on the follow-up vehicle for fruit and vegetable collection and transport to record environmental information to be laid, including relative position, speed and direction information of the following vehicle, distance, shape and size information of surrounding obstacles and perimeter and position information of road boundaries; collect and process images from cameras, lanes and obstacles in front of the road identify, send light beams through laser radars, receive synchronously reflected light, measure a distance to the obstacles using a time difference and a point cloud map draw to determine the shape, size and distance of the obstacles; sending the relevant data from the collected environmental information to a electronic control unit of the vehicle for pre-treatment, including denoising, filtering, data format conversion and calibration and synchronization of the data collected by various sensors; and merge data from different sensors to provide comprehensive environmental information to obtain, and the position, velocity, acceleration of the obstacles and geometric extract road boundary features from the merged data to create a to obtain data set with characteristics. 3. Automatic tracking and autonomous navigation system for fruit and vegetable collection and transportation according to claim 2, in which a process of distinguishing the lead vehicle from other interfering objects within the target recognition and positioning module includes the steps of: going through all the feature data in the feature dataset one by one obtained by the environmental perception module, and extracting the feature data which consist of the size, shape and color of the leading vehicle using image processing technology; matching the extracted feature data of the lead vehicle with a pre-trained feature model of the lead vehicle using an algorithm for machine learning, screening candidate targets that match the characteristics of the lead vehicle from the matching results, comparing the characteristics of these candidate targets with the obstacle characteristics specified by the environmental perception module are provided, and eliminating interference objects that deviate from the characteristics of the leading vehicle; calculating the coordinate position of the leader vehicle in space using the image processing technology of image recording based on the selected characteristics of the leader vehicle, and estimating the attitude information from an orientation and a slope angle of the leader vehicle by combining the shape and size information from the leader vehicle; and associating the coordinated position, orientation and tile angle information of the lead vehicle with point cloud data obtained from the sensors, continuous tracking and locking the lead vehicle using a tracking algorithm, and the dynamically updating the position and attitude information of the leading vehicle based on of the real-time data provided by the environmental perception module to adapt adapt to changes in the environment. 4. Automatic tracking and autonomous navigation system for fruit and vegetable collection and transport according to requirement 3, in which a calculation formula of the coordinate position of the lead vehicle in space is designated as: =(, )= < / 2+(%)2, ( , +%)), where is a coordinate position of the lead vehicle in space, and represent coordinate positions of the leading vehicle on a horizontal plane, a horizontal distance between the leading vehicle and the following vehicle is, A is a vertical distance between the leading vehicle and the following vehicle is, a slope angle of the leading vehicle relative to the horizontal plane, and a maximum value in the vertical direction is for limiting a range of ; a calculation formula of the orientation of the leading vehicle is denoted as: = 2 (L), where is an orientation of the leading vehicle, A is the vertical distance between the leading vehicle and the following vehicle, and the horizontal distance between the leading vehicle and the following vehicle, and a calculation formula of the lead vehicle's lean angle is indicated as: = (Al) in this is a tilt angle of the leading vehicle, Ah a change in height of the leading vehicle and the horizontal distance between the leading vehicle and the follow-up vehicle. 5. Automatic tracking and autonomous navigation system for fruit and vegetable collection and transportation according to claim 4, wherein an implementation process of controlling the follow-up vehicle within the propulsion and control implementation module includes the steps by: receiving the adjustment strategy as updated by the deviation adjustment decision making module via a communication interface within the drive and control execution module, and extracting the control instruction in the instruction parse adaptation strategy; extracting specific control parameters of steering angle, acceleration or braking force based on analysis results of the control instruction, and determining types to be perform actions, including steering, accelerating or braking, based on the analysis results; controlling a steering mechanism based on the steering angle parameters to steer if the control instruction is a control action during the execution of the control action; the control an engine power based on the acceleration parameters to accelerate if the control instruction is an acceleration action during the execution of the acceleration action; and controlling a braking system to apply a corresponding braking force to generate based on the braking force parameters to slow down or stop if the control instruction is a braking action during the execution of the braking action, and simultaneously or sequentially perform steering, acceleration and braking actions to achieve precise to realize control of the following vehicle; and monitoring the actual condition of the vehicles in real time via the sensors while performing steering, acceleration and braking actions, comparing the actual situation with an expected goal, evaluating the implementation effects, the feedback of the implementation results to the decision-making module for the adjustment of the deviation via the communication interface and further adjustment of the control strategy based on the feedback within the decision-making module for the adjustment of the deviation. 6. Automatic tracking and autonomous navigation system for collecting and transporting of fruits and vegetables according to claim 5, which involves a process of advancing the follow-up vehicle using the autonomous navigation planning within the autonomous navigation planning module includes the steps of: processing the environmental information collected by the following vehicle, converting in a raster map for display, and showing the obstacles and feasible areas in the area using the raster map; estimating the current position of the following vehicle using a visual Simultaneous Localization and Mapping (SLAM) algorithm based on the environmental information and the well-known map, and obtaining information about the position of the follower vehicle in real time using an inertial measurement unit (IMU); the real-time detection of obstacles in the environment based on the sensor data and the map information, formulating strategies to avoid obstacles, including avoid, stop and wait based on the position, shape and size of the obstacles, introducing a costing option for obstacles in route planning and evaluating the degree of influence of obstacles on the path to avoid the obstacles and follow the path optimize; find the shortest route to avoid obstacles and follow the traffic rules on the map tracking using a search algorithm based on target position and current position, and perform dynamic adjustments and optimizations based on the real-time status of the follow-up vehicle and changes in the environment based on a global route plan; and generating the control instruction based on the results of following the path and the real-time status of the follower vehicle, and control the steering, acceleration and and braking actions of the following vehicle to achieve precise steering. 7. Automatic tracking and autonomous navigation system for collecting and transporting fruits and vegetables according to claim 6, which includes an expression of the obstacle cost function is denoted as: 2 () > = < >< (1, _), where an obstacle cost function is for evaluating the degree of influence of obstacles on the path, a distance from the follow vehicle to a jth obstacle, a standard deviation of obstacle distance for adjusting the sensitivity of the cost action for the distance, m is a total number of obstacles on the path, a constant is for fitting a curvature of an exponential function, a maximum angle deviation is that which is allowed in the path, an angle deviation is between the follower vehicle and the jth obstacle, and a value range of is limited between 0 and 1, where O indicates that the path completely avoids the obstacles, and 1 indicates that the path runs straight through the obstacles. 1 / 2 FIG. 1 FIG. 2