Path tracking control method for hinged tractor in hilly and mountainous areas
By adopting deep learning and fuzzy control methods in hilly mountains, considering the projection polygon side length of the articulated tractor body on the target path segment, the navigation accuracy and real-time problems are solved, and the stable path of the articulated tractor in hilly mountains is realized. Tracking improves agricultural production efficiency and safety.
Patent Information
- Application Number
- CN202411286097.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-13
- Publication Date
- 2024-12-06
AI Technical Summary
In hilly and mountainous areas, the navigation accuracy of articulated tractors is low and the real-time performance is poor. Traditional vehicle navigation control algorithms fail to effectively consider the errors caused by tail flutter or vehicle deflection that articulated tractors may occur.
By taking the projected polygon side length of the articulated tractor body on the target path segment as the input for deep learning training, combined with the control instructions of artificial division, a fuzzy control method is adopted to consider the front and rear body deviations to achieve stable and reliable path tracking control. The specific steps include: placing the initial position, calibrating the sensor, expanding the local target path, decomposing the side length of the body projection polygon, deep learning training, storing neural network data, and completing the full path tracking through the sensor's real-time tracking information.
The autonomous navigation and precise operation of articulated tractors in hilly and mountains has been realized, which has improved agricultural production efficiency and safety, reduced resource waste and negative environmental impacts, and provided technical support for the sustainable development of modern agriculture.
Smart Images

Figure CN119088036A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of unmanned vehicle navigation path tracking, and relates to a path tracking control method for an articulated tractor in hilly mountainous areas. Background Art
[0002] Navigation path tracking control of unmanned tractors is of great significance in modern agriculture. Due to the complexity of hilly and mountainous terrain, traditional tractors have low operating efficiency and great safety risks in these areas. Articulated tractors are ideal for adapting to complex terrains due to their superior maneuverability and flexibility. However, in order to achieve efficient and safe autonomous driving of tractors, a suitable path tracking control algorithm has become one of the key research technologies.
[0003] In order to enable this articulated tractor to navigate and operate autonomously, an accurate path tracking control algorithm is essential. Conventional vehicle navigation control algorithms basically use four-wheel passenger cars as a control model reference, without considering the serious errors caused by the possible tail swing or vehicle deflection of articulated tractors. Therefore, this method takes the deviations of the front and rear parts of the vehicle into control considerations, and uses the projection of the vehicle body on the target path segment as the deviation measurement standard, avoiding the lack of control considerations when selecting a single point as the error reference point. The application of automated agricultural machinery not only improves operating efficiency, but also significantly reduces labor costs and labor intensity. Automatic navigation driving tractors can operate continuously for a long time, greatly improving agricultural production efficiency. In addition, high-precision path tracking control technology can indirectly reduce resource waste, such as precise fertilization and spraying, thereby reducing negative impacts on the environment. At the same time, automation technology can also improve the safety of agricultural production and reduce accidents that may occur during manual operation.
[0004] In summary, the path tracking control algorithm for articulated tractors in hilly and mountainous areas is one of the core technologies of agricultural automation. This method uses deep learning and the projected polygon edge length of the articulated tractor on the path segment, taking the deviation of the front and rear bodies into account in navigation control, and can achieve autonomous navigation and precise operation of tractors in hilly and mountainous areas. This not only improves agricultural production efficiency and safety, but also provides strong technical support for the sustainable development of modern agriculture. Summary of the invention
[0005] The purpose of the present invention is to solve the problems of low navigation accuracy and poor real-time performance of articulated tractors in hilly and mountainous areas and forests, and to incorporate the front and rear vehicle body deviations into deviation control considerations, and to provide a fuzzy control method that uses the side length of the polygon projected by the vehicle body on the target path segment as the deep learning training input and artificially divided control instructions as the training output, so as to achieve stable and reliable path tracking of articulated tractors in working conditions with slopes such as hilly and mountainous areas.
[0006] To achieve the above object, the present invention provides the following technical solutions: Place the articulated tractor in the initial position and confirm the basic status of the vehicle; confirm the status of the body sensor equipment, test and calibrate the positioning and navigation system; expand the local target tracking path through the loaded host computer to obtain the virtual target path segment microelement; decompose the typical position relationship between the articulated tractor and the target path segment to obtain the side length of the projection polygon of the articulated tractor body on the target path segment; obtain the side length data of the projection polygon of the articulated tractor body after the deep learning association control strategy is trained; store the trained neural network data and burn it into the host computer chip; the articulated tractor receives the tracking start command, enters the established target path and transmits real-time tracking information back through the sensor; combined with the current articulated tractor positioning information, the host computer determines the current control command to complete the entire path tracking. BRIEF DESCRIPTION OF THE DRAWINGS
[0007] Figure 1 It is a schematic diagram of an articulated tractor; Figure 2 It is the logical diagram of the vehicle path tracking control algorithm; Figure 3 It is the articulated tractor model used in deep neural network training and the schematic diagram of the widened target path segment; Figure 4 It is a schematic diagram of the projection polygon generated by different positional relationships between the vehicle body and the target path segment when there is no swing; Figure 5 It is a schematic diagram of the projection polygon generated by the different positional relationships between the front axle part of the vehicle and the target path segment when only the front axle part of the vehicle swings left; Figure 6 It is a schematic diagram of the projection polygon generated by the different positional relationships between the front axle part of the vehicle and the target path segment when only the front axle part of the vehicle swings right; Figure 7 It is a schematic diagram of the projection polygon generated by the different positional relationships between the rear axle of the vehicle and the target path segment when only the rear axle of the vehicle swings left; Figure 8 It is a schematic diagram of the projection polygon generated by the different positional relationships between the rear axle of the vehicle and the target path segment when only the rear axle of the vehicle swings right; Fig. 9 It is a schematic diagram of projection polygons generated by different positional relationships between the front axle part of the vehicle and the target path segment when the rear axle part swings right and left; Fig.10 It is a schematic diagram of the projection polygon generated by the different positional relationships between the right swing of the rear axle part, the left swing of the front axle part and the target path segment. DETAILED DESCRIPTION
[0008] In order to make the purpose, technical scheme and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. The various embodiments listed in the present invention are only used to illustrate the present invention and are not used to limit the scope of the present invention. Any obvious modifications or changes made to the present invention do not depart from the spirit and scope of the present invention.
[0009] A path tracking control method for an articulated tractor in hilly mountainous areas comprises the following steps: S10, placing the articulated tractor in an initial position and confirming the basic state of the vehicle; S20, confirm the status of the vehicle body sensor equipment, test and calibrate the positioning and navigation system; S30, expanding the local target tracking path through the loaded host computer to obtain a virtual target path segment microelement; S40, decomposing the typical positional relationship between the articulated tractor and the target path segment, and obtaining the side length of the projection polygon of the articulated tractor body on the target path segment; S50, side length data of the projection polygon of the articulated tractor body after the association control strategy is trained through deep learning; S60, store the trained neural network data and burn it into the host computer chip; S70, the articulated tractor receives the tracking start instruction, enters the predetermined target path and transmits real-time tracking information back through the sensor; S80. Combine the current articulated tractor positioning information, determine the current control instruction through the host computer, and complete the whole path tracking.
[0010] Preferably, the step S20 further includes: S21, check all sensors installed on the vehicle body, pass the power-on test and return the test value; S22. Determine whether the current working state of the sensor is normal by comparing the measured sensor data with the empirical measured value.
[0011] Preferably, the step S30 further includes: S31, dividing the target path into a number of micro-elements according to the length of the vehicle body; S32, widening each path segment microelement through software simulation, and the widening width is equal to the vehicle body width; S33. As the articulated vehicle continues to move, a path segment microelement closest to the vehicle body articulation point is selected as a current tracking deviation calculation standard.
[0012] Preferably, the step S40 further includes: S41. Analyze all possible positions of the articulated tractor during the tracking process and draw corresponding schematic diagrams; S42, placing each possible different tractor position at a different position of the virtual path segment, so as to simulate all possible tracking situations that may actually occur; S43, marking a shadow on the overlapping portion of the tractor and the virtual path segment, where the shadow is the projection of the vehicle body on the virtual path segment.
[0013] Preferably, the step S50 further includes: S51, classifying the side lengths of the projection polygons of the articulated tractors in all cases, and associating and assigning different navigation control strategies; S52, measuring the edges of the projected polygon, sorting them in a certain order, and finally importing them into a classification deep learning neural network for training; S53, expand the training data set as needed, and store the trained neural network data for future use.
[0014] Preferably, the step S70 further includes: S71, receiving a driving instruction and obtaining current location information; S72, comparing the target path coordinate information, obtaining the nearest virtual path segment, and calculating the length of each side of the current tractor projection area; S73, input the lengths of the sides of the tractor projection area, and obtain the corresponding navigation control strategy through deep neural network calculation; S74, looping and calculating the navigation control strategy until the tractor approaches the end point.
Claims
1. A path tracking control method for an articulated tractor in hilly and mountainous areas, comprising the following steps: S10, placing the articulated tractor in an initial position and confirming the basic state of the vehicle; S20, confirm the status of the vehicle body sensor equipment, test and calibrate the positioning and navigation system; S30, expanding the local target tracking path through the loaded host computer to obtain a virtual target path segment microelement; S40, decomposing the typical positional relationship between the articulated tractor and the target path segment, and obtaining the side length of the projection polygon of the articulated tractor body on the target path segment; S50, side length data of the projection polygon of the articulated tractor body after the association control strategy is trained through deep learning; S60, store the trained neural network data and burn it into the host computer chip; S70, the articulated tractor receives the tracking start instruction, enters the predetermined target path and transmits real-time tracking information back through the sensor; S80. Combine the current articulated tractor positioning information, determine the current control instruction through the host computer, and complete the whole path tracking.
2. The path tracking control method for an articulated tractor in hilly and mountainous areas according to claim 1, characterized in that: The step S20 also includes: S21, check all sensors installed on the vehicle body, pass the power-on test and return the test value; S22. Determine whether the current working state of the sensor is normal by comparing the measured sensor data with the empirical measured value.
3. The path tracking control method for an articulated tractor in hilly and mountainous areas according to claim 2, characterized in that: The step S30 also includes: S31, dividing the target path into a number of micro-elements according to the length of the vehicle body; S32, widening each path segment microelement through software simulation, and the widening width is equal to the vehicle body width; S33. As the articulated vehicle continues to move, a path segment microelement closest to the vehicle body articulation point is selected as a current tracking deviation calculation standard.
4. The path tracking control method for an articulated tractor in hilly and mountainous areas according to claim 3, characterized in that: The step S40 also includes: S41. Analyze all possible positions of the articulated tractor during the tracking process and draw corresponding schematic diagrams; S42, placing each possible different tractor position at a different position of the virtual path segment, so as to simulate all possible tracking situations that may actually occur; S43, marking a shadow on the overlapping portion of the tractor and the virtual path segment, where the shadow is the projection of the vehicle body on the virtual path segment.
5. The path tracking control method for an articulated tractor in hilly and mountainous areas according to claim 4, characterized in that: The step S50 also includes: S51, classifying the side lengths of the projection polygons of the articulated tractors in all cases, and associating and assigning different navigation control strategies; S52, number the edges at different positions, sort them in a certain order, and finally import them into the classification deep learning neural network for training; S53, expand the training data set as needed, and store the trained neural network data for future use.
6. The path tracking control method for an articulated tractor in hilly and mountainous areas according to claim 5, characterized in that: The step S70 also includes: S71, receiving a driving instruction and obtaining current location information; S72, comparing the target path coordinate information, obtaining the nearest virtual path segment, and calculating the length of each side of the current tractor projection area; S73, input the lengths of the sides of the tractor projection area, and obtain the corresponding navigation control strategy through deep neural network calculation; S74, looping and calculating the navigation control strategy until the tractor approaches the end point.
Citation Information
Cited By
Remote control type mountain pepper harvester and method
CN120029259A
Remote control mountain pepper harvester and method
CN120029259B
