Autonomous tracking system of electric hinged four-crawler tractor

Abstract

The invention discloses an autonomous tracking system of an electric hinged four-crawler tractor. The autonomous tracking system comprises a sensing module, a path planning module and a self-adaptive walking module, according to the system, through multi-source perception fusion and dynamic environment modeling, the environment cognition precision and robustness in a complex farmland environment are remarkably improved; a passable path conforming to the dynamic characteristics of the articulated tracked vehicle is further generated through global planning in combination with dynamic window local optimization, slippage area filtering, landform traffic weight evaluation and an energy consumption optimization model, and the adaptive defect of a traditional wheel type path planning method in an unstructured terrain is effectively overcome; and meanwhile, path deviation caused by the difference between the hinging steering characteristic and the crawler speed is remarkably inhibited through a dual control mechanism of solving an optimal control sequence and dynamically adjusting PID parameters.

Description

Technical Field

[0001] This invention relates to the field of digital agricultural machinery technology, and more specifically to an autonomous tracking system for an electric articulated four-track tractor. Background Technology

[0002] With the continuous improvement of modern agricultural mechanization, farmland operations have placed higher demands on the intelligence and autonomy of agricultural machinery. Especially in complex terrain and variable working conditions, traditional wheeled or tracked tractors have many limitations in terms of working path accuracy, energy consumption control, and steering stability. Articulated tracked tractors, due to their strong obstacle-crossing ability, low ground pressure characteristics, and good mobility, have gradually become important equipment for adapting to complex farmland operations.

[0003] However, most existing articulated tracked tractors still rely on manual operation or semi-automatic control methods, lacking an efficient and reliable autonomous tracking control system, which makes it difficult to meet the needs of precision agriculture for automation, high-precision path tracking and energy efficiency optimization.

[0004] Furthermore, existing path planning and control methods are mostly applied to wheeled vehicles, lacking systematic consideration for tracked vehicles, especially articulated tracked vehicles, in terms of terrain adaptability and steering coordination. Since the steering characteristics of the articulation points and the speed difference between the inner and outer tracks significantly affect the overall vehicle path tracking performance, traditional control methods struggle to achieve accurate path tracking, especially on complex paths or sections with significant curvature changes, resulting in significant path deviations, poor steering stability, and even track slippage or increased energy consumption.

[0005] Therefore, there is an urgent need for a method that can accurately track and control the path of an electric articulated four-track tractor to improve work efficiency and quality, reduce energy consumption, and further promote the autonomy and intelligence of agricultural machinery. Summary of the Invention

[0006] In view of this, the present invention provides an autonomous tracking system for an electric articulated four-track tractor. To at least partially solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0007] An autonomous tracking system for an electric articulated four-track tractor includes,

[0008] The sensing module is used to acquire environmental information and vehicle status in real time;

[0009] The path planning module is used to construct a local environment map based on environmental information and vehicle status and generate a dynamic, traceable path.

[0010] The adaptive walking module coordinates the control of the track drive system and the articulated steering system based on the output instructions of the path planning module.

[0011] In one optional embodiment, the sensing module includes a binocular camera, a lidar, an RTK-GNSS positioning device, an inertial measurement unit, and a track rotation speed sensor; and preprocesses the acquired data through the following steps:

[0012] Distortion correction and depth reconstruction are performed on the binocular images, and the laser point cloud is filtered and denoised.

[0013] RTK position data and IMU attitude data are fused using an extended Kalman filter;

[0014] Differential analysis is performed on real-time rotational speed data to identify nonlinear motion states, such as slippage and lateral displacement. In one optional embodiment, data fusion is performed through the following steps:

[0015] Based on timestamps and sensor calibration results, multi-source data are uniformly converted to the vehicle coordinate system;

[0016] Spatial matching is performed between the preprocessed binocular images and laser point clouds to output an obstacle grid map and the boundary of the passable area.

[0017] Dynamic target trajectories are generated based on continuous frame point cloud and image difference.

[0018] In one optional embodiment, the sensing module performs the following anomaly detection and fault tolerance mechanism:

[0019] Periodically monitor sensor status indicators, and when the lidar malfunctions, activate VSLAM with vision-IMU fusion as a redundancy scheme.

[0020] Dynamically weighted trust levels for each sensor;

[0021] The watchdog mechanism of the edge computing platform enables rapid recovery after the perception module crashes.

[0022] In one optional embodiment, the execution steps of the path planning module include:

[0023] (1) Based on the obstacle grid map, passable area boundary and dynamic target trajectory output by the perception module, a local cost map is constructed; high slip areas are filtered according to the nonlinear motion state of the tracks;

[0024] (2) Generate a global initial path using the A* algorithm based on the current task target point and real-time vehicle status;

[0025] (3) Adjust the local paths in the global initial path in real time using the dynamic window method, and update the safety buffer distance and avoidance method based on the adjusted local paths;

[0026] (4) Optimize the path based on terrain access weight, slope limitation and energy consumption;

[0027] (5) Optimize the path by taking into account terrain changes, obstacle locations and the characteristics of the tractor's articulated structure;

[0028] (6) Perform curvature constraint and smoothing on the optimized path;

[0029] (7) Output the path point sequence of the optimized global path.

[0030] In one alternative embodiment, each waypoint includes coordinates, target speed, and desired heading angle; wherein the target speed is dynamically generated by using dynamic constraints to perform speed profile planning, integrating task requirements and environmental factors.

[0031] In one optional embodiment, the adaptive walking module includes a path tracking control unit, which performs the following steps:

[0032] 1) Synchronize the path point sequence with the real-time vehicle status using Kalman filtering;

[0033] 2) Select a continuous path segment as the reference trajectory and solve for the optimal control sequence through quadratic programming;

[0034] 3) Convert the control inputs into track speed commands and articulation angle commands.

[0035] In one optional embodiment, the adaptive walking module further includes a parameter coordination unit, which dynamically adjusts the PID parameters (KP / KI / KD) with the path deviation as input, and coordinates the differential speed of the inner and outer tracks and the deflection angle of the articulation point through fuzzy rules.

[0036] In one optional embodiment, the adaptive walking module further includes a data receiving unit and an execution unit;

[0037] The data receiving unit is used to receive the path point sequence from the path planning module and the real-time vehicle status information obtained by the perception module.

[0038] The execution unit is used to send track speed commands and articulation angle commands to the track drive motor controller and the articulation motor controller, respectively.

[0039] As can be seen from the above technical solution, this invention discloses an autonomous tracking system for an electric articulated four-track tractor. This system significantly improves the accuracy and robustness of environmental perception in complex farmland environments through multi-source perception fusion and dynamic environment modeling. High-precision obstacle maps are constructed using spatial matching of binocular vision and lidar. Combined with RTK-IMU tightly coupled positioning and track speed differential analysis, it identifies terrain undulations, dynamic obstacles, and nonlinear states such as track slippage / lateral displacement in real time, providing accurate vehicle-environment coupling state data for path planning.

[0040] At the path generation level, the system innovatively integrates the motion constraints of tracked vehicles with the terrain adaptation characteristics: A* global planning combined with dynamic window local optimization, through slip region filtering, terrain access weight evaluation and energy consumption optimization model, generates a passable path that conforms to the dynamic characteristics of articulated tracked vehicles, and dynamically adjusts the speed profile and safety buffer distance, effectively solving the adaptability defects of traditional wheeled path planning methods in unstructured terrain.

[0041] For the cooperative control mechanism, the adaptive walking module transforms the path tracking problem into a cooperative optimization of track differential speed and articulated turning angle: based on quadratic programming to solve for the optimal control sequence, PID parameters are dynamically adjusted through fuzzy rules to coordinate the speed difference between the inner and outer tracks and the deflection angle at the articulation point in real time. This dual control mechanism significantly suppresses path deviation caused by the difference in articulated steering characteristics and track speed, especially in sections with sudden curvature changes, by predictively adjusting attitude to greatly reduce track slip rate and energy loss during turning. At the same time, the system's built-in multi-level fault-tolerant design (such as VSLAM redundancy in case of sensor failure, dynamic trust weighting, and rapid recovery through edge computing) ensures continuous and reliable operation under complex farmland conditions, ultimately achieving high-precision, low-energy consumption, and stable steering fully autonomous operation capabilities, meeting the core requirements of precision agriculture for automated equipment.

[0042] Compared with existing technologies, the present invention can effectively reduce the deviation between the actual travel path of an articulated tracked tractor and the preset path, and achieve path tracking control with low energy consumption and improved steering stability. Attached Figure Description

[0043] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0044] Figure 1 This is a schematic diagram of the autonomous tracking system structure of the articulated tracked tractor provided in an embodiment of the present invention;

[0045] Figure 2 This is a schematic diagram of the adaptive walking module structure provided in an embodiment of the present invention;

[0046] Figure 3 This is a schematic diagram of path tracking deviation for an articulated tracked tractor. Detailed Implementation

[0047] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0048] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0049] This invention discloses an autonomous tracking system for an electric articulated four-track tractor, comprising three parts: a sensing module, a path planning module, and an adaptive walking control module; wherein,

[0050] The sensing module is used to acquire environmental information and vehicle status in real time;

[0051] The path planning module is used to construct a local environment map and generate a dynamic, trackable path based on environmental information and vehicle status, combined with the current posture of the articulated tracked tractor and the work target; at the same time, the path planning information is transmitted to the adaptive travel control module.

[0052] The adaptive walking module coordinates the control of the track drive system and the articulated steering system based on the output instructions of the path planning module and the current state of the vehicle to reduce the deviation from the preset path and achieve tractor path tracking.

[0053] Figure 1 This is a schematic diagram of the autonomous tracking system structure of an articulated tracked tractor provided in an embodiment of the present invention. The following is in conjunction with... Figure 1 The following is an explanation through specific embodiments.

[0054] Example 1:

[0055] In this embodiment, the perception module is a multi-sensor fusion perception system, including a binocular camera, lidar, RTK-GNSS positioning device, inertial measurement unit and track speed sensor, to perceive the surrounding ground environment and the walking status of the articulated tracked tractor, and provide multi-modal fusion information for the path planning module.

[0056] The binocular camera is used to acquire image information of the area in front of and around the tractor;

[0057] LiDAR is used to acquire point cloud data of ground obstacles and environmental structures;

[0058] The RTK positioning device, combined with the inertial measurement unit, provides high-precision positioning and attitude information. Specifically, the RTK positioning device provides positioning information to ensure that the walking device can acquire positioning data in real time in complex environments; the inertial measurement unit provides ground information and vehicle vibration conditions to ensure accurate perception of dynamic states in path planning.

[0059] The speed sensor is used to obtain the current running speed of the track.

[0060] In one embodiment, the setup of the above-mentioned multi-sensor system includes:

[0061] (1) Binocular camera system: installed at the front or top center of the tractor, used to acquire stereoscopic image information of the road scene ahead, and reconstruct the depth information of the terrain ahead through stereoscopic vision algorithm, used to detect ridges, potholes, obstacles and passage boundaries.

[0062] (2) LiDAR system: It adopts solid-state LiDAR, which is installed on the top of the vehicle body and has a 360° or 270° horizontal field of view. It generates point cloud data of the surrounding environment in real time and is used to identify static obstacles (such as tree stumps and rocks) and dynamic targets (such as people or animals).

[0063] (3) High-precision RTK-GNSS positioning device: The vehicle-mounted GNSS receiver performs real-time differential positioning with the external base station to achieve centimeter-level position acquisition and provide the tractor's heading angle information;

[0064] (4) Inertial Measurement Unit (IMU): Used to sense the vehicle’s attitude changes (including roll, pitch, and yaw) and acceleration and angular velocity signals, and to work with RTK to perform attitude fusion and short-term inertial navigation compensation.

[0065] (5) Track speed sensor: installed on the drive unit of each of the four tracks to detect the real-time speed of each track, determine the slippage situation, and calculate the vehicle linear speed and steering trend.

[0066] The above sensors are connected to the edge computing platform via CAN bus or Ethernet, and have a time synchronization mechanism to ensure consistent and timely data acquisition.

[0067] In one embodiment, after data acquisition, preprocessing is performed as follows:

[0068] 1. Image data preprocessing: The left and right images acquired by the binocular camera are distorted and calibrated, and a dense disparity map is generated using the SGBM algorithm to obtain a depth image;

[0069] 2. Point cloud data filtering: The raw point cloud acquired by the lidar is downsampled using a Voxel filter to remove ground points and outliers, and to extract obstacle point clouds;

[0070] 3. RTK and IMU data fusion: The extended Kalman filter (EKF) is used to fuse and compensate the position data of RTK and the attitude data of IMU to improve the continuous stability of the system in occlusion environment.

[0071] 4. Speed ​​data processing: Perform differential analysis on the real-time speed data of the four tracks to determine whether there are nonlinear motion states such as slippage or lateral displacement, and provide a basis for correction of the control algorithm.

[0072] The preprocessed data is uniformly converted into spatial information in the ontological coordinate system for subsequent mapping and navigation.

[0073] Furthermore, the information from the aforementioned multi-source sensors is fused to utilize the complementary characteristics of different sensors, thereby enabling real-time monitoring of ground environment information and vehicle driving status. After fusion, a unified ground environment model and vehicle status information are generated for use by the path planning module.

[0074] In this embodiment, the fusion process includes:

[0075] Spatiotemporal alignment and coordinate transformation: Using timestamp synchronization and sensor calibration results, all sensor data are uniformly transformed to the vehicle coordinate system or the world coordinate system;

[0076] Image and point cloud fusion: Spatial matching of depth map obtained by binocular camera with LiDAR point cloud improves obstacle boundary recognition accuracy;

[0077] Real-time raster map construction: Based on the fused obstacle point cloud, a local cost raster map is constructed, where each raster represents an environment occupancy probability;

[0078] Accessible area extraction: Image edge detection and point cloud clustering algorithms are used to identify the boundaries of accessible areas and filter out unwalkable areas in the terrain;

[0079] Dynamic target tracking: By combining continuous frame point clouds with image difference, dynamic target (such as pedestrians and livestock) trajectories can be detected and predicted, thereby improving the obstacle avoidance capability of path planning.

[0080] The resulting map structure can be used as the decision input for the path planning module and updated at different time periods to maintain its adaptability to complex farmland environments.

[0081] To further optimize the above technical solution, this application enables the perception module to perform the following anomaly detection and fault tolerance mechanism:

[0082] (1) Sensor status monitoring: The system periodically checks whether the sensors are online, including indicators such as voltage, signal strength, and data frame rate. If any abnormality is found, an alarm will be triggered immediately.

[0083] (2) Redundancy switching of perception data: When the data of the main sensor (such as LiDAR) is abnormal, the backup perception scheme (such as VSLAM built by vision + IMU) can be activated;

[0084] (3) Dynamic weighting of multi-sensor trust: The weight of each sensor in data fusion is dynamically adjusted according to the current environmental conditions (e.g., the image trust is reduced under low light conditions);

[0085] (4) Edge computing platform fault tolerance: The edge computing platform has a watchdog mechanism and a restart mechanism to ensure rapid recovery when the sensing module partially crashes.

[0086] Finally, the perception module of this application outputs the following structured output information:

[0087] The tractor's current precise pose (X,Y,θ);

[0088] Spatial distribution of surrounding obstacles (point cloud or grid form);

[0089] Real-time accessible area boundaries;

[0090] Trajectory prediction for dynamic targets;

[0091] Track slippage status and speed feedback.

[0092] All outputs are sent as structured data to the path planning module via a communication interface.

[0093] Example 2:

[0094] In this embodiment, the path planning module constructs a local working environment map based on the environmental information provided by the perception module and the current attitude state of the articulated tracked tractor. The map uses the articulated tracked tractor as the reference coordinate system and is dynamically updated according to the environmental data. Based on the local map, and according to the work requirements (such as the work destination, direction of travel, boundary line constraints, etc.) and the current motion state of the tractor, a path planning algorithm is used to generate a feasible path from the current position to the target position.

[0095] The path planning process in this application takes into account terrain changes, obstacle locations, and the characteristics of the tractor's articulated structure to ensure the feasibility and steering smoothness of the planned path.

[0096] Further considering the tractor's turning radius, articulation mechanism limitations, and the non-holonomic constraints of the tracked chassis, the path is subjected to curvature constraints and smoothing to ensure good trackability and stability, and to avoid sharp turns or sudden increases in energy consumption.

[0097] Finally, the planned path is output as a discrete path point sequence, which includes control parameters such as coordinates, target speed, and desired heading angle, and then sent to the adaptive travel control module.

[0098] The path planning module in this application possesses dual capabilities of global path planning and dynamic local path adjustment. It adopts a modular design, exhibiting strong robustness and real-time responsiveness. Its specific implementation is as follows:

[0099] (1) Based on the obstacle grid map, passable area boundary and dynamic target trajectory output by the perception module, a local cost map is constructed; high slip areas are filtered according to the nonlinear motion state of the tracks;

[0100] In this embodiment, the path planning module first receives structured data from the perception module, including environmental obstacle information, passable area, current pose and speed information, and performs map construction and state initialization based on this data. Input data types include the tractor's current precise pose (including X, Y coordinates and heading angle θ); point cloud or grid map of surrounding obstacles (such as a cost map); a set of boundary points or contour map of the passable area; dynamic obstacle prediction information; and track slippage status.

[0101] Then, a local map is constructed based on local perception information and the current task target point (such as the end point of the planned path or intermediate operation point). The map structure can be in the form of a two-dimensional grid map or a three-dimensional cost map.

[0102] All environmental and target point information is uniformly converted to the tractor body coordinate system or world coordinate system to maintain consistency; obstacle distance cost function and driving cost function are established to provide constraint boundaries for path optimization; based on the track slippage state, theoretically feasible but practically unexecutable paths are planned in real time, automatically avoiding high slippage rate areas.

[0103] (2) Generate a global initial path using the A* algorithm based on the current task target point and real-time vehicle status;

[0104] The path planning module of this application, based on the starting point and ending point, first performs global path planning and outputs an initial driving trajectory as a benchmark for local path adjustment. Then, it uses the A* path search algorithm to construct a heuristic search graph from the starting point to the ending point; avoids high-cost (obstacle) areas, outputs the shortest cost path, and generates a sequence of path points (including coordinates, direction, and target speed). 。

[0105] (3) Adjust the local paths in the global initial path in real time using the dynamic window method, and update the safety buffer distance and avoidance method based on the adjusted local paths;

[0106] The farmland working environment is complex and frequently changes dynamically (e.g., livestock crossing, farm implement interference), therefore the path planning module must have real-time local path adjustment capabilities. It needs to dynamically detect and update obstacles, and listen to obstacle update information sent by the perception module in real time. The local map refresh time is set to 0.5s–1s, dynamically updating the value of local obstacle areas. DWA (Dynamic Window Method) is used for short-term path planning, constructing a drivable space search area around the preset path and replanning feasible local paths. To avoid large path jumps and enhance path continuity and vehicle control stability, dynamic target prediction and avoidance are implemented. The future path of the target is predicted based on parameters such as target speed and direction, a safe distance buffer is set, and the path avoidance method is dynamically adjusted (deceleration, detour, etc.). A path tracking point update mechanism is adopted, with tracking points set as look-ahead points on the global or local path, and the look-ahead distance is dynamically adjusted according to speed (e.g., 0.5–2m).

[0107] (4) Optimize the path based on terrain access weight, slope limitation and energy consumption;

[0108] For special areas in farmland environments (such as wetlands, slopes, and gullies), this embodiment introduces terrain recognition and path adaptation mechanisms into the path planning module to improve off-road capability. This includes...

[0109] Terrain classification and identification: The perception module assigns different passage weights to different types of terrain (such as soft mud, hard road surface, gravel); in route planning, high-risk areas are set as high-cost or impassable areas;

[0110] Slope restrictions and terrain matching: Slope information is extracted based on point cloud maps; the maximum longitudinal and transverse slope values ​​of the path are limited (e.g., not greater than 15°) to prevent tractors from tipping over;

[0111] Energy consumption optimization path selection: Introduce a total energy consumption estimation model for the path, and prioritize low energy consumption paths (such as paths with flat terrain, few turns, and few braking); optimize the balance between path length and energy consumption while meeting the operational objectives.

[0112] (5) Optimize the path by taking into account terrain changes, obstacle locations and the characteristics of the tractor's articulated structure;

[0113] In the path planning process, in order to take into account the dynamic impact of the articulated structure on the overall vehicle travel path, a dynamic model of the tractor's articulated structure is established in the initial stage of path planning to consider the turning angle and angular velocity limits between the front and rear vehicle bodies, as well as the mechanical constraints of the structural connection points.

[0114] This invention treats the electric articulated four-track tractor as a two-body system consisting of a front body and a rear body, using the articulation point as the connection between the two rigid bodies, and describes the motion state of the entire vehicle by establishing the following simplified dynamic model:

[0115]

[0116] Where x and y represent the coordinates of the vehicle's center of gravity, and θ represents the vehicle's heading angle. Represents the hinge angle, v R With v L The speed of the tracks on the left and right sides of the vehicle body is represented by d, and the track spacing is w. j The rotational speed is the hinge angle.

[0117] This model can dynamically predict the response characteristics and ultimate steering capability of the articulation point when the tractor executes a path with a certain curvature. During the local path generation and adjustment stage, the maximum deflection angle, steering rate and response delay of the articulation mechanism are modeled to dynamically set the steerable area of ​​the path, preventing the path planning from exceeding the working limits of the entire vehicle's articulation mechanism.

[0118] During path optimization, the constraints of the articulation angle and the speed difference between the inner and outer tracks are combined to consider the track differential speed and articulation control in a coordinated manner. This ensures that the speed and heading changes in the output path point sequence are consistent with the vehicle's steering mechanism, and that the articulation structure can transition naturally in the transition section of the path, avoiding attitude oscillation and track slippage, and improving tracking stability.

[0119] Specifically, in the path optimization process, a cooperative constraint relationship between the speed difference between the inner and outer tracks and the articulation angle is introduced:

[0120] |v R -v L |≤Δv max

[0121]

[0122] The following cost function is designed to jointly penalize the deviation between track differential speed and articulation angle:

[0123]

[0124] in Let Δv be the target hinge angle obtained from the inverse curvature solution. ref The target track differential speed is defined as w1 and w2 as cost weights. The optimizer aims to minimize this cost function and, in conjunction with constraints, generates controllable, stable, and low-energy path point speed and steering control quantities in real time.

[0125] (6) Curvature constraints and smoothing of the path: Based on the articulated structure model of the vehicle and the nonholonomic constraint characteristics of the tracked chassis, the maximum allowable yaw angle Δθ between every two consecutive path points is set, and the maximum allowable path curvature k is calculated according to the vehicle structural parameters. max Its constraint form is:

[0126]

[0127] Where Δs is the distance between path points, and k max Based on the vehicle's minimum turning radius R min The calculation yielded:

[0128]

[0129] After the path points are initially generated, the spline interpolation algorithm is used to optimize the continuity of the path, ensuring the continuity of the path in the first derivative (direction) and the second derivative (curvature), avoiding sharp turns or curvature abrupt changes, and ensuring that the path planning process automatically tends to a smooth path with small curvature and low energy consumption.

[0130] (7) Output the path point sequence of the optimized global path.

[0131] In this embodiment, the path planning module dynamically assigns a target speed v to each path point when generating the path point sequence. t This process comprehensively considers the radius of curvature, obstacle distance, terrain type, and mission requirements. The specific steps are as follows:

[0132] 1) Set the maximum permissible speed v of the vehicle. max and minimum permissible speed v min Set the initial task speed v according to the task type (such as sowing, transportation). task .

[0133] 2) Calculate the local curvature k based on the current path point:

[0134]

[0135] Based on the vehicle dynamics model, a curvature speed limit constraint is introduced. This constraint, derived from the vehicle dynamics model, aims to ensure that the vehicle does not slip, become unstable, or fail to steer in time due to excessive speed when executing a certain curvature path.

[0136]

[0137] Among them, a lat_max To maximize permissible lateral acceleration and ensure that skidding does not occur during cornering.

[0138] 3) If there are obstacles around the waypoint, calculate the minimum distance d between the obstacles using the local grid map output by the perception module. obs Define a safe speed limit:

[0139]

[0140] Where D safe The threshold value set for the impact of obstacles.

[0141] 4) If the area along the path has a slope α (generated from lidar point clouds or maps), limit the driving speed on the slope:

[0142] v slope =v max ·cos(α)

[0143] 5) The final target speed is the minimum value constrained by the above multiple factors:

[0144] v t =min(v task ,v curve ,v obs ,v slope )

[0145] The path planning module will v t The output is appended to the path point and used for speed execution and feedforward adjustment in the adaptive walking control module.

[0146] Example 3:

[0147] The adaptive travel control module is used to coordinate the track drive system and articulated steering system of the electric articulated four-track tractor based on the path points, vehicle status information and target tracking instructions output by the path planning module, so as to achieve high-precision path tracking control and optimize energy consumption and driving stability.

[0148] In one exemplary embodiment, reference is made to Figure 2 This module includes a data receiving unit, a path tracking control unit, a parameter coordination unit, and an execution unit; the functional modules communicate with each other via a control bus to collaboratively complete the path tracking task.

[0149] The data receiving unit is used to receive the path point sequence (including coordinates, speed, heading angle and other information) from the path planning module, as well as the real-time vehicle status information provided by the perception module, including the current pose, track speed, articulation angle and other information.

[0150] The path tracking control unit is used to calculate the target rotational speed of the inner and outer tracks and the steering angle of the articulation point based on the path points provided by the path planning module and the current speed and posture of the vehicle, using a model predictive control path tracking algorithm; the specific steps include:

[0151] 1) Synchronize the path point sequence with the real-time vehicle status using Kalman filtering to ensure that all information used by the control module is on the synchronized timeline.

[0152] 2) Select a continuous path segment as the reference trajectory, and solve for the optimal control sequence through quadratic programming; the control unit matches the optimal path reference point P from the path sequence based on the vehicle's current position. t Using a Model Predictive Controller (MPC), the trajectory segment is selected from P based on multiple consecutive path points as reference points. t A continuous path {P} at the beginning t ,P t+1 ,...,P t+N The optimization objective is to minimize the future prediction error to calculate the optimal control input at the current time. The optimization problem is transformed into a constrained quadratic programming problem by a real-time optimizer, which solves online to obtain the optimal control sequence {u0,...,u}. Np-1 The control sequence {u0,...,u} Np-1 Each control vector u in} k The target speeds of the left and right tracks and the articulated steering angle are obtained by a real-time optimizer and satisfy the vehicle dynamics model and input constraints.

[0153] The constraint design of the MPC controller includes constraints on tracked drive characteristics, control capability constraints on articulated structures, and synchronization constraints on the four-track system. Specifically, the constraints include tracked drive differential constraints, which limit the maximum speed difference between the left and right tracks based on the tendency of tracked vehicles to slip during turning; articulated mechanism limit constraints, which set upper and lower limits on the articulation angle and its rate of change based on the maximum allowable deflection angle and response rate of the structure; and constraints on the speed difference between the front, rear, left, and right tracks of the four-track system to ensure the stability of the vehicle's attitude.

[0154] 3) Convert the control input into track speed and articulation angle commands. Specifically, the target command is sent to the left and right track speed controllers and the articulation mechanism controller using CAN. This process runs periodically at a fixed frequency. First, the state estimate is updated, path points are matched, the reference trajectory is updated, the optimal control is solved, the execution command is converted and sent, feedback is awaited, and the next cycle begins.

[0155] The parameter coordination module is used to dynamically adjust the output of the track drive system and the articulated system based on the controller's calculation results and the path deviation as input. It uses fuzzy rules to coordinate the differential speed of the inner and outer tracks and the deflection angle of the articulation point, ensuring that the tractor travels accurately along the set path and achieving low-energy consumption and high-stability path tracking control.

[0156] In one specific embodiment, to minimize the distance deviation between the actual path and the preset path of the articulated track tractor, enabling it to accurately track the preset path, the articulated track tractor often deviates from the original planned path during automatic travel along the predetermined path due to its own inertia or external conditions. (Reference) Figure 3 S is the starting point of the planned path, A is the ending point of the planned path, and the articulated tracked tractor is set to follow the arc. The tractor is moving. However, due to a deviation in its path, after time t, the actual position of the tracked tractor is point C, and the current position and attitude of the tractor are represented by Wc = [xc, yc, θc]T. The theoretical position should be point R, and the theoretical position and attitude are represented by Wr = [xr, yr, θr]T. The path deviation P between the desired target position and the current actual position on the predetermined path of the articulated tracked tractor at any given time is... e It can be represented as:

[0157]

[0158] This module utilizes fuzzy inference rules to input path deviation as a parameter to the fuzzy controller. Based on the articulated track travel parameters, fuzzy rules are edited to adjust the output parameters KP, KI, and KD of the fuzzy controller. The controller outputs corresponding control commands to the actuators based on the deviation during travel and the real-time changes of the three parameters, controlling the rotation speed of each track and the deflection angle of the articulation point, thereby controlling the position and attitude of the four-track tractor and reducing the deviation from the preset path.

[0159] This application improves path tracking accuracy and steering stability by matching the control parameters of the track drive system and the steering system, effectively reducing the deviation between the actual driving path and the planned path.

[0160] Furthermore, the execution unit is used to send the control commands output by the path tracking control unit and the parameter coordination unit to the track drive motor controller and the articulated motor controller.

[0161] The adaptive walking control module of this invention is responsible for real-time coordinated control of the tractor's track drive system and steering articulation structure based on the target path information output by the path planning module, achieving goals such as path tracking, posture stability, and energy-saving walking. This module combines vehicle dynamics modeling, track speed adjustment, articulation angle control, and steering coordination strategies, possessing excellent adaptability and control precision, making it suitable for stable operation in complex farmland environments.

[0162] The autonomous tracking system for electric articulated four-track tractors proposed in this invention achieves high-precision, low-energy-consumption, and highly adaptable autonomous path tracking control of electric articulated track tractors in complex farmland or operating environments through the coordinated operation of multiple modules. Specifically, the perception module integrates data from various sensors, including binocular cameras, RTK positioning devices, lidar, inertial measurement units (IMUs), and speed sensors, to construct a complete environmental and vehicle state perception system with real-time perception, data fusion, and obstacle recognition capabilities. The path planning module dynamically generates a safe and smooth target driving path based on the multimodal information provided by the perception module, combined with the tractor's dynamic constraints and terrain features, and provides path look-ahead points and trajectory reference information. The adaptive travel control module controls the track drive system and articulated structure based on the path planning results, achieving precise matching control of the inner and outer track speeds and articulation angles, dynamically correcting the deviation between the actual and desired paths, and improving the overall vehicle path tracking performance and driving stability.

[0163] This invention has a reasonable structure, advanced system design, and practical control strategy, and has broad prospects for promotion. It is applicable to various scenarios such as agricultural cultivation, forest transportation, and mountain operations, and has strong engineering application value and economic benefits.

[0164] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0165] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An autonomous tracking system for an electric articulated four-track tractor, characterized in that, include: The sensing module is used to acquire environmental information and vehicle status in real time; The path planning module is used to construct a local environment map based on environmental information and vehicle status and generate a dynamic, traceable path. The adaptive walking module coordinates the control of the track drive system and the articulated steering system based on the output instructions of the path planning module.

2. The autonomous tracking system according to claim 1, characterized in that, The sensing module includes a binocular camera, a lidar, an RTK-GNSS positioning device, an inertial measurement unit, and a track rotation speed sensor; and the acquired data is preprocessed through the following steps: Distortion correction and depth reconstruction are performed on the binocular images, and the laser point cloud is filtered and denoised. RTK position data and IMU attitude data are fused using an extended Kalman filter; Differential analysis is performed on real-time rotational speed data to identify nonlinear motion states.

3. The autonomous tracking system according to claim 2, characterized in that, Data fusion is performed through the following steps: Based on timestamps and sensor calibration results, multi-source data are uniformly converted to the vehicle coordinate system; Spatial matching is performed between the preprocessed binocular images and laser point clouds to output an obstacle grid map and the boundary of the passable area. Dynamic target trajectories are generated based on continuous frame point cloud and image difference.

4. The autonomous tracking system according to claim 1, characterized in that, The perception module performs the following anomaly detection and fault tolerance mechanism: Periodically monitor sensor status indicators, and when the lidar malfunctions, activate VSLAM with vision-IMU fusion as a redundancy scheme. Dynamically weighted trust levels for each sensor; The watchdog mechanism of the edge computing platform enables rapid recovery after the perception module crashes.

5. The autonomous tracking system according to claim 3, characterized in that, The execution steps of the path planning module include: (1) Based on the obstacle grid map, passable area boundary and dynamic target trajectory output by the perception module, a local cost map is constructed; high slip areas are filtered according to the nonlinear motion state of the tracks; (2) Generate a global initial path using the A* algorithm based on the current task target point and real-time vehicle status; (3) Adjust the local paths in the global initial path in real time using the dynamic window method, and update the safety buffer distance and avoidance method based on the adjusted local paths; (4) Optimize the path based on terrain access weight, slope limitation and energy consumption; (5) Optimize the path by taking into account terrain changes, obstacle locations and the characteristics of the tractor's articulated structure; (6) Perform curvature constraint and smoothing on the optimized path; (7) Output the path point sequence of the optimized global path.

6. The autonomous tracking system according to claim 5, characterized in that, Each waypoint contains coordinates, target velocity, and desired heading angle.

7. The autonomous tracking system according to claim 1, characterized in that, The adaptive walking module includes a path tracking control unit, which is used to perform the following steps: 1) Synchronize the path point sequence with the real-time vehicle status using Kalman filtering; 2) Select a continuous path segment as the reference trajectory and solve for the optimal control sequence through quadratic programming; 3) Convert the control inputs into track speed commands and articulation angle commands.

8. The autonomous tracking system according to claim 1, characterized in that, The adaptive walking module also includes a parameter coordination unit, which uses path deviation as input to dynamically adjust PID parameters and coordinately control the differential speed of the inner and outer tracks and the deflection angle of the articulation point through fuzzy rules.

9. The autonomous tracking system according to claim 1, characterized in that, The adaptive walking module also includes a data receiving unit and an execution unit; The data receiving unit is used to receive the path point sequence from the path planning module and the real-time vehicle status information obtained by the perception module. The execution unit is used to send track speed commands and articulation angle commands to the track drive motor controller and the articulation motor controller, respectively.

Images