A mower course self-correction system and method based on vision and rudder cooperation

Abstract

This application discloses a self-correcting system and method for lawnmower heading based on vision and steering wheel collaboration, relating to the field of lawnmower technology. The system includes: a vision module for acquiring environmental images and processing them to obtain lateral and / or heading deviations; a front wheel steering execution module, located at the front wheel position of the lawnmower, for driving the front wheel to actively deflect; and a controller for generating control commands based on the deviations to drive the front wheel steering execution module. The method includes corresponding control steps. This application, by replacing the traditional passive front swivel wheel with an actively controlled steering wheel and forming a closed loop with visual perception, achieves real-time, accurate, and feedforward self-correction of the lawnmower's heading. It effectively solves the problems of response lag, trajectory oscillation, and poor heading maintenance on slopes / uneven ground caused by traditional rear wheel differential steering, improving the accuracy, efficiency, and adaptability of lawnmower operations.

Description

Technical Field

[0001] This application relates to the field of lawnmower technology, and more specifically, to a lawnmower heading self-correction system and method based on vision and steering wheel coordination. Background Technology

[0002] Traditional smart lawnmowers typically employ a mechanical structure of "front swivel casters + rear drive wheels." The front wheels are passive swivel casters, serving only as supports and for steering; the machine's primary steering function relies on controlling the differential speed between the two rear drive wheels. When combined with a purely visual navigation system, this traditional steering mechanism reveals two significant drawbacks: 1. Steering Response Lag: Differential steering requires the front casters to be passively turned by shifting the machine's center of gravity before the direction of travel can be changed. This indirect control method results in a significant delay in the execution of corrective commands after the system visually detects a heading deviation. Consequently, the lawnmower's actual trajectory is prone to "S"-shaped oscillations, resulting in poor path straightness and affecting mowing performance and work efficiency.

[0003] 2. Poor heading stability: When operating on slopes or uneven grass, the passive casters are easily disturbed by external ground forces, resulting in uncontrolled "self-excited swaying" or sideslip. This instability in chassis attitude directly interferes with the visual SLAM (Simultaneous Localization and Mapping) system, causing cumulative errors in its estimated heading angle. Over time, this can lead to positioning drift, and in severe cases, may result in repeated compaction or missed areas.

[0004] Therefore, the steering method based on the combination of rear wheel differential and passive front wheel in the existing technology has become a key bottleneck restricting the realization of high-precision and high-smoothness autonomous operation of visual navigation lawnmowers. Summary of the Invention

[0005] In view of this, this application provides a vision-based and steering wheel-coordinated lawnmower heading self-correction system and method, which aims to solve the problems of low path control accuracy and insufficient stability caused by the lag in steering response and poor heading maintenance of traditional lawnmowers.

[0006] In a first aspect, this application provides a self-correcting heading system for a lawnmower based on vision and steering wheel coordination, wherein the lawnmower has two rear wheels and at least one front wheel, characterized in that it further includes: The vision module is used to acquire and process environmental images in front of the lawnmower, and output lateral deviation and / or heading deviation. A front wheel steering actuator module is located at the front wheel position of the lawnmower and is used to drive the front wheel to actively deflect. The controller is communicatively connected to both the vision module and the front wheel steering execution module. The controller is configured to generate control commands to drive the front wheel steering execution module to perform actions based on the lateral deviation and / or heading deviation output by the vision module, so as to replace or assist the rear wheel differential in achieving heading correction.

[0007] The core design concept of this application lies in "dynamic collaboration and force control balance." By upgrading the front wheels to active steering "steering wheels" and directly closing the loop with the vision perception system, millisecond-level direct-drive response to steering commands is achieved, fundamentally eliminating the mechanical lag of traditional differential steering. More importantly, the system constructs a dynamic balance between heading stability and correction: the vision module provides real-time path and pose feedback as a control reference, the front wheel steering execution module serves as a fast and precise execution terminal, and the controller acts as a collaborative hub, integrating information from multiple sensors (such as angle sensors and IMUs) and dynamically calculating and outputting the optimal front wheel steering angle under various operating modes (such as feedforward, drift correction, and slope compensation) to actively counteract external disturbances (such as sideslip and slope), maintaining heading stability and accuracy. This strong coupling of "perception-decision-execution" and "active disturbance rejection" design represents a fundamental breakthrough from the traditional passive steering technology path in this field, producing unexpected improvements in path straightness and heading stability.

[0008] In some implementations, the controller is configured to perform a feedforward control mode: when the vision module detects the boundary of the uncut area or an obstacle ahead, it calculates the target turning angle of the front wheel steering execution module through a kinematic model based on the lateral and directional deviations of the visual output, and controls it to complete the front wheel pre-deflection before the lawnmower enters a curve or approaches an obstacle.

[0009] By adopting the above technical solutions, the system has "foresight" and can act in advance to make smooth transitions, completely eliminating steering delay and trajectory overshoot when entering curves, achieving a smooth steering experience like that of an "experienced driver", and improving the smoothness and efficiency of operation.

[0010] In some implementations, the system further includes: an angle sensor for sensing the actual steering angle of the front wheel steering actuation module; the controller is also configured to perform a drift correction mode: when there is a difference between the heading angle estimated by the vision module using the SLAM algorithm and the heading angle determined based on the feedback from the angle sensor, the visual odometry is constrained based on the feedback value of the angle sensor.

[0011] By adopting the above technical solution, a highly reliable heading fusion and correction mechanism is provided. The steering angle provided by the angle sensor is a direct, short-term, and highly accurate physical quantity. This is used to periodically correct the long-term accumulated error of visual SLAM, which can effectively suppress "heading drift" and ensure that the global positioning accuracy of the system remains stable over a long period of time, even in complex grassy environments with sparse texture or drastic lighting changes.

[0012] In some embodiments, the angle sensor includes a magnetic encoder and / or potentiometer disposed within the front wheel steering actuation module.

[0013] By adopting the above technical solution, a reliable and easily integrated angle sensing method is provided. Magnetic encoders or potentiometers can be directly integrated into the steering servo, enabling real-time and accurate measurement of the front wheel deflection angle. This provides crucial status feedback signals for closed-loop control and drift correction, ensuring the accuracy and stability of system control.

[0014] In some implementations, the system further includes an inertial measurement unit for detecting the body posture of the lawnmower; the controller is also configured to perform a slope compensation mode: based on the slope direction detected by the inertial measurement unit, the controller controls the front wheel steering module to generate a deflection angle to counteract the downward trend caused by gravity.

[0015] By adopting the above technical solution, the system has the ability to actively cope with complex terrain. By sensing the slope and actively causing the front wheels to make a slight "uphill" deflection, a compensating lateral force can be generated to directly counteract the tendency of sideslip caused by the component of gravity, thereby ensuring that the lawnmower can still travel along the predetermined straight line on the slope, avoiding missed mowing or going over the boundary.

[0016] In some embodiments, the controller is also configured to execute a precision crossing mode and an energy-saving control mode: the precision crossing mode is used to calculate the target heading when the vision module identifies a narrow passage entrance and control the front wheel steering execution module to complete a precise steering alignment in one go; the energy-saving control mode is used to dynamically adjust the response frequency of the front wheel steering execution module according to the straightness of the path identified by the vision module.

[0017] By adopting the above technical solutions, intelligent optimization of functionality and energy efficiency has been achieved. The precision crossing mode solves the problem of "stuttering" and repeated adjustments when the machine passes through narrow areas in courtyards (such as fence gates), improving passing efficiency and success rate. The energy-saving control mode reduces the frequency of active fine-tuning of the steering wheel on long straight paths, reducing energy consumption and extending the range while ensuring basic accuracy.

[0018] Secondly, this application provides a self-correcting method for lawnmower heading based on vision and steering wheel coordination, applicable to lawnmowers with two rear wheels and at least one front wheel, characterized by including the following steps: The vision module acquires environmental images, identifies the position of the lawnmower relative to a predetermined path or boundary, and calculates lateral deviation and / or heading deviation. The controller generates front wheel steering control commands based on the lateral deviation and / or heading deviation; The front wheel steering module, which is located at the front wheel position of the lawnmower, actively deflects the front wheel according to the control command to replace or assist the rear wheel differential in achieving heading correction.

[0019] By adopting the above technical solution, the method transforms the collaborative concept of "visual perception-front wheel active control" into an executable operational process. This method features clear steps and a closed-loop logic, systematically guiding lawnmowers to achieve more precise and efficient heading self-correction compared to traditional rear-wheel differential steering. It represents the concrete manifestation of the aforementioned system solution at the control logic level.

[0020] In some implementations, in the step of acquiring environmental images through the vision module, pose estimation is performed using simultaneous localization and mapping (SLAM) technology; the method further includes: obtaining the actual steering angle feedback of the front wheel steering execution module; comparing the heading angle estimated by SLAM with the heading angle determined based on the actual steering angle feedback; when the difference between the two exceeds a preset threshold, correcting the heading estimation of visual SLAM based on the actual steering angle feedback.

[0021] By adopting the above technical solution, the method incorporates multi-sensor fusion correction logic. By introducing the highly reliable mechanical feedback of the steering wheel angle into the SLAM odometry correction stage, a complementary pose correction strategy is formed, which effectively improves the accuracy and robustness of the navigation and positioning system in long-term operation. This is a key step in ensuring the long-term effectiveness of the method.

[0022] In some implementations, a feedforward control step is also included: when the vision module anticipates that it is about to enter a curve or approach an obstacle, before the lawnmower actually arrives, the target turning angle is calculated in advance based on the lateral deviation and heading deviation, and the front wheel steering execution module is controlled to complete the front wheel pre-deflection.

[0023] By adopting the above technical solution, the method introduces a forward-looking control approach. This step breaks through the passive mode of traditional feedback control, which involves "detecting deviations and then correcting them." By acting in advance to offset system inertial delays, it is the core feature of achieving smooth, overshoot-free trajectory control, significantly improving the driving quality of the lawnmower.

[0024] In some implementations, a slope operation compensation step is also included: the machine attitude and slope are detected in real time by an inertial measurement unit; the front wheel compensation deflection angle required to counteract the gravity sideslip component is calculated based on the slope direction and angle; and the front wheel steering execution module is controlled to generate the compensation deflection angle.

[0025] By adopting the above technical solution, the method enhances terrain adaptability. This step enables the lawnmower's control algorithm to sense and actively compensate for the influence of terrain slope, incorporating environmental disturbances into the feedforward compensation scope, thereby ensuring the consistency and stability of the operating trajectory under different slope conditions and improving the universality and reliability of the method.

[0026] In summary, this application has at least one of the following beneficial technical effects: 1. Achieved high-precision and fast-response heading control: By replacing the traditional passive omnidirectional wheel + rear wheel differential mode with active steering of the front wheel (steering wheel), direct drive and millisecond-level response of heading correction are achieved, fundamentally eliminating steering lag, making the mowing path straight and smooth, reducing "S"-shaped oscillations, and greatly improving mowing effect and work efficiency.

[0027] 2. A robust system with dynamic collaboration has been constructed: By integrating information from multiple sources such as vision, angle sensors, and IMU, and combining various intelligent modes such as feedforward control, drift correction, and slope compensation, the system can actively predict and compensate for various internal and external disturbances (such as terrain, slippage, and visual error accumulation) in real time, which significantly enhances the heading maintenance capability and operational stability in complex and uneven grass environments.

[0028] 3. Expanded the scenario adaptability and functionality of intelligent lawnmowers: The precision passage mode enables them to smoothly pass through narrow passages, while the energy-saving mode optimizes energy utilization. This solution provides a high-performance underlying execution platform for intelligent lawnmowers to achieve more complex operational logic (such as precise obstacle avoidance and efficient reciprocating mowing), expanding their application scenarios and value. Attached Figure Description

[0029] Figure 1 This is a schematic diagram of the internal structure of the lawnmower of this application; Figure 2 This is a schematic diagram showing the assembly location of the front wheel steering actuator module; Figure 3 This is an exploded view of the front wheel steering actuator module and its base. Figure 4 This is a schematic diagram of the front wheel steering actuator module and the front wheel assembly structure; Figure 5 This is a schematic diagram showing the exploded structure of the front wheel steering actuator module and the front wheel.

[0030] Explanation of reference numerals in the attached diagram: 1. Vision module; 2. Controller; 3. Front wheel steering actuation module; 31. Servo motor; 32. Servo motor steering component; 33. First link; 34. Second link; 35. Spring; 36. Torsion spring; 37. Front wheel steering component; 38. Drive rod; 4. Front wheel; 5. Base; 6. Rear wheel. Detailed Implementation

[0031] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0032] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0033] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0034] In the description of this application, it should be understood that the terms "upper", "lower", "left", "right", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this application.

[0035] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings. Unless otherwise specified, the features in the following embodiments can be combined with each other.

[0036] Example 1: A Lawn Mower Heading Self-Correction System Based on Vision and Steering Wheel Collaboration Combination Figures 1-5As shown, this embodiment provides a heading self-correction system applied to an intelligent lawnmower. The lawnmower includes a base 5, two rear wheels 6, and a front wheel 4. The two rear wheels 6 are rotatably connected to the base 5 and can be driven independently, serving as the power source for the lawnmower's movement. The front wheel 4 is connected to the base 5 via a front wheel steering execution module 3, and is rotatably connected to a drive rod 38 of the front wheel steering execution module 3, and is driven by the drive rod 38 to achieve directional rotation. A controller 2 is mounted on the base 5 and is communicatively connected to both the vision module 1 and the front wheel steering execution module 3.

[0037] Combination Figure 4 and Figure 5 The front wheel steering module 3 includes a servo motor 31, a servo steering component 32, a first link 33, a second link 34, a spring 35, a torsion spring 36, a front wheel steering component 37, and a drive rod 38. The servo motor 31 is controlled by the controller 2 to achieve forward and reverse rotation. The drive shaft of the servo motor 31 is splinedly connected to the servo steering component 32. The two ends of the servo steering component 32 are connected to the two ends of the front wheel steering component 37 through the first link 33 and the second link 34. The servo steering component 32 is splinedly connected to the drive rod 38. The rotation of the servo motor 31 can drive the servo steering component 32 to rotate, which in turn drives the front wheel steering component 37 to rotate through the first link 33 and the second link 34. The front wheel steering component 37 drives the drive rod 38 to rotate, which in turn drives the front wheel 4 to rotate, thereby achieving directional adjustment.

[0038] Combination Figures 2-5 As shown, the servo motor 31 is mounted on the base 5. The drive rod 38 is rotatably connected to the base 5 and is capped in the rotational axis, thus limiting axial movement. A torsion spring 36 is provided on the base 5, with both ends of the torsion spring 36 passing through two through holes in the front wheel steering component 37, providing a restoring force to the front wheel steering component 37. The front wheel steering component 37 and the drive rod 38 are circumferentially fixed by a spline and axially slidable. The drive rod 38 has a step for supporting the lower part of the front wheel steering component 37, and a limiter is provided at its end to limit the axial sliding range of the front wheel steering component 37. A spring 35 is provided between the front wheel steering component 37 and the limiter, thereby enabling the front wheel steering component 37 to float to a certain extent when driving the drive wheel, reducing the risk of jamming.

[0039] Combination Figures 1-5 As shown, the system mainly includes: Vision Module 1: A monocular or binocular camera deployed in the waterproof housing at the front of the lawnmower. It works with the onboard processor to run visual SLAM and path recognition algorithms, and outputs in real time the lateral deviation (d) and heading deviation (θ) of the lawnmower relative to the preset electronic boundary or the boundary of the mower's mowed / unmowed area.

[0040] Front wheel steering actuation module 3: Integrated within the front wheel 4 assembly, it includes a DC servo motor or stepper motor, a reduction gear set, and a steering mechanism directly coupled to the wheel axle. This module receives commands from controller 2 and precisely drives the front wheel 4 to turn left or right by a specific angle. The module integrates a high-precision magnetic encoder or potentiometer as an angle sensor to provide real-time feedback on the actual steering angle of the front wheel 4.

[0041] Controller 2: As the core of the system, it is usually the main control MCU of the lawnmower or a dedicated motion control chip. It receives deviation signals from vision module 1, actual turning angle feedback from the front wheel 4 from the angle sensor, and pitch, roll, and yaw rate / angle information from the inertial measurement unit.

[0042] Control Mode and Process: Basic feedback control: Based on the lateral and heading deviations fed back in real time by the vision module 1, the controller 2 calculates the required target turning angle of the front wheels 4 through control algorithms such as PID, and drives the front wheel steering execution module 3 to return the lawnmower to the predetermined path.

[0043] Feedforward control: When the vision module 1 detects the curve boundary approximately 0.5 meters ahead, the controller 2, based not only on the current deviation but also on the lawnmower's current speed and kinematic model, pre-calculates the required front wheel 4 turning angle for a smooth curve and instructs the front wheel steering execution module 3 to begin turning in advance. When the lawnmower reaches the start of the curve, the front wheel 4 has essentially reached the target turning angle, thus achieving a smooth, delay-free curve entry.

[0044] Drift Correction: During system operation, controller 2 runs two heading estimation channels in parallel. Channel A estimates the heading angle change based on inter-frame data from visual SLAM. Channel B calculates the heading angle change based on the front wheel steering angle feedback from the angle sensor (4) and the rear wheel encoder odometer information (6), using the Ackerman steering geometry model. Controller 2 continuously compares the results of the two channels. If the cumulative difference in heading angle between the two channels exceeds 2 degrees during a 10-second straight-line driving period, a slight drift is determined in the visual SLAM, and the heading state of the visual SLAM is immediately reset and corrected using the heading angle of channel B (directly fed back by the sensor) as the reference.

[0045] Slope Compensation: When the IMU detects a sustained lateral tilt angle in the mower (e.g., the left side is 5 degrees higher than the right), controller 2 determines that the mower is on a lateral slope. According to the mechanics model, a component of gravity will cause the machine to tend to slide downhill. To compensate for this tendency, controller 2 calculates and outputs a small, uphill-oriented (in this case, the right side) front wheel 4 compensation yaw angle (e.g., 1-3 degrees). This yaw generates a lateral grip component, precisely counteracting the sliding force, allowing the mower to resist the slope's effects and continue traveling in a straight line.

[0046] Precision Crossing Mode: When the lawnmower is performing autonomous tasks, the vision module 1 identifies in real time whether there are narrow passage entrances ahead (such as yard fence gates, gaps between two flower beds, etc.). The determination of a narrow passage is based on the following: the vision module 1 identifies the two side boundaries (such as fences, flower bed edges) through semantic segmentation or object detection algorithms, and calculates the lateral distance between the two side boundaries. When this lateral distance is less than 1.5 times the width of the lawnmower body (for example, if the body width is 50cm, the distance is less than 75cm) but greater than 1.05 times the body width, it is determined to be a narrow passage entrance. If the distance is less than 1.05 times the body width, it is determined to be impassable, the system issues an alarm and stops moving forward.

[0047] When a narrow passage entrance is identified, controller 2 performs the following precise crossing steps: 1. Establish local target heading: The vision module 1 uses the central axis of the channel entrance as a reference to calculate the angular deviation Δψ between the current lawnmower head orientation and the central axis, as well as the lateral distance deviation Δd from the center point of the lawnmower to the central axis of the channel entrance.

[0048] 2. Calculating the one-time steering angle: Controller 2, based on the current travel speed v of the lawnmower (measured by the encoder of rear wheel 6), the maximum steering rate ω_max of the steering execution module (determined by the specifications of servo motor 31), and the aforementioned deviation, uses a time-optimal control strategy to calculate the target steering angle δ_target of front wheel 4. The specific formula is as follows: δ_target = arctan( L * (k1·Δψ + k2·Δd) / v² ) Where L is the wheelbase of the lawnmower, and k1 and k2 are pre-calibrated control coefficients (value range: k1=0.5~2.0, k2=0.1~0.5). If the calculated δ_target exceeds the maximum mechanical turning angle δ_max of the front wheel 4, then the limit is δ_max.

[0049] 3. One-time steering execution: Controller 2 sends a steering angle command to the front wheel steering execution module 3, causing it to deflect from the current angle to δ_target in the shortest possible time (usually no more than 0.5 seconds), and maintain this angle until the lawnmower has fully entered the passage (i.e., the vision module 1 detects that the machine has passed the passage entrance). During the entry into the passage, no feedback adjustments based on real-time path deviation are performed to avoid wobbling due to repeated corrections that could prevent smooth passage. After the machine has fully entered the passage, the system resumes the normal feedback control mode.

[0050] By following the steps above, the lawnmower can accurately aim at the entrance of a narrow passage and pass through smoothly in one continuous, vibration-free motion, significantly improving the success rate of crossing.

[0051] Energy-saving control mode: While the lawnmower travels along the planned path, vision module 1 continuously identifies the straightness of the path ahead. The straightness is quantified by performing curve fitting on the path or boundary line extracted from the current image and calculating the average curvature C_avg (in m⁻¹) of the fitted curve. Alternatively, in applications with electronic maps, the curvature of the path is calculated based on discrete points of the preset navigation path.

[0052] Controller 2 dynamically adjusts the response frequency (i.e., the control cycle of the position loop) of the front wheel steering execution module 3 based on straightness, with the specific strategy as follows: When the path is a straight line or approximately a straight line (judgment condition: C_avg ≤ 0.05 m⁻¹, or vision module 1 does not detect a heading deviation exceeding 0.5 degrees for 3 consecutive seconds), controller 2 reduces the response frequency of front wheel steering execution module 3 from 50Hz in normal mode to 5Hz~10Hz. Simultaneously, it allows the heading deviation to accumulate to a large quiet zone threshold (e.g., ±1.5 degrees) before triggering a correction action, avoiding frequent fine-tuning on straight lines.

[0053] When the path is a gentle curve (judgment condition: 0.05 m⁻¹ < C_avg ≤ 0.2 m⁻¹), the controller 2 will restore the response frequency to the normal mode (e.g. 20Hz~50Hz) to ensure normal tracking accuracy.

[0054] When the path is a sharp bend (judgment condition: C_avg > 0.2 m⁻¹), the controller 2 increases the response frequency to the highest level (e.g., 100Hz) and enables the feedforward control mode (see claim 2) to achieve fast and precise steering response.

[0055] In addition, when the lawnmower is in slope compensation mode or drift correction mode, it automatically exits the energy-saving control mode to prioritize heading stability. The energy-saving control mode reduces switching losses in the servo motor 31 drive circuit and motor power consumption on straight paths, and actual measurements show it can extend battery life by approximately 10% to 20%.

[0056] Through the above implementation method, the system can adaptively adjust the control frequency according to the path characteristics, significantly reduce energy consumption and extend the mowing time after a single charge while ensuring operational accuracy.

[0057] Example 2: A method for self-correcting the heading of a lawnmower based on vision and steering wheel coordination Combination Figures 1-5 As shown, this embodiment elaborates on the heading self-correction method applied to the above system. This method provides complete control logic for achieving high-precision autonomous mowing.

[0058] S1: System initialization. After power-on, the vision module 1, front wheel steering actuation module 3, angle sensor, IMU and controller 2 are started to perform self-test and parameter calibration.

[0059] S2: Real-time perception and deviation calculation. Vision module 1 continuously acquires images of the environment ahead, extracts path or boundary features through image recognition algorithms, and calculates the lawnmower's current lateral deviation d and heading deviation θ.

[0060] S3: Multi-mode decision-making and target angle calculation. This is the core step of the method, with controller 2 processing multiple logic steps in parallel: Main feedback loop: Based on d and θ calculated by S2, the basic front wheel 4 steering angle δ_fb is calculated by the feedback control algorithm (such as fuzzy PID) in controller 2.

[0061] Feedforward Decision: Vision module 1 analyzes the curvature of the path ahead. If it detects that the vehicle is about to enter a curve (curvature greater than a threshold) or is approaching an obstacle, feedforward control is triggered. Combining the current vehicle speed and the path geometry model, the feedforward compensation steering angle δ_ff is calculated. The final target steering angle δ_target = δ_fb + δ_ff.

[0062] Drift correction judgment: Periodically (e.g., once per second), compare the visual SLAM heading angle Φ_v with the heading angle Φ_o calculated based on the front wheel 4-turn angle / odometer. If |Φ_v - Φ_o| > the threshold (e.g., 1.5 degrees), visual drift is judged, correction is performed: Φ_v = Φ_o, and some state variables of SLAM are reset.

[0063] Slope compensation determination: The roll angle γ of the IMU is read in real time. If |γ| > the slope threshold (e.g., 3 degrees), the slope compensation angle δ_slope is obtained by querying the preset compensation table or calculating it using a formula based on the magnitude and direction of γ. δ_target = δ_fb + δ_slope.

[0064] S4: Command Issuance and Execution. Controller 2 sends the final calculated target steering angle δ_target command to the front wheel steering execution module 3. The motor in the execution module drives the front wheels 4 to steer, while the high-precision angle sensor feeds back the actual steering angle δ_actual to controller 2, forming a closed-loop control.

[0065] S5: Repeat steps S2 to S4 until the lawn mowing task is completed.

[0066] The key to this method lies in multi-mode parallel processing and decision fusion. It is not a simple conditional branching, but allows multiple control variables such as feedback, feedforward, and compensation to be intelligently fused at the decision level (e.g., weighted summation or selection of the dominant mode through a state machine), thereby ensuring that the system can output the optimal front-wheel 4-steering command under any operating condition, achieving a balance between accuracy, smoothness, and robustness.

[0067] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments, and the embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes, modifications, substitutions, and variations can be made to the present invention without departing from its spirit and scope, and all such changes, modifications, substitutions, and variations fall within the scope of the present invention as claimed.

Claims

1. A vision-based and steering wheel-coordinated self-correction system for a lawnmower, the lawnmower having two rear wheels and at least one front wheel, characterized in that, Also includes: The vision module is used to acquire and process environmental images in front of the lawnmower, and output lateral deviation and / or heading deviation. A front wheel steering actuator module is located at the front wheel position of the lawnmower and is used to drive the front wheel to actively deflect. The controller is communicatively connected to both the vision module and the front wheel steering execution module. The controller is configured to generate control commands based on the lateral deviation and / or heading deviation output by the vision module to drive the front wheel steering execution module to perform actions, thereby replacing or assisting the rear wheel differential in correcting the heading.

2. The lawnmower heading self-correction system based on vision and steering wheel coordination according to claim 1, characterized in that, The controller is configured to perform feedforward control mode: When the vision module detects the boundary of the uncut area or an obstacle ahead, it calculates the target turning angle of the front wheel steering execution module through a kinematic model based on the lateral and directional deviations of the visual output, and controls it to complete the front wheel pre-deflection before the lawnmower enters a curve or approaches an obstacle.

3. The lawnmower heading self-correction system based on vision and steering wheel coordination according to claim 1, characterized in that, Also includes: An angle sensor is used to sense the actual steering angle of the front wheel steering actuation module; The controller is also configured to perform a drift correction mode: When there is a difference between the heading angle estimated by the vision module using the SLAM algorithm and the heading angle determined based on the feedback from the angle sensor, the visual odometry is constrained based on the feedback value of the angle sensor.

4. The lawnmower heading self-correction system based on vision and steering wheel coordination according to claim 3, characterized in that, The angle sensor includes a magnetic encoder and / or potentiometer disposed within the front wheel steering actuation module.

5. The lawnmower heading self-correction system based on vision and steering wheel coordination according to claim 1, characterized in that, Also includes: An inertial measurement unit is used to detect the attitude of the lawnmower. The controller is also configured to perform a ramp compensation mode: Based on the slope direction detected by the inertial measurement unit, the front wheel steering module is controlled to generate a deflection angle to counteract the downward trend caused by gravity.

6. The lawnmower heading self-correction system based on vision and steering wheel coordination according to claim 1, characterized in that, The controller is also configured to perform a precision traversal mode and an energy-saving control mode; The precision crossing mode is used to calculate the target heading when the vision module detects the entrance to a narrow passage, and to control the front wheel steering execution module to complete the precise steering alignment in one go; The energy-saving control mode is used to dynamically adjust the response frequency of the front wheel steering execution module based on the straightness of the path identified by the vision module.

7. A self-correcting method for lawnmower heading based on vision and steering wheel coordination, applied to lawnmowers with two rear wheels and at least one front wheel, characterized in that, Includes the following steps: The vision module acquires environmental images, identifies the position of the lawnmower relative to a predetermined path or boundary, and calculates lateral deviation and / or heading deviation. The controller generates front wheel steering control commands based on the lateral deviation and / or heading deviation; The front wheel steering module, which is located at the front wheel position of the lawnmower, actively deflects the front wheel according to the control command to replace or assist the rear wheel differential in achieving heading correction.

8. The method according to claim 7, characterized in that, In the step of acquiring environmental images through a vision module, pose estimation is performed using Simultaneous Localization and Mapping (SLAM) technology; the method further includes: Obtain the actual steering angle feedback from the front wheel steering execution module; The heading angle estimated by SLAM is compared with the heading angle determined based on the actual steering angle feedback; When the difference between the two exceeds a preset threshold, the heading estimation of visual SLAM is corrected based on the actual steering angle feedback.

9. The method according to claim 7, characterized in that, It also includes feedforward control steps: When the vision module anticipates that it is about to enter a curve or approach an obstacle, it calculates the target turning angle in advance based on the lateral deviation and heading deviation before the lawnmower actually arrives, and controls the front wheel steering execution module to complete the front wheel pre-deflection.

10. The method according to claim 7, characterized in that, It also includes slope operation compensation steps: The fuselage attitude and slope are detected in real time through an inertial measurement unit; Calculate the front wheel compensation deflection angle required to counteract the gravity sideslip component based on the slope direction and angle; The front wheel steering execution module is controlled to generate the compensation deflection angle.

Images