Inter-ridge strawberry disease inspection robot structure and control method thereof

Abstract

The invention discloses an inter-ridge strawberry disease inspection robot structure and a control method thereof. The inter-ridge strawberry disease inspection robot structure comprises a lifting platform structure, a lateral telescopic mechanism, a camera pitching mechanism, a steering wheel moving mechanism and an independent suspension structure. The steering wheel moving mechanism is located at the bottom of the robot and makes direct contact with the ground. The independent suspension structures are installed on the two sides of the steering wheel moving mechanism and can move relative to the steering wheel moving mechanism. The lifting platform structure is located in the middle of the robot and connected with the independent suspension structure through bolts, a lateral telescopic mechanism is installed on the lower portion of the lifting platform structure, and the extension direction of a telescopic arm is perpendicular to the advancing direction of the robot. The camera pitching mechanism is installed at the tail ends of the outer sides of two telescopic arms of the lateral telescopic mechanism, and the two cameras detect the outer side of the robot. According to the invention, through real-time calculation and feedback of errors, the performance of the robot in an unstructured field environment is significantly improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of agricultural machinery automation and robot control, and particularly relates to a ridge-to-ridge inspection robot structure for disease detection in strawberry fields, which is particularly suitable for agricultural scenes such as facility strawberry fields that have high requirements for the environment. BACKGROUND

[0002] Strawberry is a high-value crop in facility agriculture, and its entire growth period is in a closed environment with high temperature and high humidity, which is prone to diseases such as powdery mildew, gray mold, and red spider mites. At present, the production completely relies on manual inspection in the narrow and high ridge with a width of 20 to 30 cm and a distance of only about 50 cm between adjacent ridges, which has high labor intensity, low efficiency, and only relies on visual inspection, often missing early disease spots due to visual obstruction or fatigue, delaying the prevention and treatment opportunity. With the aging of rural labor force, this mode has been difficult to continue.

[0003] In view of the low efficiency of manual inspection, CN113691114A proposes a greenhouse intelligent inspection robot, which adopts a four-wheel differential chassis to carry a visible light camera. However, its minimum turning radius is about one meter, which is much larger than the distance between strawberry ridges, and the machine cannot turn around in the ridge, which will crush the ridge and the plants, and the structure has poor adaptability, which cannot be put into practical use.

[0004] Another scheme CN201910395967.7 uses a tracked chassis to improve the passability, but the overall width exceeds 600 mm, which still cannot enter the standard strawberry ridge; its visual system only relies on ordinary visible light imaging, which has insufficient accuracy in identifying early powdery mildew spots in the dense leaf-shielded and variable-light greenhouse environment, and cannot complete the fine early disease detection. Therefore, there is an urgent need for a ridge-to-ridge strawberry disease inspection robot with narrow body width, small turning radius, and high-precision disease identification capability. SUMMARY

[0005] The present application aims to solve the problems existing in the prior art and provide a ridge-to-ridge strawberry disease inspection robot structure and a control method thereof.

[0006] The technical scheme of the present application includes: a ridge-to-ridge strawberry disease inspection robot structure, which comprises a lifting platform structure, a lateral telescopic mechanism, a camera tilting mechanism, a steering wheel moving mechanism, and an independent suspension structure. The steering wheel moving mechanism is located at the bottom of the robot and directly contacts the ground; the independent suspension structure is installed on both sides of the steering wheel moving mechanism, and relative displacement can occur between them; the lifting platform structure is located in the middle of the robot and is connected to the independent suspension structure by bolts, and the lower part is provided with the lateral telescopic mechanism, and the extension direction of the telescopic arm is perpendicular to the forward direction of the robot; the camera tilting mechanism is installed at the outer end of the two telescopic arms of the lateral telescopic mechanism, and the detection direction of the two cameras is the outer side of the robot.

[0007] The lifting platform structure of the inspection robot comprises support columns (5), cross beams (13), lifting motors (2), lifting sliding rails (6), lifting sliding blocks (7), lead screws (16), lead screw nuts (18), couplings (20), platform cross beams (19) and platform connecting plates (17). The lifting frame is composed of four vertical support columns (5) and two horizontal cross beams (13). The two ends of each cross beam (13) are connected with the top ends of the two support columns (5) respectively, thereby forming two parallel lifting frames in front and back. Four lifting sliding rails (6) are fixed to the side surfaces of the support columns (5), and the installation positions of the lifting sliding rails (6) are inside the lifting frame structure. Two lifting sliding blocks (7) placed in an up-down manner are installed on one lifting sliding rail (6), and a total of eight lifting sliding blocks (7) are fixed to the two ends of the two horizontally installed platform cross beams (19). The top end of the lead screw nut (18) is flush with the top end of the platform cross beam (19), and the platform cross beam (19) is clamped and fixed through the platform connecting plate (17). The lifting motor (2) is connected with the lead screw (16) through the coupling (20), realizes the vertical direction movement of the lifting platform, makes the multifunctional inspection platform structure have the plant detection capability at different heights, and ensures that the camera (28) images the strawberry canopy in the best angle range.

[0008] Further, the lateral telescopic mechanism comprises telescopic cross beams (32), racks (25), cylindrical gears (34), bevel gears (42), gear fixing frames (41) and horizontal movement motors (40). The lateral telescopic mechanism comprises a set of bevel gear sets engaged at an angle of 90 degrees, including an input bevel gear connected with the horizontal movement motor (40) and two output bevel gears meshed with the input bevel gear and having mutually parallel output shafts. The two side bevel gears (42) are connected with the cylindrical gears (34), the rack (25) is installed below the telescopic cross beam (32), and the cylindrical gear (34) is meshed with the rack (25), thereby realizing the reverse movement of the two side telescopic cross beams (32) at the same speed, and making the camera have the ability to identify crops in different row widths.

[0009] Further, the camera pitching structure is installed at the end of the telescopic cross beam (32) and comprises a rudder (38), a camera support (29) and a camera (28). The camera (28) is connected with the rudder (38) through the camera support (29), realizes the adjustment of the camera pose, thereby frontally shooting the strawberry leaves and fruits at different heights and angles, and improves the disease identification accuracy.

[0010] Further, the rudder wheel moving mechanism comprises a tension sleeve (43), a tire outer fixing part (45), an outer bearing (46), an inner bearing (49), a tire inner fixing part (50), a tire (48), a clamping plate fixing part (51), an outer clamping plate (47), an inner clamping plate (52), a bearing outer clamping plate (63), a rotary bearing (64), a bearing inner clamping plate (65), a right-angle connecting part (66), a rotary motor (61) and a wheel set motor (53), the rudder wheel rotation and the rudder wheel movement are independent of each other, the omnidirectional movement of the inspection robot can be realized, including forward movement, backward movement, transverse movement, oblique movement and in-place turning, and the complex path requirements of ridge-to-ridge narrow road, U-turn and end-to-end row changing can be met.

[0011] Further, the independent suspension structure comprises a compression spring shock absorber (68), a suspension sliding block (57) and a suspension connecting plate (4). The rudder wheel moving structure is fixed on the suspension connecting plate (4) through the suspension sliding block (57) and a suspension sliding rail (70), and the compression spring shock absorber (68) is fixed between the chassis and the rudder wheel, so that the ridge-to-ridge multifunctional inspection platform structure has the ability to move on rugged terrain, effectively isolates the ground impact and ensures the clear and stable vision imaging.

[0012] Further, the sensor integration module comprises a temperature and humidity sensor (71), a carbon dioxide sensor (72), an ultrasonic sensor (24) and a laser radar module (78), which are connected with the single-chip microcomputer controller. The laser radar module (78) is installed on the top of the lifting platform structure, four ultrasonic sensors (24) are arranged in the left-right direction of the robot, two ultrasonic sensors (24) are arranged in the front-rear direction, and the temperature and humidity sensor (71) and the carbon dioxide sensor (72) are installed on the front part of the vehicle body, so that the vehicle pose can be adjusted according to the environment, and the environmental data such as carbon dioxide concentration and temperature and humidity can be detected in real time.

[0013] The control method of the ridge-to-ridge strawberry disease inspection robot structure comprises the following steps: Step 1: After the robot is powered on, the system performs self-checking on the lifting motor (2), the horizontal movement motor (40), the rotary motor (61), the wheel set motor (53) and the temperature and humidity sensor (71), the carbon dioxide sensor (72), the ultrasonic sensor (24) and the laser radar module (78) in the sensor integration module in sequence, if any module feedback is abnormal, the red warning light is immediately turned on and the buzzer is sounded, and the fault code is uploaded through the wireless module at the same time; after the self-checking is passed, the user can switch between the remote control mode and the autonomous mode.

[0014] Step 2: After the robot enters autonomous mode, the lidar module (78) performs a ring scan of the inter-row environment. The point cloud is rasterized to generate the center line of the furrow. The embedded processor runs A* sparse search based on the center line to obtain a smooth trajectory. The lateral offset of the trajectory is dynamically corrected according to the ridge width measured in real time by the ultrasonic sensor (24) to ensure that the robot center line is consistent with the furrow center line.

[0015] Step 3: During the walking process, the gyroscope and accelerometer continuously output the vehicle body attitude. The outer ring position PID takes the trajectory deviation as input, and the inner ring angular velocity PID takes the gyroscope angular velocity as feedback, and together calculate the steering wheel rotation angle and wheel group differential speed. The stroke signal of the compression spring damper (68) is read in real time and used to compensate for ground undulations, so that the platform beam (19) remains horizontal and the camera (28) maintains stable imaging.

[0016] Step 4: The camera (28) is triggered at a fixed frequency during the uniform speed segment, and the servo motor (38) adjusts the pitch angle in real time according to the height of the plant to ensure that the center of the field of view is aligned with the strawberry canopy; the image is inferred at the edge end through the YOLOv11 network, and the lesion category and confidence are output. The results are bound to the current parking position pose and stored locally, and then sent back to the background for farmers to view.

[0017] Step 5: After the task is completed, the robot records the coordinates of the destination, turns around 180° in place, and travels back to the starting point along the trajectory it came from. During the journey, the lifting motor (2) drives the lead screw (16) to lower the platform beam (19) to the lowest position, thereby lowering the center of gravity and improving driving stability. After reaching the starting point, the system automatically cuts off power and enters sleep mode, waiting for the next task to be triggered.

[0018] The present invention has the following beneficial effects: This invention proposes a disease inspection robot for strawberry farms, integrating a lifting platform structure, a lateral telescopic mechanism, a camera tilting mechanism, a steering wheel movement mechanism, and an independent suspension structure. It can autonomously inspect and diagnose diseases in complex greenhouse environments with high ridges, narrow rows, and numerous bends. By improving the YOLOv11 visual model for real-time disease segmentation of strawberry leaves and fruits, and combining multi-source data from temperature and humidity sensors, carbon dioxide sensors, ultrasonic sensors, and LiDAR modules, the system can perform initial disease and pest screening at the edge, significantly reducing reliance on cloud computing power and communication latency, and improving detection timeliness and accuracy. The robot uses a 5G image transmission link to synchronously transmit key images and diagnostic results back to the cloud big data platform, forming a closed loop of "perception-diagnosis-decision," reducing the frequency of manual greenhouse inspections and lowering labor input.

[0019] Compared to traditional manual inspections and existing fixed monitoring solutions, this invention has significant advantages: For precision operations, the system deploys YOLOv11n on an NVIDIA Jetson Xavier NX edge processor. The entire system consumes only 10 watts and can complete a single-frame inference within 7 milliseconds, maintaining an average detection rate of 0.876 for early-stage lesions down to the millimeter level; specifically, it achieves 0.904 for leaf powdery mildew and 0.934 for flower rot, significantly outperforming models of similar scale. Xintian Technology's NX3 5G image transmission module consumes only 1.5 watts and has a latency of less than 100 milliseconds. Its compact size allows for direct embedding in vehicles, transmitting disease images back to mobile phones in seconds, enabling growers to instantly obtain lesion locations and provide targeted treatment. The lateral telescopic mechanism is linked to the camera's tilt mechanism, ensuring the camera is always directly facing the target and avoiding missed detections due to viewing angle deviations.

[0020] In terms of efficient operation, the four sets of independent steering wheel movement mechanisms support on-the-spot turning, diagonal movement and U-turns, and the time for row switching between rows is short; the lifting platform structure can adapt to the plant height within the stroke, and can quickly put into the next round of inspection without changing the vehicle body during the crop rotation period; the independent suspension structure effectively attenuates the impact of uneven furrows, ensuring clear images on high-speed sections, and improving the efficiency of single-row inspection compared with manual operation.

[0021] In terms of intelligent control, the central processing unit module, based on the PID control algorithm, significantly improves the robot's performance in unstructured field environments through real-time error calculation and feedback. The vehicle body not only responds quickly to path and speed deviations, ensuring real-time inspection, but also effectively eliminates continuous steady-state errors, avoids accumulated positioning deviations, and guarantees long-term operational accuracy. Simultaneously, the vehicle body's obstacle avoidance and start-stop maneuvers are smoother, effectively protecting delicate strawberry plants and providing a stable platform for the onboard precision sensing equipment.

[0022] In terms of environmental protection and energy conservation, the robot adopts a composite structure of aluminum profile frame and carbon fiber plate, which makes it lightweight and reduces energy consumption for driving with the same battery capacity; accurate diagnosis reduces blind use of pesticides, which can save pesticides and meet the requirements of green facility agriculture development. Attached Figure Description

[0023] Figure 1 This is a structural diagram of the inter-row strawberry disease inspection robot; Figure 2 This is a structural diagram of the lateral telescopic mechanism; Figure 3 This is an exploded view of the steering wheel movement mechanism parts; Figure 4 This is a structural diagram of the steering wheel movement mechanism; Figure 5 Detailed diagram of the independent suspension structure; Figure 6 Location diagram of the sensor integration module; Figure 7 This is a schematic diagram of the inspection robot structure; Figure 8 A schematic diagram of the inspection robot in operation; Figure 9 This is a schematic diagram of the minimum limit position of the lateral telescopic mechanism; Figure 10 This is a schematic diagram of the maximum limit position of the lateral telescopic mechanism; Figure 11 This is a schematic diagram of the overall control logic of the system. Figure 12 This is a schematic diagram of the system modules of the inspection robot; Figure 13 This is a block diagram illustrating the PID control principle of the system. Detailed Implementation

[0024] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0025] like Figure 1 The diagram shown is of the overall structure of the strawberry disease inspection robot between rows. It mainly includes a lifting platform structure, a lateral telescopic mechanism, a camera pitch mechanism, a steering wheel movement mechanism, and an independent suspension structure.

[0026] Figure 1 The lifting platform structure includes 5-support column, 13-crossbeam, 2-lifting motor, 6-lifting slide rail, 7-lifting slider, 16-lead screw, 18-lead screw nut, 20-coupling, 19-platform crossbeam, and 17-platform connecting plate.

[0027] like Figure 2 The lateral telescopic mechanism shown includes 32-telescopic crossbeam, 25-rack, 34-cylindrical gear, 42-conical gear, 41-gear fixing frame, and 40-horizontal motion motor.

[0028] Figure 2 The camera pitch structure includes 26-servo mount, 38-servo, 29-camera bracket, and 28-camera.

[0029] like Figure 3 The steering wheel moving mechanism shown includes 43-tensioning sleeve, 45-external tire fixing component, 46-external bearing, 49-inner bearing, 50-inner tire fixing component, 48-tire, 51-clamping plate fixing component, 47-outer clamping plate, 52-inner clamping plate, 63-outer bearing clamping plate, 64-rotary bearing, 61-rotary motor, 53-wheel set motor, 65-inner bearing clamping plate, and 66-right angle connector.

[0030] like Figure 4 The set of independent suspension structures shown includes 68-compression spring damper, 57-suspension slider, and 4-suspension connecting plate.

[0031] In the entire structure, the movement of the inspection robot is accomplished by four sets of independently moving steering wheel mechanisms. For example... Figure 3 The diagram shows the disassembled steering wheel mechanism. The steering wheel mechanism is driven by a rotary motor 61 and a wheel assembly motor 53. The wheel assembly motor 53 is sequentially connected to an inner clamping plate 52, a clamping plate fixing component 51, a tire inner fixing component 50, an inner bearing 49, a tire 48, an outer clamping plate 47, an outer bearing 46, a tire outer fixing component 45, and a tensioning sleeve 43. The rotary motor 61 is sequentially connected to a bearing clamping plate 63 and a rotary bearing 64. The two inner bearing clamping plates 65 clamp the rotary bearing 64 and are bolted to the inner clamping plate 52 and the outer clamping plate 47 via right-angle connectors 66. The rotary motor controls the platform's running direction, and the wheel assembly motor controls the platform's running speed, enabling omnidirectional movement and posture adjustment of the platform.

[0032] In the entire structure, the moving lifting mechanism of the lifting platform completes the task. For example... Figure 1 The lifting mechanism shown has four lifting slide rails 6 that restrict the direction of movement of the lifting platform, and a lifting motor 2 that drives the lead screw to rotate, so as to realize the vertical movement of the lifting platform and thus adapt to plants of different heights.

[0033] In the entire structure, the lateral telescopic mechanism transmits torque through a bevel gear set and a rack. For example... Figure 5 The detailed diagram of the bevel gear set shows that the horizontal motion motor 40 is connected to the central bevel gear and meshes with two identical side bevel gears, enabling the side bevel gears to rotate in opposite directions. The side bevel gears are connected to the cylindrical gears, and the synchronous extension and retraction of the side cameras is achieved through the rack, allowing the inspection platform to adapt to different ridge widths.

[0034] like Figure 6 As shown, the sensor integration module is carefully arranged in key positions on the inspection robot. Temperature and humidity sensor 71 is located at the top, monitoring the temperature and humidity of the crop growth area in real time; carbon dioxide sensor 72 is installed in the middle, responsible for measuring the carbon dioxide concentration in the air; and ultrasonic sensors 24 are distributed at the bottom, used to detect the distance between the robot and obstacles. These sensors, through precise placement, together form an efficient environmental monitoring network. Data from temperature and humidity sensor 71 helps the robot optimize its inspection path, ensuring it operates in the most suitable environment; carbon dioxide sensor 72 provides a basis for optimizing plant growth conditions; and ultrasonic sensors 24 ensure the robot's safe navigation in complex environments. The data collected by these sensors is processed by an embedded processor, supporting the robot's autonomous decision-making, enabling it to adapt to different row widths and maintain efficient and accurate performance in changing environments.

[0035] In the entire structure, the movement of the camera pitch mechanism is controlled by a servo motor, which can drive the camera to rotate at a certain angle in the vertical plane, thereby achieving fine adjustment of the detection angle.

[0036] like Figure 7 The schematic diagram of the robot structure shows six ultrasonic modules symmetrically distributed on the side of the platform, numbered C1, C2, C3, C4, C5, and C6 in a counter-clockwise order. Four sets of steering wheels are mounted on the same circumference, with adjacent sets spaced 90 degrees apart, and numbered D1, D2, D3, and D4 in a counter-clockwise order. When the inspection platform executes a linear movement command, the four steering wheels point in the predetermined direction of movement, and the four rotary motors are linked for precise rotation. When the inspection platform executes a rotational movement command, the four steering wheels are tangent to the same circumference, and the four wheel motors move in the same direction to achieve in-place rotation of the platform.

[0037] like Figure 8 As shown in the diagram, the inspection robot of the present invention operates in a strawberry field, clearly demonstrating its autonomous inspection process. The robot utilizes an ultrasonic sensor (24) mounted on its bottom to detect and maintain a safe distance from the ridge walls, while simultaneously monitoring environmental conditions through a temperature and humidity sensor (71) and a carbon dioxide sensor (72) mounted on its top. In autonomous mode, the robot adjusts itself to adapt to ridges of varying widths via its lateral telescopic mechanism, while the lifting platform structure ensures that the camera can capture images of the strawberry plants from the optimal perspective. A steering wheel movement mechanism gives the robot omnidirectional mobility, including forward, backward, and lateral movement, enabling it to navigate flexibly in complex ridge environments.

[0038] Furthermore, during the autonomous inspection process, the lidar module (78) performs a ring scan of the inter-row environment. The point cloud data is rasterized to generate the center line of the furrow. The system runs the A* sparse search algorithm based on this center line to obtain a smooth driving trajectory; where the evaluation function is: Inside, This represents the actual cumulative cost from the starting node to the current node n. This represents the heuristically estimated cost from node n to the target node, which is the endpoint of the sparsed furrow centerline. To balance path smoothness and computational efficiency, the actual cost is taken as...

[0039] In the formula, The cost has been accumulated for the parent node. The Euclidean distance between the two points is... For the curvature of the current segment, =1.0、 =5.0 is the weight coefficient; the heuristic cost is the distance from the current node to the endpoint. Straight-line distance: Therefore, a minimum-heap priority search is performed among sparse track points to quickly obtain a smooth, collision-free initial trajectory. The lateral offset of the trajectory is dynamically corrected by the ridge width measured in real time by the ultrasonic sensor (24): let the distance measurement values ​​on the left and right sides be... , Then the lateral error is: via PD controller ( =0.8, =0.5, =0.1 s) After generating the correction value, it is superimposed along the trajectory normal to ensure that the robot's centerline is consistent with the centerline of the furrow, thus achieving precise navigation.

[0040] Meanwhile, the camera (28) is triggered at a fixed frequency during the constant speed segment, and the servo motor (38) adjusts the pitch angle in real time according to the plant height to ensure that the center of the field of view is aligned with the strawberry canopy; the acquired images are inferred at the edge end through the YOLOv11n network to output the lesion category and confidence level. Network loss function In the middle, classification loss weights =1.0, confidence loss weight =1.0; The regression branch uses CIoU loss: in For prediction boxes With truth box The intersection and union ratio, The Euclidean distance between the center points of the two frames. The minimum length of the outer diagonal of the two frames. =0.5 is the curvature penalty coefficient. As a measure of aspect ratio consistency, =1.0 is the regression loss weight. The inference confidence threshold is 0.35, the NMS IoU threshold is 0.45, and the output vector contains normalized bounding box parameters x, y, w, h, category c, and confidence s. The recognition results, combined with the coordinate transformation to obtain the true coordinates of the lesions, are stored locally along with the image path and timestamp, and then transmitted back to the cloud via a 5G image transmission module. The end-to-end latency is less than 100 ms, allowing farmers to view the data in real time.

[0041] like Figure 9 and Figure 10 As shown, the lateral telescopic mechanism of the present invention can be precisely adjusted at different extreme positions to adapt to ridges of different widths. Figure 9The diagram shows the mechanism in its minimum limit position, where the telescopic beam 32 is retracted to its shortest length, ensuring the robot can operate flexibly in confined spaces while maintaining structural stability. Figure 10 This diagram shows the mechanism in its maximum limit position, where the telescopic beam 32 is fully extended, enabling the robot to cover a wider ridge for disease detection.

[0042] like Figure 11 The diagram shows the overall control logic of the system. After startup, the system first enters the initialization and self-test process, with each functional module performing status verification in sequence. After initialization, the system offers two operating modes: remote control and automatic. In remote control mode, the operator directly controls the equipment's actions via remote commands. In automatic mode, the system enters an autonomous operation process, scanning the environment with a lidar and constructing a digital map. The embedded processor then performs path planning and generates motion commands to drive the chassis to move along a predetermined trajectory. During movement, the ultrasonic ranging module monitors the relative position to the ridge wall in real time, while a PID control algorithm is used to correct the platform's position and speed in real time, ensuring stable movement along the path. The vision module synchronously acquires crop images during movement, and the embedded processor analyzes the images to extract data on crop physiological status and pest and disease characteristics, providing a basis for agricultural decision-making. After the operation is completed, the platform can receive a return command and autonomously return to the starting position along the original path.

[0043] like Figure 12 The diagram shows the system modules of this intelligent inspection robot. The system uses a microcontroller as its core scheduling hub, with a power module providing stable energy to each unit. The perception layer includes multiple sensor modules: an image acquisition module to obtain visual information about crops, and ultrasonic and infrared tracking modules to detect obstacles and track paths, respectively, forming an environmental perception network. The control layer receives external commands through a remote control module, and the microcontroller integrates sensor data to output decision signals to the motor drive module. The execution layer uses a geared motor for precise movement control, and finally, a display module enables human-machine interaction and status monitoring. All modules are electrically interconnected to form a closed-loop control system, supporting the platform in completing autonomous inspection and intelligent operation tasks in agricultural scenarios.

[0044] Furthermore, such as Figure 13The diagram shows the PID control principle block diagram of the system. Its core is a closed-loop control structure based on multi-sensor feedback: The system first detects the deviation between the current position and the target path through the tracking sensor, and calculates the desired speed and angular velocity commands by the planning algorithm; then it enters the dual-loop PID adjustment - the angle PID controller handles the heading deviation, and the speed PID controller handles the speed deviation. The two outputs work together to perform fine adjustment; the generated control signal drives the left and right motors to perform specific actions. At the same time, the wheel speed sensor monitors the actual speed of the motor in real time and feeds it back to the speed control loop, while the attitude sensor collects dynamic parameters such as vehicle tilt angle and acceleration and feeds them back to the angle control loop, thus forming a precision closed-loop control system with multiple feedback loops, ensuring that the vehicle can quickly and smoothly automatically correct its position and accurately track the predetermined trajectory.

Claims

1. A structure for a strawberry disease inspection robot between rows, characterized in that, The system includes a lifting platform structure, a lateral telescopic mechanism, a camera pitch mechanism, a steering wheel movement mechanism, and an independent suspension structure. The steering wheel movement mechanism is located at the bottom of the robot and is in direct contact with the ground. The independent suspension structure is installed on both sides of the steering wheel movement mechanism, and relative displacement can occur between the two. The lifting platform structure is located in the middle of the robot and is connected to the independent suspension structure by bolts. A lateral telescopic mechanism is installed at its lower part, with the extension direction of the telescopic arms perpendicular to the robot's forward direction. The camera pitch mechanism is installed at the outer ends of the two telescopic arms of the lateral telescopic mechanism, and the detection direction of the two cameras is to the outside of the robot.

2. The structure of the strawberry disease inspection robot between rows according to claim 1, characterized in that, The lifting platform structure includes support columns (5), crossbeams (13), lifting motors (2), lifting slide rails (6), lifting sliders (7), lead screws (16), lead screw nuts (18), couplings (20), platform crossbeams (19), and platform connecting plates (17). The lifting frame consists of four vertical support columns (5) and two horizontal crossbeams (13). The two ends of each crossbeam (13) are connected to the tops of the two support columns (5) respectively, thus forming two parallel lifting frames. The four lifting slide rails (6) are fixed to the sides of the support columns (5), and the lifting slide rails (6) are installed at the lifting position. Inside the lifting frame structure; two lifting sliders (7) are installed on a lifting slide rail (6), and a total of eight lifting sliders (7) are fixed at both ends of two horizontally installed platform beams (19). The top of the screw nut (18) is flush with the top of the platform beam (19) and is clamped and fixed to the platform beam (19) through the platform connecting plate (17); the lifting motor (2) is connected to the screw (16) through the coupling (20) to realize the vertical movement of the lifting platform, so that the multi-functional inspection platform structure has the ability to detect plants at different heights and ensure that the camera (28) can image the strawberry canopy within the best viewing angle range.

3. The structure of the strawberry disease inspection robot between rows according to claim 1, characterized in that, The lateral telescopic mechanism includes a telescopic beam (32), a rack (25), a cylindrical gear (34), a bevel gear (42), a gear fixing frame (41), and a horizontal motion motor (40). The lateral telescopic mechanism includes a set of bevel gears meshing at a 90-degree angle, including an input bevel gear connected to the horizontal motion motor (40) and two output bevel gears meshing with the input bevel gear and whose output shafts are parallel to each other. The two bevel gears (42) on both sides are connected to the cylindrical gear (34). The rack (25) is installed below the telescopic beam (32). The cylindrical gear (34) meshes with the rack (25), thereby realizing that the two telescopic beams (32) move in opposite directions at the same speed, enabling the camera to identify crops with different ridge widths.

4. The structure of the strawberry disease inspection robot between rows according to claim 1, characterized in that, The camera tilt structure is installed at the end of the telescopic beam (32) and includes a servo motor (38), a camera bracket (29), and a camera (28). The camera (28) is connected to the servo motor (38) through the camera bracket (29) to realize the camera posture adjustment, thereby taking frontal shots of strawberry leaves and fruits at different heights and angles, and improving the accuracy of disease identification.

5. The structure of the strawberry disease inspection robot between rows according to claim 1, characterized in that, The steering wheel moving mechanism consists of a moving mechanism and a rotating mechanism; the moving mechanism consists of a tensioning sleeve (43), a tire outer fixing component (45), an outer bearing (46), an inner bearing (49), a tire inner fixing component (50), a tire (48), a clamp fixing component (51), an outer clamp (47), an inner clamp (52), and a wheel assembly motor (53); the tire outer fixing component (45) and the outer bearing (46) are coaxially mounted and have a transition fit, the tire outer fixing component (45), the tire (48), and the tire inner fixing component (50) are fixed together by bolts, the inner bearing (49) and the tire inner fixing component (50) are coaxially mounted and have an interference fit, the inner bearing (49) and the clamp fixing component (51) are coaxially mounted and have a transition fit, and the tensioning sleeve (43) is mounted on the tire outer fixing component. Between the drive shaft of component (45) and wheel motor (53), the motor torque is transmitted to the tire; the rotating mechanism consists of bearing clamp (63), rotating bearing (64), and rotating motor (61); the two outer bearing clamps (63) are placed horizontally to clamp the outer ring of the rotating bearing (64), the two inner bearing clamps (65) are placed horizontally to clamp the inner ring of the rotating bearing (64), the inner bearing clamps (65) are bolted to the inner clamps (52) and outer clamps (47) through right-angle connectors (66), the rotating motor (61) transmits rotational torque and drives the moving mechanism to rotate; the rotation of the steering wheel and the movement of the steering wheel are independent of each other, which can realize the omnidirectional movement of the inspection robot, including forward, backward, lateral, diagonal and in-place turning, to meet the complex path requirements of narrow lanes, U-shaped turns and end-point changes.

6. The structure of the strawberry disease inspection robot between rows according to claim 1, characterized in that, The independent suspension structure includes a compression spring damper (68), a suspension slider (57), a suspension connecting plate (4), and a suspension slide rail (70). The steering wheel moving structure is fixed to the suspension connecting plate (4) through the suspension slider (57) and the suspension slide rail (70). The compression spring damper (68) is fixed between the chassis and the steering wheel. This structure enables the multi-functional inspection platform between ridges to move on rugged terrain and effectively isolates ground impacts, ensuring clear and stable visual imaging.

7. The structure of the strawberry disease inspection robot between rows according to claim 1, characterized in that, It also includes a sensor integration module, including a temperature and humidity sensor (71), a carbon dioxide sensor (72), an ultrasonic sensor (24), and a lidar module (78), all of which are connected to a single-chip microcomputer controller. The lidar module (78) is installed on the top of the lifting platform structure. The robot is equipped with four ultrasonic sensors (24) in the left and right directions and two ultrasonic sensors (24) in the front and rear directions. The temperature and humidity sensor (71) and the carbon dioxide sensor (72) are installed at the front of the vehicle body. The vehicle position can be adjusted according to the environment to detect carbon dioxide concentration and temperature and humidity environmental data in real time.

8. A control method for the structure of a strawberry disease inspection robot between rows, characterized in that, Includes the following steps: Step 1: After the robot is powered on, the system performs self-tests on the lifting motor (2), horizontal motion motor (40), rotary motor (61), wheel motor (53), and the temperature and humidity sensor (71), carbon dioxide sensor (72), ultrasonic sensor (24), and lidar module (78) in the sensor integration module in sequence. If any module reports an abnormality, a red warning light will be lit immediately and a buzzer will be emitted. At the same time, a fault code will be uploaded through the wireless module. After the self-test is passed, the user can switch between remote control and autonomous mode. Step 2: After the robot enters autonomous mode, the lidar module (78) performs a ring scan of the inter-row environment. The point cloud is rasterized to generate the center line of the furrow. The embedded processor runs A* sparse search based on the center line to obtain a smooth trajectory. The lateral offset of the trajectory is dynamically corrected according to the ridge width measured in real time by the ultrasonic sensor (24) to ensure that the robot center line is consistent with the furrow center line. Step 3: During the walking process, the gyroscope and accelerometer continuously output the vehicle body attitude. The outer ring position PID takes the trajectory deviation as input, and the inner ring angular velocity PID takes the gyroscope angular velocity as feedback to jointly calculate the steering wheel rotation angle and wheel group differential speed. The stroke signal of the compression spring damper (68) is read in real time and used to compensate for ground undulations, so that the platform crossbeam (19) remains horizontal and the camera (28) maintains stable imaging. Step 4: The camera (28) is triggered at a fixed frequency during the uniform speed segment, and the servo motor (38) adjusts the pitch angle in real time according to the height of the plant to ensure that the center of the field of view is aligned with the strawberry canopy; the image is inferred at the edge end through the YOLOv11 network, and the lesion category and confidence are output. The results are bound to the current parking position pose and stored locally, and then sent back to the background for farmers to view. Step 5: After the task is completed, the robot records the coordinates of the destination, turns around 180° in place, and travels back to the starting point along the trajectory it came from. During the journey, the lifting motor (2) drives the lead screw (16) to lower the platform beam (19) to the lowest position, thereby lowering the center of gravity and improving driving stability. After reaching the starting point, the system automatically cuts off power and enters sleep mode, waiting for the next task to be triggered.

Images