348 results about "Agricultural robot" patented technology

The invention discloses an operation method of a humanoid agricultural robot based on path planning, which relates to the related technical field of agricultural robots and comprises the steps of multi-sensor data acquisition, data processing and map generation, map updating, target crop image recognition and positioning, global path planning, local path obstacle avoidance and picking action execution. The invention further discloses an operation system of the humanoid agricultural robot based on path planning. The operation system comprises a humanoid robot body, a visual sensor, a laser radar, a depth camera, an inertial measurement unit, an auxiliary sensor, an actuator module, a control system, a communication module and a power supply system. The inventor-shaped robot can recognize the positions and shapes of the tools in advance by using a sensitive visual perception system of the inventor-shaped robot, can skillfully cross over or bypass the tools by virtue of the flexibility of the inventor-shaped robot, can accurately determine the positions and distribution conditions of crops in a three-dimensional space, and generates a comprehensive picking environment map.

The invention relates to a multi-sensor cross-scene dynamic preferential fusion positioning and mapping method, and the method comprises the steps: obtaining the data of a plurality of sensors, and completing the unification of the time-space relation of the data of the plurality of sensors; processing the data, carrying out loopback detection on image key frame data acquired by a camera, constructing to obtain an I MU pre-integration factor, a visual inertial odometer factor, a laser radar odometer factor, a GPS inertial odometer factor, a UWB factor, a GNSS factor and a loopback detection factor, and adding the factors into a factor graph for optimization; a global positioning pose and a map are obtained; and optimizing the multi-sensor data fusion strategy based on a deep fuzzy neural network. According to the multi-sensor cross-scene dynamic preferential fusion positioning and mapping method provided by the invention, high-precision positioning and navigation of an agricultural robot in different scenes are realized through real-time fusion of various sensor data, and the problems of scene dependence and insufficient precision of an existing single sensor scheme are solved.

The invention relates to the technical field of intelligent agricultural robots, and provides a multi-agent collaborative anti-collision picking method based on digital twinning and deep reinforcement learning, which comprises the following steps: constructing a digital twinning model of a picking scene, and generating environmental geometric parameters, agent kinetic parameters and fruit position parameters through three-dimensional point cloud reconstruction; acquiring environment state data in real time and inputting the environment state data into the digital twin model for space-time alignment processing to generate synchronous state data; based on the synchronous data, a collaborative strategy containing a collision avoidance priority matrix, a path planning sequence and a task allocation weight is generated through a deep reinforcement learning network; an action instruction set is generated according to the strategy, and multiple agents are controlled to execute a picking task after virtual-physical space bidirectional verification of the digital twin model. According to the invention, efficient collision avoidance and accurate picking of multiple agents in a dynamic environment can be realized, and the picking efficiency, safety and system robustness are improved.

The invention relates to the field of agricultural robots, and discloses an agricultural spraying robot system with visual navigation, and the system comprises a multi-mode sensing module, a dynamic environment modeling module, a time sequence prediction module, a multi-target path planning module, a spraying control module, and a hardware acceleration unit. The method comprises the steps that a dynamic semantic map is constructed through multi-modal perception, agricultural semantics and an attention mechanism are fused to predict an obstacle trajectory, a safe path is generated in combination with multi-objective optimization, the spraying amount is adjusted based on crop density, and real-time operation is ensured by means of hardware acceleration. According to the method, a dynamic semantic map is constructed through multi-modal perception, agricultural semantics is combined to enhance trajectory prediction, a safe path is generated by utilizing multi-objective optimization, the spraying amount is adaptively adjusted according to the crop density, and real-time control is realized by adopting heterogeneous hardware acceleration; the defects of a traditional agricultural robot in the aspects of navigation precision, obstacle avoidance safety, resource utilization and response speed are overcome, and the intelligent level of complex farmland operation is improved.

The invention relates to the technical field of agricultural robot motion control, in particular to a mechanical arm posture stable adjusting method based on vibration feedback, which comprises the following steps: S1, collecting a vibration signal of an end effector of a mechanical arm in real time; s2, frequency domain analysis and amplitude-phase decomposition are conducted on the vibration signals, and vibration characteristic parameters of the mechanical arm in the X / Y / Z axis direction are extracted; s3, generating a multi-axis vibration vector in combination with the terrain inclination angle data; s4, attitude compensation parameters of all the joints are calculated; s5, a mechanical arm joint angle adjusting instruction is generated according to the posture compensation parameters; and S6, circularly executing the steps S1 to S5 until the vibration amplitude is lower than a preset threshold value. According to the method, the multi-axis vibration characteristics are extracted in real time, the joint compensation angle is calculated through the anti-offset and anti-overturning matrix, rapid and stable attitude control of the mechanical arm in the high-frequency disturbance environment is achieved, and therefore the tail end positioning precision and the overall operation reliability are remarkably improved.

The invention relates to the field of intelligent agricultural robots, in particular to a multi-robot collaborative harvesting method and system for woody oil plants based on swarm intelligence, and provides efficient harvesting of the woody oil plants by constructing a distributed sensing network and a semantic enhanced digital twinborn model, and the system adopts a three-layer decision architecture to formulate a harvesting strategy. The working efficiency is improved through self-adaptive task allocation and collaborative harvesting, multiple robots share experience through federated learning, group collaborative learning is achieved, harvesting performance is optimized, the reliability of the system is improved by 85%, single-point fault risks do not exist, the recognition accuracy is improved by 50% through a multi-modal sensing system, and the system is suitable for popularization and application. And an innovative harvesting solution is provided for the field of intelligent agricultural robots.

The invention relates to a tracking navigation method for positioning a crop center point based on laser radar and image fusion, and belongs to the technical field of navigation, and the method comprises the steps: carrying out the marking of a farmland image after distortion correction, generating a training sample, inputting the training sample into a U-net model, carrying out the training, obtaining a crop segmentation model, and carrying out the classification of crops and weeds; performing feature fusion on the filtered laser radar point cloud data and the farmland image after distortion correction to obtain crop point cloud with semantic information; calculating a crop center point according to the crop point cloud with the semantic information; and providing navigation for the agricultural robot by taking the crop center point as a preview point. The spatial structure information of the target crop is obtained through the laser radar and the camera, the semantic point cloud is generated through feature fusion, the crop center point is dynamically extracted from the point cloud to serve as a real-time preview target, a fixed preview distance mechanism in a tracking algorithm is replaced, centimeter-level precise navigation is achieved, and the precision of the target crop is improved. And the operation precision and the operation safety of the agricultural robot are greatly improved.

The invention provides an agricultural robot path navigation method based on a vision-language model, and the method comprises the steps: adjusting the constraint condition of path planning according to the real-time result of risk level evaluation, and if the risk level exceeds a preset threshold value, preferentially selecting a path for avoiding a high-risk region, and determining an optimized navigation route; a dynamic sensing capability module is driven through a refined operation instruction, environment feedback data in the operation process is continuously collected, whether sudden change occurs or not is judged, and if abnormity is detected, an updating mechanism of path planning adjustment is triggered; according to an updating mechanism of path planning adjustment, recalculating the matching degree of the navigation route and the behavior strategy, obtaining a newest operation execution scheme, and determining a final operation path and a final action sequence; and aiming at the final operation path and action sequence, generating a real-time execution log, recording whole-process data of emotion semantic recognition, operation situation modeling and dynamic sensing ability, and obtaining complete operation state tracking information.

The invention belongs to the technical field of agricultural robots, and discloses a visual perception digital twinborn cotton weeding robot identification obstacle avoidance method, which comprises the following specific steps: step 1, establishing a digital cotton and weed database, cotton field image information and cotton field environment information of real scenes in the same area under the condition of changing illumination in different time periods of the whole day are collected through a data collection system, and timestamps during collection and collected real-time data are transmitted to a cloud server. Through mutual cooperation of the data acquisition system, the cloud server and the virtual simulation platform, a cotton and weed virtual scene is established, through simulation of a cotton field and a weeding robot on the simulation platform, time-varying digital scene construction is perfected, and the time-varying digital scene construction efficiency is improved. The problem that an existing digital scene is not perfect in reduction of the time-varying process of the influence of illumination on the cotton field in the real environment is solved, and therefore the consistency of the virtual scene and the real environment and the weeding accuracy are further improved.

The invention discloses an agricultural scene three-dimensional reconstruction method based on multi-modal sensor fusion, and belongs to the technical field of three-dimensional modeling. The method comprises the following steps: acquiring multi-modal data of a target scene based on a tracked robot, wherein the multi-modal data comprises image data acquired by a binocular camera, point cloud data acquired by a laser radar, time sequence inertial data acquired by an inertial measurement unit and position data acquired by a Beidou positioning system; extracting different modal data features in the multi-modal data, and carrying out heterogeneous fusion on the different modal data features to obtain multi-modal data fusion features; and agricultural scene three-dimensional reconstruction is carried out by using a three-dimensional dynamic construction method based on space-time-semantic joint embedding. According to the invention, the environment scene modeling precision of the agricultural robot under the interference of muddy, raised dust and vibration can be improved.

The invention discloses a cotton field robot fusion navigation path generation method, and relates to the field of agricultural robot positioning navigation, and the method comprises the steps: 1, installing an RTK positioning module on a cotton field robot for position navigation; a deep learning visual camera is installed in front of the cotton field robot and used for recognizing and keeping lane lines, and visual navigation is achieved; a walking device, a turning device and an in-situ turning device are installed on the cotton field robot and used for achieving the control effect of walking, turning and in-situ turning. 2, constructing a visual navigation model based on semantic segmentation, and deploying the visual navigation model to an automatic driving system of the cotton field robot; step 3, dotting and marking land parcel information by the RTK dotting device, and importing the land parcel information into an automatic driving system of the cotton field robot; and step 4, starting the cotton field robot to walk and work in the sown cotton field, generating a fusion navigation path, and synchronously recording RTK position track coordinates. According to the invention, an operation path can be automatically planned on a sown standard cotton field.

The invention relates to an agricultural robot, and discloses an animal husbandry pushing robot autonomous navigation method based on a laser radar, which comprises a microsecond-level multi-sensor time synchronization architecture based on an FPGA (Field Programmable Gate Array); in a navigation layer, seamless switching between the GNSS and the laser SLAM is realized by constructing a virtual fusion observation value by utilizing extended Kalman filtering and a self-adaptive dynamic weight algorithm; in the sensing layer, an RGB-I multi-mode model fused with laser reflectivity and a biological micro-motion detection technology are adopted, and feed and biological obstacles are accurately distinguished; and in the execution layer, flexible self-adaptive material pushing is achieved in combination with material groove edge fitting and force-position hybrid control, and the navigation precision, the operation quality and the safety of the material pushing robot in the all-weather and unstructured pasture environment are effectively improved.

The invention relates to the technical field of agricultural robot navigation, and discloses an agricultural robot autonomous navigation system based on a three-dimensional point cloud model.The agricultural robot autonomous navigation system comprises a three-dimensional point cloud collection module, a data processing center and a navigation control unit, robot path planning and obstacle avoidance are achieved by collecting environment data in real time and constructing a precise three-dimensional map, and the navigation efficiency is improved. The three-dimensional point cloud acquisition module acquires point cloud data and image information in an environment, and the data processing center adopts a multi-source data fusion algorithm, and performs dynamic environment perception and path optimization, so as to improve the robustness and the real-time performance of the navigation system. The system adopts multi-module cooperative control, realizes dynamic self-adaption of global and local paths, improves the flexibility and reliability of robot operation, combines a PID control algorithm with real-time feedback adjustment, accurately controls the advancing track of the robot, and ensures the operation precision and stability.

The invention discloses an agricultural robot control method based on multi-agent cooperation, and the method comprises the following steps: S1, collecting environment data of a farmland region, and carrying out the preprocessing of the environment data; s2, constructing a job task graph, and forming structured task graph data; s3, inputting the operation task graph into the graph neural network model, and performing task initialization on all agricultural robots; s4, sharing the local operation sub-graphs among the agricultural robots, and fusing all the local operation sub-graphs based on a graph neural network model; s5, executing a multi-agent cooperative task division operation, and generating a task allocation matrix; s6, after the agricultural robot receives the operation instruction, executing a path planning and obstacle avoidance algorithm; and S7, summarizing operation states of the agricultural robots in real time, and dynamically updating weight parameters of the graph neural network model according to operation feedback. Efficient cooperation and intelligent scheduling of the agricultural robot are realized, and the operation efficiency and the coverage rate are improved.

The invention relates to the technical field of robot control, and discloses a mechanical arm specified performance fault-tolerant control method based on a fuzzy logic system, and the method comprises the steps: building a kinematics and dynamics model and a fault model of an agricultural robot; establishing a relative error dynamic equation between the actual position and the expected position; designing a specified performance function, and constraining the tracking error within an expected range by using the specified performance function; on the basis, designing a virtual control law by using a backstepping method under the condition of meeting error constraints; designing an adaptive control law by adopting a fuzzy logic system based on an uncertain item, unknown environment interference and a fault model in the kinetic model; a specified performance controller is designed based on a virtual control law, fuzzy logic adaptive control and full-state feedback agricultural robot, and specified performance fault-tolerant control of the agricultural robot is realized. The stability proves that the tracking error of the agricultural robot converges to any small neighborhood near zero within fixed time.

The invention relates to the technical field of fault early warning, in particular to an agricultural robot self-learning fault early warning method based on multi-source data fusion, and the method comprises the steps: obtaining the operation data and environment data of an agricultural robot, and generating a multi-source fusion tensor through time sequence synchronization and normalization processing; constructing a neural network comprising a self-supervised check branch and a minimum decoupling branch, and generating a reconstruction tensor; mapping operation and environment vectors by using an independent encoder, stripping linear and nonlinear environment features through an antagonism prediction path established by an orthogonal penalty term and a gradient inversion layer, and generating operation features; and constructing a feature topological graph, calculating local relative density, and generating a fault early warning signal. According to the method, deep association of multi-source data is mined through self-supervised verification, the problem of false alarm caused by environmental noise in an unstructured environment is effectively solved by forcibly eliminating environmental interference in a feature space by using orthogonal constraint and antagonism training, and the robustness and accuracy of fault early warning are remarkably improved.

The invention discloses an agricultural robot laser strike positioning method and device based on a virtual mapping surface, and the method comprises the following steps: S1, constructing a virtual mapping surface for the projection of a coordinate system of a visual system through a calibration board in a to-be-operated scene, establishing a mapping function between projection coordinates on the virtual mapping surface and instruction coordinates of the laser galvanometer through data fitting; s2, dynamically compensating target coordinates acquired in real time according to a built-in attitude sensor of the visual system; and S3, projecting the target coordinate after dynamic compensation to the virtual mapping surface, converting the target coordinate into a preliminary instruction coordinate of the laser galvanometer according to the mapping function, and performing dynamic compensation on the preliminary instruction coordinate to obtain a final instruction coordinate to control the laser galvanometer to act. The calibration process is simple, the coordinate conversion calculation amount is small, and fluctuation generated in the robot movement process can be dynamically compensated to ensure the positioning accuracy.

The invention belongs to the technical field of agricultural robots, and particularly relates to a multi-degree-of-freedom double-arm pear fruit picking robot based on an agricultural machinery and agricultural technology fusion idea, which comprises a walking chassis, a picking system, a fruit collecting system, a visual system, a control system and an endurance system, the walking chassis is mainly used for bearing mechanical parts and picked fruits, walking in a field and providing power for the whole picking robot, the picking system is carried on a picking robot body and mainly used for grabbing recognized pears capable of being picked, and the fruit collecting system is carried on the picking robot body and used for collecting the fruit collected by the fruit collecting system. The picking robot body is used for conveying the picked pears into the fruit collecting box, and the visual system is carried on the picking robot body. The device can adapt to the picking operation of standardized pear garden planting agriculture in different planting modes, is highly integrated and automatic, can effectively improve the pear fruit picking operation efficiency, greatly reduces the dependence on the labor cost, and achieves the intelligent upgrading of agricultural production.

The invention discloses an inter-ridge autonomous navigation method and system for an asparagus picking robot based on machine vision, and belongs to the technical field of agricultural automation. The method comprises the following steps: processing a field image acquired by a robot in real time by using a pre-trained lightweight inter-ridge path segmentation network to generate a path mask; extracting and smoothing left and right boundary points of the path based on the mask; fitting a middle navigation line; and finally, a control error is calculated according to the navigation line and the robot pose, and a steering instruction is generated to drive the robot to autonomously travel. Through lightweight model design and a robust path fitting algorithm, the problems that an existing visual navigation method is large in calculation amount, poor in real-time performance and insufficient in precision are solved, an efficient, accurate and easy-to-deploy autonomous navigation solution is provided for an agricultural robot, and the method is particularly suitable for complex field environments of ridge crops such as asparagus.

The present invention provides a graphic recognition method and obstacle avoidance system for intelligent agricultural robots, which relates to the technical field of image pattern recognition. By introducing polarization imaging, the present invention effectively suppresses water droplet highlights, avoids misjudging optical artifacts as diseases, and significantly improves recognition accuracy. For low-confidence targets, the original micro-area thermal excitation physical review mechanism can confirm their true physical properties online, solving the problem of recognition uncertainty; in addition, the dynamic risk volume constructed based on multi-dimensional information, combined with the adaptive speed control of path risk foresight, realizes differentiated intelligent obstacle avoidance for rigid and flexible obstacles. While ensuring safety, this method greatly improves the robot's travel efficiency and smoothness in complex environments.

The invention discloses an orchard crawler differential robot trajectory tracking method based on adaptive incremental control, and relates to the technical field of agricultural robot control. The method comprises the following steps: receiving environmental data collected by a robot sensor, constructing a probability grid map of an orchard environment according to the environmental data, and aligning the probability grid map with a pre-constructed map inertial coordinate system; the course angle error and the distance error between the current position and a target point are calculated based on the pose of the robot, an error vector is formed, and the target point is dynamically planned according to the fruit maturity and reachability priority detected by a visual system and a laser radar point cloud. According to the method, coordinate transformation is not needed, and different from traditional tracking control, a control instruction is generated directly based on pose increment, the course and distance error of a target point is calculated directly based on an inertial coordinate system, and real-time matrix projection operation is avoided.

The invention discloses a tomato picking point positioning and feasibility analysis system based on RGB-D attitude estimation, and belongs to the technical field of agricultural robot visual perception, and the system comprises an RGB-D data collection module which is used for synchronously collecting an RGB image and a depth image of a tomato cluster; a data preprocessing module; a multi-scale feature extraction and fusion module; the multi-task prediction module comprises a detection and key point prediction head, a picking point depth regression prediction head and a picking feasibility analysis module which are arranged in parallel; the tomato cluster three-dimensional picking system has the advantages that synchronous output of tomato cluster structure attitude information, picking point three-dimensional positions and picking feasibility judgment is achieved, tomato cluster attitude sensing stability and picking point three-dimensional positioning precision are improved, and the tomato cluster three-dimensional picking system is high in practicability and high in practicability. And the robot is endowed with an intelligent picking decision-making capability, and the operation success rate of the tomato picking robot is improved.

The utility model belongs to the technical field of agricultural robots, and relates to a nursery stock cuttage robot which comprises a base and a linear module which is installed on the base and can move in multiple degrees of freedom of X-axis, Y-axis and Z-axis, a vibration disc is arranged in the middle of the base, a feeder is arranged on one side of the vibration disc, and the output end of the feeder is located above the vibration disc. The feeder is used for storing and providing nursery stocks to be cut for the vibration disc, the other side of the vibration disc is provided with a plurality of hole trays, the hole trays are used for cutting the nursery stocks to be cut according to an array, the bottom of the linear module can be leveled, the linear module is provided with a tail end gripper, and the linear module drives the tail end gripper to move left and right, front and back and up and down. And the rotating hand at the tail end grabs the to-be-cut seedlings in the vibration disc and accurately cuts the seedlings into the hole disc. According to the utility model, the running speed and the running precision of the robot are improved, the gripper at the tail end can quickly and accurately stop at any position, cutting seedlings at all angles can be grabbed and cut into a hole tray, the accuracy of the cutting operation is improved, and the application range is expanded.

The invention discloses an end effector device applied to broccoli picking operation and a harvesting machine thereof, and belongs to the technical field of agricultural robots and vegetable harvesting equipment. Comprising a main sleeve, a square sleeve, a round sleeve, a track plate, a positioning device, a shearing mechanism, an extension spring, a linear guide rail module and a cover plate. And the main sleeve is used for covering and guiding the broccoli balls, so that the broccoli balls enter the picking channel and are matched with a shearing mechanism arranged at the lower part to complete the picking action. The square sleeve is connected with a sliding block on the linear guide rail in a screwed mode through a screw and can move up and down relative to the main sleeve, and a wedge-shaped groove is formed in the square sleeve and used for fixing the track plate. Through the integrated design of sleeve coating and an unpowered shearing structure, the precise picking process without additional motor driving is achieved. Compared with a traditional harvesting device, the broccoli harvesting device has the advantages that selective harvesting of broccoli is achieved through the sleeve structure, the ball-flower damage rate is reduced, automatic shearing without electric drive is achieved, and certain innovativeness is achieved.

The invention discloses a light-weight tea leaf tender shoot detection method and device for tea leaf picking, belongs to the technical field of image recognition and target detection, and aims to solve the problems of sensing and shielding of tender shoot targets for tea leaf picking. Comprising the steps of obtaining a to-be-detected image in a tea picking scene in real time; processing the to-be-detected image based on a pre-trained lightweight tender shoot detection model to obtain position information and a category identification result of the tender shoot small target; the training process of the lightweight tender shoot detection model comprises the following steps: acquiring an image data set containing a tea tender shoot target; performing data division on the image data set containing the tea tender shoot target to obtain a training set; and training a YOLO RGS detection model taking YOLOv8n as a basic network by using the training set to obtain a lightweight tea tender shoot detection model. The method can improve the perception level of tender shoot targets in a tea leaf picking scene, and provides efficient and stable visual support for intelligent equipment such as agricultural robots.

The utility model belongs to the technical field of agricultural robots, and relates to a pneumatic quantitative fertilization, powder fixation and deep fertilization robot, which is characterized in that a conveying component, a deep scarification fertilization component, an air compression spray explosion component and an automatic following component are mounted on a crawler chassis, and are connected through a three-way valve; the material conveying assembly is used for providing fertilizer for the deep scarification and fertilization assembly, the air compression spray explosion assembly is used for providing compressed air for the deep scarification and fertilization assembly, the deep scarification and fertilization assembly is used for conveying air and fertilizer for soil, and the automatic following assembly is used for detecting and identifying people and obstacles in the environment; a discharging motor of the material conveying assembly drives a rotating shaft and a partition plate to rotate, a discharging baffle is opened, fertilizer enters a material storage bin from a discharging bin, and the material storage bin is communicated with the deep scarification and fertilization assembly. Quantitative discharging and fertilizing are achieved, discharging is conducted in a gas pushing mode, fertilizer particles are evenly distributed, and damage caused by low-frequency vibration to equipment is reduced; and the automatic following assembly automatically follows the working personnel to walk.

The invention relates to the technical field of agricultural robots, in particular to an agricultural robot computing task collaboration method and system, and the method comprises the following steps: S1, environment perception and data collection: collecting environment perception data of a farmland operation region in real time; s2, generating a dynamic task priority queue: generating the dynamic task priority queue; s3, heterogeneous computing resource adaptation model construction: constructing a heterogeneous computing resource adaptation model, and outputting a task-resource matching matrix; s4, generating an initial task allocation scheme: generating the initial task allocation scheme; s5, environment interference monitoring and compensation: adjusting the initial task allocation scheme; and S6, task execution and queue updating: executing the adjusted initial task allocation scheme, and continuously updating the dynamic task priority queue. According to the method, the working efficiency can be remarkably improved while the working precision is ensured, the working time is shortened, the resource allocation is optimized, and the manual intervention is reduced.

The invention relates to the technical field of agricultural robots, and discloses a shearing mechanism and a fruit picking end effector, the shearing mechanism comprises a mounting seat, a bionic structure, a shear assembly and a first driving assembly; the bionic structure comprises a plurality of bionic fingers which are mounted on the mounting seat at intervals, and connecting brackets are mounted on the bionic fingers; the scissor assemblies are installed between every two adjacent bionic fingers, each scissor assembly comprises a fixed blade, a movable blade and an elastic piece, the fixed blades are fixed to the connecting support, the middles of the movable blades are hinged to the middles of the fixed blades, and the elastic pieces are installed between the fixed blades and the movable blades; the first driving assembly is used for driving the movable blade to rotate relative to the fixed blade. According to the fruit picking device, the bionic structure is matched with the active shearing mode, the fruit stem can be easily sheared, and therefore the fruit picking efficiency is improved.

The invention relates to an agricultural inspection robot navigation decision planning system and a planning method thereof, and belongs to the technical field of agricultural robots, and the agricultural inspection robot navigation decision planning system comprises a multi-sensor fusion sensing module, a navigation decision planning module, a dynamic obstacle processing module, a transverse-longitudinal combined controller and a task management platform. The multi-sensor fusion sensing module comprises a laser radar, a visual sensor, a Beidou / GPS positioning module and an IMU and is used for collecting environment images, point cloud data and robot poses and recognizing crop row navigation lines through an improved YOLOX-Tiny model. According to the method, high-precision navigation can be achieved, high-precision tracking with the transverse error being smaller than or equal to 0.04 m and the longitudinal error being smaller than or equal to 0.06 m is achieved through transverse-longitudinal combined control and multi-sensor fusion positioning, the self-adaptive capacity is high, the reinforcement learning and path decomposition technology and a dynamic risk evaluation mechanism are combined, and the reliability of the system is improved. The robot can effectively deal with sudden obstacles and complex terrains in a farmland environment.

The invention relates to an agricultural robot comprising a straddle-type chassis, equipped with a plurality of rows of rolling elements parallel to a main axis of movement and allowing the agricultural robot to move on ground, the chassis having at the center a ground clearance greater than a predetermined height. The agricultural robot also comprises a tool holder having at least one arm carrying at least one tool, the arm pivoting about an axis while being connected by a cable to a winch fixed to the chassis, the height of the tool relative to the chassis being adjusted by said winch, the arm being configured to be fully raised above the ground clearance in a so-called high position. Figure for abstract: figure 1