A method and system for non-continuous spot cleaning based on discrete topological nodes

Abstract

The application discloses a kind of discontinuity fixed-point cleaning method and system based on discrete topological node, belong to industrial robot technical field, its method includes: constructing environment map, marking out work anchor point on environment map, each work anchor point binds an attribute vector;Global path planning is executed, and each work anchor point is cleaned along the planned path;When moving along the planned path, ultrasonic detection is executed and water pump circuit remains disconnected;Robot calculates current pose in real time by itself sensor, and calculates the space-time state deviation between current pose and current target work anchor point preset pose;Robot is checked by double threshold;When ultrasonic detection is qualified and double threshold check passes, trigger to position locking signal;After receiving to position locking signal, prepare cleaning injection scheme according to attribute vector;Make water pump circuit remain closed and execute cleaning injection scheme, execute after completion and release lock and switch to next target work anchor point.

Description

Technical Field

[0001] This invention belongs to the field of industrial robot technology, and in particular relates to a non-continuous fixed-point cleaning method and system based on discrete topological nodes. Background Technology

[0002] Most existing industrial cleaning robots employ a continuous spraying operation mode along a path. However, in building material storage warehouses, equipment operation areas, or electrical equipment rooms, the ground environment is complex, often containing water-sensitive precision equipment, easily corroded metal coils, or breathable packaged goods.

[0003] Existing continuous operation modes cannot flexibly switch between the "cleaning area" and other areas, including areas for precision instruments. If the robot continuously sprays water, the splashing water mist can easily cause corrosion or other damage to equipment such as machine tools, or even electrical short circuits. Furthermore, existing robots are prone to deviation from the spraying area in dusty environments due to positioning drift. Therefore, there is an urgent need for an intelligent control method that can precisely control the water spraying area and strictly distinguish between "driving mode" and "operating mode." Summary of the Invention

[0004] The purpose of this invention is to provide a non-continuous fixed-point cleaning method and system based on discrete topology nodes, thereby solving the problems existing in the prior art.

[0005] To achieve the above objectives, this invention provides a non-continuous fixed-point cleaning method based on discrete topology nodes, comprising the following steps: Step S1: Construct an environment map, mark each job anchor point on the environment map, and bind an attribute vector to each job anchor point; Set the area in the environmental map, excluding the work anchor point, as a no-spray zone; Step S2: Obtain the task requirements and the location information of the work anchor points, perform global path planning, and clean each work anchor point in sequence along the planned path; As the robot moves along the planned path toward the target work anchor point, ultrasonic detection is performed while the water pump circuit remains physically disconnected. Step S3: When the robot moves along the planned path, the robot calculates the current pose in real time through its own sensors, and calculates the spatiotemporal state deviation between the current pose and the preset pose of the current target work anchor point, and generates a pose state deviation vector. Step S4: Based on the pose state deviation vector, perform dual threshold verification on the robot's position deviation and attitude deviation respectively; When the ultrasonic test is qualified and the dual threshold verification is passed, the position locking signal is triggered; Step S5: After receiving the positioning lock signal, the system triggers the positioning lock state, the chassis locks up, and reads the attribute vector of the current target work anchor point, and compiles a cleaning spraying plan based on the attribute vector; Keep the water pump circuit closed and execute the cleaning spray program. Once completed, unlock and switch to the next target work anchor point.

[0006] Preferably, an environment map is constructed, and each task anchor point is marked on the environment map. Each task anchor point is bound to an attribute vector including: Step S11: Start the lidar to scan the site environment. The industrial control computer runs the SLAM mapping algorithm to build a high-precision environmental grid map covering the entire area. Step S12: The operator views the constructed grid map on the interactive interface of the industrial control computer and selects and marks all the work anchor points in sequence in the grid map interface. In step S13, the industrial control computer automatically records the spatial coordinates and preset cleaning parameters of each work anchor point, and integrates the spatial coordinates and preset cleaning parameters of the work anchor points into an attribute vector.

[0007] Preferably, the steps of performing global path planning and cleaning each work anchor point sequentially along the planned path include: Step S21: Generate a discrete task topology graph based on the task anchor points; In step S22, the industrial control computer plans a global navigation path connecting each work anchor point based on the discrete task topology map, and generates a real-time obstacle avoidance path using a local dynamic window algorithm. Step S23: The industrial control computer sends the path data to the STM32 main control board via serial port as navigation instructions, and performs cleaning according to the navigation instructions; As the robot moves along the path, the STM32 main control board executes navigation-mode physical fuse logic.

[0008] Preferably, executing the navigation-state physical circuit breaker logic specifically includes: Step S231: Set the water pump operation enable to control the circuit opening and closing state of the water pump; Step S232, set the pump operation enable at time t to Its expression is: ; in, This is the positioning and locking signal of the target operation anchor point at time t. ; The host computer operation request at time t. ; When no position lock signal is received 0, regardless Any value is possible. 0, at this time the water pump circuit remains disconnected; when the positioning lock signal is received... 1. At this point, the host computer's operation request is determined, and then the on / off status of the water pump circuit is determined.

[0009] Preferably, the current pose is calculated in real time, and the spatiotemporal state deviation between the current pose and the preset pose of the current target work anchor point is calculated, generating a pose state deviation vector including: Step S31: Extract the preset pose of the current target work anchor point. And calculate the robot's current pose. ; Step S32: Calculate the pose state deviation vector at time t of the i-th target work anchor point. The calculation formula is: ; in, The preset center X coordinate of the i-th target operation anchor point; The preset center Y coordinate of the i-th target operation anchor point; This refers to the robot's heading angle deviation, i.e., the robot's heading angle at time t. Preset optimal water spray direction for the i-th target work anchor point The absolute value of the difference between them; Let t be the X coordinate of the robot at the current time t; Let t be the Y-coordinate of the robot at the current time t.

[0010] Preferably, the specific steps for performing dual threshold checks on the robot's position deviation and attitude deviation are as follows: Step S41: Perform distance determination and set the distance threshold to a preset value. Calculate the Euclidean distance between the robot's current position and the center coordinates of the target work anchor point. The calculation formula is: ; like ≥ If the distance condition is not met, the verification fails; ≤ If the distance condition is met, then the attitude determination is performed. Step S42: Perform attitude determination and set a preset attitude threshold. Calculate the absolute value of the angular deviation between the robot's current heading angle and the preset optimal water spray direction; ; like ≥ If the attitude conditions are not met, the verification fails; ≤ If the attitude condition is met, the double threshold check passes.

[0011] Preferably, the step of reading the attribute vector of the current target operation anchor point and compiling a cleaning spraying plan based on the attribute vector includes: Step S51: After the robot arrives at the target work anchor point and receives the position locking signal, it uses the lidar to scan the point cloud data of the current target work anchor point, extracts the reflection intensity feature value of the ground at the anchor point, and compares the reflection intensity feature value with the preset cleanliness benchmark value to determine the stain level of the current anchor point. In step S52, the industrial control computer calls the preset parameter mapping model according to the stain level and automatically matches and generates the initial cleaning spraying plan for the anchor point. The initial cleaning spray program includes the set values ​​for basic spray duration, number of pulse cycles, and water pump pressure; Step S53: When executing the initial cleaning spraying plan, the ultrasonic detection array collects the ground echo signal after water spraying in real time, and the industrial control computer analyzes the acoustic characteristics of the echo signal to obtain real-time cleanliness feedback data. In step S54, the industrial control computer compares the real-time cleanliness feedback data with the preset cleanliness threshold in real time. If the feedback data indicates that the area has been cleaned, the dynamic cutoff logic is executed to terminate the spraying in advance and release the lock. If the target is not met, the spraying time is dynamically extended or the number of pulse cycles is increased until the cleanliness target is met.

[0012] Preferably, step S6 is autonomous resupply: Step S61: The system monitors the battery level in real time. When the battery level is lower than the preset battery level threshold, the current cleaning task is paused. In step S62, the industrial control computer plans the optimal path back to the contactless charging station, and the robot returns along the path. Step S63: When the distance between the receiving coil and the transmitting coil is not greater than the preset distance, start contactless charging and cut off the motor drive to enter low power standby mode.

[0013] Preferably, ultrasonic testing includes: Step S201: The STM32 main control board independently reads the distance data of obstacles around the vehicle body fed back by the ultrasonic detection array; Step S202: Compare the read distance to the nearest obstacle with the preset safe distance threshold to determine whether the current environment meets the safe working conditions; Step S203: If the distance to the nearest obstacle is determined to be less than the preset safe distance threshold, the ultrasonic detection is deemed unqualified. The STM32 main control board directly triggers the bottom-level emergency stop, forcibly disconnects the motor and water pump circuits, and immediately terminates the water spraying operation. Step S204: When the STM32 main control board detects that the distance to the nearest obstacle is greater than the preset safe distance threshold, it determines that the ultrasonic detection is qualified.

[0014] The present invention also provides a discontinuous fixed-point cleaning system based on discrete topology nodes, comprising: The multi-source environmental perception module includes a lidar mounted on the top and an ultrasonic detection array around the vehicle body; LiDAR is used to construct an environmental grid map containing operational anchor points and provide positioning data, while ultrasonic detection arrays are used to scan obstacles in the blind zone of LiDAR in real time. The industrial control computer is connected to the multi-source environmental perception module. It is used to run the particle filter-based positioning algorithm and task status, navigation algorithm, and calculate the robot pose and output operation control commands according to the spatiotemporal coupling constraint logic. The STM32 main control board connects to the industrial computer via a serial port. It is used to parse navigation commands, independently read ultrasonic feedback data, and execute the underlying safety fuse logic based on the nearest neighbor obstacle distance. The work execution mechanism is used to perform pulsed water spraying operations when the system is in the locked position and there are no obstacles within the ultrasonic detection range; The energy supply module is used to detect the system's energy reserve usage, provide continuous power to the system, and guide autonomous recharging.

[0015] Therefore, the present invention employs the above-described discontinuous fixed-point cleaning method and system based on discrete topology nodes, which has the following beneficial effects: 1. Discontinuous operation mode based on discrete topology nodes. The system divides the environment into two attributes: "operation anchor points" and "navigation paths". When the robot passes through non-operation areas such as precision instruments, power distribution boxes, or permeable goods along the path, the system is in "navigation mode", and the underlying controller forcibly locks the water pump circuit to an open circuit state. This mechanism does not rely on complex visual recognition algorithms to determine the position of instruments, but instead uses a "whitelist" mechanism (i.e., no water spraying if not at the anchor point), fundamentally eliminating water spraying accidents caused by software misjudgment or drift when passing through precision instruments, greatly improving the system's safety in complex industrial environments composed of expensive equipment.

[0016] 2. A dual threshold determination mechanism of "position + attitude" based on lidar positioning is introduced. Compared with the crude mode of traditional robots that only care about position and not direction, this invention ensures that the nozzle is strictly aligned with the center of the stain, avoiding ineffective spraying caused by the tilt of the vehicle body.

[0017] 3. It combines the global perception of LiDAR with the low-level fracturing of ultrasound. By using an ultrasonic detection array to fill the near-ground blind spot of LiDAR, it can effectively identify and avoid low obstacles such as steel coils, wooden blocks, and forklift pallets, significantly improving the robot's robustness in complex dusty environments.

[0018] 4. The fixed-point pulse cleaning strategy significantly reduces water consumption and energy consumption compared to full-coverage blind cleaning, and extends the runtime of a single operation.

[0019] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0020] Figure 1 This is a flowchart of a non-continuous fixed-point cleaning method based on discrete topology nodes according to an embodiment of the present invention; Figure 2 This is an architectural block diagram of a non-continuous fixed-point cleaning system based on discrete topology nodes, according to an embodiment of the present invention. Detailed Implementation

[0021] The following detailed description of embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0022] Example 1 This embodiment uses a fixed-point cleaning operation in a building materials storage warehouse as an example. When the robot moves along the planned path towards the target work anchor point, the high-pressure water pump remains off, and the ultrasonic detector continuously detects low obstacles within a 30cm radius around the vehicle. When an obstacle is detected to be less than 20cm away, the STM32 main control board immediately triggers an emergency stop and cuts off the motor and water pump control circuits. When the robot reaches the target work anchor point and simultaneously meets the position deviation threshold and attitude deviation threshold, the system enters a locked position state, the chassis stops moving, and the high-pressure water pump performs pulse-style fixed-point cleaning according to a sequence of "3s spraying—1s pause—3s spraying again." After completing a single-point operation, the system unlocks and proceeds to the next work anchor point. When the battery level is below 20%, the robot stops performing the cleaning task and automatically returns to the contactless charging station; when the distance between the receiving coil and the transmitting coil is no greater than 3cm, contactless charging is initiated.

[0023] Example 2 Please see Figure 1 A non-continuous fixed-point cleaning method based on discrete topology nodes includes the following steps: Step S1: Construct an environment map, mark each operation anchor point on the environment map, and bind an attribute vector to each operation anchor point; set the area in the environment map other than the operation anchor points as a no-spray zone.

[0024] The environmental map is a high-precision two-dimensional environmental grid map covering the entire operating area. The map consists of multiple grid cells representing obstacles, free space, and reflectivity.

[0025] Among them, the work anchor point is the specific spatial coordinate point of the center of the area to be cleaned, which is manually marked by the operator on a high-precision environmental grid map. This point has a unique identifier, and the system defaults to all areas other than the anchor point being no-spray zones.

[0026] The attribute vector is a data structure bound to the work anchor point. It is an array composed of the spatial coordinates (X coordinate, Y coordinate and preset optimal water spray direction angle) of the anchor point and preset cleaning parameters (basic spray duration, number of pulse cycles and water pump pressure setting value).

[0027] Step S11: Start the lidar to scan the field environment. The industrial control computer runs the SLAM mapping algorithm to build a high-precision environmental grid map covering the entire area. Step S12: The operator views the constructed grid map on the interactive interface of the industrial control computer, and selects and marks all the work anchor points in the grid map interface in sequence. Step S13: The industrial control computer automatically records the spatial coordinates and preset cleaning parameters of each work anchor point, and integrates the spatial coordinates and preset cleaning parameters of the work anchor points into an attribute vector.

[0028] Among them, the high-precision environmental grid map is an environmental model represented by a two-dimensional grid data structure. Each grid cell records the probability of obstacle occupancy, reflectivity and passability, providing reference information for navigation and positioning.

[0029] The spatial coordinates of the work anchor point are the X-axis coordinates and Y-axis coordinates of the anchor point in the global coordinate system, as well as the preset optimal water spraying direction angle of the anchor point, which constitute a three-dimensional pose vector.

[0030] The preset cleaning parameters are: a set of control parameters predefined by the user, including the basic spray duration (e.g., 3 seconds), the number of pulse cycles (e.g., 3 times), the water pump pressure setting (e.g., 5MPa), and the spray interval time (e.g., 1 second).

[0031] Specifically, integrating the spatial coordinates of the work anchor point with the preset cleaning parameters into an attribute vector involves concatenating the aforementioned spatial coordinates and cleaning parameters (duration, pressure, number of times) into a one-dimensional array or structure according to a preset field order, and storing it in the node data of the task topology graph.

[0032] Step S2: Obtain the task requirements and the location information of the work anchor points, perform global path planning, and clean each work anchor point in sequence along the planned path; when the robot moves towards the target work anchor point along the planned path, perform ultrasonic detection and keep the water pump circuit physically disconnected.

[0033] Performing ultrasonic testing includes: Step S201: The STM32 main control board independently reads the distance data of obstacles around the vehicle body fed back by the ultrasonic detection array; Step S202: Compare the read distance to the nearest obstacle with the preset safe distance threshold to determine whether the current environment meets the safe working conditions; Step S203: If the distance to the nearest obstacle is determined to be less than the preset safe distance threshold, the ultrasonic detection is deemed unqualified. The STM32 main control board directly triggers the bottom-level emergency stop, forcibly disconnects the motor and water pump circuits, and immediately terminates the water spraying operation. Step S204: When the STM32 main control board detects that the distance to the nearest obstacle is greater than the preset safe distance threshold, it determines that the ultrasonic detection is qualified.

[0034] The preset safe distance threshold is the alarm value of the distance to obstacles around the vehicle detected by the ultrasonic detection array, preferably 20 centimeters.

[0035] Among them, the safe operating condition is that the ultrasonic testing is qualified.

[0036] The task requirements are the cleaning scheduling instructions issued by the user and received by the system, including the list of job anchors that need to be cleaned in sequence, the cleaning priority, and the mode of task execution.

[0037] Global path planning involves using the A* pathfinding algorithm or Dijkstra's algorithm on the environmental grid map to calculate the globally optimal navigation path from the current robot position to the next target task anchor point, while ensuring no collisions.

[0038] The process of executing global path planning and cleaning each work anchor point sequentially along the planned path includes: Step S21: Generate a discrete task topology graph based on the task anchor points; Step S22: The industrial control computer plans a global navigation path connecting each work anchor point based on the discrete task topology map, and generates a real-time obstacle avoidance path using a local dynamic window algorithm. Step S23: The industrial control computer sends the path data to the STM32 main control board via serial port as navigation instructions, and performs cleaning according to the navigation instructions; As the robot moves along the path, the STM32 main control board executes navigation-mode physical fuse logic.

[0039] Specifically, forming a discrete task topology graph based on task anchor points involves discretizing the environment into a graph model, using task anchor points as "nodes" and the planned paths connecting each anchor point as "edges" to construct a directed or undirected topology graph structure to guide the robot's movement sequence.

[0040] Specifically, the real-time obstacle avoidance path generation using the local dynamic window algorithm involves the industrial control computer using the DWA (Dynamic Window) algorithm to calculate the optimal speed command within a local range while the robot moves along the global path, taking into account robot speed limits, acceleration limits, and real-time sensor (LiDAR) data. .

[0041] Among them, the navigation command sent to the STM32 main control board is a speed command frame containing linear velocity and angular velocity information sent by the industrial control computer to the underlying STM32 main control board through the serial communication protocol, which is then converted into PWM control signals to drive the left and right track motors.

[0042] Step S231: Set the water pump operation enable to control the circuit opening and closing state of the water pump; Step S232: Set the pump operation enable at time t to... Its expression is: ; in, This is the positioning and locking signal of the target operation anchor point at time t. ; The host computer operation request at time t. ; When no position lock signal is received 0, regardless Any value is possible. 0, at this time the water pump circuit remains disconnected; when the positioning lock signal is received... 1. At this point, the host computer's operation request is determined, and then the on / off status of the water pump circuit is determined.

[0043] Among them, the pump operation enable is a logic variable used by the underlying execution module (STM32) to control the on / off state of the high-pressure water pump circuit. It is typically used... This indicates that the value is 0 (circuit open, spraying prohibited) or 1 (circuit closed, spraying allowed).

[0044] Among them, the positioning lock signal is triggered when the robot simultaneously meets the preset position deviation threshold ( ≤ ) and preset attitude deviation threshold ( ≤ The logic signal sent by the industrial control computer, when the value is 1, indicates that the lock is successful and the water pump is allowed to start.

[0045] Among them, the host computer operation request is the water spray control command signal sent by the industrial control computer to the underlying STM32 main control board after determining that the robot has reached and locked in place. It is typically used... This means that a value of 1 indicates a request to spray water, and a value of 0 indicates a request to stop.

[0046] Step S3: When the robot moves along the planned path, the robot calculates the current pose in real time through its own sensors, and calculates the spatiotemporal state deviation between the current pose and the preset pose of the current target work anchor point, and generates a pose state deviation vector.

[0047] The calculation of the current pose involves the industrial control computer using the AMCL (Adaptive Monte Carlo Localization) algorithm, combining LiDAR scan data with a grid map to estimate the robot's current global coordinates in real time. ) and heading angle .

[0048] Step S31: Extract the preset pose of the current target work anchor point. And calculate the robot's current pose. ; Step S32: Calculate the pose state deviation vector at time t of the i-th target work anchor point. The calculation formula is: ; in, The preset center X coordinate of the i-th target operation anchor point; The preset center Y coordinate of the i-th target operation anchor point; This refers to the robot's heading angle deviation, i.e., the robot's heading angle at time t. Preset optimal water spray direction for the i-th target work anchor point The absolute value of the difference between them; Let t be the X coordinate of the robot at the current time t; Let t be the Y-coordinate of the robot at the current time t.

[0049] The preset center of the work anchor point is the ideal position and ideal water spray direction set by the operator when marking the anchor point.

[0050] Step S4: Based on the pose state deviation vector, perform dual threshold checks on the robot's position deviation and attitude deviation respectively; when the ultrasonic detection is qualified and the dual threshold checks are passed, trigger the position locking signal.

[0051] Step S41: Perform distance determination and set the distance threshold to a preset value. Calculate the Euclidean distance between the robot's current position and the center coordinates of the target work anchor point. The calculation formula is: like ≥ If the distance condition is not met, the verification fails; ≤ If the distance condition is met, then the attitude determination is performed. Step S42: Perform attitude determination and set the preset attitude threshold as follows. Calculate the absolute value of the angular deviation between the robot's current heading angle and the preset optimal water spray direction; ; like ≥ If the attitude conditions are not met, the verification fails; ≤ If the attitude condition is met, the double threshold check passes. Step S43: After the dual threshold verification passes, a position lock signal is generated; otherwise, the system remains prohibited from operation.

[0052] The preset attitude threshold is the maximum allowable angular deviation between the robot's current heading angle and the preset optimal water spray direction, preferably ±5 degrees.

[0053] The preset distance threshold is the maximum allowable Euclidean distance between the robot's current position and the center coordinates of the target anchor point, as set by the system, preferably 10 centimeters.

[0054] Step S5: After receiving the positioning lock signal, the system triggers the positioning lock state, the chassis locks up, and reads the attribute vector of the current target work anchor point, and compiles a cleaning spraying plan based on the attribute vector; Keep the water pump circuit closed and execute the cleaning spray program. Once completed, unlock and switch to the next target work anchor point.

[0055] Step S51: After the robot arrives at the target work anchor point and receives the position locking signal, it uses the lidar to scan the point cloud data of the current target work anchor point, extracts the reflection intensity feature value of the ground at the anchor point, and compares the reflection intensity feature value with the preset cleanliness benchmark value to determine the stain level of the current anchor point. Step S52: The industrial control computer calls the preset parameter mapping model according to the stain level and automatically matches and generates the initial cleaning spray plan for the anchor point. The initial cleaning spray program includes the set values ​​for basic spray duration, number of pulse cycles, and water pump pressure; Step S53: When executing the initial cleaning spraying plan, the ultrasonic detection array collects the ground echo signal after water spraying in real time, and the industrial control computer analyzes the acoustic characteristics of the echo signal to obtain real-time cleanliness feedback data. Step S54: The industrial control computer compares the real-time cleanliness feedback data with the preset cleanliness threshold in real time. If the feedback data indicates that the area has been cleaned, the dynamic cutoff logic is executed to terminate the spraying in advance and release the lock. If the standard is not met, the spraying time is dynamically extended or the number of pulse cycles is increased until the cleanliness standard is met.

[0056] Specifically, extracting the reflection intensity feature value of the ground at the anchor point involves using a lidar mounted on top of the robot to perform point cloud scanning on the ground at the anchor point, reading the laser reflection intensity signal of each data point in the ground grid, and calculating the mean and variance of the reflection intensity in that area as the basis for stain assessment.

[0057] The preset cleanliness benchmark value is a standard reflection intensity value pre-set for the original ground material of the site (such as cement or epoxy flooring), preferably the average reflection intensity (85% reflectivity) when the ground is clean.

[0058] The method for determining the stain level of the current anchor point involves comparing the extracted real-time reflection intensity feature value with the preset cleanliness benchmark value, and classifying the stain into heavy (difference > 40%), moderate (difference 20%-40%), and light (difference < 20%) levels based on the difference.

[0059] Among them, the preset parameter mapping model automatically matches and generates the initial cleaning spray scheme for the anchor point; the pre-trained lookup table or linear regression model automatically matches and outputs the corresponding basic spray duration (e.g., 10 seconds for heavy stains, 5 seconds for moderate stains), pulse cycle count (e.g., 3 times, 2 times) and water pump pressure setting value (e.g., 8MPa, 5MPa) based on the detected stain level.

[0060] Specifically, the industrial control computer analyzes the acoustic characteristics of the echo signal to obtain real-time cleanliness feedback data. The industrial control computer receives the echo signal of water hitting the ground collected by the ultrasonic detection array, analyzes its signal amplitude attenuation rate, frequency spectrum characteristics and return waveform distortion rate, and judges whether the cleaning is thorough by comparing the acoustic differences between the dirty ground and the clean ground.

[0061] Step S61: The system monitors the battery level in real time. When the battery level is lower than the preset threshold, the current cleaning task is paused. Step S62: The industrial control computer plans the optimal path back to the contactless charging station, and the robot returns along the path; Step S63: When the distance between the receiving coil and the transmitting coil is not greater than the preset distance, start contactless charging and cut off the motor drive to enter low power standby mode.

[0062] The preset power threshold is the percentage of remaining battery power set by the system, preferably 20%.

[0063] Specifically, the optimal path for returning to the contactless charging station is calculated by the industrial control computer using the A* algorithm on the constructed grid map based on the current robot position and the fixed coordinates of the charging station, and obstacle avoidance is performed in real time along the way to achieve automatic return.

[0064] The preset distance is the physical distance required for the robot to dock with the contactless charging station, preferably 3 centimeters.

[0065] The multi-source environmental perception module includes a lidar mounted on the top and an ultrasonic detection array around the vehicle body; LiDAR is used to construct an environmental grid map containing operational anchor points and provide positioning data, while ultrasonic detection arrays are used to scan obstacles in the blind zone of LiDAR in real time. The industrial control computer is connected to the multi-source environmental perception module. It is used to run the particle filter-based positioning algorithm and task status, navigation algorithm, and calculate the robot pose and output operation control commands according to the spatiotemporal coupling constraint logic. The STM32 main control board connects to the industrial computer via a serial port. It is used to parse navigation commands, independently read ultrasonic feedback data, and execute the underlying safety fuse logic based on the nearest neighbor obstacle distance. The work execution mechanism is used to perform pulsed water spraying operations when the system is in the locked position and there are no obstacles within the ultrasonic detection range; The energy supply module is used to detect the system's energy reserve usage, provide continuous power to the system, and guide autonomous recharging.

[0066] The underlying safety circuit breaker refers to the STM32 main control board bypassing the industrial control computer and independently reading the distance data of obstacles around the vehicle body fed back by the ultrasonic detection array. When the nearest obstacle is determined to be less than a preset safety threshold, an emergency stop is directly triggered, forcibly cutting off the power circuit between the motor and the water pump and locking the chassis motor, immediately terminating the water spraying operation. When the obstacle distance returns to above the safety threshold, the emergency stop is released and the safety status is re-reported, awaiting instructions from the system.

[0067] Discrete operation control process such as Figure 1 As shown, the control method in this embodiment does not follow the traditional continuous operation mode, but instead executes the following discretized logic: Step S1: Environmental Semantic Modeling and Topology Definition. After the robot is powered on, the STM32 main control board completes a self-test. An environmental grid map is constructed using LiDAR, and operators or algorithms mark "work anchor points" (i.e., the center of the area to be cleaned) on the map. Each work anchor point is bound to an attribute vector, defining the water spray duration at that point.

[0068] Step S2: Safe Path Planning (Navigation State) The industrial control computer performs global path planning (using the A* algorithm) on the existing map according to task requirements, and uses the DWA algorithm for local dynamic obstacle avoidance. Forced safety logic: As the robot moves along the path towards the target anchor point, the high-pressure water gun circuit is physically disconnected. At this time, the ultrasonic detection array continuously monitors a 30cm radius around the vehicle; once an obstacle is detected at a distance less than the safety threshold (20cm), the STM32 immediately triggers a low-level emergency stop, achieving a "safety circuit breaker".

[0069] Step S3: Spatiotemporal Coupling Constraint Determination (Determination State) As the robot approaches the target anchor point, the industrial control computer calculates in real-time the deviation between the robot's current pose and the preset pose of the anchor point. The system performs a "dual threshold" check, including a preset position threshold and a preset attitude threshold: Position check: Euclidean distance is less than the threshold; Attitude check: Heading angle deviation is less than the threshold. Only when both position and attitude simultaneously meet the thresholds does the industrial control computer issue a "position lock" command; otherwise, it controls the chassis to perform fine adjustments or in-situ rotation corrections.

[0070] Step S4: Fixed-point pulse operation (operating state) After receiving the locking command and confirming safety via ultrasonic waves, the STM32 controller locks the chassis motor (to prevent recoil displacement) and adjusts the gimbal angle to the preset value. It then activates the high-pressure water pump to perform pulse cleaning (following preset logic, e.g., 3s high-pressure spray → 1s pause → 3s spray again). This mode effectively removes stubborn oil stains from the ground using the water hammer effect, while avoiding excessive water accumulation that could damage surrounding equipment.

[0071] Step S5: After the task cycle and single-point supply operation are completed, the system releases the lock and switches back to navigation mode to proceed to the next anchor point.

[0072] Automatic recharging logic: The system monitors the battery voltage in real time. When the battery level drops below a set threshold (e.g., 20%), it automatically plans a path to return to the contactless charging station. Within a certain range of the wireless charging station, the robot can additionally use infrared guidance to reverse or approach from the side. When the distance between the receiving coil and the transmitting coil is ≤3cm, contactless charging is initiated, and the motor drive is cut off to enter a low-power standby mode.

[0073] Hardware sensing and control architecture such as Figure 2 As shown, this embodiment proposes a discontinuous fixed-point cleaning system based on discrete topology nodes. Its hardware architecture adopts a hierarchical control design, specifically including: Perception layer: LiDAR: Mounted on top of the robot, it scans the surrounding environment at frequencies above 10Hz. It is used not only to build high-precision grid maps, but also to provide real-time feedback on the robot's global pose, providing a spatiotemporal reference for "precise point water spraying".

[0074] Ultrasonic detection array: Distributed at the front, rear, left, and right corners of the tracked vehicle, it is used to detect low-lying obstacles within the blind spots of the lidar. This detector can provide feedback on the absolute distance (with millimeter-level accuracy) and signal-to-noise ratio (SNR) during time of flight, ensuring the reliability of obstacle avoidance at lower levels in dusty environments.

[0075] Decision-making level: Industrial control computer: As the control center of the system, it is located inside the tracked vehicle. It runs the ROS operating system, is responsible for processing point cloud data, executing SLAM mapping algorithms, and running other programs such as task state navigation. It also calculates spatiotemporal coupling constraints, performs pose calculations, runs particle filter-based localization algorithm (AMCL) and task state, and performs spatiotemporal coupling constraint determination (i.e., comparing the robot's position and attitude deviations with preset position and attitude thresholds for dual threshold verification).

[0076] Execution layer: The STM32 main control board communicates with the industrial computer via a serial port. It is responsible for executing the underlying safety fuse mechanism; it also controls the on / off state of the track drive motor and the chassis movement based on the verification results. After all verifications pass, it performs a chassis lock-up operation. In the cleaning process, it controls the high-pressure water pump and high-pressure water gun to perform pulsed water spraying.

[0077] In addition, the industrial control computer is connected to the contactless charging pile to detect the real-time power level and control the automatic recharging when the power level is low.

[0078] Therefore, this invention employs a discontinuous fixed-point cleaning method and system based on discrete topological nodes. By constructing an environmental map and marking each work anchor point on it, each anchor point is bound to an attribute vector. The environment is divided into two categories of attributes: work anchor points and navigation paths. When the robot passes through non-work areas along the path, the underlying controller forcibly locks the water pump circuit to an open circuit state. This eliminates the need for complex visual recognition algorithms to determine instrument positions, preventing water spraying accidents caused by software misjudgments or drift when passing precision instruments. Furthermore, by setting dual threshold checks including positional and attitude deviations, it ensures that the nozzle is precisely aligned with the center of the stain, avoiding ineffective spraying due to vehicle tilt. Ultrasonic detection fills the near-ground blind zone of the lidar, improving the robot's robustness in complex dusty environments. Finally, the flexible and controllable fixed-point pulse cleaning spray scheme significantly reduces water consumption and energy consumption, improving the safety and resource utilization of industrial site cleaning.

[0079] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A non-continuous fixed-point cleaning method based on discrete topological nodes, characterized in that, Includes the following steps: Step S1: Construct an environment map, mark each job anchor point on the environment map, and bind an attribute vector to each job anchor point; Set the area in the environmental map, excluding the work anchor point, as a no-spray zone; Step S2: Obtain the task requirements and the location information of the work anchor points, perform global path planning, and clean each work anchor point in sequence along the planned path; As the robot moves along the planned path toward the target work anchor point, ultrasonic detection is performed while the water pump circuit remains physically disconnected. Step S3: When the robot moves along the planned path, the robot calculates the current pose in real time through its own sensors, and calculates the spatiotemporal state deviation between the current pose and the preset pose of the current target work anchor point, and generates a pose state deviation vector. Step S4: Based on the pose state deviation vector, perform dual threshold verification on the robot's position deviation and attitude deviation respectively; When the ultrasonic test is qualified and the dual threshold verification is passed, the position locking signal is triggered; Step S5: After receiving the positioning lock signal, the system triggers the positioning lock state, the chassis locks up, and reads the attribute vector of the current target work anchor point, and compiles a cleaning spraying plan based on the attribute vector; Keep the water pump circuit closed and execute the cleaning spray program. Once completed, unlock and switch to the next target work anchor point.

2. The method for non-continuous fixed-point cleaning based on discrete topology nodes according to claim 1, characterized in that, In step S1, an environment map is constructed, and each job anchor point is marked on the environment map. Each job anchor point is bound to an attribute vector including: Step S11: Start the lidar to scan the field environment. The industrial control computer runs the SLAM mapping algorithm to build a high-precision environmental grid map covering the entire area. Step S12: The operator views the constructed grid map on the interactive interface of the industrial control computer, and selects and marks all the work anchor points in the grid map interface in sequence. Step S13: The industrial control computer automatically records the spatial coordinates and preset cleaning parameters of each work anchor point, and integrates the spatial coordinates and preset cleaning parameters of the work anchor points into an attribute vector.

3. The method for non-continuous fixed-point cleaning based on discrete topology nodes according to claim 2, characterized in that, Step S2, which involves performing global path planning and cleaning each work anchor point sequentially along the planned path, includes: Step S21: Generate a discrete task topology graph based on the task anchor points; Step S22: The industrial control computer plans a global navigation path connecting each work anchor point based on the discrete task topology map, and generates a real-time obstacle avoidance path using a local dynamic window algorithm. Step S23: The industrial control computer sends the path data to the STM32 main control board via serial port as navigation instructions, and performs cleaning according to the navigation instructions; As the robot moves along the path, the STM32 main control board executes navigation-mode physical fuse logic.

4. The method for non-continuous fixed-point cleaning based on discrete topology nodes according to claim 3, characterized in that, The execution of the navigation-state physical circuit breaker logic specifically includes: Step S231: Set the water pump operation enable to control the circuit opening and closing state of the water pump; Step S232: Set the pump operation enable at time t to... Its expression is: ； in, This is the positioning and locking signal of the target operation anchor point at time t. ; for The host computer sends a request for operation. ; When no position lock signal is received 0, regardless Any value is possible. 0, at this time the water pump circuit remains disconnected; when the positioning lock signal is received...

1. At this point, the host computer's operation request is determined, and then the on / off status of the water pump circuit is determined.

5. The method for non-continuous fixed-point cleaning based on discrete topology nodes according to claim 1, characterized in that, Step 3 involves real-time calculation of the current pose and the spatiotemporal state deviation between the current pose and the preset pose of the current target work anchor point, generating a pose state deviation vector including: Step S31: Extract the preset pose of the current target work anchor point. And calculate the robot's current pose. ; Step S32, calculate the first... Target operation anchor points Pose state deviation vector at time t The calculation formula is: ； in, For the first The preset center X coordinate of each target operation anchor point; The preset center Y coordinate of the i-th target operation anchor point; This represents the robot's heading angle deviation, indicating the robot's current... Heading angle at time With the The preset optimal water spray direction for each target operation anchor point The absolute value of the difference between them; For the robot at the current moment The X coordinate; For the robot at the current moment The Y-coordinate.

6. The method for non-continuous fixed-point cleaning based on discrete topology nodes according to claim 5, characterized in that, The specific steps for performing dual threshold checks on the robot's position deviation and attitude deviation in step S4 are as follows: Step S41: Perform distance determination and set a preset distance threshold. Calculate the Euclidean distance between the robot's current position and the center coordinates of the target work anchor point. The calculation formula is: ； like ≥ If the distance condition is not met, the verification fails; ≤ If the distance condition is met, then the attitude determination is performed. Step S42: Perform attitude determination and set a preset attitude threshold. Calculate the absolute value of the angular deviation between the robot's current heading angle and the preset optimal water spray direction; ； like ≥ If the attitude conditions are not met, the verification fails; ≤ If the attitude condition is met, the double threshold check passes.

7. The method for non-continuous fixed-point cleaning based on discrete topology nodes according to claim 1, characterized in that, Step S5, which involves reading the attribute vector of the current target work anchor point and formulating a cleaning spraying plan based on the attribute vector, includes the following steps: Step S51: After the robot arrives at the target work anchor point and receives the position locking signal, it uses the lidar to scan the point cloud data of the current target work anchor point, extracts the reflection intensity feature value of the ground at the anchor point, and compares the reflection intensity feature value with the preset cleanliness benchmark value to determine the stain level of the current anchor point. Step S52: The industrial control computer calls the preset parameter mapping model according to the stain level and automatically matches and generates the initial cleaning spray plan for the anchor point. The initial cleaning spray program includes the set values ​​for basic spray duration, number of pulse cycles, and water pump pressure; Step S53: When executing the initial cleaning spraying plan, the ultrasonic detection array collects the ground echo signal after water spraying in real time, and the industrial control computer analyzes the acoustic characteristics of the echo signal to obtain real-time cleanliness feedback data. Step S54: The industrial control computer compares the real-time cleanliness feedback data with the preset cleanliness threshold in real time. If the feedback data indicates that the area has been cleaned, the dynamic cutoff logic is executed to terminate the spraying in advance and release the lock. If the standard is not met, the spraying time is dynamically extended or the number of pulse cycles is increased until the cleanliness standard is met.

8. The method for non-continuous fixed-point cleaning based on discrete topology nodes according to claim 7, characterized in that, It also includes step S6, autonomous resupply: Step S61: The system monitors the battery level in real time. When the battery level is lower than the preset threshold, the current cleaning task is paused. Step S62: The industrial control computer plans the optimal path back to the contactless charging station, and the robot returns along the path; Step S63: When the distance between the receiving coil and the transmitting coil is not greater than the preset distance, start contactless charging and cut off the motor drive to enter low power standby mode.

9. A non-continuous fixed-point cleaning method based on discrete topology nodes according to claim 8, characterized in that, Ultrasonic testing includes: Step S201: The STM32 main control board independently reads the distance data of obstacles around the vehicle body fed back by the ultrasonic detection array; Step S202: Compare the read distance to the nearest obstacle with the preset safe distance threshold to determine whether the current environment meets the safe working conditions; Step S203: If the distance to the nearest obstacle is determined to be less than the preset safe distance threshold, the ultrasonic detection is deemed unqualified. The STM32 main control board directly triggers the bottom-level emergency stop, forcibly disconnects the motor and water pump circuits, and immediately terminates the water spraying operation. Step S204: When the STM32 main control board detects that the distance to the nearest obstacle is greater than the preset safe distance threshold, it determines that the ultrasonic detection is qualified.

10. A discontinuous fixed-point cleaning system based on discrete topology nodes, characterized in that, A method for performing a discontinuous fixed-point cleaning method based on discrete topology nodes as described in any one of claims 1-9 includes: The multi-source environmental perception module includes a lidar mounted on the top and an ultrasonic detection array around the vehicle body; the lidar is used to construct an environmental grid map containing operational anchor points and provide positioning data, while the ultrasonic detection array is used to scan obstacles in the lidar blind spot in real time. The industrial control computer is connected to the multi-source environmental perception module. It is used to run the particle filter-based positioning algorithm and task status, navigation algorithm, and calculate the robot pose and output operation control commands according to the spatiotemporal coupling constraint logic. The STM32 main control board connects to the industrial computer via a serial port. It is used to parse navigation commands, independently read ultrasonic feedback data, and execute the underlying safety fuse logic based on the nearest neighbor obstacle distance. The work execution mechanism is used to perform pulsed water spraying operations when the system is in the locked position and there are no obstacles within the ultrasonic detection range; The energy supply module is used to detect the system's energy reserve usage, provide continuous power to the system, and guide autonomous recharging.

Images