An unmanned sanitation vehicle missing scanning area identification method and system of local path deviation

Abstract

The present application relates to a kind of local path deviation's unmanned sanitation vehicle leak scanning area identification method and system, the method comprises: S1.working area entry detection, when unmanned sanitation vehicle travels from parking point to the starting point of pre-defined working area, control system detects to enter working area, start to open upper installation and prepare to start cleaning work;S2.variable initialization;S3.planning period integral calculation;S4.working area exit detection, when vehicle drives out of working area, control system detects to exit working area, starts to retract upper installation and stops operation, while stopping the execution of operation route length statistics module;S5.leak scanning rate result output.The present application not only effectively cleans total length and the total length of leak scanning caused by detouring to avoid obstacle, but also calculates leak scanning rate in real time, and uploads vehicle cloud monitoring platform, so that relevant staff can pay attention to cleaning quality in real time during the operation of unmanned cleaning vehicle, and promptly find the working area of leak scanning serious and take corresponding solution measures.

Description

Technical Field

[0001] This invention relates to the field of unmanned sanitation vehicle technology, and in particular to a method and system for identifying missed areas by unmanned sanitation vehicles due to local path deviation. Background Technology

[0002] As urban roads continue to improve, road cleaning has become increasingly important. Traditional manual cleaning methods face problems such as high labor costs, low efficiency, and high safety risks. Meanwhile, fewer and fewer people are willing to engage in road cleaning work. Therefore, unmanned cleaning has been developing rapidly in recent years thanks to innovations in autonomous driving technology. Modern unmanned cleaning vehicles, based on high-precision maps, multi-sensor fusion positioning, path planning, and motion control technologies, are being used more and more widely in scenarios such as closed parks and urban roads.

[0003] The actual operating scenarios of unmanned sanitation vehicles are more unique than those of ordinary intelligent driving vehicles. In order to ensure that the sweeping brushes can clean the edge areas of the road, sanitation vehicles often drive along the edge of the road or the curb. In order to ensure that it can complete precise edge cleaning operations, unmanned sanitation vehicles need a high-precision positioning system, high-precision map, multi-sensor environmental perception, and path planning and control modules to work together.

[0004] The workflow of unmanned sanitation vehicles in actual operation scenarios is as follows: Workers first delineate the cleaning area and work points on a high-precision map. Then, the unmanned sanitation vehicle's autonomous driving system plans a global path covering all work points and a real-time reference line. During actual operation, the vehicle follows this reference line for more refined local path planning and control. However, actual operation scenarios are often complex and varied. During edge cleaning, the vehicle often encounters various static obstacles such as stationary vehicles, trash cans, and roadblocks, as well as dynamic obstacles such as pedestrians, non-motorized vehicles, and other vehicles. In these situations, the vehicle needs to deviate from the reference line to detour, resulting in uncollected trash in areas obscured by obstacles, leading to missed areas.

[0005] Regarding the identification of missed areas, some existing methods use visual perception to detect the coverage of the cleaned road surface. Specifically, a camera is installed behind the vehicle, and then image segmentation algorithms are used to identify the boundaries between cleaned and uncleaned areas. However, visual perception is highly susceptible to factors such as light intensity, weather conditions, and the consistency of road texture, resulting in low detection accuracy in complex environments such as rain, nighttime, and strong backlighting. Furthermore, image segmentation methods using rear-view cameras suffer from significant detection lag. This means that it's only possible to determine whether a section of road has been cleaned after the unmanned sanitation vehicle has already passed through it, making it impossible to provide proactive warnings of missed areas while the vehicle is in motion.

[0006] Besides the visual perception methods mentioned above, there are also methods that determine the sweeping coverage rate by analyzing the control signals and movement trajectories of unmanned sanitation vehicles (UGVs). This method specifically calculates the length of the travel path by recording the UUV's position coordinate sequence, thereby estimating the sweeping coverage rate. However, this method has a problem: it cannot distinguish between the UUV's normal edge-hugging trajectory and its obstacle avoidance trajectory. When the UUV performs obstacle avoidance maneuvers, a longer obstacle avoidance trajectory is generated. This method includes the distance of the bypassed section in the total working area length, thus overestimating the sweeping coverage rate. This type of method has a significant error in calculating the missed sweeping rate under large-angle obstacle avoidance conditions.

[0007] In existing unmanned sanitation vehicle systems, the path planning module and the operation monitoring module are typically two independent subsystems. The planning module is used for trajectory generation and tracking control, while the operation monitoring module records vehicle operating status such as battery level, speed, and sweeping brush status. There is no data collaboration between the two modules. This separate architecture means that the reference line information and obstacle avoidance decision-making information held by the planning module are not effectively used for calculating sweeping coverage, nor can the superstructure be dynamically controlled based on the unmanned vehicle's driving status.

[0008] In summary, the existing technology has the following main drawbacks: (1) Lack of quantitative identification methods for missed areas. Existing unmanned sanitation vehicle autonomous driving systems mainly focus on the accuracy of path tracking and the safety of obstacle avoidance, and the assessment of cleaning coverage is limited to simple qualitative judgment. When the unmanned sanitation vehicle deviates from the reference line path due to obstacle avoidance, the system cannot accurately identify which areas are missed, nor can it quantify the length of the missed areas, resulting in the inability to objectively assess the quality of the cleaning operation.

[0009] (2) Lack of real-time quantitative statistical methods for missed sweeping rate. Current technical methods usually rely on manual inspection or reviewing playback videos after cleaning operations to judge the cleaning quality, without calculating the missed sweeping rate in real time during the cleaning operation. Staff cannot know whether the missed sweeping rate of the current work area meets the standard, nor can they dynamically adjust and optimize the cleaning operation.

[0010] (3) Lack of a quantitative method for the length of missed sweeping in the detour area. When the unmanned sanitation vehicle detours due to obstacles, the existing technology does not effectively project the detour path onto the reference line, and cannot distinguish the boundary between the cleaned area and the missed area. Therefore, it is impossible to calculate the proportion of the extra missed sweeping length caused by the detour to the total length of the working area.

[0011] (4) Lack of a unified and reliable mileage integration benchmark. Existing methods lack a mathematical model that projects the actual driving path of the vehicle onto the reference line. This makes it impossible to accurately align and decompose the real-time pose of the vehicle, including its position and heading, with the reference line in each planning cycle. Consequently, it is impossible to establish a spatial mapping relationship between the detour area and the reference line, and therefore, it is impossible to accurately calculate the missed scan rate. Summary of the Invention

[0012] In view of the above problems, the present invention provides a method and system for identifying missed areas of unmanned sanitation vehicles due to local path deviation. It not only effectively measures the total cleaning length and the total missed length caused by detours and obstacle avoidance, but also calculates the missed sweeping rate in real time and uploads it to the vehicle cloud monitoring platform. This allows relevant personnel to monitor the cleaning quality in real time during the operation of the unmanned sanitation vehicle, promptly identify work areas with serious missed sweeping, and take corresponding solutions.

[0013] To achieve the above and other related objectives, the present invention provides the following technical solution: A method for identifying missed areas by unmanned sanitation vehicles due to local path deviation, the method comprising: S1. Work area entry detection: When the unmanned sanitation vehicle travels from the parking point to the predefined work area starting point, the control system detects that it has entered the work area and starts to open the upper body to prepare for cleaning work. S2. Variable initialization; S3. Integral calculation within the planning period; S4. Work area exit detection: When the vehicle leaves the work area, the control system detects that it has exited the work area, starts to retract the superstructure and stops working, and at the same time stops the execution of the work route length statistics module. S5. Output of missed scan rate results.

[0014] Furthermore, in step S5, the missed sweep rate result is output as the total effective sweeping length g of the vehicle along the reference line within the working area. totals The total length of unscanned areas (g) within this work area due to obstacle avoidance and detours. detours The missed scan rate is calculated using these two variable values. , The results of missed scans in this work area are reported to the cloud management platform and saved to the local log for reference in subsequent job quality assessment and path optimization.

[0015] Furthermore, in step S2, the variable initialization includes the total length g accumulated along the reference line. totals Zeroing out the cumulative detour length g along the reference line detours Set to zero, last timestamp of the previous frame timestamp Clear and first frame flag is firstframeSet to true.

[0016] Further, in step S3, the integral calculation within the planning period includes: S31. Check the validity of the reference line, obtain the reference line of the current frame obtained from the high-precision map and routing request information, and then check whether its reference point sequence is empty. If it is empty, return directly and exit the calculation of the current velocity integral accumulation; if it is not empty, continue to execute the current velocity integral accumulation calculation. S32. Obtain the status and coordinate projection of the unmanned sweeper; S33. Calculate the time difference dt and detect dt anomalies; S34. Extract heading angle information and calculate the included angle; S35. Calculate the displacement increment; S36. Anomaly detection of displacement increment; S37. Total length accumulation; S38. Detour determination and detour length accumulation; S39. Record a new timestamp.

[0017] Further, in step S32, obtaining the unmanned sweeper status and coordinate projection includes... S321. Obtain the vehicle's pose information for the current frame from the autonomous driving system of the unmanned sweeper, including the vehicle's coordinates, velocity v, and heading angle; S322. Convert the Cartesian coordinates (x, y) of the vehicle's position to the Frenet coordinate system (s, l) using the XYToSL() method of the reference line, where s is the arc length coordinate along the reference line and l is the lateral distance from the reference line. At the same time, determine whether the conversion is successful. If the conversion fails, return directly and do not perform the following operations.

[0018] Further, in step S33, the calculation of the time difference dt and the dt anomaly detection include dt being derived by subtracting the timestamp of the current frame from the timestamp of the previous frame, as shown in the formula: dt=current timestamp -last timestamp Then perform anomaly detection on dt, if is firstframe If true, it means the user has just entered the current workspace, and then is... firstfram Set to false, record the current timestamp, and then return; if dt<0, it is abnormal data, indicating that there is an abnormal situation of clock synchronization jump or clock reset after module restart, update and return; if dt>1.0 seconds, it indicates that there is a recovery after module lag or suspension, update and return.

[0019] Further, in step S34, the extraction of heading angle information and calculation of the included angle includes: S341. Obtain the vehicle's heading angle Car from the status of the unmanned sanitation vehicle. heading ; S342. Obtain the reference line heading angle Ref on the current position of the unmanned sanitation vehicle projected onto the reference line. heading ; S343. Calculate the heading deviation angle Δθ between the two. Δθ=Car heading -Ref heading ; Then normalize Δθ to the interval [-π, π]. Δθ=atan2(sin(Δθ), cos(Δθ)); In step S35, the calculation of the displacement increment includes the displacement calculation for each small segment. s increment =v×dt×cos(Δθ), Where v is the current speed of the unmanned sweeper obtained in step S32, dt is the time difference calculated in step S33, and Δθ is the angle between the vehicle's heading and the reference line's heading calculated in step S34. The physical meaning of the formula is: the actual distance traveled by the vehicle in time dt is v×dt, and then the projection length s of this vector onto the tangent direction of the reference line is... increment As an increment in effective cleaning length.

[0020] Furthermore, in step S36, the anomaly determination of the displacement increment includes determining the anomaly of the displacement increment value calculated in S35, therefore it is necessary to... increment Perform a reasonableness check: if s increment <0, negative displacement, indicates that the vehicle is reversing or the angle between its heading and the reference line is too large, exceeding 90°. This data result is invalid, and the process should be returned directly, skipping the current accumulation; if s increment If v×1.0 exceeds the theoretical maximum possible projection length within one cycle, it is caused by sensor noise or data anomaly. Return directly and skip this accumulation. In step S37, after the total length is accumulated and all the above outlier checks pass, s is... increment Added to the total length accumulated along the reference line, g totals +=s increment .

[0021] Further, in step S38, the detour determination and detour length accumulation include using the lateral distance l of the vehicle's deviation from the reference line in the current frame obtained in S32 to determine whether |l| is greater than the detour determination threshold D. threshold D threshold It is usually set to 0.15m, if |l|>D thresholdThis indicates that the unmanned sweeper is currently in obstacle avoidance mode, and this displacement represents the distance missed during obstacle avoidance. The displacement s needs to be adjusted. increment Added to the total detour length: g detours += s increment If |l|≤D threshold This indicates that the unmanned sweeper is driving close to the reference line and is currently performing normal sweeping operations, without affecting the s increment Add it to the total detour length.

[0022] To achieve the above and other related objectives, the present invention also provides a system for identifying missed areas by unmanned sanitation vehicles due to local path deviation, used to implement the aforementioned method for identifying missed areas by unmanned sanitation vehicles due to local path deviation. The system includes: The data perception layer includes a multi-sensor fusion positioning system: This system outputs centimeter-level accuracy in the pose of the unmanned sanitation vehicle via RTK and IMU combined with a navigation filter. It also utilizes feature point matching based on information from wheel speed encoders, LiDAR, and cameras to output the vehicle's coordinates, speed (v), and heading angle (θ). A high-precision map provides road traffic information, including road topology, lane boundaries, and driving boundaries. The environmental perception subsystem uses LiDAR point cloud target detection and tracking, visual target detection, and millimeter-wave radar target detection to output the location, size, and motion status of obstacles. The planning and decision-making layer, connected to the data perception layer, includes: a reference line generation module (based on high-precision maps and routing requests, generating one or more reference driving paths represented by a discrete reference point sequence, each point containing attributes such as position, heading angle, and curvature); a planning and decision-making module (operating at 10Hz throughout the entire lifecycle of the autonomous driving system, combining the aforementioned reference lines, obstacle prediction, and traffic rule factors to generate a safe and feasible driving trajectory in each cycle); and a work route length statistics module (a new addition to this invention, embedded in the execution chain of the planning and decision-making module, performing integral accumulation calculations in each planning cycle).

[0023] The execution layer is connected to the planning and decision-making layer. The motion control module receives the planned trajectory from the planning and decision-making layer and outputs steering wheel angle, throttle or braking control values ​​to drive the unmanned sanitation vehicle. The monitoring and evaluation layer, connected to the execution layer, provides a missed scan rate statistics and output interface: when a vehicle leaves the work area, the g value of the work area is obtained through the acquisition interface. totals and g detours Calculate the missed scan rate η=g detours / g totals ×100%, and then report the statistical results to the management platform.

[0024] The present invention has the following positive effects: 1. This invention enables real-time statistics of the total effective sweeping length along the reference line and the total missed sweeping length due to obstacle avoidance in each work area during the operation of unmanned sanitation vehicles. The missed sweeping rate is calculated in real time and uploaded to the vehicle cloud monitoring platform, allowing relevant personnel to monitor the sweeping quality in real time during the operation of unmanned sanitation vehicles, promptly identify work areas with serious missed sweeping, and take corresponding solutions.

[0025] 2. This invention is based entirely on the existing autonomous driving system of unmanned sanitation vehicles, utilizing the sensor configuration and planning modules therein. It does not require the addition of additional sensor hardware such as cameras and lidar, as well as computing units. It can achieve missed scan identification and quantitative statistics simply through software algorithms, thus saving hardware costs.

[0026] 3. This invention retains and analyzes historical information on the missed sweep rate of each work area, allowing the system to identify which road sections are prone to serious detours and missed sweeps. This enables optimization and adjustments during the offline path planning stage, and also allows for the pre-planning of better obstacle avoidance paths during the online stage to reduce the impact of detours on sweep coverage.

[0027] 4. Based on the quantitative index of missed sweeping rate provided by this invention, sanitation operation and management parties can formulate quantitative cleaning quality assessment standards, realize the transformation from qualitative management to quantitative management, and improve the standardization level of sanitation operations. Attached Figure Description

[0028] Figure 1 This is a schematic diagram of the method flow of the present invention; Figure 2 Here is the algorithm flowchart (a) for the planning period of this invention; Figure 3 Here is the algorithm flowchart (b) within the planning period of this invention; Figure 4 This is a flowchart illustrating the specific process of integral calculation within the planning period of this invention. Figure 5 This is a schematic diagram showing the angle between the heading of the unmanned sanitation vehicle of the present invention and the heading of the reference line; Figure 6 This is a schematic diagram of the length increment along the reference line direction within each planning cycle of the present invention; Figure 7 This is a schematic diagram of the system framework of the present invention; Figure 8 This is a schematic diagram of the working area and the bypass missed scan area of ​​the present invention; Figure 9 This is a schematic diagram showing the relationship between the modules of the autonomous driving system of the present invention. Detailed Implementation

[0029] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0030] Example: Figure 1 or Figure 2 or Figure 3 As shown, a method for identifying missed areas by unmanned sanitation vehicles due to local path deviation is described, the method comprising: S1. Work area entry detection: When the unmanned sanitation vehicle travels from the parking point to the predefined work area starting point, the control system detects that it has entered the work area and starts to open the upper body to prepare for cleaning work. S2. Variable initialization; S3. Integral calculation within the planning period; S4. Work area exit detection: When the vehicle leaves the work area, the control system detects that it has exited the work area, starts to retract the superstructure and stops working, and at the same time stops the execution of the work route length statistics module. S5. Output of missed scan rate results.

[0031] In this embodiment, in step S5, the missed sweep rate result is output as the total effective sweeping length g of the vehicle along the reference line within the working area. totals The total length of unscanned areas (g) within this work area due to obstacle avoidance and detours. detours The missed scan rate is calculated using these two variable values. , The results of missed scans in this work area are reported to the cloud management platform and saved to the local log for reference in subsequent job quality assessment and path optimization.

[0032] In this embodiment, in step S2, the variable initialization includes the total length g accumulated along the reference line. totals Zeroing out the cumulative detour length g along the reference line detours Set to zero, last timestamp of the previous frame timestamp Clear and first frame flag is firstframe Set to true.

[0033] In this embodiment, as Figure 4 As shown, in step S3, the integral calculation within the planning period includes: S31. Check the validity of the reference line, obtain the reference line of the current frame obtained from the high-precision map and routing request information, and then check whether its reference point sequence is empty. If it is empty, return directly and exit the calculation of the current velocity integral accumulation; if it is not empty, continue to execute the current velocity integral accumulation calculation. S32. Obtain the status and coordinate projection of the unmanned sweeper; S33. Calculate the time difference dt and detect dt anomalies; S34. Extract heading angle information and calculate the included angle; S35. Calculate the displacement increment; S36. Anomaly detection of displacement increment; S37. Total length accumulation; S38. Detour determination and detour length accumulation; S39. Record a new timestamp.

[0034] In this embodiment, in step S32, obtaining the unmanned sweeper status and coordinate projection includes... S321. Obtain the vehicle's pose information for the current frame from the autonomous driving system of the unmanned sweeper, including the vehicle's coordinates, velocity v, and heading angle; S322. Convert the Cartesian coordinates (x, y) of the vehicle's position to the Frenet coordinate system (s, l) using the XYToSL() method of the reference line, where s is the arc length coordinate along the reference line and l is the lateral distance from the reference line. At the same time, determine whether the conversion is successful. If the conversion fails, return directly and do not perform the following operations.

[0035] In this embodiment, in step S33, the calculation of the time difference dt and the anomaly detection of dt include dt being derived by subtracting the timestamp of the current frame from the timestamp of the previous frame, as shown in the formula: dt=current timestamp -last timestamp Then perform anomaly detection on dt, if is firstframe If true, it means the user has just entered the current workspace, and then is... firstfram Set to false, record the current timestamp, and then return; if dt<0, it is abnormal data, indicating that there is an abnormal situation of clock synchronization jump or clock reset after module restart, update and return; if dt>1.0 seconds, it indicates that there is a recovery after module lag or suspension, update and return.

[0036] In this embodiment, as Figure 5 As shown, in step S34, the extraction of heading angle information and calculation of the included angle includes: S341. Obtain the vehicle's heading angle Car from the status of the unmanned sanitation vehicle. heading ; S342. Obtain the reference line heading angle Ref on the current position of the unmanned sanitation vehicle projected onto the reference line. heading ; S343. Calculate the heading deviation angle Δθ between the two. Δθ=Car heading -Ref heading ; Then normalize Δθ to the interval [-π, π]. Δθ=atan2(sin(Δθ), cos(Δθ)); like Figure 6 As shown, in step S35, the calculation of the displacement increment includes the displacement calculation for each small segment. s increment =v×dt×cos(Δθ), Where v is the current speed of the unmanned sweeper obtained in step S32, dt is the time difference calculated in step S33, and Δθ is the angle between the vehicle's heading and the reference line's heading calculated in step S34. The physical meaning of the formula is: the actual distance traveled by the vehicle in time dt is v×dt, and then the projection length s of this vector onto the tangent direction of the reference line is... increment As an increment in effective cleaning length.

[0037] In this embodiment, in step S36, the anomaly determination of the displacement increment includes determining the anomaly of the displacement increment value calculated in S35. Therefore, it is necessary to determine the anomaly of s. increment Perform a reasonableness check: if s increment <0, negative displacement, indicates that the vehicle is reversing or the angle between its heading and the reference line is too large, exceeding 90°. This data result is invalid, and the process should be returned directly, skipping the current accumulation; if s increment If v×1.0 exceeds the theoretical maximum possible projection length within one cycle, it is caused by sensor noise or data anomaly. Return directly and skip this accumulation. In step S37, after the total length is accumulated and all the above outlier checks pass, s is... increment Added to the total length accumulated along the reference line, g totals +=s increment .

[0038] In this embodiment, in step S38, the detour determination and detour length accumulation include using the lateral distance l of the vehicle's deviation from the reference line in the current frame obtained in S32 to determine whether |l| is greater than the detour determination threshold D. threshold D threshold It is usually set to 0.15m, if |l|>D thresholdThis indicates that the unmanned sweeper is currently in obstacle avoidance mode, and this displacement represents the distance missed during obstacle avoidance. The displacement s needs to be adjusted. increment Added to the total detour length: g detours += s increment If |l|≤D threshold This indicates that the unmanned sweeper is driving close to the reference line and is currently performing normal sweeping operations, without affecting the s increment Add it to the total detour length.

[0039] In this embodiment, as Figure 7 As shown, this invention provides a system for identifying missed areas by unmanned sanitation vehicles due to local path deviation, used to implement the aforementioned method for identifying missed areas by unmanned sanitation vehicles with local path deviation. The system includes: The data perception layer includes a multi-sensor fusion positioning system: This system outputs centimeter-level accuracy in the pose of the unmanned sanitation vehicle via RTK and IMU combined with a navigation filter. It also utilizes feature point matching based on information from wheel speed encoders, LiDAR, and cameras to output the vehicle's coordinates, speed (v), and heading angle (θ). A high-precision map provides road traffic information, including road topology, lane boundaries, and driving boundaries. The environmental perception subsystem uses LiDAR point cloud target detection and tracking, visual target detection, and millimeter-wave radar target detection to output the location, size, and motion status of obstacles. The planning and decision-making layer, connected to the data perception layer, includes: a reference line generation module (based on high-precision maps and routing requests, generating one or more reference driving paths represented by a discrete reference point sequence, each point containing attributes such as position, heading angle, and curvature); a planning and decision-making module (operating at 10Hz throughout the entire lifecycle of the autonomous driving system, combining the aforementioned reference lines, obstacle prediction, and traffic rule factors to generate a safe and feasible driving trajectory in each cycle); and a work route length statistics module (a new addition to this invention, embedded in the execution chain of the planning and decision-making module, performing integral accumulation calculations in each planning cycle).

[0040] The execution layer is connected to the planning and decision-making layer. The motion control module receives the planned trajectory from the planning and decision-making layer and outputs steering wheel angle, throttle or braking control values ​​to drive the unmanned sanitation vehicle. The monitoring and evaluation layer, connected to the execution layer, provides a missed scan rate statistics and output interface: when a vehicle leaves the work area, the g value of the work area is obtained through the acquisition interface. totals and g detours Calculate the missed scan rate η=g detours / g totals ×100%, and then report the statistical results to the management platform.

[0041] In this embodiment, as Figure 9 This invention proposes a method for identifying and quantifying missed areas by unmanned sanitation vehicles based on local path deviation. This method relies on the existing autonomous driving system architecture of unmanned sanitation vehicles, which mainly includes an environmental perception module, a multi-sensor fusion positioning module, a high-precision map module, a path planning and decision-making module, and a motion control module. The method adds a sub-module for calculating the length of the work route within the path planning and decision-making module.

[0042] In this embodiment, as Figure 8 As shown, the core of this invention lies in the execution of a route length statistics module within each planning cycle of the path planning and decision-making module. The input to this module is the unmanned sweeper's status information and reference line information obtained by the planning module. Then, through coordinate projection, difference calculation, and cosine projection of the included angle, the incremental sweeping displacement of the unmanned sweeper along the reference line direction within the current cycle is calculated in real time. Simultaneously, the lateral offset of the unmanned sweeper relative to the reference line is used to determine whether this displacement constitutes a missed sweeping section. Finally, the displacement increments for the entire working area cycle are accumulated, outputting the total sweeping length and the missed sweeping length of the working area, thereby calculating the missed sweeping rate for this segment of the working area.

[0043] The route length statistics module defines four core static member variables during the planning process: the total length accumulated along the reference line: g totals Cumulative detour length along the reference line: g detours The timestamp of the previous frame: last timestamp and the first frame flag: is firstframe These four variables collectively ensure the statistical status of the operational route length across planning cycles, forming a global length accumulator. If the unmanned sweeper enters a new work area, these four variables are initialized to zero. In each planning cycle, calculations are performed based on the attitude information of the current frame and the previous frame, updating the increment for each planning cycle. When the unmanned sweeper leaves the work area, the final total length and detour length are read from the variables, thus completing the calculation of the missed sweep rate.

[0044] In summary, this invention not only effectively cleans the total sweeping length and the total sweeping length missed due to obstacle avoidance, but also calculates the missed sweeping rate in real time and uploads it to the vehicle cloud monitoring platform. This allows relevant personnel to monitor the sweeping quality in real time during the operation of the unmanned sweeper, promptly identify work areas with serious missed sweeping, and take corresponding solutions.

[0045] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for identifying missed areas by unmanned sanitation vehicles due to local path deviation, characterized in that, The method includes: S1. Work area entry detection: When the unmanned sanitation vehicle travels from the parking point to the predefined work area starting point, the control system detects that it has entered the work area and starts to open the upper body to prepare for cleaning work. S2. Variable initialization; S3. Integral calculation within the planning period; S4. Work area exit detection: When the vehicle leaves the work area, the control system detects that it has exited the work area, starts to retract the superstructure and stops working, and at the same time stops the execution of the work route length statistics module. S5. Output of missed scan rate results.

2. The method for identifying missed areas by unmanned sanitation vehicles due to local path deviation as described in claim 1, characterized in that, In step S5, the missed sweep rate result is output as the total effective sweeping length g of the vehicle along the reference line within the working area. totals The total length of unscanned areas (g) within this work area due to obstacle avoidance and detours. detours The missed scan rate is calculated using these two variable values. ， The results of missed scans in this work area are reported to the cloud management platform and saved to the local log for reference in subsequent job quality assessment and path optimization.

3. The method for identifying missed areas by unmanned sanitation vehicles due to local path deviation as described in claim 1, characterized in that, In step S2, the variable initialization includes the total length g accumulated along the reference line. totals Zeroing out the cumulative detour length g along the reference line detours Set to zero, last timestamp of the previous frame timestamp Clear and first frame flag is firstframe Set to true.

4. The method for identifying missed areas by unmanned sanitation vehicles due to local path deviation as described in claim 1, characterized in that, In step S3, the integral calculation within the planning period includes: S31. Check the validity of the reference line, obtain the reference line of the current frame obtained from the high-precision map and routing request information, and then check whether its reference point sequence is empty. If it is empty, return directly and exit the calculation of the current velocity integral accumulation; if it is not empty, continue to execute the current velocity integral accumulation calculation. S32. Obtain the status and coordinate projection of the unmanned sweeper; S33. Calculate the time difference dt and detect dt anomalies; S34. Extract heading angle information and calculate the included angle; S35. Calculate the displacement increment; S36. Anomaly detection of displacement increment; S37. Total length accumulation; S38. Detour determination and detour length accumulation; S39. Record a new timestamp.

5. The method for identifying missed areas by unmanned sanitation vehicles due to local path deviation as described in claim 4, characterized in that, In step S32, obtaining the status and coordinate projection of the unmanned sweeper includes... S321. Obtain the vehicle's pose information for the current frame from the autonomous driving system of the unmanned sweeper, including the vehicle's coordinates, velocity v, and heading angle; S322. Convert the Cartesian coordinates (x, y) of the vehicle's position to the Frenet coordinate system (s, l) using the XYToSL() method of the reference line, where s is the arc length coordinate along the reference line and l is the lateral distance from the reference line. At the same time, determine whether the conversion is successful. If the conversion fails, return directly and do not perform the following operations.

6. The method for identifying missed areas by unmanned sanitation vehicles due to local path deviation as described in claim 4, characterized in that, In step S33, the calculation of the time difference dt and the anomaly detection of dt include dt being derived by subtracting the timestamp of the current frame from the timestamp of the previous frame, as shown in the formula: dt=current timestamp -last timestamp , Then perform anomaly detection on dt, if is firstframe If true, it means the user has just entered the current workspace, and then is... firstfram Set to false, record the current timestamp, and then return; if dt<0, it is abnormal data, indicating that there is an abnormal situation of clock synchronization jump or clock reset after module restart, update and return; if dt>1.0 seconds, it indicates that there is a recovery after module lag or suspension, update and return.

7. The method for identifying missed areas by unmanned sanitation vehicles due to local path deviation as described in claim 4, characterized in that, In step S34, the extraction of heading angle information and calculation of the included angle includes: S341. Obtain the vehicle's heading angle Car from the status of the unmanned sanitation vehicle. heading ; S342. Obtain the reference line heading angle Ref on the current position of the unmanned sanitation vehicle projected onto the reference line. heading ; S343. Calculate the heading deviation angle Δθ between the two. Δθ=Car heading -Ref heading ; Then normalize Δθ to the interval [-π, π]. Δθ=atan2(sin(Δθ), cos(Δθ)); In step S35, the calculation of the displacement increment includes the displacement calculation for each small segment. s increment =v×dt×cos(Δθ), Where v is the current speed of the unmanned sweeper obtained in step S32, dt is the time difference calculated in step S33, and Δθ is the angle between the vehicle's heading and the reference line's heading calculated in step S34. The physical meaning of the formula is: the actual distance traveled by the vehicle in time dt is v×dt, and then the projection length s of this vector onto the tangent direction of the reference line is... increment As an increment in effective cleaning length.

8. The method for identifying missed areas by unmanned sanitation vehicles due to local path deviation as described in claim 7, characterized in that, In step S36, the anomaly determination of the displacement increment includes determining the anomaly of the displacement increment value calculated in S35, therefore it is necessary to... increment Conduct a reasonableness check: If s increment <0, negative displacement, indicates that the vehicle is reversing or the angle between its heading and the reference line is too large, exceeding 90°. This data result is invalid, and the process should be returned directly, skipping the current accumulation; if s increment If v×1.0 exceeds the theoretical maximum possible projection length within one cycle, it is caused by sensor noise or data anomaly. Return directly and skip this accumulation. In step S37, after the total length is accumulated and all the above outlier checks pass, s is... increment Added to the total length accumulated along the reference line, g totals +=s increment .

9. The method for identifying missed areas by unmanned sanitation vehicles due to local path deviation as described in claim 4, characterized in that, In step S38, the detour determination and detour length accumulation include using the lateral distance l of the vehicle's deviation from the reference line in the current frame obtained in S32 to determine whether |l| is greater than the detour determination threshold D. threshold D threshold It is usually set to 0.15m, if |l|>D threshold This indicates that the unmanned sweeper is currently in obstacle avoidance mode, and this displacement represents the distance missed during obstacle avoidance. The displacement s needs to be adjusted. increment Added to the total detour length: g detours += s increment ; If |l|≤D threshold This indicates that the unmanned sweeper is driving close to the reference line and is currently performing normal sweeping operations, without affecting the s increment Add it to the total detour length.

10. A system for identifying missed areas by unmanned sanitation vehicles due to local path deviation, characterized in that, The system is used to implement the method for identifying missed areas by unmanned sanitation vehicles with local path deviations as described in any one of claims 1-9, the system comprising: The data perception layer includes a multi-sensor fusion positioning system: This system outputs centimeter-level accuracy in the pose of the unmanned sanitation vehicle via RTK and IMU combined with a navigation filter. It also utilizes feature point matching based on information from wheel speed encoders, LiDAR, and cameras to output the vehicle's coordinates, speed (v), and heading angle (θ). A high-precision map provides road traffic information, including road topology, lane boundaries, and driving boundaries. The environmental perception subsystem uses LiDAR point cloud target detection and tracking, visual target detection, and millimeter-wave radar target detection to output the location, size, and motion status of obstacles. The planning and decision-making layer, connected to the data perception layer, includes: a reference line generation module (based on high-precision maps and routing requests, generating one or more reference driving paths, represented by a discrete reference point sequence, each point containing attributes such as position, heading angle, and curvature); a planning and decision-making module (operating at 10Hz throughout the entire lifecycle of the autonomous driving system, combining the aforementioned reference lines, obstacle prediction, and traffic rule factors to generate a safe and feasible driving trajectory in each cycle); and a work route length statistics module (a new module added to this invention, embedded in the execution chain of the planning and decision-making module, performing integral accumulation calculations in each planning cycle). The execution layer is connected to the planning and decision-making layer. The motion control module receives the planned trajectory from the planning and decision-making layer and outputs steering wheel angle, throttle or braking control values ​​to drive the unmanned sanitation vehicle. The monitoring and evaluation layer, connected to the execution layer, provides a missed scan rate statistics and output interface: when a vehicle leaves the work area, the g value of the work area is obtained through the acquisition interface. totals and g detours Calculate the missed scan rate η=g detours / g totals ×100%, and then report the statistical results to the management platform.

Images