Dynamic switching system for unmanned sanitation vehicle operation modes

Abstract

This invention discloses a dynamic switching system for unmanned sanitation vehicle operation modes, relating to the field of vehicle driving technology. It includes a waste sensing module, a central processing module, and a road surface monitoring module. The waste sensing module includes a lidar and camera installed on the sanitation vehicle to obtain environmental information around the vehicle and identifies waste spatial distribution information using a built-in waste spatial distribution recognition algorithm. The road surface monitoring module includes roadside base stations and roadside sensors distributed along both sides of the road. The roadside base stations are used to locate the sanitation vehicle's position and traffic flow. The roadside sensors are used to collect real-time road weather data. The central processing module combines waste spatial distribution information, sanitation vehicle position, traffic flow, road weather data, and an operation mode reference model for comprehensive analysis, outputting the corresponding operation mode. This enables dynamic switching of the unmanned sanitation vehicle's operation modes, improving the efficiency of sanitation vehicle operations.

Description

Technical Field

[0001] This invention relates to the field of vehicle driving technology, specifically to a dynamic switching system for the operation modes of unmanned sanitation vehicles. Background Technology

[0002] With the accelerating pace of urbanization and the continuous increase in the total amount of highway paving, the amount of road litter has also increased, seriously affecting vehicle driving safety. The sanitation industry is facing enormous pressure. Because cleaning vehicles travel at relatively slow speeds and follow relatively fixed routes, they are prone to causing traffic congestion and affecting vehicle driving safety when cleaning operations are carried out on urban ring roads and highways, especially expressways, where speeds are high and traffic volume is large. This can easily lead to traffic accidents. If the cleaning speed is too fast, the interaction time between road litter and airflow is too short, and the suction capacity of the nozzle is insufficient to completely remove the road litter, resulting in missed areas. Based on the above shortcomings, this invention proposes a dynamic switching system for the operation mode of unmanned sanitation vehicles. Summary of the Invention

[0003] This invention aims to at least solve one of the technical problems existing in the prior art. To this end, this invention proposes a dynamic switching system for the operation modes of unmanned sanitation vehicles.

[0004] To achieve the above objectives, an embodiment of the first aspect of the present invention provides a dynamic switching system for the operation mode of an unmanned sanitation vehicle, comprising a waste sensing module, a central processing module, a road surface monitoring module, a mode switching module, a data acquisition module, a database, and a model creation module;

[0005] The waste sensing module includes a lidar and a camera installed on the sanitation vehicle to obtain environmental information around the vehicle and to identify waste spatial distribution information through a built-in waste spatial distribution recognition algorithm. The lidar is used to sense whether there are obstacles around the vehicle, and the camera is used to collect video images of the vehicle's surroundings in real time.

[0006] The road surface monitoring module includes roadside base stations and roadside sensors distributed on both sides of the road; the vehicle terminal is used to send the vehicle model to the roadside base station, sense the vehicle speed, sense the distance between the vehicle and surrounding vehicles, and combine the roadside base station to locate the sanitation vehicle's position and traffic flow.

[0007] The roadside sensors are used to collect road surface meteorological data in real time and transmit the collected data to the central processing module through the roadside base station; the road surface meteorological data includes road surface temperature, road surface humidity, and road surface water or snow accumulation.

[0008] The data acquisition module is used to acquire and review the original operation mode data, and store the approved original operation mode data as explicit operation mode data in the database; the explicit operation mode data refers to the original operation mode data where the operation quality coefficient ZP is greater than the preset quality threshold.

[0009] The model creation module is used to use the explicit operation mode data as a parameter training set and analyze it based on AI deep learning recognition algorithm to obtain a reference model of the operation mode of sanitation vehicles.

[0010] The central processing module is used to perform comprehensive analysis by combining information on the spatial distribution of garbage, the location of sanitation vehicles, traffic flow, road weather data, and the reference model of the operation mode, and then outputs the corresponding operation mode to the mode switching module to realize the dynamic switching of the operation mode of the unmanned sanitation vehicle; the operation mode includes the sweeping speed of the sanitation vehicle, the rotation speed of the sweeping brush, and the suction power of the suction nozzle.

[0011] Furthermore, the specific working steps of the data acquisition module are as follows:

[0012] First, the operating modes of all sanitation vehicles in the system are obtained, and the potential related operating data of all sanitation vehicles are analyzed and mined. The potential related operating data are represented by the spatial distribution information of garbage, the location of sanitation vehicles, traffic flow and road weather data when sanitation vehicles are working in this operating mode.

[0013] Then, the operation quality data of the sanitation vehicle after working in this operation mode is obtained, and the operation quality coefficient ZP is evaluated based on the operation quality data; the operation quality data includes the congestion caused to other vehicles during the working period and the cleaning of road garbage;

[0014] The operation quality coefficient ZP is compared with the preset quality threshold. If the operation quality coefficient ZP is greater than the preset quality threshold, the corresponding operation mode of the sanitation vehicle and the potential related operation data are stored in the database as explicit operation mode data.

[0015] Furthermore, the work quality coefficient ZP is evaluated based on the work quality data; specifically:

[0016] When the sanitation vehicle is working in the aforementioned operating mode, the vehicle speed, the distance between the sanitation vehicle and surrounding vehicles, and the speed of surrounding vehicles are collected.

[0017] Using the sanitation vehicle as the center, all vehicles traveling within a radius r1 are marked as associated vehicles; where r1 is a preset value; the number of associated vehicles is counted as L1.

[0018] The speed of the sanitation vehicle is denoted as CVt; the distance between the associated vehicle and the sanitation vehicle is denoted as vehicle spacing CLI; the speed of the associated vehicle is denoted as CVi; where i = 1, 2, ..., n; CLI and CVi correspond one-to-one; i represents the i-th associated vehicle;

[0019] Using formula The vehicle speed deviation PZ is calculated.

[0020] The maximum vehicle spacing is marked as CLmax, and the minimum vehicle spacing is marked as CLmin; the spacing difference ratio Gb is calculated using the formula Gb=(CLmax-CLmin) / CLmin;

[0021] The average vehicle spacing GDz is calculated according to the average value calculation formula; the congestion coefficient Ds is calculated using the formula Ds=f×L1×(1+PZ) / [GDz×(1-Gb)×a1], where a1 is the preset coefficient factor and f is the preset balance coefficient.

[0022] Compare the congestion coefficient Ds with the preset congestion threshold; if the congestion coefficient Ds is greater than the preset congestion threshold, it means that the sanitation vehicle is causing congestion to other vehicles at this time.

[0023] The percentage of congestion time of the sanitation vehicles is Tb, and the percentage of garbage cleaning area of ​​the sanitation vehicles is Mb. The operation quality coefficient ZP is calculated using the formula ZP=(Mb×a2) / (Tb×a3), where a2 and a3 are preset coefficient factors.

[0024] Furthermore, the specific analysis steps of the model creation module are as follows:

[0025] The explicit job pattern data obtained from the database will be used as the parameter training set, and the parameter training set will be split into training set, validation set and test set according to a preset ratio.

[0026] Establish an LSTM neural network model; wherein, the number of input nodes of the Long Short-Term Memory (LSTM) neural network is specified according to the number of input variables; set an appropriate number of hidden layer nodes and the number of output nodes representing the operating mode;

[0027] The training set, validation set, and test set are used as historical feature values ​​to input into the LSTM neural network model for model training. The model is then evaluated using a loss function to obtain the optimal operating mode reference model that minimizes the overall error of the training samples.

[0028] Furthermore, the specific steps of the waste spatial distribution identification algorithm include:

[0029] Road surface image information is extracted from video images of the vehicle's surroundings; the road surface image information is converted into a grayscale image, and the grayscale image is converted into a standard image through image preprocessing; the image preprocessing includes Gaussian filtering, image segmentation, and image enhancement;

[0030] Obtain the total number of pixels in a standard image within a preset region and mark it as the region area X1; identify each pixel and pinpoint the corresponding garbage pixels, specifically:

[0031] First, the gray values ​​of pixels in the standard image are marked as H1; then, the gray values ​​of each pixel are compared with the preset standard gray values, and the difference result is marked as C1; if C1 is greater than the difference threshold, the pixel is considered a garbage pixel.

[0032] Connected garbage pixels are merged. If the total number of garbage pixels is greater than a preset threshold, the area where the corresponding garbage pixel is located is marked as a garbage accumulation area. All garbage accumulation areas are integrated to obtain garbage spatial distribution information.

[0033] Compared with the prior art, the beneficial effects of the present invention are:

[0034] 1. The waste sensing module of this invention includes a lidar and a camera installed on the sanitation vehicle to obtain environmental information around the vehicle and identify waste spatial distribution information through a built-in waste spatial distribution recognition algorithm; the road monitoring module includes roadside base stations and roadside sensors distributed on both sides of the road. The roadside base stations are used to locate the position of the sanitation vehicle and the traffic flow; the roadside sensors are used to collect road weather data in real time; the central processing module is used to integrate the waste spatial distribution information, the position of the sanitation vehicle, the traffic flow and the road weather data into the input data of the operation mode reference model, and output the corresponding operation mode to realize the dynamic switching of the unmanned sanitation vehicle operation mode and improve the operation efficiency of the sanitation vehicle.

[0035] 2. In this invention, the data acquisition module is used to acquire and review the original operation mode data. First, it acquires the operation modes of all sanitation vehicles in the system and analyzes and mines the potential related operation data of all sanitation vehicles. Then, it acquires the operation quality data of the sanitation vehicles after they work in the operation mode, and evaluates the operation quality coefficient ZP based on the operation quality data. If the operation quality coefficient ZP is greater than the preset quality threshold, the corresponding original operation mode data is stored in the database as explicit operation mode data. The model creation module is used to use the explicit operation mode data as a parameter training set, and analyzes it based on the AI ​​deep learning recognition algorithm to obtain the operation mode reference model of the sanitation vehicles, so as to reduce the error of the operation mode reference model and improve the model accuracy. Attached Figure Description

[0036] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 This is a system block diagram of the unmanned sanitation vehicle operation mode dynamic switching system of the present invention. Detailed Implementation

[0038] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0039] like Figure 1 As shown, the unmanned sanitation vehicle operation mode dynamic switching system includes a garbage sensing module, a central processing module, a road surface monitoring module, a mode switching module, a data acquisition module, a database, and a model creation module.

[0040] The waste sensing module includes a lidar and a camera installed on the sanitation vehicle to obtain environmental information around the vehicle and to identify waste spatial distribution information through a built-in waste spatial distribution recognition algorithm.

[0041] LiDAR is used to detect whether there are obstacles around the vehicle, and cameras are used to collect video images of the vehicle's surroundings in real time.

[0042] It should be noted that the garbage spatial distribution identification algorithm in this invention specifically includes:

[0043] Road surface image information is extracted from video images of the vehicle's surroundings; the road surface image information is converted into a grayscale image, and the grayscale image is converted into a standard image through image preprocessing; image preprocessing includes Gaussian filtering, image segmentation, and image enhancement;

[0044] Obtain the total number of pixels in a standard image within a preset region and mark it as the region area X1; identify each pixel and pinpoint the corresponding garbage pixels, specifically:

[0045] First, the gray values ​​of pixels in the standard image are marked as H1; then, the gray values ​​of each pixel are compared with the preset standard gray values, and the difference result is marked as C1; if C1 is greater than the difference threshold, the pixel is considered a garbage pixel.

[0046] Connected garbage pixels are merged. If the total number of garbage pixels is greater than a preset threshold, the area where the corresponding garbage pixel is located is marked as a garbage accumulation area. All garbage accumulation areas are integrated to obtain garbage spatial distribution information.

[0047] The waste sensing module is used to transmit waste spatial distribution information to the central processing module;

[0048] The road surface monitoring module includes roadside base stations and roadside sensors distributed on both sides of the road; the on-board terminal on the sanitation vehicle is used to send the vehicle model to the roadside base station, sense the vehicle speed, sense the distance between the vehicle and surrounding vehicles, and combine the roadside base station to locate the sanitation vehicle's position and traffic flow.

[0049] Roadside sensors are used to collect real-time road weather data and transmit the collected data to the central processing module via roadside base stations. Road weather data includes road surface temperature, road surface humidity, and road surface water or snow accumulation. The data transmitted via roadside base stations here includes sanitation vehicle location, traffic flow, and road weather data.

[0050] The central processing module is used to comprehensively analyze information on the spatial distribution of garbage, the location of sanitation vehicles, traffic flow, road weather data, and the reference model of the operation mode, so as to realize the dynamic switching of the operation mode of unmanned sanitation vehicles and improve the operation efficiency of sanitation vehicles; the specific analysis steps are as follows:

[0051] Once the central processing module receives information on the spatial distribution of garbage, the location of sanitation vehicles, traffic flow, and road weather data, it integrates these information into the input data for the operation mode reference model.

[0052] Input data is entered into the operation mode reference model, and the corresponding operation mode is output; the operation mode includes the sanitation vehicle's sweeping speed, brush rotation speed, and suction power of the nozzle;

[0053] The central processing module is used to transmit the output operating mode to the mode switching module, which is used to control the switching of the sanitation vehicle's operating mode in order to improve the sanitation vehicle's operating efficiency.

[0054] In one embodiment of the present invention, the data acquisition module is used to acquire and review the original operation mode data, and store the approved original operation mode data as explicit operation mode data in the database; the specific working steps are as follows:

[0055] First, the operating modes of all sanitation vehicles in the system are obtained, and the potential related operating data of all sanitation vehicles are analyzed and mined. The potential related operating data are represented by the spatial distribution information of garbage, the location of sanitation vehicles, traffic flow and road weather data when sanitation vehicles are working in this operating mode.

[0056] Then, the operational quality data of the sanitation vehicle after working in this operation mode is obtained, and the operational quality coefficient ZP is evaluated based on the operational quality data; the operational quality data includes the congestion caused to other vehicles during the working period and the cleaning of road garbage;

[0057] Since not all sanitation vehicle operation modes are reasonable, the data acquisition module has a preset quality threshold. When the operation quality coefficient ZP of the sanitation vehicle is less than the preset quality threshold, the corresponding operation mode is not considered.

[0058] The operation quality coefficient ZP is compared with the preset quality threshold. If the operation quality coefficient ZP is greater than the preset quality threshold, the corresponding operation mode of the sanitation vehicle and the potential related operation data are stored in the database as explicit operation mode data.

[0059] In one embodiment of the present invention, the job quality coefficient ZP is evaluated based on job quality data; the specific evaluation process is as follows:

[0060] When the sanitation vehicle is working in operation mode, the vehicle speed, the distance between the sanitation vehicle and surrounding vehicles, and the speed of surrounding vehicles are collected; with the sanitation vehicle as the center, all vehicles traveling within a radius r1 are marked as associated vehicles; where r1 is a preset value.

[0061] The number of associated vehicles is counted as L1; the speed of the sanitation vehicles is marked as CVt;

[0062] The distance between the associated vehicle and the sanitation vehicle is denoted as vehicle spacing CLI; the speed of the associated vehicle is denoted as CVi; where i = 1, 2, ..., n; CLI and CVi are in one-to-one correspondence; i represents the i-th associated vehicle; using the formula... The vehicle speed deviation PZ is calculated.

[0063] The maximum vehicle spacing is marked as CLmax, and the minimum vehicle spacing is marked as CLmin; the spacing difference ratio Gb is calculated using the formula Gb=(CLmax-CLmin) / CLmin;

[0064] The average vehicle spacing GDz is calculated according to the average value calculation formula; the congestion coefficient Ds is calculated using the formula Ds=f×L1×(1+PZ) / [GDz×(1-Gb)×a1], where a1 is the preset coefficient factor and f is the preset balance coefficient.

[0065] Compare the congestion coefficient Ds with the preset congestion threshold; if the congestion coefficient Ds is greater than the preset congestion threshold, it means that the sanitation vehicle is causing congestion to other vehicles at this time.

[0066] The percentage of congestion time for sanitation vehicles is denoted as Tb, and the percentage of garbage cleaning area for sanitation vehicles is denoted as Mb; where Mb is obtained from the spatial distribution information of garbage before and after cleaning.

[0067] The work quality coefficient ZP is calculated using the formula ZP=(Mb×a2) / (Tb×a3), where a2 and a3 are preset coefficient factors;

[0068] The model creation module is used to analyze and obtain a reference model of the operation mode of sanitation vehicles based on AI deep learning recognition algorithms. The data source for the model creation module is data stored in the database. The specific analysis steps are as follows:

[0069] The explicit job pattern data obtained from the database will be used as the parameter training set, and the parameter training set will be split into training set, validation set and test set according to a preset ratio.

[0070] Establish an LSTM neural network model; wherein, the number of input nodes of the Long Short-Term Memory (LSTM) neural network is specified according to the number of input variables; set an appropriate number of hidden layer nodes and the number of output nodes representing the operating mode;

[0071] The training set, validation set, and test set are used as historical feature values ​​to input into the LSTM neural network model for model training. The model is then evaluated using a loss function to obtain the optimal operating mode reference model that minimizes the overall error of the training samples.

[0072] The above formulas are all numerical calculations after removing dimensions. The formulas are obtained by software simulation based on a large amount of data and are closest to the real situation. The preset parameters and preset thresholds in the formulas are set by those skilled in the art according to the actual situation or obtained by simulation based on a large amount of data.

[0073] Working principle of the invention:

[0074] The unmanned sanitation vehicle operation mode dynamic switching system consists of a waste sensing module, including a lidar and camera mounted on the sanitation vehicle, used to obtain environmental information around the vehicle and identify waste spatial distribution information through a built-in waste spatial distribution recognition algorithm; a road monitoring module, including roadside base stations and roadside sensors distributed on both sides of the road, where the roadside base stations are used to locate the sanitation vehicle's position and traffic flow; and the roadside sensors are used to collect real-time road weather data; and a central processing module, which integrates waste spatial distribution information, sanitation vehicle position, traffic flow, and road weather data into input data for an operation mode reference model, outputs the corresponding operation mode, realizes dynamic switching of unmanned sanitation vehicle operation modes, and improves the efficiency of sanitation vehicle operation.

[0075] The data acquisition module is used to acquire and review the original operation mode data. First, it acquires the operation modes of all sanitation vehicles in the system and analyzes and mines the potential related operation data of all sanitation vehicles. Then, it acquires the operation quality data of the sanitation vehicles after they work in the operation mode, and evaluates the operation quality coefficient ZP based on the operation quality data. If the operation quality coefficient ZP is greater than the preset quality threshold, the corresponding original operation mode data is stored in the database as explicit operation mode data. The model creation module is used to analyze and obtain the operation mode reference model of sanitation vehicles based on AI deep learning recognition algorithm. The data source for the analysis of the model creation module is the data stored in the database, so as to reduce the error of the operation mode reference model and improve the model accuracy.

[0076] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0077] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A dynamic switching system for the operation mode of an unmanned sanitation vehicle, characterized in that, It includes a waste sensing module, a central processing module, a road surface monitoring module, a mode switching module, a data acquisition module, a database, and a model creation module; The waste sensing module includes a lidar and a camera installed on the sanitation vehicle to obtain environmental information around the vehicle and to identify waste spatial distribution information through a built-in waste spatial distribution recognition algorithm. The lidar is used to sense whether there are obstacles around the vehicle, and the camera is used to collect video images of the vehicle's surroundings in real time. The road surface monitoring module includes roadside base stations and roadside sensors distributed on both sides of the road; the vehicle terminal is used to send the vehicle model to the roadside base station, sense the vehicle speed, sense the distance between the vehicle and surrounding vehicles, and combine the roadside base station to locate the sanitation vehicle's position and traffic flow. The roadside sensor is used to collect real-time road weather data and transmit the collected data to the central processing module through the roadside base station; The road meteorological data includes road surface temperature, road surface humidity, and road surface water or snow accumulation. The data acquisition module is used to acquire and review the original operation mode data, and store the approved original operation mode data as explicit operation mode data in the database; the explicit operation mode data refers to the original operation mode data where the operation quality coefficient ZP is greater than the preset quality threshold. The model creation module is used to use the explicit operation mode data as a parameter training set and analyze it based on AI deep learning recognition algorithm to obtain a reference model of the operation mode of sanitation vehicles. The central processing module is used to perform comprehensive analysis by combining information on the spatial distribution of garbage, the location of sanitation vehicles, traffic flow, road weather data, and the reference model of the operation mode, and then outputs the corresponding operation mode to the mode switching module to realize the dynamic switching of the operation mode of the unmanned sanitation vehicle; the operation mode includes the sweeping speed of the sanitation vehicle, the rotation speed of the sweeping brush, and the suction power of the suction nozzle. The specific working steps of the data acquisition module are as follows: First, the operating modes of all sanitation vehicles in the system are obtained, and the potential related operating data of all sanitation vehicles are analyzed and mined. The potential related operating data are represented by the spatial distribution information of garbage, the location of sanitation vehicles, traffic flow and road weather data when sanitation vehicles are working in this operating mode. Then, the operation quality data of the sanitation vehicle after working in this operation mode is obtained, and the operation quality coefficient ZP is evaluated based on the operation quality data; the operation quality data includes the congestion caused to other vehicles during the working period and the cleaning of road garbage; The operation quality coefficient ZP is compared with the preset quality threshold. If the operation quality coefficient ZP is greater than the preset quality threshold, the corresponding operation mode of the sanitation vehicle and the potential related operation data are stored in the database as explicit operation mode data. The work quality coefficient ZP is evaluated based on the aforementioned work quality data; specifically: When the sanitation vehicle is working in the aforementioned operating mode, the vehicle speed, the distance between the sanitation vehicle and surrounding vehicles, and the speed of surrounding vehicles are collected. Using the sanitation vehicle as the center, all vehicles traveling within a radius r1 are marked as associated vehicles; where r1 is a preset value; the number of associated vehicles is counted as L1. The speed of the sanitation vehicle is denoted as CVt; the distance between the associated vehicle and the sanitation vehicle is denoted as vehicle spacing CLI; the speed of the associated vehicle is denoted as CVi. Where i = 1, 2, ..., n; CLi and CVi correspond one-to-one; i represents the number of the i-th associated vehicle; Using formula The vehicle speed deviation PZ is calculated. The maximum vehicle spacing is marked as CLmax, and the minimum vehicle spacing is marked as CLmin; The spacing difference ratio Gb is calculated using the formula Gb=(CLmax-CLmin) / CLmin; The average vehicle spacing GDz is calculated using the average value calculation formula. The congestion coefficient Ds is calculated using the formula Ds=ƒ×L1×(1+PZ) / [GDz×(1-Gb)×a1], where a1 is a preset coefficient factor. ƒ is the preset balance coefficient; Compare the congestion coefficient Ds with the preset congestion threshold; if the congestion coefficient Ds is greater than the preset congestion threshold, it means that the sanitation vehicle is causing congestion to other vehicles at this time. The percentage of congestion time of the sanitation vehicles is Tb, and the percentage of garbage cleaning area of ​​the sanitation vehicles is Mb. The work quality coefficient ZP is calculated using the formula ZP=(Mb×a2) / (Tb×a3), where a2 and a3 are preset coefficient factors.

2. The unmanned sanitation vehicle operation mode dynamic switching system according to claim 1, characterized in that, The specific analysis steps of the model creation module are as follows: The explicit job pattern data obtained from the database will be used as the parameter training set, and the parameter training set will be split into training set, validation set and test set according to a preset ratio. Establish an LSTM neural network model; wherein, the number of input nodes of the Long Short-Term Memory (LSTM) neural network is specified according to the number of input variables; set an appropriate number of hidden layer nodes and the number of output nodes representing the operating mode; The training set, validation set, and test set are used as historical feature values ​​to input into the LSTM neural network model for model training. The model is then evaluated using a loss function to obtain the optimal operating mode reference model that minimizes the overall error of the training samples.

3. The unmanned sanitation vehicle operation mode dynamic switching system according to claim 1, characterized in that, The specific steps of the garbage spatial distribution identification algorithm include: Extract road surface image information from video images of the vehicle's surroundings; convert the road surface image information into a grayscale image; and convert the grayscale image into a standard image through image preprocessing. Obtain the total number of pixels in the standard image within the preset area and mark it as the area X1; Each pixel is identified, and the corresponding garbage pixels are identified; connected garbage pixels are merged to obtain the total number of garbage pixels; if the total number of garbage pixels is greater than a preset threshold, the area where the corresponding garbage pixel is located is marked as a garbage accumulation area. By integrating all the garbage accumulation areas, we can obtain information on the spatial distribution of garbage.

4. The unmanned sanitation vehicle operation mode dynamic switching system according to claim 3, characterized in that, Each pixel is identified, and the corresponding garbage pixels are identified; specifically: First, the gray values ​​of the pixels in the standard image are marked as H1; then, the gray values ​​of each pixel are compared with the preset standard gray values ​​to obtain the difference result, which is marked as C1; if C1 is greater than the difference threshold, the pixel is considered a garbage pixel.

5. The unmanned sanitation vehicle operation mode dynamic switching system according to claim 3, characterized in that, The image preprocessing includes Gaussian filtering, image segmentation, and image enhancement.

Images