A Smart Control Method for Environmentally Friendly Trash Cans

Abstract

This invention relates to the field of garbage bin collection and control technology, specifically to an intelligent control method for environmentally friendly garbage bins. The method includes: obtaining a time difference characterization value between the current garbage filling sequence and the target reference sequence based on DDTW matching data pairs between the current garbage filling sequence and the target reference sequence; aligning the current garbage filling sequence and the target reference sequence based on the time difference characterization value; and obtaining the target predicted overflow time of the intelligent garbage bin at the current monitoring time based on the alignment result and the difference between the current garbage filling sequence and the target reference sequence; and controlling the garbage collection of the intelligent garbage bin based on the target predicted overflow time. Furthermore, this invention controls the garbage collection of the intelligent garbage bin by predicting the time it will take for the intelligent garbage bin to reach overflow, thereby improving the timeliness and efficiency of garbage collection.

Description

Technical Field

[0001] This invention relates to the field of garbage bin collection and control technology, specifically to an intelligent control method for environmentally friendly garbage bins. Background Technology

[0002] With the rapid pace of urbanization, smart trash cans are gradually replacing traditional trash cans and are being used in urban public areas and residential communities. These smart trash cans typically utilize the Internet of Things (IoT), wireless communication, and big data analytics to achieve overflow detection, remote management, and intelligent scheduling. While current smart trash cans do have overflow detection capabilities—that is, they issue warnings when the remaining capacity reaches a preset threshold to remind staff to collect the trash—this method does not consider the amount of trash deposited during the time the garbage truck travels to the smart trash can. This can lead to problems such as untimely or ineffective trash collection, where the smart trash can is already overflowing before it is emptied, causing environmental pollution, health hazards, and a series of other issues. Therefore, improving the timeliness and effectiveness of smart trash collection has become an urgent problem to be solved. Summary of the Invention

[0003] To address the aforementioned problems, this invention provides an intelligent control method for environmentally friendly trash cans, the specific technical solution of which is as follows:

[0004] One embodiment of the present invention provides an intelligent control method for an environmentally friendly trash can, comprising the following steps:

[0005] Obtain the reference sequence for each historical monitoring unit time period of the smart trash can, and the current trash filling volume sequence at the current monitoring moment;

[0006] Based on the DDTW distance between any two reference sequences, all reference sequences are clustered to obtain various clusters. Reference clusters are obtained based on the distance between the current waste filling quantity sequence and each cluster. Based on the DDTW matching data pairs between the current waste filling quantity sequence and the target reference sequence, the time difference characterization value between the current waste filling quantity sequence and the target reference sequence is obtained. The sequences in the reference clusters are all target reference sequences. The current waste filling quantity sequence and the target reference sequence are aligned based on the time difference characterization value. Based on the alignment result and the difference between the current waste filling quantity sequence and the target reference sequence, the target predicted overflow time of the smart trash can at the current monitoring time is obtained.

[0007] Based on the predicted overflow time, the garbage collection of the smart trash can is controlled.

[0008] Beneficial effects: This invention first obtains the reference sequence corresponding to each historical monitoring unit time period of the smart trash can and the current trash filling volume sequence corresponding to the current monitoring time; then, based on the DDTW distance between any two reference sequences, all reference sequences are clustered to obtain various clusters, and reference clusters are obtained based on the distance between the current trash filling volume sequence and each cluster; based on the DDTW matching data pair between the current trash filling volume sequence and the target reference sequence, the time difference characterization value between the current trash filling volume sequence and the target reference sequence is obtained; based on the time difference characterization value, the current trash filling volume sequence and the target reference sequence are aligned, and based on the alignment result and the difference between the current trash filling volume sequence and the target reference sequence, the target predicted overflow time of the smart trash can at the current monitoring time is obtained; finally, the trash collection of the smart trash can is controlled based on the target predicted overflow time. Furthermore, this invention controls the garbage collection of smart trash cans by predicting the time it takes for the smart trash cans to overflow, which can improve the timeliness and effectiveness of garbage collection, thus minimizing the possibility of garbage not being collected in time when the trash cans are overflowing. Attached Figure Description

[0009] To more clearly illustrate the technical solutions and advantages 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.

[0010] Figure 1 This is a flowchart of an intelligent control method for an environmentally friendly trash can according to the present invention. Detailed Implementation

[0011] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention are within the protection scope of the embodiments of the present invention.

[0012] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art.

[0013] This embodiment provides an intelligent control method for environmentally friendly trash cans, which is described in detail below:

[0014] like Figure 1 As shown, the intelligent control method for this environmentally friendly trash can includes the following steps:

[0015] Step S001: Obtain the reference sequence corresponding to each historical unit monitoring time period of the smart trash can and the current trash filling volume sequence corresponding to the current monitoring time.

[0016] The main purpose of this embodiment is to improve the timeliness and efficiency of smart trash can collection by predicting the time it takes for the smart trash can to reach full capacity. The smart trash can in this embodiment is equipped with sensors that transmit real-time remaining capacity data and location information to a remote terminal via an internal IoT module. The remote terminal then analyzes and calculates this data to predict the time it takes for the smart trash can to reach full capacity. Furthermore, for ease of understanding, this embodiment will use the control process of smart trash can collection on any collection route in any city as an example, and this collection route will be designated as the collection route to be analyzed. All subsequent collection routes belong to the same route, and all subsequent smart trash cans are on the same route. A collection route typically consists of multiple trash cans, and smart trash cans located further forward on the route are collected first.

[0017] This embodiment will next predict the time it takes for each smart trash can on the collection route to reach full capacity. Since the method for predicting the time it takes for each smart trash can on the collection route to reach full capacity is the same, this embodiment will take the prediction process of the time it takes for any smart trash can A on the collection route to reach full capacity as an example. That is, the purpose of this embodiment is to obtain the target predicted full capacity time of smart trash can A at the current monitoring time. Before obtaining the target predicted full capacity time of smart trash can A at the current monitoring time, it is necessary to obtain the reference sequence corresponding to smart trash can A in each historical unit monitoring time period and the current trash filling volume sequence corresponding to smart trash can A at the current monitoring time. Subsequently, the prediction of the full capacity time will be realized based on the reference sequence and the current trash filling volume sequence. The specific process of obtaining the reference sequence corresponding to smart trash can A in each historical unit monitoring time period and the current trash filling volume sequence corresponding to smart trash can A at the current monitoring time is as follows:

[0018] First, obtain each complete day preceding the current monitoring time, and treat each complete day as a historical unit monitoring period. That is, the time period from the start to the end of a day is considered a historical unit monitoring period, with a duration of 24 hours. For any given historical unit monitoring period, obtain the waste filling amount corresponding to each monitoring moment within that period. The time sequence formed by the waste filling amounts corresponding to all monitoring moments within that historical unit monitoring period is denoted as the reference sequence for smart trash can A within that historical unit monitoring period. In other words, the reference sequence for smart trash can A within that historical unit monitoring period is composed of all the waste filling amounts obtained within that historical unit monitoring period.

[0019] Then, the current day is recorded as the current unit monitoring time period, and the amount of garbage filled at each monitoring time in the time period from the start time of the current unit monitoring time period to the current monitoring time is obtained. The time sequence consisting of the amount of garbage filled at all monitoring times in the time period from the start time of the current unit monitoring time period to the current monitoring time is recorded as the current garbage filling amount sequence of smart trash can A at the current monitoring time. The current monitoring time belongs to the current day, that is, the current garbage filling amount sequence of smart trash can A at the current monitoring time is composed of all the garbage filling amounts obtained in the time period from the start time of the current unit monitoring time period to the current monitoring time.

[0020] In this embodiment, the method for obtaining the amount of waste filling corresponding to each monitoring time is as follows: For any monitoring time: First, the monitoring time adjacent to this monitoring time and located before this monitoring time in time is recorded as the adjacent historical monitoring time of this monitoring time. Then, the remaining capacity in smart trash can A at this monitoring time and the remaining capacity in smart trash can A at the adjacent historical monitoring time of this monitoring time are obtained. If smart trash can A has not been emptied between the adjacent historical monitoring time of this monitoring time and this monitoring time, then the remaining capacity in smart trash can A at the adjacent historical monitoring time of this monitoring time minus the remaining capacity in smart trash can A at this monitoring time is taken as the amount of waste filling corresponding to this monitoring time. If smart trash can A has not been emptied between the adjacent historical monitoring time of this monitoring time and this monitoring time, then the remaining capacity in smart trash can A at this monitoring time is taken as the amount of waste filling corresponding to this monitoring time. If trash can A is being emptied, then the remaining capacity of trash can A before emptying is obtained, and the sum of the first filling amount difference and the second filling amount difference is recorded as the trash filling amount corresponding to the monitoring time. The first filling amount difference is the result of subtracting the remaining capacity of trash can A before emptying from the remaining capacity of trash can A at the adjacent historical monitoring time of the monitoring time. The first filling amount difference is the result of subtracting the remaining capacity of trash can A at the monitoring time from the total capacity of trash can A. If there is no adjacent historical monitoring time, then the result of subtracting the remaining capacity of trash can A at the monitoring time from the total capacity of trash can A is also used as the trash filling amount corresponding to the monitoring time. In addition, the method for obtaining the remaining capacity of trash can A is a known technology, so it will not be described in detail in this embodiment.

[0021] In addition, this embodiment sets the time interval between adjacent monitoring times based on the monitoring frequency typically set for smart trash cans used in different scenarios. For example, if the smart trash can in this embodiment is a public trash can in the city, the time interval between adjacent monitoring times is generally set to between 15 and 30 minutes. However, it is required that the time interval between all adjacent monitoring times in this embodiment be consistent, and the time corresponding to the monitoring times on different days should also be consistent. For example, if 9:00 AM is a monitoring time, then a trash filling amount needs to be obtained at 9:00 AM every day.

[0022] Therefore, this embodiment can obtain the reference sequence corresponding to each historical unit monitoring time period and the current garbage filling volume sequence corresponding to the current monitoring time through the above process.

[0023] Step S002: Based on the DDTW distance between any two reference sequences, cluster all reference sequences to obtain various clusters. Reference clusters are obtained based on the distance between the current waste filling quantity sequence and each cluster. The time difference representation value between the current waste filling quantity sequence and the target reference sequence is obtained based on the DDTW matching data pairs between the current waste filling quantity sequence and the target reference sequence. All sequences in the reference clusters are target reference sequences. The current waste filling quantity sequence and the target reference sequence are aligned based on the time difference representation value. Based on the alignment result and the difference between the current waste filling quantity sequence and the target reference sequence, the target predicted overflow time of the smart trash can at the current monitoring time is obtained.

[0024] This embodiment, after obtaining the reference sequence corresponding to each historical monitoring unit time period and the current waste filling volume sequence corresponding to the current monitoring time for smart trash can A, analyzes the reference sequence and the current waste filling volume sequence to obtain the target predicted overflow time of smart trash can A at the current monitoring time. Before obtaining the target predicted overflow time of smart trash can A at the current monitoring time, it is necessary to perform clustering based on the similarity between the reference sequences, and obtain the target reference cluster corresponding to the current monitoring time based on the clustering results. That is, this embodiment will next utilize the derivative dynamic time warping algorithm (Derivative DTW). The DDTW distance between any two reference sequences is obtained by using DDTW. The smaller the DDTW distance between the reference sequences, the higher the similarity between the two reference sequences, and the easier it is to classify them into the same cluster during subsequent clustering. Moreover, in this embodiment, the DDTW distance between any two reference sequences is the DDTW distance between the normalized sequences of any two reference sequences. That is, the DDTW distance between reference sequence 1 and reference sequence 2 is the DDTW distance between the normalized sequences of reference sequence 1 and reference sequence 2. The process of calculating the DDTW distance between two sequences is a well-known technique, so it will not be described in this embodiment. Furthermore, the normalized sequence of the reference sequence refers to the sequence of data obtained after normalizing all data in the reference sequence. In this embodiment, the y-th data in the normalized sequence of any sequence is the result of normalizing the y-th data in that sequence. For example, the y-th data in the normalized sequence of a certain sequence is Norm(C), where C is the y-th data in that sequence, that is, Norm(C) is the normalized value of the y-th data in that sequence, and Norm() is the normalization function. The purpose of normalizing the obtained waste filling amount is to avoid deviations in subsequent prediction and analysis due to the different amounts of waste disposed of at different times, even though the disposal patterns are similar.

[0025] After obtaining the DDTW distance between any two reference sequences, the mean-moving clustering algorithm is used to cluster all reference sequences to obtain individual clusters. Then, based on the distance between the current waste filling amount sequence and each cluster, the reference cluster corresponding to the current monitoring time is obtained, and subsequent predictions will be made using sequences in the reference clusters. The specific process for obtaining the reference cluster corresponding to the current monitoring time is as follows: First, the cluster center of each cluster is obtained, and the sequence corresponding to the cluster center of each cluster is recorded as the center sequence of the corresponding cluster. Then, the DDTW distance between the normalized sequence of the center sequence of each cluster and the normalized sequence of the current waste filling amount sequence is calculated, and the cluster corresponding to the smallest DDTW distance is taken as the reference cluster corresponding to the current monitoring time. All reference sequences in the reference cluster are recorded as target reference sequences. Finally, the new sequence generated by averaging all reference sequences in the cluster using the DBA algorithm is the center sequence of the corresponding cluster. Furthermore, the reason for obtaining reference clusters is that the amount of garbage disposed of in trash cans located in different places on different days varies. For example, if trash cans are placed downstairs in an office building, the amount or frequency of garbage disposal may be higher during working hours, such as Monday and Friday when the number of office workers is larger. On Saturday and Sunday, the number of office workers may be smaller, and the amount or frequency of garbage disposal may be lower. Clustering can group together reference sequences corresponding to historical monitoring time periods with similar disposal patterns. Since the historical monitoring time periods corresponding to the sequences in the reference clusters are similar to the monitoring time periods of the current garbage filling sequence, the difference between the time when smart trash can A reaches fullness and the current monitoring time can be accurately predicted based on the sequences in the reference clusters. In other words, the target predicted fullness time of smart trash can A at the current monitoring time can be accurately obtained.

[0026] After obtaining the reference cluster, all DDTW matching data pairs between the current garbage filler sequence and each target reference sequence in the reference cluster are obtained. A DDTW matching data pair between any two sequences refers to the matching data pair generated by backtracking the path after finding the optimal path through dynamic programming when calculating the DDTW distance between any two sequences. Since the process of obtaining DDTW matching data pairs when calculating the DDTW distance between any two sequences is well-known, it will not be described in this embodiment. However, since this embodiment calculates the DDTW distance between normalized sequences, if the k-th data in the normalized sequence of the current garbage filler sequence and the z-th data in the normalized sequence of a target reference sequence form a matching data pair during backtracking, then the combination of the k-th data in the current garbage filler sequence and the z-th data in the target reference sequence constitutes a DDTW matching data pair between the current garbage filler sequence and the target reference sequence. Then, based on all DDTW matching data pairs between the current garbage fill sequence and any target reference sequence G in the reference cluster, the time difference representation value between the current garbage fill sequence and the target reference sequence G is obtained. That is, the process of obtaining the time difference representation value between the current garbage fill sequence and the target reference sequence G is as follows:

[0027] First, obtain all DDTW matching data pairs between the current garbage filler sequence and the target reference sequence G. Denote the set of all DDTW matching data pairs between the current garbage filler sequence and the target reference sequence G as the matching data pair set corresponding to the target reference sequence G. Then, obtain the time difference corresponding to each DDTW matching data pair in the matching data pair set. Use the average of the time differences corresponding to all DDTW matching data pairs in the matching data pair set as the time difference representation value between the current garbage filler sequence and the target reference sequence G. Furthermore, the time difference corresponding to any DDTW matching data pair in the matching data pair set corresponding to the target reference sequence G is the time difference corresponding to the first data point in that DDTW matching data pair. The time difference is calculated by subtracting the time of the monitoring time corresponding to the second data in the corresponding DDTW matching data pair from the monitoring time of the target reference sequence G. The first data in each DDTW matching data pair in the matching data pair set corresponding to the target reference sequence G belongs to the current garbage filling volume sequence, and the second data in each DDTW matching data pair in the matching data pair set corresponding to the target reference sequence G belongs to the target reference sequence G. Furthermore, the date is not considered when calculating the time difference. That is, if the monitoring time corresponding to the first data in a DDTW matching data pair is 7:00 AM and the monitoring time corresponding to the second data is 6:00 AM, then the time difference for the DDTW matching data pair is the result of subtracting 6:00 AM from 7:00 AM. The reason for calculating the time difference is that although the overall waste disposal pattern or characteristics of the current unit monitoring time period to which the current waste filling volume sequence belongs are similar to those of the historical unit monitoring time period corresponding to the target reference sequence G, in actual disposal, natural factors such as weather may affect the waste disposal pattern in the current unit monitoring time period, causing a certain time deviation or misalignment between the waste disposal pattern in the current unit monitoring time period and the waste disposal pattern in the historical unit monitoring time period corresponding to the target reference sequence G. For example, if 8:00 AM is the peak time for waste disposal in the historical unit monitoring time period corresponding to the target reference sequence G, but the peak time for waste disposal is delayed due to rain in the current unit monitoring time period, becoming 8:30 AM, but both of these peak times belong to the early disposal peak times, that is, these two peak times have the same attribute, then there is a time difference or misalignment between these two peak times with the same attribute on different dates. The existence of this time difference or misalignment will cause prediction bias when using the target reference sequence G to predict the time until the smart trash can A is full. Therefore, it is necessary to perform time alignment by calculating the time difference characterization value, and make predictions based on the aligned results.

[0028] Furthermore, if the time difference between the current waste filling volume sequence and the target reference sequence G is equal to 0, it means that the waste disposal change pattern within the current unit monitoring time period corresponding to the current waste filling volume sequence is consistent with the waste disposal change pattern within the historical unit monitoring time period corresponding to the target reference sequence G in time, and no alignment is required. For example, if the time difference between the current waste filling volume sequence and the target reference sequence G is equal to 0, and the first disposal peak period within the historical unit monitoring time period corresponding to the target reference sequence G is 9:00 AM, then the first disposal peak period within the current unit monitoring time period is generally also 9:00 AM. If the time difference between the current waste filling volume sequence and the target reference sequence G is greater than 0, it indicates that the waste disposal pattern in the current monitoring period lags behind the waste disposal pattern in the historical monitoring period corresponding to the target reference sequence G, and alignment is required. For example, if the time difference between the current waste filling volume sequence and the target reference sequence G is greater than 0, the time difference between the current waste filling volume sequence and the target reference sequence G is 30 minutes, and the first peak disposal period in the historical monitoring period corresponding to the target reference sequence G is 9:00 AM, then the first peak disposal period in the current monitoring period may be 9:30 AM. If the time difference between the current waste filling volume sequence and the target reference sequence G is less than 0, it indicates that the waste disposal change pattern in the historical unit monitoring time period corresponding to the target reference sequence G lags behind the waste disposal change pattern in the current unit monitoring time period and needs to be aligned. For example, if the time difference between the current waste filling volume sequence and the target reference sequence G is less than 0, the time difference between the current waste filling volume sequence and the target reference sequence G is 30 minutes, and the first disposal peak in the historical unit monitoring time period corresponding to the target reference sequence G is 9:00 AM, then the first disposal peak in the current unit monitoring time period may be 8:30 AM.

[0029] After obtaining the time difference characterization value, the current waste filling volume sequence is aligned with the target reference sequence based on the obtained time difference characterization value. Based on the alignment result and the difference between the current waste filling volume sequence and the target reference sequence, the target predicted overflow time of smart trash can A at the current monitoring time is obtained. The specific process for obtaining the target predicted overflow time of smart trash can A at the current monitoring time is as follows:

[0030] First, based on the DDTW distance between the current garbage filler sequence and each target reference sequence, and the matching data pair sets corresponding to the current garbage filler sequence and each target reference sequence, the difference feature values ​​between the current garbage filler sequence and each target reference sequence are obtained. Then, a negative correlation mapping is performed on the difference feature values ​​between the current garbage filler sequence and each target reference sequence, and the mapping result is used as the initial prediction confidence under the corresponding target reference sequence. That is, the result obtained by negatively correlating the difference feature values ​​between the current garbage filler sequence and any target reference sequence is the initial prediction confidence under that target reference sequence, and the initial prediction confidence under any target reference sequence is exp. (-L), where L is the difference feature value between the current garbage filling sequence and the target reference sequence, and exp() is an exponential function with a base of constant e; then, the sum of the initial prediction confidence of all target reference sequences in the reference cluster is obtained and recorded as the comprehensive prediction confidence; the ratio of the initial prediction confidence to the comprehensive prediction confidence of each target reference sequence is obtained and recorded as the target prediction confidence of the corresponding target reference sequence; and the larger the target prediction confidence of the target reference sequence, the higher the confidence of the initial prediction overflow duration obtained based on the corresponding target reference sequence, and the higher the participation of the initial prediction overflow duration with higher confidence in obtaining the target prediction overflow duration.

[0031] Next, based on the time difference representation values ​​between the current waste filling volume sequence and each target reference sequence, the current waste filling volume sequence is aligned with each target reference sequence, and the initial predicted overflow duration under each target reference sequence is obtained based on the alignment results. Then, the product of the initial predicted overflow duration under each target reference sequence and the target prediction confidence under the corresponding target reference sequence is obtained, and used as the weighted predicted overflow duration under the corresponding target reference sequence. That is, the weighted predicted overflow duration under any target reference sequence is the product of the initial predicted overflow duration under that target reference sequence and the target prediction confidence under that target reference sequence. After that, the sum of the weighted predicted overflow durations under all target reference sequences in the reference cluster is obtained and recorded as the target predicted overflow duration of smart trash can A at the current monitoring time.

[0032] In this embodiment, the specific process of obtaining the difference feature values ​​between the current garbage filling sequence and each target reference sequence based on the DDTW distance between the current garbage filling sequence and each target reference sequence, and the matching data pair set corresponding to the current garbage filling sequence and each target reference sequence, is as follows:

[0033] For any target reference sequence G, firstly, obtain the absolute value of the difference between the normalized values ​​of the two data points in each DDTW matching data pair in the matching data pair set corresponding to the target reference sequence G, and denot it as the data difference value corresponding to the corresponding DDTW matching data pair. That is, the data difference value corresponding to any DDTW matching data pair is the absolute value of the difference between the normalized value of the first data point and the normalized value of the second data point in the DDTW matching data pair. The normalization function used here is the Norm() function. Then, obtain the mean of the data difference values ​​corresponding to all DDTW matching data pairs in the matching data pair set corresponding to the target reference sequence G, and use it as the mean of the data difference. Then, the absolute value of the difference between the data difference value and the mean data difference for each DDTW matching data pair in the matching data pair set corresponding to the target reference sequence G is obtained, and denoted as the first difference value for the corresponding DDTW matching data pair. The mean of the first difference values ​​for all DDTW matching data pairs in the matching data pair set corresponding to the target reference sequence G is obtained and denoted as the feature mean. Next, the result of multiplying the DDTW distance between the current garbage filling sequence and the target reference sequence G by the feature mean is obtained, and this result is used as the difference feature value between the current garbage filling sequence and the target reference sequence G. That is, the difference feature value between the current garbage filling sequence and the target reference sequence G is... ,in, M represents the DDTW distance between the current garbage fill sequence and the target reference sequence G, and M is the total number of DDTW matching data pairs in the matching data pair set corresponding to the target reference sequence G. Let be the data difference value corresponding to the m-th DDTW matching data pair in the set of matching data pairs corresponding to the target reference sequence G. This represents the mean difference in the data.

[0034] In addition, when The smaller the value, the higher the similarity between the current garbage filling sequence and the target reference sequence G. Therefore, the reliability and accuracy of the subsequent prediction of the time to overflow based on the target reference sequence G—that is, the initial predicted overflow time under the obtained target reference sequence G—should be higher. The smaller the value, the smaller the data difference among all DDTW matching data pairs obtained based on the current garbage fill volume sequence and the target reference sequence G, or the better the stability among all DDTW matching data pairs obtained based on the current garbage fill volume sequence and the target reference sequence G. Therefore, the reliability and accuracy of the subsequent prediction of the time to overflow based on the target reference sequence G, i.e., the initial predicted overflow time under the obtained target reference sequence G, should be higher; and also because... and The smaller the value of the difference between the current garbage filling sequence and the target reference sequence G, the smaller the value of the difference between the current garbage filling sequence and the target reference sequence G. Therefore, if the value of the difference between the current garbage filling sequence and the target reference sequence G is smaller, it indicates that the reliability and accuracy of the initial predicted overflow duration under the target reference sequence G should be higher. Conversely, if the value of the difference between the current garbage filling sequence and the target reference sequence G is larger, it indicates that the reliability and accuracy of the initial predicted overflow duration under the target reference sequence G should be lower.

[0035] In this embodiment, the process of aligning the current garbage filling sequence with each target reference sequence based on the time difference representation value between the current garbage filling sequence and each target reference sequence, and obtaining the initial predicted overflow duration under each target reference sequence based on the alignment result, is as follows:

[0036] For any target reference sequence G: Based on the time difference representation value between the current waste filling volume sequence and the target reference sequence G, align the current waste filling volume sequence with the target reference sequence G to obtain the subsequence to be analyzed corresponding to the target reference sequence G. This subsequence to be analyzed is denoted as the subsequence to be analyzed H, where the subsequence to be analyzed H is the predicted waste filling volume corresponding to a future monitoring time within the current unit monitoring time period. Furthermore, within the current unit monitoring time period, all monitoring times following the current monitoring time are considered future monitoring times within the current unit monitoring time period. Then, when... The remaining capacity in the smart trash can at the previous monitoring time is recorded as the current remaining capacity. Then, it is determined whether the result of subtracting the second accumulated value from the current remaining capacity is not greater than a preset remaining capacity threshold, and whether the result of subtracting the first accumulated value from the current remaining capacity is greater than the preset remaining capacity threshold. If the result of subtracting the second accumulated value from the current remaining capacity is not greater than the preset remaining capacity threshold, but the result of subtracting the first accumulated value from the current remaining capacity is greater than the preset remaining capacity threshold, then the time span between the monitoring time corresponding to the first data point in the subsequence H to be analyzed and the monitoring time corresponding to the b data points is used as the initial value under the target reference sequence G. The predicted overflow duration refers to the predicted duration from the current monitoring time when the smart trash can reaches full capacity. The first accumulated value is the result of accumulating the first b-1 data points in the subsequence H to be analyzed, and the second accumulated value is the result of accumulating the first b data points in the current subsequence H to be analyzed, where b is greater than 1. The preset remaining capacity threshold is generally set to 0, meaning that when the remaining capacity in the trash can is 0, it indicates that the trash can has reached full capacity. For example, if the monitoring time corresponding to the first data point in the subsequence H to be analyzed is the previous monitoring time period corresponding to the historical unit monitoring time period of the target reference sequence G... At 9:00 AM, the monitoring time corresponding to the b-th data in the subsequence H to be analyzed is 10:00 AM in the historical unit monitoring time period corresponding to the target reference sequence G. Then, the time span between the monitoring time corresponding to the first data in the subsequence H to be analyzed and the monitoring time corresponding to the b-th data is 1 hour. This indicates that the initial predicted overflow duration based on the target reference sequence G is 1 hour. Moreover, if the current monitoring time is 3:00 PM in the current unit monitoring time period, then the prediction result obtained by the target reference sequence G is that the smart trash can may reach full capacity at 4:00 PM in the current unit monitoring time period.

[0037] In this embodiment, the specific process of aligning the current waste filling volume sequence with the target reference sequence G based on the time difference characterization value between the current waste filling volume sequence and the target reference sequence G, and obtaining the subsequence to be analyzed corresponding to the target reference sequence G, is as follows:

[0038] First, let T be the time corresponding to the current monitoring time, and let t be the time difference between the current garbage filling sequence and the target reference sequence G. Then, determine whether the time difference between the current garbage filling sequence and the target reference sequence G is 0. If it is, it means that there is no need to align the current garbage filling sequence with the target reference sequence G. It also means that the garbage filling volume after T in the target reference sequence G is the predicted garbage filling volume corresponding to the future monitoring time in the current unit monitoring time period. Then, the time T is obtained in the historical unit monitoring time period corresponding to the target reference sequence G and recorded as the feature monitoring time in the historical unit monitoring time period corresponding to the target reference sequence G. Continue to determine whether the time difference representation value between the current garbage filling volume sequence and the target reference sequence G is less than 0. If it is, it indicates that the current garbage filling volume sequence needs to move to the right by a time length of t relative to the target reference sequence G. It also indicates that the garbage filling volume corresponding to the monitoring time in the target reference sequence G that is after (T+t) is the predicted garbage filling volume corresponding to the future monitoring time in the current unit monitoring time period. Then, the time (T+t) is obtained in the historical unit monitoring time period corresponding to the target reference sequence G and recorded as the feature monitoring time in the historical unit monitoring time period corresponding to the target reference sequence G. Next, determine if the time difference between the current garbage filling sequence and the target reference sequence G is greater than 0. If so, it indicates that the current garbage filling sequence has shifted to the left by a time length of t relative to the target reference sequence G. This also indicates that the garbage filling amounts corresponding to the monitoring time (Tt) in the target reference sequence G are all predicted garbage filling amounts corresponding to future monitoring times within the current unit monitoring time period. In this case, the time (Tt) is obtained from the historical unit monitoring time period corresponding to the target reference sequence G and recorded as the characteristic monitoring time within the historical unit monitoring time period corresponding to the target reference sequence G. Furthermore, if the time difference between the current garbage filling sequence and the target reference sequence G is less than 0, and the current monitoring time is 8:00 AM with t being 30 minutes, then the characteristic monitoring time within the historical unit monitoring time period corresponding to the target reference sequence G is 8:30 AM.

[0039] Then, within the historical unit monitoring time period corresponding to the target reference sequence G, the sequence consisting of the waste filling amount corresponding to all monitoring times that are later than the characteristic monitoring time is recorded as the subsequence to be analyzed corresponding to the target reference sequence G.

[0040] For example, if t is 30 minutes, the time interval between adjacent monitoring times is 10 minutes, and the current monitoring time corresponds to 8:00 AM, then after shifting the current waste filling volume sequence to the left relative to the target reference sequence G by time length t, the v-th data in the current waste filling volume sequence is aligned with the (v-3)-th data in the target reference sequence G. If the waste filling volume corresponding to 8:00 AM in the historical unit monitoring time period of the target reference sequence G is the U-th data in the target reference sequence G, then the (U-3+x)-th data in the target reference sequence G is the predicted data corresponding to the x-th future monitoring time in the current unit monitoring time period. That is, if the current monitoring time corresponds to 8:00 AM and the time interval between adjacent monitoring times is 10 minutes, then the first future monitoring time in the current unit monitoring time period is 8:00 AM. The garbage filling amount corresponding to the time of 7:40 in the target reference sequence G is the predicted garbage filling amount corresponding to the first future monitoring time in the current unit monitoring time period. In other words, if the characteristic monitoring time in the historical unit monitoring time period corresponding to the target reference sequence G is 6:00 PM, then the sequence consisting of the garbage filling amounts corresponding to all monitoring times after 6:00 PM in the historical unit monitoring time period corresponding to the target reference sequence G is denoted as the subsequence to be analyzed corresponding to the target reference sequence G. That is, if the monitoring time corresponding to the Jth data in the target reference sequence G is the characteristic monitoring time, then the time series consisting of the remaining data after the Jth data in the target reference sequence G and belonging to the target reference sequence G is the subsequence to be analyzed corresponding to the target reference sequence G.

[0041] Therefore, this embodiment obtains the target predicted overflow time of smart trash can A at the current monitoring time through the above process.

[0042] Step S003: Based on the predicted overflow time, control the garbage collection of the smart trash can.

[0043] This embodiment will next control the garbage collection of smart trash cans on the collection route based on the target predicted overflow time of all smart trash cans on the collection route at the current monitoring time. The specific process is as follows:

[0044] First, the garbage truck responsible for collecting garbage along the route to be analyzed is identified and designated as the target garbage truck. Then, the estimated travel time of the target garbage truck to each smart trash can along the route is obtained. It is then determined whether the estimated travel time of the target garbage truck to each smart trash can along the route is less than the predicted overflow time of the corresponding smart trash can at the current monitoring time. If so, it is determined that the target garbage truck does not need to immediately depart to collect garbage from the smart trash cans along the route at the current monitoring time. Otherwise, an overflow warning is issued, and it is determined that the target garbage truck needs to immediately depart to collect garbage from the smart trash cans along the route in sequence at the current monitoring time. That is, if the garbage truck travels... If the estimated time to any smart trash can on the analyzed collection route is not less than the target predicted overflow time of the corresponding smart trash can at the current monitoring time, then it is determined that the target collection vehicle must immediately depart at the current monitoring time to collect the smart trash cans on the analyzed collection route in sequence. The collection process is as follows: if the smart trash cans on the analyzed collection route are W1, W2, W3, and W4 in sequence, then the target collection vehicle will first arrive at W1 for collection, then depart from W1 to W2 for collection, then depart from W2 to W3 for collection, and then depart from W3 to W4 for collection. After collection, the collection vehicle will generally transport the trash to a landfill, waste incineration plant, composting plant, etc. for processing.

[0045] In this embodiment, the process of obtaining the estimated time for the target waste collection vehicle to travel to each smart trash can on the waste collection route to be analyzed is as follows: each time the target waste collection vehicle collects waste from the smart trash cans on the waste collection route to be analyzed, the time taken for the target waste collection vehicle to travel to the f-th smart trash can on the waste collection route to be analyzed is obtained and recorded as the historical time corresponding to the f-th smart trash can. The set of all historical times corresponding to the f-th smart trash can is recorded as the set of times to be analyzed corresponding to the f-th smart trash can. The median of the set of times to be analyzed is used as the estimated time for the target waste collection vehicle to travel to the f-th smart trash can on the waste collection route to be analyzed. As another implementation method, the mean of the set of times to be analyzed corresponding to the f-th smart trash can can also be used as the estimated time for the target waste collection vehicle to arrive at the f-th smart trash can on the waste collection route to be analyzed. For example: If the garbage collection route to be analyzed was collected 3 times before the current monitoring time, and the smart trash cans on the route to be analyzed are W1, W2, W3, and W4 respectively, and the time taken for the target garbage truck to reach smart trash can W2 during the first collection is 30 minutes, the time taken for the target garbage truck to reach smart trash can W2 during the second collection is 30 minutes, and the time taken for the target garbage truck to reach smart trash can W2 during the third collection is 32 minutes, then the estimated time for the target garbage truck to travel to smart trash can W2 is 30 minutes. Moreover, the time taken for the target garbage truck to travel to W3 during collection refers to the time taken for the target garbage truck to travel from the beginning to W1, collect the garbage in W1, then travel from W1 to W2, collect the garbage in W1, and then travel from W2 to W3.

[0046] Thus, this embodiment completes the control of waste collection from the smart trash cans on the analyzed collection route.

[0047] In summary, this embodiment first obtains the reference sequence corresponding to each historical monitoring unit time period of the smart trash can and the current trash filling volume sequence corresponding to the current monitoring time. Then, based on the DDTW distance between any two reference sequences, all reference sequences are clustered to obtain various clusters. Reference clusters are obtained based on the distance between the current trash filling volume sequence and each cluster. Based on the DDTW matching data pairs between the current trash filling volume sequence and the target reference sequence, the time difference representation value between the current trash filling volume sequence and the target reference sequence is obtained. The current trash filling volume sequence and the target reference sequence are aligned based on the time difference representation value. Based on the alignment result and the difference between the current trash filling volume sequence and the target reference sequence, the target predicted overflow time of the smart trash can at the current monitoring time is obtained. Finally, the trash collection of the smart trash can is controlled based on the target predicted overflow time. Furthermore, this embodiment controls the garbage collection of smart trash cans by predicting the time it takes for the smart trash cans to overflow, which can improve the timeliness and effectiveness of garbage collection, thus minimizing the possibility of garbage not being collected in time when the trash cans are overflowing.

[0048] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A smart control method for an environmentally friendly trash can, characterized in that, The method includes the following steps: Obtain the reference sequence for each historical monitoring unit time period of the smart trash can, and the current trash filling volume sequence at the current monitoring moment; Based on the DDTW distance between any two reference sequences, all reference sequences are clustered to obtain various clusters. Reference clusters are obtained based on the distance between the current waste filling quantity sequence and each cluster. Based on the DDTW matching data pairs between the current waste filling quantity sequence and the target reference sequence, the time difference characterization value between the current waste filling quantity sequence and the target reference sequence is obtained. The sequences in the reference clusters are all target reference sequences. The current waste filling quantity sequence and the target reference sequence are aligned based on the time difference characterization value. Based on the alignment result and the difference between the current waste filling quantity sequence and the target reference sequence, the target predicted overflow time of the smart trash can at the current monitoring time is obtained. Based on the predicted overflow time, the garbage collection of the smart trash can is controlled.

2. The intelligent control method for an environmentally friendly trash can as described in claim 1, characterized in that, Methods for obtaining the reference sequence and the current garbage fill volume sequence include: Each complete day preceding the current monitoring time is considered as a historical unit monitoring period. For any historical unit monitoring period, the sequence of garbage filling amounts corresponding to all monitoring times within the historical unit monitoring period is recorded as the reference sequence of the smart trash can for that historical unit monitoring period. The current day is designated as the current monitoring time period. The sequence of waste filling amounts corresponding to each monitoring moment within the time period from the start of the current monitoring time period to the current monitoring moment is designated as the current waste filling amount sequence of the smart trash can at the current monitoring moment. The current monitoring moment belongs to the current day. The waste filling amount corresponding to the monitoring moment is the result of subtracting the remaining capacity of the smart trash can at the current monitoring moment from the remaining capacity at the adjacent historical monitoring moment. The adjacent historical monitoring moment refers to the monitoring moment that is adjacent to the current monitoring moment and is located before the current monitoring moment in time.

3. The intelligent control method for an environmentally friendly trash can as described in claim 2, characterized in that, The method for obtaining the reference cluster includes: The sequence corresponding to the cluster center of each cluster is recorded as the center sequence of the corresponding cluster. The DDTW distance between the normalized sequence of the center sequence of each cluster and the normalized sequence of the current waste filling amount sequence is calculated, and the cluster corresponding to the smallest DDTW distance is taken as the reference cluster.

4. The intelligent control method for an environmentally friendly trash can as described in claim 3, characterized in that, The method for obtaining the time difference representation value between the current waste filling volume sequence and the target reference sequence includes: For any target reference sequence, the set of all DDTW matching data pairs between the current waste filling volume sequence and the target reference sequence is denoted as the matching data pair set corresponding to the target reference sequence. The time difference corresponding to each DDTW matching data pair in the matching data pair set is obtained. The mean of the time differences corresponding to all DDTW matching data pairs in the matching data pair set is used as the time difference characterization value between the current waste filling volume sequence and the target reference sequence. The time difference corresponding to each DDTW matching data pair is the time of the monitoring time corresponding to the first data in the corresponding DDTW matching data pair minus the time of the monitoring time corresponding to the second data in the corresponding DDTW matching data pair. The first data in each DDTW matching data pair belongs to the current waste filling volume sequence, and the second data in each DDTW matching data pair belongs to the target reference sequence.

5. The intelligent control method for an environmentally friendly trash can as described in claim 4, characterized in that, The method for obtaining the target predicted overflow duration of the smart trash can at the current monitoring time includes: Based on the DDTW distance between the current garbage filling sequence and the target reference sequence and the matching data pair set corresponding to the target reference sequence, the difference feature value between the current garbage filling sequence and the target reference sequence is obtained. The negative correlation mapping value of the difference feature value between the current garbage filling sequence and the target reference sequence is used as the initial prediction confidence under the corresponding target reference sequence. The sum of the initial prediction confidence under all target reference sequences is recorded as the comprehensive prediction confidence. The ratio of the initial prediction confidence under the target reference sequence to the comprehensive prediction confidence is recorded as the target prediction confidence under the corresponding target reference sequence. The current waste filling sequence is aligned with the target reference sequence based on the time difference characterization value. The initial predicted overflow duration under each target reference sequence is obtained based on the alignment result. The product of the initial predicted overflow duration under the target reference sequence and the target prediction confidence level under the corresponding target reference sequence is taken as the weighted predicted overflow duration under the corresponding target reference sequence. The sum of the weighted predicted overflow durations under all target reference sequences is recorded as the target predicted overflow duration of the smart trash can at the current monitoring time.

6. The intelligent control method for an environmentally friendly trash can as described in claim 5, characterized in that, The method for obtaining the difference feature values ​​between the current waste filling volume sequence and the target reference sequence includes: For any target reference sequence, the data difference value corresponding to each DDTW matching data pair in the matching data pair set corresponding to the target reference sequence is obtained. The data difference value corresponding to the DDTW matching data pair is the absolute value of the difference between the normalized value of the first data and the normalized value of the second data in the corresponding DDTW matching data pair. The mean of the data difference values ​​corresponding to all DDTW matching data pairs in the matching data pair set is taken as the data difference mean. The absolute value of the difference between the data difference value corresponding to the DDTW matching data pair and the data difference mean is recorded as the first difference value corresponding to the corresponding DDTW matching data pair. The mean of the first difference values ​​corresponding to all DDTW matching data pairs in the matching data pair set is recorded as the feature mean. The result of multiplying the DDTW distance between the current garbage filling sequence and the target reference sequence by the feature mean is taken as the difference feature value between the current garbage filling sequence and the target reference sequence.

7. The intelligent control method for an environmentally friendly trash can as described in claim 5, characterized in that, A method for aligning the current garbage filling volume sequence with the target reference sequence based on the time difference characterization value, and obtaining the initial predicted overflow duration for each target reference sequence based on the alignment result, includes: The remaining capacity in the smart trash can at the current monitoring time is recorded as the current remaining capacity. For any target reference sequence: Based on the time difference characterization value between the current waste filling volume sequence and the target reference sequence, the subsequence to be analyzed corresponding to the target reference sequence is obtained, and the subsequence to be analyzed corresponding to the target reference sequence is denoted as the subsequence to be analyzed H; the sum of the first b-1 data in the subsequence to be analyzed H is denoted as the first accumulated value, and the sum of the first b data in the subsequence to be analyzed H is denoted as the second accumulated value, where b is greater than 1; if the result of the current remaining capacity minus the second accumulated value is not greater than a preset remaining capacity threshold, and the result of the current remaining capacity minus the first accumulated value is greater than the preset remaining capacity threshold, then the time span between the monitoring time corresponding to the first data in the subsequence to be analyzed H and the monitoring time corresponding to the b data is taken as the initial predicted overflow duration under the target reference sequence G.

8. The intelligent control method for an environmentally friendly trash can as described in claim 7, characterized in that, The method for obtaining the subsequence to be analyzed corresponding to the target reference sequence includes: Let T be the time corresponding to the current monitoring moment, and let t be the time difference between the current waste filling volume sequence and the target reference sequence. If the time difference between the current waste filling volume sequence and the target reference sequence is 0, then the moment with time T in the historical unit monitoring time period corresponding to the target reference sequence is recorded as the characteristic monitoring moment in the historical unit monitoring time period corresponding to the target reference sequence. If the time difference between the current waste filling volume sequence and the target reference sequence is less than 0, then the moment with time (T+t) in the historical unit monitoring time period corresponding to the target reference sequence is recorded as the characteristic monitoring moment in the historical unit monitoring time period corresponding to the target reference sequence. If the time difference between the current waste filling volume sequence and the target reference sequence is greater than 0, then the moment with time (Tt) in the historical unit monitoring time period corresponding to the target reference sequence is recorded as the characteristic monitoring moment in the historical unit monitoring time period corresponding to the target reference sequence. Within the historical unit monitoring time period corresponding to the target reference sequence, the sequence consisting of the waste filling amounts corresponding to all monitoring times that are later than the characteristic monitoring time is recorded as the subsequence to be analyzed corresponding to the target reference sequence.

9. The intelligent control method for an environmentally friendly trash can as described in claim 1, characterized in that, A method for controlling the garbage collection of the smart trash can based on the predicted overflow duration, including: The collection route where the smart trash can is located is recorded as the collection route to be analyzed. The estimated time for the collection vehicle to travel to each smart trash can on the collection route to be analyzed is obtained. If the estimated time for the collection vehicle to travel to any smart trash can on the collection route to be analyzed is not less than the target predicted overflow time of the corresponding smart trash can at the current monitoring time, it is determined that the collection vehicle should immediately depart at the current monitoring time to collect the smart trash cans on the collection route to be analyzed in sequence.

10. The intelligent control method for an environmentally friendly trash can as described in claim 9, characterized in that, A method for obtaining the estimated time for a waste collection vehicle to travel to each smart trash can on the waste collection route to be analyzed includes: Each time the garbage truck collects garbage from the smart garbage bins on the garbage collection route to be analyzed, the time taken for the garbage truck to travel to the f-th smart garbage bin on the garbage collection route to collect garbage is obtained and recorded as the historical travel time corresponding to the f-th smart garbage bin. The set of all historical travel times corresponding to the f-th smart garbage bin is recorded as the historical travel time set corresponding to the f-th smart garbage bin. The median of the historical travel time set is used as the estimated time for the garbage truck to travel to the f-th smart garbage bin on the garbage collection route to be analyzed.

Images