How pay as you throw waste collection data can be analysed with the wearable RFID technology
In PAYT waste collection, reading the RFI tags on trash bins or bags using a wearable device provides information in a normalised, standard string. The string, read from left to right, contains the following information:
♦ the ID of the device: the wearable wristband bears the ID of the device, never the name of the waste collector. As provided for by the EU Regulation on the protection of data 2016/679 (GDPR), it is not possible to associate the name of the person who uses the device or the device serial number on any type of server;
♦ the battery level;
♦ the date and time the RFID tag is scanned;
♦ the date and time the reading is received by the server;
♦ the tag ID, a personal 24 digit code associated to each user.
Double codings, swapping of bags and bins, collection route inconsistencies
Since each RFID ID corresponds to a specific user, a rough analysis of the CSV (Comma-Separated Values) file, format used to export data files to electronic spreadsheets and that provides the tag reading, coding errors can be immediately noted: the most common error is double coding, that is, the presence of two identical tag IDs – e.g. the same tag is associated to two different streets in the same town; this indicates incorrect coding. Therefore, for correct application of the PAYT system, it is vital that the wearable device be precise and that it correctly identifies the tag read in that precise moment.
The co-existence of two identical tag IDs can also be caused by the swapping of bags or bins used for PAYT waste collection; this can happen when a user uses the same bin for two houses – for example, their habitual residence and their holiday home, or in the case of accidental swapping of bags between neighbours. These elements allow the waste collection company to “clean” the user’s data records, immediately eliminating errors.
Another factor emerging from preliminary analysis is the possible inconsistency in the waste collectors’ routes: if, at the end of a shift in the same town, there is a significant difference in the number of tags read by two different workers, and the length of the route is similar, then the workloads are clearly not balanced. This information allows the company that manages the PAYT service to re-balance and better distribute the work loads among the workers.
Input filters: white list and black list
To perform a more detailed analysis and to correctly calculate the PAYT tariff, first and foremost the waste collection company has to understand how to filter the RFIDs of users’ data records. The first step is to apply an input filter; that is, “telling” the scanner to accept – for a given town where a PAYT waste collection system is used – only tags with a specific four-digit code, which is usually alphanumeric but can also be formed of only numbers or only letters. For town X, the wearable device will read all the tags but will take into consideration, that is, it will only save those that, for example, have the code 00aa or 1234 or abcd as the head.
Therefore, a white list has to be created in which only a given family of tags can be entered: this enables all the RFIDs that do not belong to this family to be filtered defining, by contrast, a dynamic black list; that is, one that can be updated, blocking specific tags, “telling” the reader not to accept them. Typical cases of “black list” tags are those regarding insolvent users that have not paid for services such as the collection of garden waste or the emptying of diaper bins, or that repeatedly make mistakes in the collection – throwing glass or plastic into the organic waste, etc. The black list includes white list tags that the reader can accept but which it decides not to accept for reasons such as insolvency or repeated incorrect separation of the waste.