WO2014187997A1

WO2014187997A1 - Self-learning method and system for automatically checking shipments for correctness

Info

Publication number: WO2014187997A1
Application number: PCT/EP2014/060752
Authority: WO
Inventors: Peter Van Eijk
Original assignee: Neopost Technologies SA
Current assignee: Quadient Technologies France SA
Priority date: 2013-05-24
Filing date: 2014-05-26
Publication date: 2014-11-27
Anticipated expiration: 2015-11-24

Abstract

A self-learning method for automatically checking shipments for correctness comprises the steps of a) based on an dispatch order assembling an shipment comprising items of different types in individual quantities; b) weighing the shipment in order to obtain the actual total mass of the shipment; c) comparing the actual total mass with an expected total mass; and d) generating an error message when the actual total and the expected total mass differ by more than a predetermined tolerance value. The expected total mass is determined for each dispatch order by e) checking a database if information on the individual specific mass of each item of the dispatch order is present to calculate the sum of the individual specific masses times the individual quantities of the items that should according to the dispatch order be contained in the shipment and, if information on the individual specific masses of all items of the dispatch order is present, calculating the expected total mass; if information on the individual specific masses of only some items or no item of the dispatch order is present, storing in said database at least information on the individual quantities of each item of the dispatch order, of which the individual specific mass is unknown, and information at least on the actual total mass of those items to obtain an equation and f) repeating steps a) - e) for further dispatch orders, wherein each time an item with an unknown specific mass is present in an dispatch order, the information in said database is automatically checked, if the information is sufficient for calculating the specific mass of said item and possibly of other items, and if the information is sufficient, automatically calculating the specific mass and storing this information in said database.

Description

SELF-LEARNING METHOD AND SYSTEM FOR

AUTOMATICALLY CHECKING SHIPMENTS FOR CORRECTNESS

TECHNICAL FIELD

The invention relates to a self-learning method and a system for automatically checking shipments for correctness, said shipments comprising several items in several quantities, which are individually assembled based on a dispatch order for the respective shipment.

TECHNICAL BACKGROUND Mail ordering has become a widely used way of buying goods. More and more companies offer virtual department stores, in which the customers can electronically put goods in a shopping cart that later will be transferred by the respective company into an dispatch order, so that in a warehouse a shipment comprising the items ordered (and sometimes additional items such as samples, vouchers, invoices, etc.) can be assembled based on the respective dispatch order.

Checking the correctness of a shipment has always been a challenge in the mail order business, even in cases where the warehouse is operated more or less fully automated, in particular when a shipment comprises several items of different types and in different quantities. A person assembling the items of the order by taking out the respective items from a storage rack may take out the wrong quantity or the wrong item. In case of automatic transport of the ordered items to a packaging station, some of the items may fall of the transport line. Numerous other causes for incorrectness of a shipment are known.

Since freight and postal charges often depend on the mass of a shipment to be transported, assembled shipments are - before or after packaging - generally weighed. If the individual masses of the items forming the shipment are known, the total mass of the shipment is generally an indicator for the correctness of the shipment, since it is then possible to calculate an expected mass of the shipment based on the masses of the individual items of the shipment and to compare this expected mass with the measured actual mass of the shipment. If the actual mass is lower than the expected mass, one or more items, which according to the dispatch order should form part of the shipment, may be missing. If the actual mass is greater than the expected mass, erroneously more items than ordered or different items than ordered may be contained in the shipment. The shipment can then be sorted out for further inspection. It should be noted that "correctness" in the sense of this application means that the dispatch order is exactly fulfilled, i.e. that not only not less than the ordered items are present, but also that not more than the items according to the dispatch order are present, that no parts of an item are missing and that the correct items are shipped.

DISCLOSURE OF THE INVENTION

In particular when a company operates a warehouse with numerous different items provided by several suppliers, it has turned out to be very complex and time consuming to maintain a database, in which for each item the individual mass is stored. Items to be shipped may be items having their own package, for example a pair of earphones in a blister pack. Often suppliers make changes only affecting the individual package of the respective item, for example change a filling material that might be present in the respective package, which does not affect in any way the quality of the item itself, but affects the mass of the individual package, in which the item is contained. Since the quality of the item is not changed, the supplier may not notify the company operating the warehouse about the change. Since, for the purpose of shipment, not the item itself, but the item with its individual package forms one "item to be sent", an error message may be generated erroneously. Also, when a company introduces a new system for automatic packaging shipments, in standard or custom-sized packages, the company may for the first time be able to effectively use the total mass of the shipment for a quality check concerning the correctness of the shipment and may therefore not be in possession of a database with the individual masses of all items (including their respective individual packaging) that the company may want to ship. It is an object of the invention to provide a self-learning method and system, which upon operation over the time automatically obtain the necessary information on the individual masses of the shipped items to perform an increasingly precise correctness check.

The object is achieved by a method as set out in claim 1 respectively by a system as set out in claim 10. The respective dependent claims are directed to advantageous ways of performing the method respectively advantageous embodiments of the system. One major advantage of the invention is, that even if initially no information on the masses of the items to be shipped is known, during operation with each new shipment more information is gathered that over the time allows to automatically calculate the individual masses of the items and thus to fill or update a corresponding database. Further objectives, advantages and details of the invention will become apparent from the following description of preferred embodiments, which is given by means of non- limiting examples.

DESCRIPTION OF PREFERRED EMBODIMENTS

A self-learning system for automatically checking shipments for correctness according to the invention will now be described. From this description, also the method according to the invention becomes apparent.

The system comprises means for weighing an i-th shipment S , wherein i is an integer used in this application to distinguish several shipments from each other. Each of those shipments comprises one or more items Aj of n different types in n individual quantities q(Aj)j, with j being an integer running from 1 to n and n being a positive integer. Hence, a first shipment Si may comprise 2 books B, 3 CDs C, a pair of earphones E and a

2

custom sized package P made from 0,5 m of corrugated cardboard. All these items are assembled together according to a first dispatch order Oi and have a total mass MTi of for example 1 ,500 g. In that case, j would run from 1 to 4 and Ai would be the books, q(Ai)i would be 2, A2 would be the CDs, q(A2)i would be 3, A3 would be the earphones, q(A3)i would be 1 and A4 would be the custom sized package, q(A4)i 2

would be 0.5 m .

It should be noted that under the term "dispatch order" machine readable information on the items that should form the shipment is understood, which will normally contain the actual list of items ordered by a customer and additional information for example on the sizes of the package to be used or, in case a custom sized package is prepared, information on the dimensions of such package, as well as information on additional items to be included in the shipment, such as for example an invoice, vouchers, samples of other products and filling material.

According to the invention, the system further comprises means for determining an expected total mass ME, for each dispatch order 0/ by checking a data base if information on the individual specific mass m(Aj) of each item Aj of the dispatch order 0/ is present to calculate the sum of all individual specific masses m(Aj) times the individual quantities q(Aj)j of the items that according to the dispatch order 0/ should be contained in the respective shipment Sj.

If the necessary information on the individual specific masses m(Aj) of all items of the dispatch order 0/ is present in the data base, the expected total mass ΛΖΕ/ is calculated and compared with the actual total mass MT, as measured. If the actual total mass MT, and the expected total mass ME; differ by more than a predetermined tolerance value Dj, an error message is generated. Such error message can be an optical and/or acoustical signal, so that an operator of the system may manually mark or sort out the respective shipment to be further examined. It should be noted that under the term

"error message" not only direct signals to an operator are to be understood, but also all forms of control signals that lead to providing the shipment for example with a label, a printing or the like, marking the shipment as (possibly) incorrect so that it should be manually inspected, and/or that lead to manually or automatically sorting the shipment out for manual inspection, without interrupting the packaging process.

In a preferred embodiment, the error message may be forwarded to means for automatically sorting out the shipment and/or marking the shipment for further examination. Such marking can be done in several ways. For example, a red label may be attached to the shipment or the lid of a package used may be left open.

If the data base does not contain information on all the individual specific masses m(Aj) of the items Aj that shall form the shipment according to the respective dispatch order Oj, at least the actual total mass RTMj of those items, of which the individual specific mass m(Aj) is unknown and information on the individual quantities q(Aj)j of those items in the respective shipment is stored and an expected mass ME, is estimated for performing at least a plausibility check, as will be explained later on. If for example the actual total mass TMj of an i-th shipment is 1 kg and the shipment is comprised by two items A-i , of which the individual specific mass m(A -i) is known to be for example 200 g, and one item A2 and one item A3, of which the individual specific masses m(A2) and m(A3) are unknown, it is not necessary to store the information that 1 kg = 2 x 200 g + 1 x m(A₂) + 1 x m(A₃).

It is fully sufficient to store the information on the "remaining" total mass after deducting the known mass, i.e. it is sufficient to store 600 g = 1 x m(A₂) + 1 x m(A₃).

In this way, over the time in the data base a set of r equations with s unknowns is accumulated, r, s e IN₊. Each time r equals s, the system is in a position to solve the equations and hence to obtain those unknown individual specific masses that are part of the set of equations.

Example:

A first shipment Si assembled according to a first dispatch order O1 may contain

2

2 items Α , 3 items A₂ and 0,5 m of item A3 (the cardboard, of which a custom sized package is made). The measured actual total mass TM-i of this first shipment may be 1 ,600 g. A second shipment S2 assembled according to a second dispatch order O2 may

2

contain 2 items A∑, 1 item A4 and 0.6 m of A3. The measured actual total mass TM2 of this shipment may be 1 ,560 g. A third shipment S3 may have a total mass TM3 of 680 g and may according to a third

2

dispatch order O3 contain 1 item Α , 1 item A2 and 0.3 m A3.

The information on TM, and on the quantities q(Aj)j of the items in each shipment is stored in the data base. The system notices that so far it has 3 equations with 4 unknowns.

Now a fourth shipment S4 may comprise according to a fourth dispatch order O4

2

4 items A2 and 0.4 m A3, having a measured actual total mass TM4 of 1 ,440 g. At this point in time, the system notices that there are now 4 equations with 4 unknowns, which can easily be solved, for example by starting with the equation

1 ,440 g = 4 m(A₂) + 0.4 [m²] A₃ and rearranging this equation to an expression for A . A2 = ^ ,440 g - 0,4 [m²] A₃ or >A₂ = 360 g - 0.1 [m²] A₃ This can then for example be entered into the equation 630 g = A-i + A₂ + 0.3 [m²] A₃ so that this equation becomes

320 g = /\f + 0.2 [m²] A₃ which can be formed into an expression for A . A ₁ = 320 g - 0,2 [m²] A₃

This can then be entered for example in the for TMi expression to read 1 ,600 g = 2 (320 g - 0.2 [m²] A₃) + 3 (360 g - 0.1 [m²] A₃) + 0.5 [m²] A₃ which leads to

-120 g = -0,2 [m²] A₃

or

A₃ = 600 g/m².

2

Having thus calculated the individual specific mass m(A₃) to be 600 g/m , it then immediately follows that the individual specific masses of the other items are m(A i) = 200 g, m(A₂) = 300 g and m(A₄) = 600 g (each per piece).

It should be noted that, as in the example case, the individual quantities q(Aj); of an item Aj can be measured in any suitable way and can for example be simply the number of respective items or, for example with respect to the cardboard from which a custom sized package is formed, a surface or, for example the case of filling material, a volume.

Accordingly, the specific individual mass m(Aj) can be given in any suitable way, for example weight per piece, weight per surface or weight per volume.

Although in the example case no information on the individual specific masses of the respective items was initially known, the system learned just "on the fly" (during operation) and acquired the necessary information, which then could be used for calculating an expected total mass Λ ΖΕ/ for each shipment Sj based on the information in the dispatch order O .

It should also be noted that, depending on the environment, in which the system and the corresponding method are used, the "dispatch order" may not be an actual existing order but may be automatically generated by corresponding means, for example scanners, barcode readers or other suitable equipment that looks at some point in time at the shipment that is being processed, for example packed, and then later compares the actual total mass of the shipment with the expected total mass.

Also depending on the environment, in which the method and the system according to the invention are used, a package for the shipment may be considered as forming part of the shipment and may be treated accordingly as an item of the shipment or may be treated completely separately. Accordingly, it is possible to obtain the actual total mass of the shipment before or after packing the items into a suitable package.

As is apparent from the above, during the initial process it is assumed that the shipments are complete, i.e. that the weighed actual total mass of a shipment indeed results from the individual specific masses and individual quantities of the items of the shipment according to the dispatch order. If this is not true, for example because in a certain shipment only 3 items Aj were included instead of 4 items, which should be included according to the corresponding dispatch order, the calculated specific mass m(Aj) would obviously not be correct. However, in a preferred embodiment of the invention such errors, as well as changes in the respective individual specific mass of a certain item, which may for example be due to changes a supplier of the respective item has made for example to the individual package of the item, are automatically corrected. For this purpose, the differences between the expected total masses ME; and the actual total masses MTj of the shipments are regularly monitored. It is apparent that in case the individual specific masses are, for which ever reason, not correct, the deviation between the measured actual total mass and the expected total mass will generally be of a certain type, i.e. the actual total mass will always be larger or always be smaller than expected. If such tendency to deviate in a certain direction is noticed, the individual specific mass m(Aj) of an item Aj, which regularly leads to deviations between the expected total mass and the actual total mass of the shipment, can be recalculated automatically and the corresponding information the data base can be updated.

It should be also noted that the tolerance value D,, by which the actual total mass MTj and the expected total mass ME; of an i-th shipment may differ from each other without resulting in the generation of an error message, is in an preferred embodiment individually set for the respective shipment, based for example on information concerning the items of the shipment. For example, it may be known that the individual masses of certain items have a strong tendency to differ from each other, so that the respective item sometimes weigh 200 g, sometimes 210 g. In such cases, a respective indicator, for example a flag, may be generated, so that the tolerance value D; is set higher than in cases where the items of the same type always have almost the same mass. It is also possible possible to assign individual tolerance values to each individual specific mass m(Aj), so that a specific tolerance value D can be calculated for a specific shipment.

When the information in the database is not sufficient to calculate an expected total mass value ME,, such value may be estimated based on certain values in order to perform at least a plausibility check for the respective shipment. For example, even in case custom sized packages are made, the respective machine may have a minimum package size, which would lead to a minimum weight even if the shipment contains only of an empty package. Hence, it is possible to set at least a minimum value higher than the weight of the smallest possible package that a shipment should have if it does not only contain an empty package.

Claims

1 . Self-learning method for automatically checking shipments for correctness, comprising the steps of

a) based on an i-th dispatch order O/, assembling an i-th shipment S/ comprising items Aj of n different types in n individual quantities q(Aj) (j = 1, n; i, n e IN₊),

b) weighing the shipment in order to obtain the actual total mass MTj of the shipment,

c) comparing the actual total mass MTj with an expected total mass ME; and

d) generating an error message when the actual total MTj and the expected total mass ME, differ by more than a predetermined tolerance value D,,

wherein said expected total mass ME; is determined for each dispatch order 0/ by

e) checking a database if information on the individual specific mass m(Aj) of each item Aj of the dispatch order is present to calculate the sum of the individual specific masses m(Aj) times the individual quantities q(Aj)j of the items that should according to the dispatch order 0/ be contained in the shipment S , and

- if information on the individual specific masses m(Aj) of all items Aj of the dispatch order 0/ is present, calculating the expected total mass ME,,

if information on the individual specific masses m(Aj) of only some items or no item Aj of the dispatch order 0/ is present, storing in said database at least information on the individual quantities q(Aj)j of each item Aj of the dispatch order O/, of which the individual specific mass m(Aj) is unknown, and information at least on the actual total mass RTM, of those items to obtain an equation of the type RTM, = (sum of all q(Aj)j times m(Aj), of the which the individual specific mass m(Aj) is unknown),

f) repeating steps a) - e) for further dispatch orders, wherein each time an item / with an unknown specific mass m(Aj) is present in an dispatch order, the information in said database is automatically checked, if the information is sufficient for calculating the specific mass m(Aj) of said item and possibly of other items, and if the information is sufficient, automatically calculating the specific mass m(Aj) and storing this information in said database.

2. Method according to claim 1 , wherein each individual quantity of q(Aj)j of an item Aj is one of a number of pieces, a volume and a surface and the specific individual mass m(Aj) is one of a mass per piece, mass per surface and mass per volume.

3. Method according to claim 1 or 2, wherein a package, in particular a custom sized package made individually for the specific shipment, forms part of the shipment and is treated like any other item Aj of the shipment.

4. Method according to one of claims 1 to 3, wherein filling material forms part of the shipment and is treated like any other item Aj of the shipment.

5. Method according to one of claims 1 to 4, wherein the error message is forwarded to a unit for automatic sorting out the respective shipment for further examination.

6. Method according to one of claims 1 to 5, wherein the error message is forwarded to a unit for automatic marking the respective shipment for further examination.

7. Method according to one of claims 1 to 6, wherein the tolerance value D,, is individually predetermined for a shipment S; based on information on the quantities q(Aj)j and the type of the items Aj that according to the respective dispatch order O . should form the shipment Sj.

8. Method according to one of claims 1 to 7, wherein the differences between the the expected total mass ME,, and the actual total mass MT, of the shipments are monitored, and, if the differences for shipments containing certain items show a tendency to deviate in a certain direction, the individual specific masses m(Aj) of those items are recalculated and the corresponding information in the database is updated.

9. Method according to one of claims 1 to 8, wherein the expected mass ME, is estimated in case the information in the database is not sufficient to calculate the expected total mass ME,.

10. Self-learning system for automatically checking shipments for correctness, comprising:

means for weighing an i-th shipment S comprising items Aj of n different types in n individual quantities q(Aj)j, (j = 1, n; i, n e IN+) based on an i-th dispatch order 0/ and obtaining the actual total mass MTj of the shipment,

means

for determining an expected total mass MEj for the dispatch order 0/ by checking a database if information on the individual specific mass m(Aj) of each item Aj of the dispatch order 0/ is present to calculate the sum of the individual specific masses m(Aj) times the individual quantities q(Aj)j of the items that should according to the dispatch order 0/ be contained in the shipment S , and

for calculating the expected total mass ME,, if information on the individual specific masses m(Aj) of all items Aj of the dispatch order 0/ is present, and for storing in said database, if information on the individual specific masses m(Aj) of only some items or no item Aj of the dispatch order 0/ is present, at least information on the individual quantities q(Aj)j of each item Aj of the dispatch order Oj, of which the individual specific mass m(Aj) is unknown, and at least on the actual total mass RTMj of those items to obtain an equation of the type RTMj = (sum of all q(Aj)j times m(Aj), of the which the individual specific mass m(Aj) unknown),

— for checking each time an item / with an unknown specific mass m(Aj) is present in an dispatch order the information in said database, if the information is sufficient for calculating the specific mass m(Aj) of said item and possibly of other items, and if the information is sufficient, automatically calculating the specific mass m(Aj) and storing this information in said database,

- means for comparing the actual total mass MTj with the expected total mass MEj of the shipment,

means for generating an error message when the actual total MTj and the expected total mass MEj differ by more than a predetermined tolerance value D,.

1 1 . System according to claim 10, further comprising means for automatically making a custom sized package.

12. System according to claim 10 or 1 1 , further comprising means for sorting out a shipment for further examination, when an error message concerning this shipment is generated.

13. System according to one of claims 10 to 12, further comprising means for marking a shipment for further examination, when an error message concerning this shipment is generated.

14. System according to one of claims 10 to 13, further comprising means for individually determining said tolerance value D for a shipment Sj based on information in the respective dispatch order O .

15. System according to one of claims 10 to 14, further comprising means for monitoring the differences between the expected total mass ME,, and the actual total mass MT, of the shipments, and for recalculating, if the differences for shipments containing certain items show a tendency to deviate in a certain direction, the individual specific masses m(Aj) of those items and updating the corresponding information in the database.