EP1651551B1 - Method for operating a detector for identifying the overlapping of flat mail in a sorting machine - Google Patents

Abstract

The invention relates to a method whereby the average costs CNDD for each unidentified overlap in a distribution and sorting process and the average costs CRF for the renewed sorting and distribution of each identified overlapped item of mail are determined. Before the sorting operation, the identification rate DR(Pi) of the detector is determined at different working points according to an adjusting parameter Pi. During the sorting operation, the respective actual overlap rate DFR'(Pi) of the sorting machine is estimated, for defined time intervals, from the identification result, the identification rate DR(Pi) and the error rate ER(Pi), according to the adjusting parameters Pi. The actual adjusting parameter Popt with the greatest utility CB'(Popt), which is used to operate the detector until the next interval, is determined by optimisation calculations from the relation for the utility CB'(Pi) of the sorting and distribution process CB'(Pi) = N{( DFR'(Pi)*DR(Pi)*(CNDD-CRF)) - (ER(Pi)*CRF*(1-DFR'(Pi)))} for the operator.

Description

The invention relates to a method for operating a detector for detecting overlaps of flat broadcasts in a transport path of a sorting machine following a separating device.

In conventional mail processing systems, mail batches are sorted in sorting machines. It may happen that several items are simultaneously withdrawn from the stack, ie several overlapping items leave the feeder and lead, if this is not recognized, to incorrect sorting. For this reason, it is necessary to recognize the overlaps in a timely manner after leaving the singulator to guide them into a rejection / reject bin (cf. US Pat. No. 6,023,034 ).

So far, corresponding detectors have been known which, based on different measuring principles, determine overlaps caused by multiple withdrawals of the singling device. The detected multiple prints are directed to the rejection endpoints of the sorting machines to avoid mis-distribution of the mailpieces. So will in the DE 43 37 004 A1 a device and a method for detecting overlaps of flexible flat transmissions described, in which at least a portion of each shipment is movable perpendicular to the conveying direction in the transport path. At least one deflection element is arranged on the transport path, by which the movable transmission sections are deflected perpendicular to the conveying direction temporarily by a predetermined amount during the conveyance in the transport path. In a detection direction, the presence of overlapping transmission sections is detected due to the recoil behavior of the transmissions.

Also known were corresponding detectors, in which the narrow sides of the singular transmissions are recorded by means of image-receiving devices and then evaluated in an image processing device by evaluating the image signals to determine whether multiple transmissions are transported overlapped, eg by counting the light-dark transitions ( FR 2 546 083 A . 2 057 309 A ).

All of these detectors have a fixed operating point with a specific error and detection rate, in which they operate independently of the process in which they are involved.

The WO 03/042082 A discloses a method according to the preamble of claim 1.

The invention has for its object to provide a method for automatic operation of a detector for detecting overlaps of flat items in a separating device following a transport path of a sorting machine, wherein the detector is operated depending on sorting and Verteilprozessbedingungen.

According to the invention the object is achieved by the features of claim 1.

The average cost C _NDD for each unrecognized overlap in the distribution and sorting process and the average cost C _RF for re-sorting and distributing each detected overlapped shipment are determined offline. Before the sorting operation, the recognition rate DR (P _i ) and the error rate ER (P _i ) of the detector are determined at different operating points as a function of a corresponding setting parameter P _i . During the sorting operation, at fixed time intervals or intervals in which a predetermined number of mail items are processed in the sorting machine, the respective current overlapping rate DFR '(P _i ) of the sorting machine is determined from the recognition result, the recognition rate DR (P _i ) and the error rate ER (P _i ) is estimated as a function of the setting parameters P _i .

wird anschließend in Optimierungsrechnungen der aktuelle Stellparameter P_opt. mit dem höchsten Nutzen CB' (P_opt.) ermittelt, mit welchem der Detektor bis zum nächsten Intervall betrieben wird. Damit kann der Detektor auch bei unterschiedlichen Prozessbedingungen und -kosten kostenoptimal automatisch betrieben werden.From the relationship for the benefit CB '(P _i ) of the operator of the sorting and distribution process $$\mathrm{CB\text{'}}\left({\mathrm{P}}_{\mathrm{i}}\right)\mathrm{=}\mathrm{N}\left\{\mathrm{(}\mathrm{DFR\text{'}}\left({\mathrm{P}}_{\mathrm{i}}\right)\mathrm{*}\mathrm{DR\text{'}}\mathrm{\left(}{\mathrm{P}}_{\mathrm{i}}\mathrm{\right)}\mathrm{*}\left({\mathrm{C}}_{\mathrm{NDD}}\mathrm{-}{\mathrm{C}}_{\mathrm{RF}}\right)\mathrm{)}\mathrm{-}\left(\mathrm{HE}\left({\mathrm{P}}_{\mathrm{i}}\right)\mathrm{*}{\mathrm{C}}_{\mathrm{RF}}\mathrm{*}\left(\mathrm{1}\mathrm{-}\mathrm{DFR\text{'}}\left({\mathrm{P}}_{\mathrm{i}}\right)\right)\right)\right\}$$

is then in optimization calculations of the current setting parameters P _opt. with the highest benefit CB '(P _opt. ) determines with which the detector is operated until the next interval. As a result, the detector can be operated automatically and cost-effectively even with different process conditions and costs.

zu ermitteln, wobei N_AD die Anzahl der vom Detektor ermittelten Überlappungen bei der Anzahl N der detektierten Sendungen in einem Intervall ist.It is advantageous, the current overlap rate DFR '(P _i ) for the different control parameters P _i from the relationship $$\mathrm{DFR\text{'}}\left({\mathrm{P}}_{\mathrm{i}}\right)\mathrm{=}\frac{\frac{{\mathrm{N}}_{\mathrm{AD}}}{\mathrm{N}}\mathrm{-}\mathrm{HE}\left({\mathrm{P}}_{\mathrm{i}}\right)}{\mathrm{DR\text{'}}\left({\mathrm{P}}_{\mathrm{i}}\right)\mathrm{-}\mathrm{HE}\left({\mathrm{P}}_{\mathrm{i}}\right)}$$

where N _{AD is} the number of overlaps detected by the detector in the number N of detected transmissions in one interval.

The invention will be explained in more detail with reference to the drawing in an embodiment.

FIG 1

ein Diagramm der Erkennungsrate und der Fehlerrate eines Detektors in Abhängigkeit vom Stellparameter P_i;

FIG 2

ein Blockdiagramm des Betriebsprozesses.

It shows

FIG. 1

a diagram of the detection rate and the error rate of a detector in dependence on the control parameter P _i ;

FIG. 2

a block diagram of the operating process.

The overlaps detected by an overlap detector, ie a plurality of programs leaving the singulator sorting device overlapping each other, are converted into a so-called Rejected and then fed once again the singulator. In this case, the shipments thus discharged consist of real overlaps and falsely recognized individual shipments. If the overlapped items were not recognized and rejected, the items associated with the item with the address read would at least be sent to the wrong next distribution center and would then have to be forwarded from there to the correct address.

Of course, this causes additional costs, which can be avoided without the cost of re-applying the overlaps by early detection of the overlaps. Since a detection process can also erroneously recognize normal individual consignments as overlaps due to misdetections, the costs for the additional sorting of these consignments in the sorting machine in question must be subtracted from the additional costs saved.

The benefit CB for the operator of the sorting and distribution process thus results from the relationship $$\mathrm{CB}\mathrm{=}\mathrm{N}\left\{\mathrm{(}\mathrm{DFR}\mathrm{*}\mathrm{DR}\mathrm{*}\left({\mathrm{C}}_{\mathrm{NDD}}\mathrm{-}{\mathrm{C}}_{\mathrm{RF}}\right)\mathrm{)}\mathrm{-}\left(\mathrm{HE}\mathrm{*}{\mathrm{C}}_{\mathrm{RF}}\mathrm{*}\left(\mathrm{1}\mathrm{-}\mathrm{DFR}\right)\right)\right\}$$

die Anzahl der im betrachteten Zeitraum detektierten Überlappungen und $$\mathrm{N}*\mathrm{ER}*\left(1-\mathrm{DFR}\right)$$

die Anzahl der falsch detektierten Einzelsendungen, mit

N =

Zahl der verarbeiteten Sendungen

DFR =

Überlappungsrate der Sortiermaschine

DR =

Erkennungsrate des Detektors

ER =

Fehlerrate des Detektors

C_NDD =

Kosten für jede nicht erkannte Überlappung

C_RF =

Kosten für jedes nochmaliges Zuführen zur jeweiligen Sortiermaschine von erkannten Überlappungen

It is $$\mathrm{N}*\mathrm{DFR}*\mathrm{DR}$$

the number of overlaps detected in the period considered and $$\mathrm{N}*\mathrm{HE}*\left(1-\mathrm{DFR}\right)$$

the number of incorrectly detected individual consignments, with

N =

Number of processed consignments

DFR =

Overlap rate of the sorting machine

DR =

Detection rate of the detector

ER =

Error rate of the detector

C _NDD =

Cost of each unrecognized overlap

C _RF =

Cost of each re-feeding to the respective sorting machine of detected overlaps

These costs depend on the particular type of shipment processing, i. For example, it is an in-sort or an out-of-order sort.

Detectors for detecting overlaps have heretofore operated at a fixed operating point with a certain recognition rate DR and a certain error rate ER, e.g. according to DE-Acc. 103 10 546.8-27 can be determined.

However, in order to influence the benefit for the operation, it is necessary to parameterize the recognition and the error rate, ie they can be changed in dependence on a control parameter P _i . This dependence is in Tab. 1 and FIG. 1 exemplified. Tab. 1 P ₁ P ₂ P ₃ P ₄ P ₅ P ₆ P ₇ P ₈ HE 0.1 0.4 0.7 1 1.3 1.7 2 2.3 DR 80.0 84.3 87.5 90.3 92.5 95.0 96.3 97.5

It can be seen that as the recognition rate increases, so does the error rate.

Thus, the benefit to the operator results from the parameterized relationship $$\mathrm{CB\text{'}}\left({\mathrm{P}}_{\mathrm{i}}\right)\mathrm{=}\mathrm{N}\left\{\mathrm{(}\mathrm{DFR\text{'}}\left({\mathrm{P}}_{\mathrm{i}}\right)\mathrm{*}\mathrm{DR}\mathrm{\left(}{\mathrm{P}}_{\mathrm{i}}\mathrm{\right)}\mathrm{*}\left({\mathrm{C}}_{\mathrm{NDD}}\mathrm{-}{\mathrm{C}}_{\mathrm{RF}}\right)\mathrm{)}\mathrm{-}\left(\mathrm{HE}\left({\mathrm{P}}_{\mathrm{i}}\right)\mathrm{*}{\mathrm{C}}_{\mathrm{RF}}\mathrm{*}\left(\mathrm{1}\mathrm{-}\mathrm{DFR\text{'}}\left({\mathrm{P}}_{\mathrm{i}}\right)\right)\right)\right\}$$

The estimated value for the current overlapping rate of the respective sorting machine is given by $$\mathrm{DFR\text{'}}\left({\mathrm{P}}_{\mathrm{i}}\right)\mathrm{=}\frac{\frac{{\mathrm{N}}_{\mathrm{AD}}}{\mathrm{N}}\mathrm{-}\mathrm{HE}\left({\mathrm{P}}_{\mathrm{i}}\right)}{\mathrm{DR\text{'}}\left({\mathrm{P}}_{\mathrm{i}}\right)\mathrm{-}\mathrm{HE}\left({\mathrm{P}}_{\mathrm{i}}\right)}$$

The entire recognition and optimization process is in FIG. 2 shown.

The block diagram in FIG. 2 clearly describes the procedure of the procedure. The overlap detector detects the overlaps during the sorting and distribution process 1, which are then passed into a reject bin and thus excluded from the regular process. This reduces the number of _{incorrect sorting} and thus the costs C _NDD for each unrecognized overlap. The shipments from the reject bins are fed to the sorting machine again, thereby causing additional costs C _RF per shipment. These costs C _NDD , C _RF serve as input data for the benefit optimization 2 for the process operator. They are derived offline from the process model. It is obvious that these costs are dependent on the nature of the respective processes, ie the Fehlsortierkosten at a so-called input sorting and distribution process according to the street and house number are different to the Fehlsortierkosten at a so-called output sorting and distribution process according to postcode and location. Table 2 shows an example of the corresponding costs for a specific machine type (FVM). Tab. 2 Type of sorting machine type C _NDD C _RF [U.S $] [U.S $] incoming sorting FVM 0.058 0.0035 default order FVM 0,040 0.0035

wird dann in einem Optimierungsprozess 2.1 der Parameter P_opt mit dem maximalen Nutzen CB' ermittelt.The sorting process 3 with the overlap detection contains the essential components singulation 3.1, subsequently the overlap detection 3.2 and then the sorting 3.3 according to the read destination addresses. The detected overlaps are again fed to the separation 3.1 via the reject compartments. The consignments detected as individual consignments (also the unrecognized overlaps) are distributed in the sort bins provided by the sorting plan. From the number of overlaps detected in the considered time period and the parameterized power values ER '(P _i ) and DR' (P _i ) 2.2, the current estimated overlap rate DFR 'is determined in the benefit optimization 2 2.4. From the utility equation / utility model 2.3 $$\mathrm{CB\text{'}}\left({\mathrm{P}}_{\mathrm{i}}\right)\mathrm{=}\mathrm{N}\left\{\mathrm{(}\mathrm{DFR\text{'}}\left({\mathrm{P}}_{\mathrm{i}}\right)\mathrm{*}\mathrm{DR}\mathrm{\left(}{\mathrm{P}}_{\mathrm{i}}\mathrm{\right)}\mathrm{*}\left({\mathrm{C}}_{\mathrm{NDD}}\mathrm{-}{\mathrm{C}}_{\mathrm{RF}}\right)\mathrm{)}\mathrm{-}\left(\mathrm{HE}\left({\mathrm{P}}_{\mathrm{i}}\right)\mathrm{*}{\mathrm{C}}_{\mathrm{RF}}\mathrm{*}\left(\mathrm{1}\mathrm{-}\mathrm{DFR\text{'}}\right)\right)\right\}$$

is then determined in an optimization process 2.1 the parameter P _opt with the maximum benefit CB '.

For the detector with the in Tab. 1 and FIG. 1 illustrated performance _parameters , error costs C _NDD = 0.058, C _RF = 0.0035 and N = 1,000,000 processed shipments results in the utility shown in Tab. 3 as a function of the overlap rate DFR 'and the parameter P _i . The maximum benefit for each overlap rate is indicated by underlining and bold. The associated setting parameter P _i is then the selected and automatically set parameter P _opt. Tab. 3 DFR P ₁ P ₂ P ₃ P ₄ P ₅ P ₆ P ₇ P ₈ 0.10% € 40.1 € 34.5 € 26.2 € 16.8 € 6.9 - € 4,8 - € 16.5 - € 26.7 0.40% € 170,9 € 180,1 € 178.3 € 172.2 € 164,0 € 153,9 € 143,9 € 134.6 0.70% € 301,7 € 325.6 € 330.5 € 327,7 € 321.1 € 312,7 € 304.4 € 295,8 1.00% € 432.5 € 471.2 € 482.6 € 483.1 € 478.2 € 471.5 € 464,8 € 457.1 1.30% € 563.3 € 616.7 € 634.7 € 638.5 € 635,3 € 630.2 € 625.2 € 618,4 1.60% € 694.2 € 762.3 € 786,9 € 794,0 € 792.3 € 789.0 € 785.7 € 779.7 1.90% € 825,0 € 907,9 € 939,0 € 949.4 € 949.4 € 947,8 € 946,1 € 941,0 2, 20% € 955,8 € 1,053.4 € 1,091.1 € 1,104.8 € 1,106.5 € 1,106.6 € 1,106.6 € 1,102.3 2.50% € 1,086.6 € 1,199.0 € 1,243.2 € 1,260.3 € 1,263.6 € 1,265.3 € 1,267.0 € 1,263.6

Claims (2)

1. Method for operating a detector for detecting overlaps of flat mail items in a transport path following on from a separation unit of a sorting machine,
   characterised in that

   a) before sorting

   a1) the average costs C_NDD for each undetected overlap in the distribution and sorting process and the average costs C_RF for sorting and distribution each detected overlapped mail item again are determined,

   a2) and as detection rate DR(P₁) of the detector the proportion of correctly detected overlaps and as error rate ER(P_i) the mail items detected incorrectly as overlap are determined as a function of an adjustment parameter P_i,

   b) and that during sorting, automatically or after an interval respectively which is defined by a specified time or all by a specified number of mail items processed in the sorting machine,

   b1) for the interval just ended, a detection result N_AD/N is determined from the number N_AD of overlaps determined in this interval by the detector and the number N of mail items determined in this interval by the detector,

   b2) from the detection result Nm/N together with the variables DR(P_i) and ER(P_i) depending on the value of the adjustment parameter P_i in this interval, an estimation DFR'(P_i) of the overlap rate which has actually occurred in this interval is undertaken,

   b3) from the utility CB'(P_i) of the operator or of the sorting and distribution process, $$\mathrm{CBʹ}\left({\mathrm{P}}_{\mathrm{i}}\right)\mathrm{=}\mathrm{N}\left\{\mathrm{(}\mathrm{DFRʹ}\left({\mathrm{P}}_{\mathrm{i}}\right)\mathrm{*}\mathrm{DR}\mathrm{\left(}{\mathrm{P}}_{\mathrm{i}}\mathrm{\right)}\mathrm{*}\left({\mathrm{C}}_{\mathrm{NDD}}\mathrm{-}{\mathrm{C}}_{\mathrm{RF}}\right)\mathrm{)}\mathrm{-}\left(\mathrm{ER}\left({\mathrm{P}}_{\mathrm{i}}\right)\mathrm{*}{\mathrm{C}}_{\mathrm{RF}}\mathrm{*}\left(\mathrm{1}\mathrm{-}\mathrm{DFRʹ}\left({\mathrm{P}}_{\mathrm{i}}\right)\right)\right)\right\}$$

   

   
   in optimisation calculations, those adjustment parameters P_opt are determined which would have produced the greatest utility CB' (P_opt) for this interval,

   b4) and for the next interval the value P_opt is determined as the value for the adjustment parameter P_i and the detector is operated for the next interval with this adjustment parameter.
2. Method according to claim 1, characterised in that the estimation DFR'(P_i) of the overlap rate which actually occurred in the interval just ended is undertaken in accordance with the equation. $$\begin{array}{cc}& \begin{array}{c}{\mathrm{N}}_{\mathrm{AD}}\\ ---\\ \mathrm{N}\end{array}\mathrm{-}\mathrm{ERʹ}\left({\mathrm{P}}_{\mathrm{i}}\right)\\ \mathrm{DFRʹ}\left({\mathrm{P}}_{\mathrm{i}}\right)=& ---------\\ & \mathrm{DRʹ}\left({\mathrm{P}}_{\mathrm{i}}\right)\mathrm{-}\mathrm{ERʹ}\left({\mathrm{P}}_{\mathrm{i}}\right)\end{array}$$

   

Images