FR2732527A1 - METHOD AND DEVICE FOR PERFORMING TRANSFERS IN CELLULAR RADIOCOMMUNICATION SYSTEMS - Google Patents

Abstract

The invention relates to transfer techniques in cellular radio communication. In a method of transferring to a candidate base station (202) a connection established between a serving base station (201) and a mobile station (203), received signal level values respectively associated with the signal are measured. connection established and connection possible; determining a variation of these received signal level values over a time interval; and transferring the connection to the candidate base station (202) based on the received signal level values and the variations determined. Application to mobile radiotelephony.

Description

I

  The present invention generally relates to

  processes and radiocommunication systems.

  and more particularly such systems in which a connection can be transferred from one channel or base station to another. In radiocommunication systems, we have

  characteristically the possibility of transferring

  establish an established connection, for example between a

  mobile station and a base station to another base station as the mobile station changes position and thereby leaves the coverage range of a base station and enters the coverage of another base station. Depending on the channel assignment plan that is used in the system, it may also be desirable to transfer a connection of a

  channel from the base station to another channel.

  In all cases, these processes are commonly

peeled transfer or "handoff".

  To achieve a transfer in good conditions, the network that controls the stations of

  basis first determines, for each station

  bile, if the need for a transfer is imminent, and then it determines which new base station or channel to transfer to. To make this last decision, it is desirable that the network control unit

  know the level of quality with which each

  that base station can receive signals from

  mobile station, which is the quality level with which the mobile station can receive signals from each station.

or both.

  Many existing radiocommunication systems are based on an access process that is

  calls Multiple Access by Frequency Distribution

  quence (FDMA), in which each mobile station transmits on a particular frequency within

  its current base station area. The modern station

  bile therefore knows nothing about signals on other frequencies from surrounding base stations. In FDMA systems, it is typically considered too expensive

  to equip mobile stations with a receiver

  which could be used to scan other frequencies of base stations. In place,

  it is common for base stations to be equi-

  scanned receptors carrying out an explo-

  concerning the signals of mobile stations

  who are approaching. The network then transfers a

  bile of a base station covering an area that the mobile is departing from, to the base station which signals the best signal reception of the

mobile station.

  More mobile phone standards

  use Time-Domain Multiple Access (TDMA), in which a fixed period of time

  (eg 20 ms) on each pair of radiofrequency

  quences (one used in the direction from

  mobile station to the base station (global link)

  the other used in the direction from the base station to the mobile station (downlink)) is divided into a number (for example 3) of short time slots (for example 6.6 ms), which are used cyclically by different mobile stations. Thus, a first mobile station transmits in the first time slot in each period, a second mobile station transmits in the second time slot in each period, and so on. Similarly, the base station transmits to a mobile station in the first time slot, to another mobile station in the second time slot, and so on. By shifting the allocation of time slots in the two directions of communication, ie from the base to the mobile (the downlink) and from the mobile to the base (the uplink), one can make that a first mobile emits in

  the first time slot and receive in the se-

  time slot; that a second mobile emits in the second time slot and receives in the third; and a third mobile emits in the

  third time slot and receive in the first

  first time slot. An advantage of this configuration

  ration is that a mobile station does not need to transmit and receive simultaneously,

which reduces the disturbance.

  In the example with three time slots

  above, each mobile station is active for transmitting

  to be or to receive in two of the three slots

  time, and it is inactive in the temporary window.

  porel remaining. Mobile stations operating in

  TDMA can therefore use this time

  active to receive signals from other

  base stations and to measure their level of

  gnal. By transmitting these level measurement of

  general to the base station, using a slow data channel multiplexed with the traffic (i.e. the voice), the network is informed with respect to the base stations that each mobile station can receive. The network can use this information

  to transfer to the best

  and such a process is called "mobile assisted handover" (MAHO). When the stations

  base perform a scan to determine the

  mobile station signal level, the process of

  It should be called "basic assisted handover" (BAHO). Systems offering MAHO also have access to

  base stations, and so they are able to

  kill transfers in better conditions and with greater reliability, because of the

  account of the signal levels of the global link

  aunt and the downlink, instead of only taking into account the levels of the link

rising in the case of the BAHO.

  In yet another access technique, which is Code Division Multiple Access (CDMA), mobile stations can share the same

  frequency band, but it is possible to distin-

  communication through the use of

  particular. Even in CDMA systems, it

  It is possible to measure a signal level of

  pilots associated with a particular base station

  die. The base station and / or the mobile station may use this information to determine when a transfer to another code, or

  another frequency band in the CDMA with

multiple, is desirable.

  As can be seen from the above, a common element in determining when a transfer is desirable is

  measuring the level of signal received at the station

  station or at the mobile station, or both. Of

  classical way, the decision to trans-

  Factor is based on this signal level information, taking into account a certain constant hysteresis value. If, for example, the signal level received by a mobile station from a base base of an adjacent cell exceeds that of the base station to which the mobile station is connected at the present moment, plus the hysteresis value, then a transfer to the base station of the adjacent cell takes place. By introducing the hysteresis value, we usually thought

  accidental transfers due to fluctuations in

  signal level could be minimized,

  so that the mobile station does not suffer a

  ping-pong between base stations,

cell boundaries.

  However, the introduction of a hys-

  constant teresis is a poor substitute for the ability to accurately determine whether received signal levels actually or not suggest

  that a particular mobile station must make the

  jet of a transfer. For example, a disadvantage

  deny is that a hysteresis value

  constant needs to be fixed relatively

  so that the large instantaneous fluctuations that are possible in the radio signal levels, for example due to fading and multipath propagation, do not trigger

  not incorrect transfers. However, using

  If a hysteresis value is set at a relatively high level, some mobile stations will not be transferred at the time when a transfer would actually optimize the signal quality.35 It would therefore be desirable to determine more exact the properties of the received signal, concerning the transfer, so as to be able to refine the criteria which have been used in a classical way

  to determine if a transfer is desirable.

  These and other problems, defects and limitations are eliminated in accordance with the

  invention, in which the variation of the

  signal is determined over a time interval, both for cells to which a mobile could be transferred (candidate base stations) and for the present cell (the station

  basic serving). In this way, decisions

  Transfers are based on information

  which guarantees that the signal level in a

target lule is not unstable.

  In accordance with embodiments

  For example, if relatively high signal level fluctuations are noted,

  use a relatively higher hysteresis value

  for transfer determinations. If on the contrary the signal level fluctuations are

  relatively low, we can then use a

  their relatively low hysteresis. From this

  certain areas (eg rural areas,

  which) have variations in the level of

  which are less abrupt than variations in other areas, receive a hysteresis value

  lower, which improves the quality of transmission

  medium speed by providing faster transfer to better base stations. The value

  hysteresis will therefore vary from one situation to another

  be. Signal intensity levels are determined using, for example, Kalman filtering,

  and a spread of the signal levels is calculated. It is then possible to make the transfer decision

  formally to this information, that is to say

  both levels of intensity of

  and signal spread during an inter-

time lapse.

  Goals will be more easily understood, because

  The features and advantages of the present invention indicated above, as well as others, on reading

  detailed description that follows, referring to

  in conjunction with the drawings in which: Figure 1 shows the levels of downlink signals that are received by a mobile station in the embodiment of the present invention shown in Figure 2; Figure 2 shows a moving body which makes a change of direction at an intersection, and which is associated with the signals shown in Figure 1; FIG. 3 is a flowchart for a transfer for the mobile station shown in FIG.

  Figure 2, when making a change of

  rection at an intersection; and Fig. 4 is a block diagram of an exemplary base station according to the present invention. In accordance with embodiments of the

  present invention which are given by way of example

  ple, the base station or the fixed network comprises processing circuits for deciding, for each mobile station, the moment at which a transfer is necessary, and for performing a transfer at such times. Transfer decisions are made in

  using as a criterion not only

  signal levels of the candidate base stations and

  but also the variations between

  are signal levels before the transfer. In one embodiment, the transfer decision is made by considering the downlink signal level of the candidate and serving base stations, as measured by the mobile station and communicated to the terrestrial system, for example using the function of "transfer assisted by the

  "mobile" (or MAHO).

  two levels of downlink signal

  over a time interval, for example 2 se-

  onds. In general, we can express the

  by any mathematical measure of dispersion, such as variance, standard deviation, and dynamics. In the following example, these variations are expressed by standard deviations (so-called

  feature using the Greek letter sig-

  ma) respective signal measurements. The level of each signal is then adjusted, for example to

  down, in a quantity that is a function of the difference

  respective type. For example, we apply a

  "penalty" to each signal by subtracting from its

  signal calf measured and filtered an amount that increases when the signal standard deviation becomes more

great.

  It is also possible to add the

  quantity at the signal level, instead of the sub

  milking the signal level. The quantity that is used

  in the adjustment of the signal level depends on

  of the variation, for example of the standard deviation, of the

  received signal calf. In general, she is

  function of the variation of the respective signal level

  tif over a period of time. A simple example of how the amount of adjustment, a, depends on the variation, dSS (equal to sigma if the

  variation is the standard deviation) is given by the expression

  following: a = k * dSS where k is a constant. In one embodiment, k is a dynamic variable whose value is in the range

0 <k <1.

  The adjusted signal level SSat of the candidate base station is then compared to the level of

  adjusted signal SSas of the base station of

  serat, and a transfer is ordered if: SSat> SSas + deltaHYl

  by deltaHY1 denoting a hysteresis of safety

  having a constant low value, for example 1 dB, to provide stability, i.e.

  to avoid oscillatory transfers from a

  mobile service, traveling back and forth between

the two base stations.

  It is important to understand that the hysteresis

  sis or safety margin, deltaHYl, is not neces-

  to provide an improved transfer. The fact that the respective signal variations vary

  course of time is usually sufficient to

  to obtain better and more reliable transfers. This-

  during, a small, constant safety margin was used to ensure stability. It should be noted that a margin of safety of 1 dB is very low in

  comparison with conventional hysteresis values

c.

  In another embodiment, the technical

  that the invention is applied to propagation attenuation instead of being applied to the signal level. The propagation attenuation between a mobile station and a base station is determined by

  subtracting the received signal level, in dB, from the

  signal level, in dB. The signal levels that are subtracted relate to the same link, i.e. the uplink or the downlink. If for example we use the connection

  downward to calculate the propagation attenuation

  tion, although the measuring equipment in the

  mobile measurement usually has a significant measurement error, it will provide propagation attenuation

  (Lat) from the candidate base station to the station

  the same measurement error as the corresponding propagation attenuation (Las) of

  the serving base station to the modern station

  bile. This is useful because the errors of

  the measurement equipment is canceled out-

  mutually when calculating the difference between

  (Lat - Las) between the two propagation attenuations. Thus, in a manner similar to that explained for the embodiment

  preceding, in accordance with the prior embodiment

  sent, a transfer is ordered if: Lat <Las - deltaHY2 by deltaHY2 denoting a safety hysteresis having a low and constant value, for example 1 dB, to provide stability, that is to say to avoid transfers oscillatory oscillations giving rise to trips back and forth between the two base stations. In the two embodiments envisaged

  above, the small security hysteresis delta-

  HY1, deltaHY2 are constant, while the standard deviations (sigma) of the respective signals vary over time. We will now describe a detailed embodiment, given by way of example, which applies the above principles. This embodiment uses the variation of the signal level of the candidate base station to choose between two possible hysteresis values, while the signal level of the serving base station.

is supposed to be constant.

  In what follows, reference is first made to FIGS. 1 and 2. Reference 101 designates the signal level function SStb of the candidate base station 202, when measured at the mobile station 203, based on time. The reference

  reference 202 indicates the prefiltered value SSre which

  This is done by prefiltering the signal level 101 in the mobile station over a time interval of 0.5 seconds. The mobile station 203 communicates twice per second its prefiltered measurements, that is to say that the interval 131 between communications is 0.5 seconds. About 10 measurements 101 are made

  during a period of 0.5 seconds. The values

  filtered which are communicated by the mobile station are filtered again in the serving base station 201, over a previous time interval, for example a period of about 2 seconds, which

  gives a signal level curve 103 more regular

  for filtered SSfi values for the candidate base station 102. The corresponding SSsb signal level of the serving base station.

  201 is supposed to be very stable and have an ampli-

  constant study, and prefiltered and filtered values

  Thus, SSsb's are both referred to as "SSsb". This signal is represented in FIG. 1 in the form of a constant level 105. The "penalty" for the transfer which is caused by

  signal variations of the base station of des-

  serte 201 is therefore assumed to be zero. In addition, to simplify FIG. 1, it has been assumed that the small constant hysteresis deltaHY1 is equal to

  zero, and she is not represented.

  Usually, that is, in systems

  my existing, we used a hysteresis 114 fixed and relatively large, leading to a transfer at time t5, because the level of filtered signal

  103 of the candidate base station exceeded the

  111 between t4, t5, point 121. Processing and transfer decisions take place after the receipt of measurement

  the terrestrial system, at times tl, t2 ,. . . t6.

  In accordance with the present invention, instead of using

  To read a constant hysteresis 114, the hysteresis or "penalty" to be used varies according to the stability of the signal levels. In the case contemplated here, a variation occurs only in the signal level 101 of the candidate base station 202, but one skilled in the art will note that variations in

  the signal level of the base station of des-

  serte. A transfer takes place respectively to the

  t4 or t6, depending on the variation that is

  associated with the signal level 103 during the two se-

previous counts.

  A transfer at time t4 is a hypo-

  the thetic in Figure 1 and this instant would be the moment for the transfer if the signal level 101 increased slowly instead of making a jump between times t3, t4, as it does in the figure. The hysteresis value 115 which is small (relative to the hysteresis 116) would be applied in this hypothetical case, and the filtered signal level 103 would cross the level 112 at point 122, thus causing an early transfer at time t4 . Instead, since the signal level 101 jumps between times t3, t4, the

  large hysteresis 116 is applied for decisions

  subsequent transfers, since the

  riation of signal levels has increased. A trans-

  then occurs at time t6, because the filtered signal level 103 passes level 113 at point 123. The jump is detected by the system

  at time t4, because the pre-existing value

  filtered filtered, 102, which is communicated at time t4, is much greater than the value 102 which has been communicated at time t3. The criterion

  for the selection of the relatively large hysteresis-

  sis 116 is that the jump of the values 102 must exceed a first predefined threshold. The criterion for returning to the small hysteresis 115 consists

  in that the changes in the values 102 must

  be less than a second predefined threshold, during the same predetermined time interval,

for example, 2 seconds.

  In another embodiment of the present invention,

  In this invention, the criterion that is used to switch to the higher hysteresis and return to the lower hysteresis is the standard deviation (sigma).

  associated with signal level 103, instead of a different

  between pre-filtered values 102. If sigma

  is above a threshold for a certain

  time, for example 2 seconds, the upper hysteresis is used, and if sigma is below another threshold during this time interval, the lower hysteresis is used again. The way to determine sigma is well known in the

  when filtering, for example

  by a method described by Kalman. Because Kalman filtering is in itself a technique

  classic, it is considered that a detailed description

  this technique is beyond the scope of this document.

  is lying. However, the interested reader may obtain

  additional context information concerning

  using Kalman techniques with reference to

  following documents: "Applied Optimal Estimation", Cha-

  4, by A. Gelb, 1974, or "Stochastic Models,

  Estimation and Control ", Volume 1, by Peter S. May-

  Beck, 1979. By choosing a lower hysteresis

  when signals are stable, and a hysteria

  resis upper 116 when signals are more unstable, instead of always choosing a hysteresis-

  sis 114 constituting a compromise between the two

  their, it is possible to accelerate the transfer in the case of stable signals and to delay it in the case of more unstable signals, in comparison with

* the delay of the prior art. This improves the quality

average transmission capacity.

  Figure 2 shows the physical conditions

  in which the described transfer takes place. A sta-

  moving 203 moves along a road 204 to

  a moment when an appeal is in progress, through

  connection with a base station of

  201. The mobile station communicates

  pre-filtered signal level res, 102, which are influenced by radio masking and / or reflections caused by buildings 206 and

  207. When the mobile station arrives at the intersec-

  tation of Highway 204 and 205, Signal 101 which

  is received by the mobile station from the station

  Candidate base 202 increases sharply

  than at a much higher value. The mechanisms

  of the invention, described above,

  would then select the hysteresis of

  for the purpose of avoiding a transfer if the

  mobile station was to continue to move in

  right and stay on Route 204, in which case the

  signal level 101 would fall back to its lower value

  of origin. However, as the mobile

  kills a turn to take Route 205, a trans-

  This is done because the signal level remains high enough for a sufficient time.

  to satisfy the condition of equality.

  Fig. 3 is a flowchart illustrating the steps involved in the handover for the mobile station shown in Fig. 2, at the time of a turn. In step 301, the signal levels 101 and 105 are measured and prefiltered to become the filtered values 102 and 105 (because of a constant value) which are communicated at 0.5 second intervals to the base station. in which these values are further filtered and the sigma parameter or some other value representing the variation is calculated. In step 302, it is decided whether the variation of the signals during a certain time interval, for example 2 seconds, warrants or not the selection of a hysteresis value different from that which is used at the present time. If the decision is YES, the change is made at step 303, after which the sequence continues at step 304. If the decision is NO, the sequence goes directly from step 302 to step 304. step 304, it is decided whether a transfer is to be made or not, on the

  basis of the signal level values and the hysteresis

  selected resis. If the decision is YES, a transfer is made at step 305, after which the sequence returns to the starting point, whereas if the decision is NO, the sequence returns to the point

  departure without making a transfer.

  Figure 4 is a functional block diagram

  showing a base station system

  according to the invention. However, the material

  shown in the diagram of Figure 4 can also

  be divided among several units, for example

  be a BTS base station and a BCS base station control unit. To cover both cases, the functional block diagram of Figure 4 has been called the "base station system", BSS. The

  block BSS 800 has three antennas, out of which two, 810 and 812, are used to receive signals, while

only one antenna, namely

  811, is used to transmit

  nals. An emitter 820 is connected to the antenna 811.

  The transmitter is also connected to the processing unit

  840. An 830 receiver is connected

  antennas 810 and 812. The receiver 830 is also

  connected to the signal processing unit 840.

  Radio equipment for modula-

  demodulation and equalization, is incorporated in blocks 820 and 830. The

  signal processing 840 includes means for co-operating

  channel definition and decoding, and it also

  function of processing speech

  trants and outgoing. The signal processing unit 840 is also connected to the PCM link adapter block 860 and the control logic 850. The foregoing techniques can be implemented by means of a program registered in the control logic 850. and this program is called "Kalman logic" and carries the designation 851. The control logic 850 performs the call control, as is known in the art. The adapter block

  MIC 860 is connected to the network by means of

  from one or more MIC lines.

  The present invention therefore continuously adapts

  circumstances, in a way that

  the criteria or criteria used to determine

  if a transfer is desirable, by determining

  and considering variations in levels of

respective receipts.

  The embodiments described above as examples are intended to illustrate to all

  the present invention, and not to limit it.

  The present invention can therefore be the object, in

  detailed implementation, numerous amendments to

  which the person skilled in the art can deduce from the

  cription presented here. As an example, it should be noted that the adjustment value, a, may depend on several ways of variation, dSS. It is considered that all changes and modifications of

  the detailed embodiment form part of the

  taken and the scope of the present invention, such

  that they are defined by the following claims:

wind.

Claims (15)

  A method for transferring to a candidate base station (202) a connection established between a serving base station (201) and a remote unit (203), characterized in that it comprises the following steps: received signal level values associated with the connection   between the serving base station (201) and the uni-   far away (203); we measure second values of   received signal level that are associated with a con-   possible connection between the candidate base station (202) and the remote unit (203); a variation of the first signal level values is determined   received, and a variation of the second values of   calf of signal received; and transferring the connection to the candidate base station (202) based on the first received signal level values, the second received signal level values, and the variations that are determined.   2. Process according to claim 1, comprising   furthermore the step of applying a   Kalman treatment at the first and second their received signal level.   3. Process according to claim 1, characterized   in that the determining step comprises Kalman filtering of the first and second values received signal.   4. Process according to claim 1, characterized   in that the determining step comprises the step of determining a standard deviation of the received signal level values over a time interval, and a standard deviation of the second values of   received signal level over a time interval.   5. Process according to claim 1, characterized   in that the transfer step includes the   following steps: a value of hys-   teresis (303) on the basis of the deter-   a relatively large variation gives a relatively large hysteresis value, and so that a relatively small variation gives a relatively high hysteresis value.   low; comparing (304) a second value of   a signal received with a sum of a first received signal value and the hysteresis value; and transferring the connection (305) to the candidate base station (202) based on the result of this comparison.   6. Process according to claim 1, characterized   in that the transfer step includes the   following steps: a value of hys-   teresis (303) on the basis of the deter-   such that a relatively large variation gives a relatively large hysteresis value, and a relatively small variation gives a relatively low hysteresis value; comparing (304) a filtered value of at least one of the second received signal level values with a sum of the hysteresis value and a   filtered value of at least one of the first   their received signal level; and transferring the connection (305) to the candidate base station   (202) based on the result of this comparison reason.   7. Process according to claim 1, characterized   in that the transfer step comprises the following steps: the expression SSat> SSas + deltaHY1 is evaluated, in which SSas is a filtered value of the first received signal level values, which is adjusted on the basis of the variation first received signal level values, SSat   is a filtered value of the second values of ni-   received signal calf, adjusted on the basis of   of the second signal level values re-   ce, and deltaHY1 is a safety margin; and transferring the connection (305) if the evaluated expression is validated.   A transfer decision system for a radio communication system for determining whether a transfer is to be performed from a serving base station (201) to a candidate base station (202), characterized in that it comprises: means (850) for measuring and filtering first received signal level values associated with a connection between the   serving base station (201) and a remote unit   aged (203); means (850) for measuring and   filtering second signal level values re-   which are associated with a possible connection between   the candidate base station (202) and the remote unit   aged (203); means (850) for determining a value   riation of the first signal level values   received, and a variation of the second values of   calf of signal received; and transfer means (850) for transferring the connection to the station   candidate (202) based on the first   received signal level values, the second received signal level values and the variations which have been determined.   9. System according to claim 8, characterized   in that the determination means (850) de-   end a standard deviation of the first values of ni-   received signal calf filtered over a range of   time, and a standard deviation of the second values of ni-   received signal calf filtered on a second time lapse.   A method for transferring to a candidate base station (202) a connection established between a serving base station (201) and a remote unit (203), characterized in that it comprises the following steps: of signal   received that are associated with the connection between the   serving basic base (201) and the remote unit (203), and a possible connection between the candidate base station (202) and the remote unit (203);   variations in the signal levels re-   ceived; the received signal levels are adjusted by   based on the calculated variations, the propagation attenuation on a link between   the remote unit (203) and the base station of   (201), and on a link between the remote unit   generated (203) and the candidate base station (202), using the adjusted received signal levels; and we   transfers the connection to the base station can-   didate (202) based on a comparison of propagation attenuations.   11. The method of claim 10, wherein   in that it also includes the step of   both to filter the received received signal levels, and that the adjustment step is performed to   received signal levels filtered.   12. Method for transferring to a station   basic candidate (202) a connection which is   between a serving base station (201) and a remote unit (203), characterized in that it comprises the following steps: measuring the received signal level which is associated with the connection between the serving base station (201) and the remote unit (203), measuring the received signal level which is associated with a possible connection between the candidate base station (202) and the remote unit (203); the variation of the received signal level over a time interval is calculated for the established connection and the possible connection, respectively; we   adjusts the measured received signal level, respectively   for the established connection and the possible connection   sible, based on the calculated variation   sponding; comparing the received received signal level adjusted for the connection established between the serving base station (201) and the remote unit (203),   with the received signal level adjusted for the   possible connection between the candidate base station (202) and the remote unit (203); and transferring the connection to the candidate base station (202) in   based on the result of this comparison.   13. The method of claim 12, wherein   in that the comparison step is carried out   killed in relation to a constant safety margin.   14. The method of claim 12, wherein   characterized in that it further comprises the step of filtering, respectively for the connection   established and the possible connection, the level of   received over a time interval, and in that the adjustment step is performed for the level of received signal filtered.   15. Device for transferring to a station   candidate base station (202) a connection which is established between a serving base station (201) and a remote unit (203), characterized by comprising: means (850) for measuring the received signal level which is associated with the connection between the serving base station (201) and the unit   distant (203); means (850) for measuring the   received signal that is associated with a possible connection between the candidate BS (202) and   the remote unit (203); means (850) for calculating   for the established connection and the   possible connection, a variation in the level of   received over a period of time; means (850) for adjusting the measured received signal level,   respectively for the established connection and the   possible, based on the cal-   corresponding abutment; means (850) for comparing the adjusted received signal level for the connection   between the serving base station (201) and the uni-   remote (203), with the received signal level adjusted for possible connection between the station   candidate base (202) and the remote unit (203); and transfer means (850) for transferring the connection to the candidate base station (202) based on the result of the comparison.