CN101026837A - Adaptive power saving method for wireless mobile packet communication system - Google Patents

Abstract

When selecting size of initial sleep window, the disclosed self-adapting sleep mechanism of comm system based on wireless mobile packets does not take a fixed value, instead select a suitable value based on previous features of window. Mean while, adjusting size of sleep window by self-adaptation, the invention reduces unnecessary conversion from sleep window to listening window of mobile station (MS) effectively so as to raise efficiency for saving power, prolong service time of batteries of MS, and raise use ratio of wireless resource. The method includes procedures: after MS awakens from a time of sleep mode, through a certain time of no service state, and enters into next time of sleep mode, the invention feeds back parameters of previous sleep state and traffic performance parameters to base station. Based on received information, base station calculates size of initial sleep window, and adjusted size of sleep window, and informs them to MS.

Description

The adaptive power saving method of wireless mobile packet communication system

Technical field

This invention belongs to wireless communication field, not only be suitable for the IEEE802.16e wireless MAN, also be fit to the of future generation mobile radio system of various employing sleep patterns, 3G long evolving system (3G LTE) for example, WLAN (wireless local area network) etc. as energy-efficient mechanism.

Background technology

In future mobile communication system, along with the enhancing that terminal mobility is required, battery has become one of focus of people's concern its useful life gradually as the topmost energy supplier of mobile station equipment.The main development trend of future mobile communication system is the Packet data service towards a large amount of non real-times.Different with traditional circuit domain business, the outstanding feature of Packet data service is the sudden transmission of a large amount of high data rates, therefore can cause terminal in most of available machine time, to be in the state that does not have business datum to transmit, so, cause unnecessary energy consumption easily.At this problem, all mentioned among IEEE802.16e and the 3G LTE by adopting sleep pattern to reduce the energy consumption of travelling carriage under the no professional transfer of data situation.Because the sleep mechanism basic principle of IEEE 802.16e and 3G LTE is identical, main difference is in the concrete protocol signaling design, so next only be introduced at the sleep mode operation method that proposes in the IEEE 802.16e agreement, the core innovative point of this method is suitable equally to 3G LTE.

Sleep pattern is a kind of travelling carriage by the periodicity off position of consulting in advance with the serving BS air interface.This section is in off position from serving BS (hereinafter to be referred as the base station) observation travelling carriage upstream or downstream link period.The resource occupation that sleep pattern is used to reduce the power consumption of travelling carriage and reduces the serving BS air interface.

Travelling carriage in sleep pattern has two states: sleep state and intercept state.Call sleeping window being in the dormant one continuous period continuously, relative, be called listen window the one period continuous time that is in the state of intercepting continuously.In sleeping window, travelling carriage is closed its air interface, neither sends upstream data to the base station, does not also receive the downlink data of any base station, and promptly for the base station, the travelling carriage of this moment is sightless.Be inserted between listen window between two sleeping windows, during travelling carriage look like to be under the non-sleep mode state and carry out the transmission and the reception (being mainly administrative messag, ranging information etc.) of data between the base station.Sleeping window and thing followed listen window are collectively referred to as a sleep cycle.Wireless communication system is mainly by constantly adjusting the relevant parameter in each sleep cycle, to reach the purpose that realizes saving energy under the prerequisite of certain delay requirement satisfying.

At the characteristic of different business, stipulated 4 kinds of service quality (QoS) grade in the IEEE 802.16e agreement, be respectively: UGS, rtPS, nrtPS and BE business for Business Stream.Sleep pattern has been stipulated following three kinds of power save classes:

Power save class 1: towards nrtPS and BE business;

Power save class 2: towards UGS and rtPS business;

Power save class 3: towards multicast connection and bookkeeping.

Because towards different types of data flow, the sleep of every kind of power save class and the parameter of listen window determine it is different, wherein power save class 3 is primarily aimed at bookkeeping but not the main business flow data; Sleeping window size in the power save class 2 in each sleep cycle is constant, and since its towards periodic traffic, its sleeping window size can simply be provided with according to the specific object of Business Stream.

Adjustment to relevant parameter in the power save class 1 is then flexible relatively, and the size of sleeping window increases with the form of index, and promptly the sleep state time in next sleep cycle is the twice of interim sleep state time of the last week, jumps out sleep state up to the generation business; Listen window then remains unchanged, in order to receive signaling and other administrative messags of basic fixed length.The major parameter that the adjustment of sleeping window in the power save class 1 is related to has: initial sleeping window, and maximum sleeping window, sleeping window increases the factor etc.But this adjustment mode also exists some problems of not properly settling as yet, will carry out brief analysis to the method for adjustment of its sleeping window below.

Definite method of home window is not clear and definite in the agreement, when business is not busy, if its value is less, then before several sleep period times too short, the shared ratio of listen window is relatively large, then will inevitably cause unnecessary energy waste; When business is busy relatively, if its value is bigger, then can cause the increase of time delay again, the while may be followed overflowing of buffer memory and cause loss of data.

To increase the factor be 2 to the regulation sleeping window in the agreement, can not adapt to the demand under the different business busy extent, when business more after a little while, sleeping window increases relatively slow, such situation is similar to the too small waste that causes energy of home window; Business more for a long time, it is too fast relatively that sleeping window increases, it is big and cause time delay to be similar to home window.

The nrtPS business has certain sensitivity to time delay, and is sudden weak relatively; And the BE business is the poorest to the sensitiveness of time delay, and is sudden the strongest, therefore can not " make no exception " to these two kinds of business, should be distinguished on the sleeping window method of adjustment according to its traffic performance.

This shows, regulation about power save class 1 in the agreement is too simple, and efficient is lower, particularly in the BE business, owing to wherein the value of initial sleeping window is not made respective specified, so, when the less and initial sleeping window of business than hour, travelling carriage will enter listen window continually, and this can cause the lot of energy waste.

Above problem shows, needs a kind of adaptive approach according to different business kind and busy extent the sleep mechanism of its application to be adjusted.

Summary of the invention

The object of the present invention is to provide the adaptive power saving method that is fit to non-real-time service in a kind of wireless mobile packet communication system, can realize the self adaptation setting that initial sleeping window and sleeping window increase the factor by this method, thereby improve the power service efficiency of broadband wireless travelling carriage, prolong end cell service time.

Technical scheme of the present invention is to having proposed a kind of adaptive sleep pattern based on non-real-time service, specifically may further comprise the steps:

The first step judges that whether travelling carriage is for entering sleep pattern first;

Second step, when travelling carriage when entering sleep pattern first, with the initial sleeping window size that pre-defines and the sleeping window increase factor come the initialization sleep pattern go forward side by side into the sleep operation room every; When travelling carriage is non-when entering sleep pattern first, by in conjunction with last when withdrawing from sleep pattern the mean value of the comparison of last sleeping window size and threshold value and the preceding several sleep operations initial sleeping window size at interval determine the size and the sleeping window increase factor of initial sleeping window jointly;

In the 3rd step, enter sleep operation at interval;

In the 4th step, when will transmitting, business enters general mode;

In the 5th step, judging whether to continue for some time does not have business to transmit;

The 6th step did not have business to transmit if continue for some time, and then turned back to the first step; Otherwise got back to for the 5th step, repeat above-mentioned steps, cycling.

That is to say that travelling carriage is after a sleep pattern is waken up,, when preparing to enter again next time sleep pattern, some parameters and the traffic performance parameter of former sleep pattern need be fed back to the base station through the professional transmission of the nothing of certain hour.The base station just can calculate the sleep operation initial sleeping window size and the sleeping window increase factor at interval next time according to this method according to receiving these sleep pattern parameters and the traffic performance parameter of coming.When determining initial sleeping window size, consider stability and real-time problem, so when initial sleeping window size is set, using three factors (is α, β, γ) control respectively before the sleep operation initial sleeping window mean value at interval and the size of last sleep operation last sleeping window at interval several times, thereby initial sleeping window is set efficient, flexible, determine that according to traffic performance sleeping window increases the factor simultaneously, so just avoided travelling carriage, avoided again for energy-conservation and cause the decline of quality of service owing to professional less and waste big energy at sleeping window and listen window conversion continually.

Be used in this method determining that the parameter of initial sleeping window size and the sleeping window increase factor are to select respectively according to the difference of type of service, and after travelling carriage enters sleep pattern, concrete calculating in the method is carried in the base station end, travelling carriage just feeds back to the base station by signaling message, receive the parameter acknowledge message of sending the base station then, come to hold consultation with the base station with this.Sleeping window increases the factor and can decide by an operation factor, and for to the insensitive business of time delay, the value of this operation factor obtains bigger; For the business to delay sensitive, the value of this operation factor obtains less.

Based on IEEE 802.16e agreement, initial sleeping window size and sleeping window increase the factor and can transmit by MOB_SLP-REQ among the IEEE802.16e and MOB_SLP-RSP message, sleeping window when wherein the initiation of sleep pattern will be withdrawed from the sleep operation interval last time by travelling carriage is issued the base station, determine the initial sleeping window size and the sleeping window increase factor by the base station again, and issue travelling carriage by MOB_SLP-RSP and make it enter sleep pattern.

Wherein, the feedback of travelling carriage is to transmit by the MOB_SLP-REQ message of agreement regulation, it is initial sleeping window (initial-sleep window) item and reservation (Reserved) item that length is 3 bits of 8 bits that a length is arranged in this message body, can utilize initial sleeping window to transmit last sleep operation last sleeping window size at interval, with keeping the parameter that item transmits relevant traffic performance.It is to feed back by the MOB_SLP-REQ message of stipulating in the agreement that the size of travelling carriage sleep operation last time last sleeping window at interval and sleeping window increase the factor, be exactly to feed back sleep operation last time last sleeping window size at interval particularly in fact, feed back sleeping window with the reservation item and increase the factor by the initial sleeping window in this message.

Same, the base station is after receiving feedback, having calculated next time by this method, sleep operation initial sleeping window size and sleeping window at interval increases the factor, just transmit initial sleeping window size by the initial sleeping window in the MOB_SLP-RSP message, transmit the sleeping window increase factor with keeping item, so just reduced modification agreement.

Beneficial effect of the present invention mainly contains:

1. reduce frequent transitions unnecessary between sleeping window and the listen window owing to adaptive power saving method, thereby improved the power utilization efficient of travelling carriage, prolonged the battery service time of travelling carriage.

2. because travelling carriage has reduced unnecessary sleeping window and the frequent transitions between the listen window, obtained length of one's sleep of travelling carriage increase with regard to making, this has just reduced unnecessary the taking to air interface resource, thereby has improved the utilance to air interface resource.

3. because the prerequisite of this adaptive approach is exactly to guarantee professional quality, to decide by each professional characteristic for choosing also of parameter in the adaptive approach, this takes place with regard to the situation of having been avoided for conserve energy too reducing quality of service, thereby has guaranteed professional quality.

4. travelling carriage can utilize energy efficiently, and this has just reduced the interference to other travelling carriage on the other hand.

5. considered professional characteristic, increased the factor, improved power save efficient thereby reach by the analysis of traffic performance being selected initial sleeping window and sleeping window.

6. when business is comparatively busy, can not increase time delay the too much length of one's sleep because of existing.

7. when business is comparatively idle, can reach the comparatively reasonably length of one's sleep fully fast.

Description of drawings

Fig. 1: sleep pattern and adaptive power saving method are on average intercepted the number of times comparison diagram in the IEEE802.16e agreement

Fig. 2: sleep pattern and adaptive power saving method are at the energy comparison diagram of sleep interval internal consumption in the IEEE802.16e agreement

Fig. 3: based on the process chart of the adaptive power saving method of IEEE 802.16e

Embodiment

Fig. 1 and Fig. 2 have reflected the conventional method property comparison that proposes in adaptive power saving method among the present invention and the IEEE802.16e agreement.

As can be seen from Figure 1 adopted to make sleeping window reach suitable size faster behind the adaptive power saving method, thereby reduced sleep operation enters the state of intercepting in the cycle number of times, the more little then performance advantage of traffic carrying capacity is obvious more.

Fig. 2 has provided corresponding energy consumption contrast, and the energy that consumes in the state is very high in the heavy shared ratio of total power consumption owing to intercepting, and therefore along with entering the minimizing of intercepting the state number of times, causes energy consumption total in the whole sleep operation cycle also obviously to reduce.

Fig. 3 is based on the process chart of the adaptive power saving method of IEEE 802.16e among the present invention.Its handling process is as follows, after the mobile platform started, because record without any sleep pattern, so, in the time of when travelling carriage is not having the professional time through one section after, will entering sleep pattern for the first time, increase the factor with initialization sleeping window that pre-defines and sleeping window and come the initialization travelling carriage, make it enter sleep pattern.After this, travelling carriage comes to life from sleep state, and when entering sleep state once more after a while, just adopts adaptive approach to determine that initial window size and sleeping window increase the factor.

At first define a sleep operation (Sleep Operation Interval at interval, SOI) notion, so-called sleep interval just is meant that travelling carriage plays the time that is next time entered normal state by activation of service from entering sleep pattern, and it should comprise that all sleeping windows before initial sleeping window and the normal state add listen window specifically.Use SW _{Init_i}Represent the initial sleeping window size of i sleep interval, use SW _{Exit_i}Represent last sleeping window size of i sleep interval, in the listen window after this window is finished, travelling carriage enters state of activation.

When travelling carriage is prepared to enter sleep pattern once more, it at first feeds back to the base station by MOB_SLP-REQ message with sleep operation last time some parameters at interval, and these parameters comprise the size and the traffic performance of last sleep operation last sleeping window at interval.The base station just determines that by following concrete mode the size of initial sleeping window and sleeping window increase the factor after receiving these feedback messages.

Determining of initial sleeping window size

When determining i sleep operation initialization sleeping window size at interval, the sleeping window size SW in the time of can withdrawing from the last time _{Exit_i-1}With a threshold value SW _{Exit_TH}Compare, if SW _{Exit_i-1}More than or equal to SW _{Exit_TH}Think that then traffic carrying capacity is less, can equal according to the initial sleeping window size that method A obtains i sleep interval

${\mathrm{SW}}_{\mathrm{init}\_i}=\min ({\mathrm{SW}}_{\mathrm{init}\_\mathrm{TH}},\frac{\underset{k=i-N+1}{\overset{i}{\mathrm{\Σ}}}{\mathrm{SW}}_{\mathrm{init}\_k}}{N}\×\mathrm{\α}+(1-\mathrm{\α})\×{\mathrm{SW}}_{\mathrm{exit}\_i-1}\×\mathrm{\β})$

If SW _Exiti-1Less than SW _{Exit_TH}, think that then traffic carrying capacity is bigger, the initial sleeping window size that can reduce i sleep interval according to method B is

SW _{init_i}＝min(SW _{init_TH}，γ×SW _{init_i-1})

Wherein, α, the value of beta, gamma is according to the difference of type of service and difference, thereby can control initial sleeping window size neatly, improves power save efficient.

Sleeping window increases determining of the factor

Certainly, the present invention also comprises an aspect, and that is exactly in the sleep operation interval, and sleeping window increases the factor and can select according to the difference of type of service.For example, for the BE business,,, under the prerequisite that does not influence QoS, increase the size of sleeping window so can allow multiple that window increases more greatly because it is very low to delay requirement as far as possible; For the nrtPS business, because it is higher than BE to delay requirement, thus can be more careful being provided with of multiple, established a little bit smaller increase multiple.Like this, use μ _BEAnd μ _NrtPSRepresent the business for BE respectively, the increase rule of sleeping window is: in same sleep operation interval, next sleeping window is the μ of a last sleeping window _BEDoubly, for the nrtPS business, next sleeping window is the μ of a last sleeping window _NnPSDoubly.

Below this adaptive approach is carried out theory analysis.

Because under sleep pattern, the length of listen window is certain, so use LW _BERepresent the listen window size under the BE business, use LW _NrtPSThe size of representing the listen window under the nrtPS business, suppose simultaneously sleep operation at interval in the number that occurs of maximum sleeping window be k+1 (k=0,1,2...) individual, so just can express the length at i sleep operation interval.Under the BE business,

${\mathrm{SOI}}_i={\mathrm{SW}}_{\mathrm{init}\_i}+{\mathrm{\μ}}_{\mathrm{BE}}\×{\mathrm{SW}}_{\mathrm{init}\_i}+{\mathrm{\μ}}_{\mathrm{BE}}^2\×{\mathrm{SW}}_{\mathrm{init}\_i}+L+{\mathrm{\μ}}_{\mathrm{BE}}^{{\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}}\×{\mathrm{SW}}_{\mathrm{init}\_i}$

$+{\mathrm{LW}}_{\mathrm{BE}}\×({\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+1)+k({\mathrm{SW}}_{\max }+{\mathrm{LW}}_{\mathrm{BE}})$

Abbreviation can get

${\mathrm{SOI}}_i={\mathrm{SW}}_{\mathrm{init}\_i}\×\frac{1-{\mathrm{\μ}}_{\mathrm{BE}}^{{\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+1}}{1-{\mathrm{\μ}}_{\mathrm{BE}}}+{\mathrm{LW}}_{\mathrm{BE}}\×({\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+k+1)+k\×T_{\max }$

Illustrate, when k＜0, illustrate also not that travelling carriage has just withdrawed from sleep operation at interval, so can make k=0 this moment, makes following formula represent sleep operation size at interval time the maximum sleeping window; When k 〉=0, illustrate that a maximum sleeping window has appearred in travelling carriage at least, can make SW this moment _{Exit_i}=SW _MaxThereby, make following formula can represent sleep operation size at interval.

Same, can obtain i the sleep operation length at interval under the nrtPS business

${\mathrm{SOI}}_i={\mathrm{SW}}_{\mathrm{init}\_i}\×\frac{1-{\mathrm{\μ}}_{\mathrm{nrtPS}}^{{\log }_{{\mathrm{\μ}}_{\mathrm{nrtPS}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+1}}{1-{\mathrm{\μ}}_{\mathrm{nrtPS}}}+{\mathrm{LW}}_{\mathrm{nrtPS}}\×({\log }_{{\mathrm{\μ}}_{\mathrm{nrtPS}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+k+1)+k\×T_{\max }$

The arrival of supposing the downlink service data bag is that the obedience parameter is the Poisson distribution of λ, now the BE business is analyzed.

For i sleep operation at interval, the prerequisite of its appearance is in this sleep operation interval, except last sleeping window, other all sleeping windows all do not have the arrival of downlink data in the time, having only in the end, a sleep window eloquence might have data to arrive, specifically arrive as for data, all can not predict or control, the problem that this neither this method be concerned about in when.Like this, consider N sleep operation altogether at interval, the probability that occurs at interval of i sleep operation is exactly so

$p_i=(1-e^{-\mathrm{\λ}\×{\mathrm{SW}}_{\mathrm{exit}\_i}})\×e^{-\mathrm{\λ}({\mathrm{SOI}}_i-{\mathrm{SW}}_{\mathrm{exit}\_i})}$

If the power of travelling carriage in sleeping window is P _S, the power in listen window is P _L, the power of travelling carriage under state of activation is P _W, travelling carriage is W at the energy of whole sleep operation interval internal consumption _SOI, the energy of the sleeping window consumption in the sleep operation interval is E _S, be E at the energy of listen window consumption _LIf the energy that is in state of activation consumption on sleep operation length blanking time is E _WSo, obvious

E _TOTAL＝E _S+E _L

So just can obtain i sleep operation at interval in travelling carriage at the energy of sleeping window consumption be

$E_{\mathrm{SOI}\_i}=({\mathrm{SW}}_{\mathrm{init}\_i}+{\mathrm{\μ}}_{\mathrm{BE}}\×{\mathrm{SW}}_{\mathrm{init}\_i}+{\mathrm{\μ}}_{\mathrm{BE}}^2\×{\mathrm{SW}}_{\mathrm{init}\_i}+L+{\mathrm{\μ}}_{\mathrm{BE}}^{{\log }_{\mathrm{PBE}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}}\×{\mathrm{SW}}_{\mathrm{init}\_i})\×P_{S\_i}+{\mathrm{LW}}_{\mathrm{BE}}\×({\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+1)\×P_{L\_i}$

$+k({\mathrm{SW}}_{\max }\×P_{S\_i}+{\mathrm{LW}}_{\mathrm{BE}}\×P_{L\_i})$

The abbreviation following formula can obtain

$E_{\mathrm{SOI}\_i}=({\mathrm{SW}}_{\mathrm{init}\_i}\×\frac{1-{\mathrm{\μ}}_{\mathrm{BE}}^{{\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+1}}{1-{\mathrm{\μ}}_{\mathrm{BE}}}+k+T_{\max })\×P_{S\_i}+{\mathrm{LW}}_{\mathrm{BE}}\×({\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+k+1)\×P_{L\_i}$

Wherein

$E_{S\_i}=({\mathrm{SW}}_{\mathrm{init}\_i}\×\frac{1-{\mathrm{\μ}}_{\mathrm{BE}}^{{\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+1}}{1-{\mathrm{\μ}}_{\mathrm{BE}}}+k+T_{\max })\×P_{S\_i}$

E _{L_i}＝LW _BE×(log _μBESW _{exit_i}/SW _{init_i}+k+1)×P _{L_i}

Can get simultaneously

$E_{W\_i}={\mathrm{SOI}}_i\×P_{W\_i}={\mathrm{SOI}}_i\×P_{W\_i}({\mathrm{SW}}_{\mathrm{init}\_i}\×\frac{1-{\mathrm{\μ}}_{\mathrm{BE}}^{{\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+1}}{1-{\mathrm{\μ}}_{\mathrm{BE}}}+{\mathrm{LW}}_{\mathrm{BE}}\×({\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+k+1)+k\×T_{\max })\×P_{W\_i}$

So just can obtain gross energy E at N sleep operation interval internal consumption _TOTALFor

$E_{\mathrm{TOTAL}}=\underset{i=0}{\overset{N-1}{\mathrm{\Σ}}}{p}_i\×E_{\mathrm{SOI}\_i}=\underset{I=0}{\overset{N-1}{\mathrm{\Σ}}}(1-e^{-\mathrm{\λ}\×{\mathrm{SW}}_{\mathrm{exit}\_i}})\×e^{-\mathrm{\λ}({\mathrm{SOI}}_i-{\mathrm{SW}}_{\mathrm{exit}\_i})}\×[({\mathrm{SW}}_{\mathrm{init}\_i}\×\frac{1-{\mathrm{\μ}}_{\mathrm{BE}}^{{\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+1}}{1-{\mathrm{\μ}}_{\mathrm{BE}}})\×P_{S\_i}+{\mathrm{LW}}_{\mathrm{BE}}\×{\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}\×P_{L\_i}]$

The ENERGY E that sleeping window in N sleep operation interval consumes _SFor

$E_S=\underset{i=0}{\overset{N-1}{\mathrm{\Σ}}}{p}_i\×E_{S\_i}=\underset{i=0}{\overset{N-1}{\mathrm{\Σ}}}[(1-e^{-\mathrm{\λ}\×{\mathrm{SW}}_{\mathrm{exit}\_i}})\×e^{-\mathrm{\λ}({\mathrm{SOI}}_i-{\mathrm{SW}}_{\mathrm{exit}\_i})}\×({\mathrm{SW}}_{\mathrm{init}\_i}\×\frac{1-{\mathrm{\μ}}_{\mathrm{BE}}^{{\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+1}}{1-{\mathrm{\μ}}_{\mathrm{BE}}}+k\×T_{\max })\×P_{S\_i}]$

The ENERGY E that listen window in N sleep operation interval consumes _LFor

$E_L=\underset{i=0}{\overset{N-1}{\mathrm{\Σ}}}{p}_i\×E_{L\_i}=\underset{i=0}{\overset{N-1}{\mathrm{\Σ}}}[(1-e^{-\mathrm{\λ}\×{\mathrm{SW}}_{\mathrm{exit}\_i}})\×e^{-\mathrm{\λ}({\mathrm{SOI}}_i-{\mathrm{SW}}_{\mathrm{exit}\_i})}\×{\mathrm{LW}}_{\mathrm{BE}}\×({\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+k+1)\×P_{L\_i}]$

If travelling carriage is in the ENERGY E that state of activation consumes on N sleep operation interlude length _WFor

$E_W=\underset{i=0}{\overset{N-1}{\mathrm{\Σ}}}{p}_i\×E_{W\_i}=\underset{i=0}{\overset{N-1}{\mathrm{\Σ}}}[(1-e^{-\mathrm{\λ}\×{\mathrm{SW}}_{\mathrm{exit}\_i}})\×e^{-\mathrm{\λ}({\mathrm{SOI}}_i-{\mathrm{SW}}_{\mathrm{exit}\_i})}\×({\mathrm{SW}}_{\mathrm{init}\_i}\×\frac{1-{\mathrm{\μ}}_{\mathrm{BE}}^{{\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+1}}{1-{\mathrm{\μ}}_{\mathrm{BE}}}{+\mathrm{LW}}_{\mathrm{BE}}\×({\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+k+1)+k\×T_{\max })\×P_{W\_i}]$

The ratio that is defined in the energy that consumes when travelling carriage is in state of activation on energy that i sleep operation consume at interval and i sleep operation length blanking time is ζ _i, then

${\mathrm{\ζ}}_i=\frac{E_{\mathrm{SOI}\_i}}{E_{W\_i}}=\frac{({\mathrm{SW}}_{\mathrm{init}\_i}\×\frac{1-{\mathrm{\μ}}_{\mathrm{BE}}^{{\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+1}}{1-{\mathrm{\μ}}_{\mathrm{BE}}}+k\×T_{\max })\×P_{S\_i}+{\mathrm{LW}}_{\mathrm{BE}}\×({\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+k+1)\×P_{L\_i}}{({\mathrm{SW}}_{\mathrm{init}\_i}\×\frac{1-{\mathrm{\μ}}_{\mathrm{BE}}^{\mathrm{lo}{g}_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+1}}{1-{\mathrm{\μ}}_{\mathrm{BE}}}+{\mathrm{LW}}_{\mathrm{BE}}\×({\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+k+1)+k\×T_{\max })\×P_{W\_i}}$

So, it is as follows to obtain the power save efficiency eta of this method:

$\mathrm{\η}=1-\underset{i=0}{\overset{N-1}{\mathrm{\Σ}}}{p}_i\×{\mathrm{\ζ}}_i$

Substitution can get

$\mathrm{\η}=1-\frac{\underset{i=0}{\overset{N-1}{\mathrm{\Σ}}}(1-e^{-\mathrm{\λ}\×{\mathrm{SW}}_{\mathrm{exit}\_i}})\×e^{-\mathrm{\λ}(\mathrm{SO}{I}_i-{\mathrm{SW}}_{\mathrm{exit}\_i})}\×[({\mathrm{SW}}_{\mathrm{init}\_i}\×\frac{1-{\mathrm{\μ}}_{\mathrm{BE}}^{{\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+1}}{1-{\mathrm{\μ}}_{\mathrm{BE}}}+k\×T_{\max })\×P_{S\_i}+{\mathrm{LW}}_{\mathrm{BE}}\×({\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+k+1)\×P_{L\_i}]}{\underset{i=0}{\overset{N-1}{\mathrm{\Σ}}}(1-e^{-\mathrm{\λ}\×{\mathrm{SW}}_{\mathrm{exit}\_i}})\×e^{-\mathrm{\λ}({\mathrm{SOI}}_i-{\mathrm{SW}}_{\mathrm{exit}\_i})}\×[({\mathrm{SW}}_{\mathrm{init}\_i}\×\frac{1-{\mathrm{\μ}}_{\mathrm{BE}}^{\mathrm{lo}{g}_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+1}}{1-{\mathrm{\μ}}_{\mathrm{BE}}}+{\mathrm{LW}}_{\mathrm{BE}}\×({\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+k+1)+k\×T_{\max })\×P_{W\_i}]}$

Below η is carried out brief analysis, suppose P _{W_i}=P _{L_i}=10P _{S_i}, then following formula can be reduced to

$\mathrm{\η}=1-\frac{\underset{i=0}{\overset{N-1}{\mathrm{\Σ}}}(1-e^{-\mathrm{\λ}\×{\mathrm{SW}}_{\mathrm{exit}\_i}})\×e^{-\mathrm{\λ}(\mathrm{SO}{I}_i-{\mathrm{SW}}_{\mathrm{exit}\_i})}\×[({\mathrm{SW}}_{\mathrm{init}\_i}\×\frac{1-{\mathrm{\μ}}_{\mathrm{BE}}^{{\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+1}}{1-{\mathrm{\μ}}_{\mathrm{BE}}}+k\×T_{\max })+{\mathrm{LW}}_{\mathrm{BE}}\×({\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+k+1)\×10]}{\underset{i=0}{\overset{N-1}{\mathrm{\Σ}}}(1-e^{-\mathrm{\λ}\×{\mathrm{SW}}_{\mathrm{exit}\_i}})\×e^{-\mathrm{\λ}({\mathrm{SOI}}_i-{\mathrm{SW}}_{\mathrm{exit}\_i})}\×[({\mathrm{SW}}_{\mathrm{init}\_i}\×\frac{1-{\mathrm{\μ}}_{\mathrm{BE}}^{\mathrm{lo}{g}_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+1}}{1-{\mathrm{\μ}}_{\mathrm{BE}}}+{\mathrm{LW}}_{\mathrm{BE}}\×({\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+k+1)+k\×T_{\max })\×10]}$

$=\frac{9}{10}\frac{\underset{i=0}{\overset{N-1}{\mathrm{\Σ}}}(1-e^{-\mathrm{\λ}\×{\mathrm{SW}}_{\mathrm{exit}\_i}})\×e^{-\mathrm{\λ}(\mathrm{SO}{I}_i-{\mathrm{SW}}_{\mathrm{exit}\_i})}\×\left[({\mathrm{SW}}_{\mathrm{init}\_i}\×\frac{1-{\mathrm{\μ}}_{\mathrm{BE}}^{{\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+1}}{1-{\mathrm{\μ}}_{\mathrm{BE}}}+k\×T_{\max })\right]}{\underset{i=0}{\overset{N-1}{\mathrm{\Σ}}}(1-e^{-\mathrm{\λ}\×{\mathrm{SW}}_{\mathrm{exit}\_i}})\×e^{-\mathrm{\λ}({\mathrm{SOI}}_i-{\mathrm{SW}}_{\mathrm{exit}\_i})}\×[({\mathrm{SW}}_{\mathrm{init}\_i}\×\frac{1-{\mathrm{\μ}}_{\mathrm{BE}}^{\mathrm{lo}{g}_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+1}}{1-{\mathrm{\μ}}_{\mathrm{BE}}}+{\mathrm{LW}}_{\mathrm{BE}}\×({\log }_{{\mathrm{\μ}}_{\mathrm{BE}}}{\mathrm{SW}}_{\mathrm{exit}\_i}/{\mathrm{SW}}_{\mathrm{init}\_i}+k+1)+k\×T_{\max })]}$

As can be seen, first of molecule and denominator is identical in the multinomial in the following formula, under the situation of not considering time delay, and SW _{Init_i}And μ _BEBig more, then can make second in the denominator more little, the η value that finally obtains is also just big more, this basic thought with the adaptive power saving method of our proposition herein is consistent.

Claims (8)

1. the adaptive power saving method of a wireless mobile packet communication system, be applicable to the sleep mechanism of IEEE802.16e and 3G long evolving system, by the relevant parameter in each sleep cycle of continuous adjustment, satisfy the saving of energy under certain delay requirement, it is characterized in that this method may further comprise the steps:

The first step judges that whether travelling carriage is for entering sleep pattern first;

Second step, when travelling carriage when entering sleep pattern first, with the initial sleeping window size that pre-defines and the sleeping window increase factor come the initialization sleep pattern go forward side by side into the sleep operation room every; When travelling carriage is non-when entering sleep pattern first, by in conjunction with last when withdrawing from sleep pattern the mean value of the comparison of last sleeping window size and threshold value and the preceding several sleep operations initial sleeping window size at interval determine the size and the sleeping window increase factor of initial sleeping window jointly;

In the 3rd step, enter sleep operation at interval;

In the 4th step, when will transmitting, business enters general mode;

In the 5th step, judging whether to continue for some time does not have business to transmit;

The 6th step did not have business to transmit if continue for some time, and then turned back to the first step; Otherwise got back to for the 5th step, repeat above-mentioned steps, cycling.

2. adaptive power saving method according to claim 1 is characterized in that,

In second step is not the situation that enters sleep pattern first at travelling carriage, when last sleeping window size is more than or equal to threshold value when the last time is withdrawed from sleep pattern, be weighted processing according to the preceding mean value of initial sleeping window several times and the size of last sleeping window, unite to determine the size of initial sleeping window; When last sleeping window size is less than threshold value when the last time is withdrawed from sleep pattern, only determine the size of initial sleeping window according to the size of last sleeping window.

3. adaptive power saving method according to claim 2 is characterized in that,

Be used in second step determining that it is that difference according to type of service is configured respectively that the parameter of initial sleeping window size and sleeping window increase the factor.

4. adaptive power saving method according to claim 1 is characterized in that,

After travelling carriage enters sleep pattern, the calculation of parameter relevant with the sleep pattern adjustment carried out at the base station end, travelling carriage is just given the base station these message feedback by signaling, receives the information that base station feedback is returned then, finishes and the negotiation of base station about sleep operation with this.

5. adaptive power saving method according to claim 2 is characterized in that,

Sleeping window increases the factor and can decide by an operation factor, and concerning the insensitive business of time delay, the value of this operation factor obtains bigger; For the business to delay sensitive, the value of this operation factor obtains less.

6. according to the described adaptive power saving method of arbitrary claim in the claim 1 to 5, it is characterized in that,

Determine the size of initial sleeping window by two weight factors, comprise a factor of being responsible for the initial sleeping window mean value of control, guaranteeing its stability, and one be responsible for the last factor that withdraws from the sleeping window size of control, to embody real-time.

7. adaptive power saving method according to claim 6 is characterized in that,

Under situation based on the IEEE802.16e agreement, initial sleeping window size and sleeping window increase the factor and can transmit by MOB_SLP-REQ among the IEEE 802.16e and MOB_SLP-RSP message, sleeping window when wherein the initiation of sleep pattern will be withdrawed from the sleep operation interval last time by travelling carriage is issued the base station, determine the initial sleeping window size and the sleeping window increase factor by the base station again, and issue travelling carriage by MOB_SLP-RSP and make it enter sleep pattern.

8. adaptive power saving method according to claim 6 is characterized in that,

Under situation based on the IEEE802.16e agreement, it is to feed back by the MOB_SLP-REQ message of stipulating in the agreement that the size of travelling carriage sleep operation last time last sleeping window at interval and sleeping window increase the factor, be exactly to feed back sleep operation last time last sleeping window size at interval particularly in fact, feed back sleeping window with the reservation item and increase the factor by the initial sleeping window in this message.

Images