method for power control is proposed for broadband, packet-switched TDMA wireless networks. By observing the temporal correlation of cochannel interference when transmitters can send data contiguously, a Kalman filter is used to predict interfence power in
nin (3), the base station instructs (via a downlinkI
channel) the terminal to transmit in slotnwith power
pn
ni_ = β__.gn
*
(11)
notation, (13a) in [FM93] for adjusting transmission powercan be re-written as
zn__pn+1=( 1 γ)pn+ γ β*_ (13)gnwhereγis a proportionality constant between 0 and 1, andznis the measured interference power in mW,corresponding toZnin dBm in (2). Evidently,γprovides"smoothing" effects as it tends to reduce changes oftransmission power from one slot to the next. Results in[FM93] reveal that settingγto 1 maximizes theconvergence rate, which is needed for bursty packet traf cin the broadband wireless networks under consideration.However, it will be inappropriate to determinetransmission power in the networks according to (13)whenγ =1. This is so because, as discussed earlier, theinterference powerin(thus its measured valuezn) uctuates greatly in the packet-switched networks, asdifferent terminals are transmitting bursts of packets attimes. Having observed the temporal correlation for theinterference when a message consists of multiple packets,the Kalman- lter method bases on the previousinterference measurements to predict interference power inthe next time slot. Then, the power is determined by thepredicted interference level according to (11). Clearly,(11) and (13) are similar. In fact, we can show thefollowing:
Proposition.If the path gain (i.e., the combined effect ofpath loss and shadowing) and thermal noise for each radiolink between a terminal and a base station does not changein time and the message length is in nitely long, theKalman- lter method is identical to the Foschini-Miljanicapproach withγ =1.
Proof.Since the path loss, shadow fading and noise foreach radio link are assumed to remain unchanged in timeand since a xed set of terminals have been transmittingfor an in nitely long time, we have a)gn=gn+1for allnand b) the interference power received at a base station_
becomes a constant. The latter impliesZi→Znasn→ ∞according to (8). Thus, bothQnandRnin (9) and (10)become zero. These zero values reveal that the Kalman
n=Znby (6). Then, the ltergainKnis 1 in (5). In turn,I
n+1=I n=Zn. Combining this with the fact a),predictsI
the methods thus adjust the power in the same way as (11)and (13) withγ =1 become identical.
Due to terminal movement, change of thermal noiseand fast fading (not considered in this paper), the path gainfor a radio link always changes in time and message lengthmay be short and bursty in the broadband, packet-switchednetworks. Nevertheless, the proposition helps understandthat the Kalman- lter method and the Foschini-Miljanicapproach indeed operate based on a similar theoreticalbasis. In particular, our performance results show that theKalman- lter method is capable of providing signi cantimprovement in term of SINR performance for nitemessage length over systems with no power control. Morespeci cally, for typical radio environments, the Kalman-
The goal of this setting of transmission power is to choosejust enough power to achieve the target SINRβ*, thusminimizing interference to others without degrading one’slink quality. In fact, different SINR targets can be used in(11) for different terminals, depending on their path gain tothe associated base stations and their application needs.For example, for a poor radio link with large path loss andunfavorable shadowing, the link can adapt to the poorquality by using a smaller size of modulation constellation.Thus, a lower SINR target is used in (11) to support adecreased data rate for the terminal. Nevertheless, weassume all terminals have an identical target SINR here.When power is selected according to (11), the actualSINRβnreceived at the base station in slotnis given by
βn
np_n___gni*_ _ __== β
inin
(12)
whereinis the actual interference power in mW for slotn.
It is clear from (12) that when the interference prediction
n~by the Kalman lter is accurate (i.e.,i~in), the target ndoes not predictinSINR is achieved. Even wheni
exactly, the method also helps in reducing the spread of
nandinare strongly correlated. In fact,βn, provided thati
the way that (11) selects the transmission power tends to
nandin. This enhance the correlation betweeniis so
because a higher predicted interferenceinyields a higher
transmission powerpn, which in turn causes the actualinterferenceinto be higher.
3.4 Steps for the Kalman-Filter Method
The Kalman- lter method for controlling transmissionpower for each time slotncan be summarized as:a. For each time slotn, each base station measures the
interference power for the time slot.b. The interference measurements are used as input to
the Kalman lter in (3) to (10) to predict the
n+1(or equivalently,i n+1) ininterference powerI
the next slotn+1.c. Based on the MAC protocol in use (which satis es
Assumption 2), the base station tracks the path gaingn+1, and selects the transmission power by (11) tomeet a given target SINR for the terminal thattransmits in slotn+1.d. The power levelpn+1is forwarded via the downlink
to the terminal for actual transmission.4. THEORETICAL BASIS FOR THE KALMAN-FILTER METHODThe Kalman- lter method for power control is basedon the theory of Foschini-Miljanic approach for powercontrol [FM93], as explained as follows. Using our
(0); //initialize with a measurement x_est_last = z_real + frand()*0.09; float sum_error_kalman = 0; float sum_error_measure = 0; for (...
% [x, V, VV, loglik] = kalman_filter(y, A, C, Q, R, init_x, init_V, ...) % % INPUTS: % y(:,t) - the observation at time t % ...
源于1957年他在哥伦比亚大学的一篇博士 论文和1960年发表的论文《A New Approach to Linear Filtering and Prediction Problems》,60年代Kalman 把状态空间模型引入滤波...
y〔n〕=Cx〔n〕 为 A = 〔1.1269-0.49400.1129 1.0000 0 B = 〔-0.3832 0.5919 0.5191〕; C = 〔1 0 0〕; 我们的目标是设计 Kalman 滤波...
卡尔曼全名 Rudolf Emil Kalman,匈牙利数学家,1930 年出生于匈牙利首都布达佩斯。...4. 简单例子 (A Simple Example) 这里我们结合第二第三节, 举一个非常简单...
(cm == 'A') mtype = 2; else error('Unkown measurement model.'); ...(sd,T.nmeas); % Kalman gain % Initialize filter L_init = L_max/2.0...
Kalman Filter介绍_数学_自然科学_专业资料。卡尔曼滤波(kalman filter)最简单...4 简单例子(A Simple Example) 这里我们结合第二第三节,举一个非常简单的例子...
(A,[B B],C,0,-1,'inputname',{'u' 'w'},'outputname','y'); Q = 1; R = 1; [kalmf,L,P,M] = kalman(Plant,Q,R); kalmf = kalmf(...
Kalman 滤波 在学习卡尔曼滤波器之前,首先看看为什么叫“卡尔曼”。跟其他著名...(k-1k-1)对应的 covariance,A’表示 A 的转置矩阵,Q 是系统过程的 ...
Kalman滤波中文版 17页 1下载券 kalman滤波的C与C++实现 10页 1下载券 ... Kalman滤波的公式推导...
热门文档