Publisher Item Identifier S 0733-8716(00)00191-8.
cars, remote hospitals, military services, homes, and office automation systems is an attractive proposition. It would free the users from cords or optical fibers tying them to particular locations within the building, thus offering true mobility which convenient and sometimes even necessary. The development ofmultimedia terminals will support the ever-growing demand for mixed data, audio, and video applications and will connect the portable pen pad and lap-top devices to backbone information resources and computational facilities. The possibility ofmultimediaservices will allow services such as dial-up video conference, video-on-demand (VOD) services, and portable PC-based applications incorporating video/audio/data transfer to any location. Moreover, a number of different mobile users can simultaneously request multimedia data from one or more multimedia servers on the network. Each multimedia server is capable of catering to multiple data requests from multiple users, simultaneously. Presentation of preorchestrated multimedia information requires synchronous playback of time-dependent multimedia data according to some prespeci- fied temporal relations. At the time of creation ofmultimedia information, a user needs a model to specify temporal con- straints among various data objects which must be observed at the time of playback. Usually, the temporal relationships ofmultimedia information may be characterized by a timeline diagram which is the commonly used tool in commercial multimedia authoring products. Fig. 1 depicts an example of a timeline diagram and its associated multimedia title generated by the most commonly used product called MacroMind Di- rector. Although the timeline diagram is a useful description tool, it has a lot of redundancies in characterizing the temporal relationships and is not suitable for further analysis and system evaluation, however. To tackle this difficulty and to obtain a more compact multimedia representation, a well-known model called object-composition Petri-net (OCPN) ,  is able to describe the temporal relationships of the various components of a multimedia document and represents them in the form of a graph. Since preorchestrated multimedia information has highly time-varying bandwidth, the fixed bandlimited constant bit rate (CBR) wireless channel may not be appropriate for the variable bit rate (VBR) multimediaservices. Therefore, it is desirable to design a dynamic mechanism to manage and allocate bandwidth according to the changing levels of concurrencies ofmultimedia data streams. Woo et al. have introduced a dynamic RF channel capacity allocation to deal with the OCPN-based multimedia data stream. In this paper, an alternative method has been proposed to provide a cost-effective resource allocation scheme for the OCPN-based multimediaservices by employing the well-known antimultipath spread
technologies: SIP and MPLS to provide good quality to users. The proposed architecture is suitable for VoIP applications. Especially, it also provides a solution to security and traffic engineering issues. This VoIP-over-MPLS architecture employs the SIP protocol for signaling to set up call connection and uses the MPLS core network to forward packet efficiently. Traffic engineering in MPLS can contribute to congestion avoiding, load balancing and fast rerouting capabilities. With the implementation of traffic engineering, VoIP packets can be forwarded fast and efficiently. Especially, we put emphasis on the necessary components to provide such interactive and secure services. The three major parts in this architecture are described in detail, which includes the edge devices, access network, and MPLS core network. The necessary elements and requirements for these major parts are also described in details in this paper as well. The modified SIP registration and signaling process are proposed to give a global view of how this architecture functions well in achieving the security purpose and maintaining the QoS for the VoIP applications.
Fig. 7 shows the handoff rate versus the calling rate per user for schemes of NFRM, FCAC, OCA, and CCA at time . It reveals that NFRM has more handoff rate than OCA by an amount of 2%. The reason is that the designof NFRM is based on the knowledge of FCAC and CCA, which combines overflow, reversible, and underflow. Fortunately, the signaling overheads for these handoffs might not cost so much as those for conventional handoffs between macrocells since most of these handoffs occur in the same macrocell. It also reveals that NFRM achieves less handoff rate than CCA and FCAC by an amount of 14.9% and 6.8%, respectively. It is not only because of more information, such as the speed of mobile station considered in NFRM, but also because of the neural fuzzy logic control that can provide decision support and expert system with powerful reasoning and learning capabilities.
In Table II, we compare the computation complexity of the two approaches at different bandwidth (chip rate) and window sizes, assuming that the processing gain is 512. We also assumed that the number of RAKE fingers (for data detection) and the number of taps of the reconstruction filter (for pilot signal re- construction) are both four for the equivalent time-domain ap- proach. Table II shows that the FFT approach does not always have a lower computation complexity than the time-domain ap- proach, but its computation advantage becomes more apparent for a wide-band system operating an environment with a large delay spread. Furthermore, this architecture is very suitable for a multicode system. As high-data-rate transmission is needed formultimediaservices, one method suggested is to use multiple code channels for a single user. For a conventional receiver, each added code channel uses three to four additional fingers to de- tect data. With the FFT-based architecture, however, each added code channel employs only two additional multipliers for data code depreading and channel matching.
As VBR MPEG video traffic is both delay sensitive and has a high degree of burstiness, it is commonly believed that a large number of spreading codes corresponding to the peak source bit rate must be reserved for the video transmission to satisfy its QoS requirements. The reservation of a large number of spreading codes assigned to each MPEG video seems extremely inefficient for the transmission of multiple MPEG videos over a narrow-band radio channel. Taking into consideration that the ratio between the peak and average bit rate for VBR MPEG video is generally high, this implies that a large portion of spreading codes remains unused when the low bit-rate B-frames are transmitted via the multicodeCDMA network. To increase the efficiency in the spreading code assignment mechanism, we designed an algorithm which dynamically allocates an appropriate number of codes to each MPEG video on the basis of its actual source rate. Furthermore, in order to avoid the self- interference that a MPEG employing multiple codes may incur, the multiple codes to/from one MPEG should be made orthogonal. This particular spreading coding scheme is called the concatenated orthogonal/pseudonoise (PN) spreading code  which is capable of subdividing a high bit-rate MPEG stream into several parallel lower basic bit-rate streams without self-interference. In addition, the
For the transmission of images over SS-CDMA AWGN channels, a subband coding scheme that divides the image information into a number of independent data streams using an analysis filter bank, each of which is multiplied by its unique signature PN code, enables the transmission of these data streams via multiple parallel virtual channels created by their correspond- ing PN codes. With a sufficiently large number of streams, the total signal is able to fit within the narrow radio channel bandwidth even though the total bandwidth of all the signals may exceed the channel bandwidth. At receiver, each received signal is separately recovered at the decoder by multiplying its PN code and integrating over the code length in order to obtain the desired subband. All the recovered subbands are then reassembled by a synthesis filter bank into a close reproduction to the original image. Additionally, for color subband image transmission, color images are first transferred to luminance (Y) and two chrominance components (I, Q). Each component is then decomposed independently into several subbands for SS-CDMA transmission. Therefore, a number of additional PN codes are required to support the transmission of the chrominance signals over the CDMA channels whereas the luminance signal was treated in the same manner as monochrome pictures. Moreover, SS-CDMA allows more than one image to be transmitted and be accessed simultaneously at the same limited channel bandwidth.
The second major factor in picture degradation is caused by the transmission error in the payload containing MPEG-based multiresolution 3-D video data, even though its associated ATM cell is successful. The impact of transmission error in MPEG video data generated from the left image sequence seems very significant in the picture quality since both the DCT coefficients and macroblock headers are encoded using run-length coding (VLC). A modification of the original MPEG codec proposed by ISO/IEC is made to localize the error in a picture frame. For instance, a DCT coefficient is lost if a single bit error occurs in that coefficient. A loss of the complete macroblock occurs when a bit error appears in its header. Moreover, a slice of sub- sequent macroblocks may be destroyed when a bit error occurs in a slice header. For P and B frames generated from the left image sequence, a bit error in a motion vector may result in the corruption of this macroblock and the following predicted mac- roblocks since motion vectors are differentially encoded for the intercoded macroblocks in the same slice. On the other hand, a number of subsequent macroblocks in the right image may be destroyed when a bit error appears in the disparity vector. The adaptive outer/inner FEC code combining has been proposed to protect the payload in accordance with the importance of data types in MPEG-based multiresolution 3-D videos including ei- ther disparity information in the low-resolution subimages or the data belonging to the detail subimages plus motion and dis- parity-compensated information.
B ANDWIDTH C OMPARISON R ESULTS B ETWEEN PSPN C ODES AND PN C ODES
Fig. 4. Plot of bandwidth versus toggle rate of the spreading codes.
The block of “frequency divider” generates the clock with data rate. PN code generator generates the PN code. Despreader is used for despreading procedure and the decision circuit detects the signal. The transistor netlist of the blocks in Fig. 6 is implemented. The circuit level simulator Hspice simulates the power consumption of the transistor netlist. All the blocks shown in Fig. 6 are included in this power consumption simulation. The circuit schematics described in Fig. 6 could be operated by different spreading codes with different code lengths. That is to say, this is a soft-coded spreadspectrum system. The simulation results of the power consumption are listed in Table IV. From Table IV, we find the percentages of reduction for power consumption range from 8% to 14% with PSPN codes compared to PN codes. The concept of low toggle rate means low-power consumption has been verified by the simulation results.
Channel modeling simulation tools that enable researchers and designers to accurately predict the performance of wire- less systems become increasingly important as personal com- munications and wireless data services evolve. A basic under- standing of the channel is important not only for designing modulation and coding schemes for robust communication over such channel, but also for investigating the channel fad- ing impact on existing networking algorithms, such as rout- ing and power adjustment which critically depend on channel attenuation. At present, most network protocol simulations and even power control algorithms are using the free space (distance) channel propagation model which is basically only function of transmitter-receiver distance. Typically, for the indoor environment, the channel characteristics are much too complex to be modeled by simple distance functions. Yet, a realistic channel model is essential for network protocol eval- uation, especially in the presence of mobility. Therefore, a more realistic channel fading model which accounts for chan- nel quality variations with movement is needed for network protocol simulation.
troller contains a power-spectrum-indexed table for managing multimedia call requests, where traffic characteristics of call re- quests are described by the power spectrum. The power spectrum can be obtained from the claimed traffic parameters of peak rate, mean rate, and peak rate duration; the power spectrum has been shown to have a dominant effect on system performance. The results show that the proposed power-spectrum-based connec- tion admission control method achieves higher system utilization and lower call-blocking probability than the equivalent-capacity allocation method.
Fig. 4. Adaptive PRNN predictor.
neurons can be divided into a number of simpler small-scale RNN modules with less computational complexity. In the following section, we are trying to apply the PRNN structure to improve the computational performance of performing the nonlinear adaptive predictor. Fig. 4 shows the system diagram of PRNN-based nonlinear adaptive filter for NBI rejection. The PRNN-based filter is placed behind the channel and receives the channel output. The network inputs at time provides the desired response of the PRNN to train the network to achieve the optimal nonlinear adaptive filter by minimizing the prediction error (residual) . The estimate of actually contains a large portion of the NBI signal component since the other SS signal and AWGN components which have a flat spectrum cannot be predicted and are then filtered out by the PRNN-based filter. In other words, the predicted value of the received signal is, in effect, equal to the estimate of the NBI. Thus, by subtracting the estimate from the received signal, the residual signal becomes a sum of an AWGN and a SS signal. Once the optimal residual signal is obtained, the resulting prediction error signal is used as the input to the SS detector, i.e., PN correlator.
However, the primary goal of embedded system is usually meeting the performance need at a minimum price, rather than achieving higher performance. Although using many HW components can speed up the running time ofmultimedia applications, it has the disadvantages of high cost, great depletion of capacity, and big measure of area. Then the drawbacks make it can’t apply to embedded system . Therefore, it is necessary to carefully decide which computation task executed by SW component and which computation task implemented by HW component for the sake of keeping the balance between cost and efficiency. Besides, it is also necessary to specify the execution order of these components, so as to speed up the period of system development.
In the first experiment, a task graph with eight functional nodes and twelve communication edges is adopted. The available number of pipeline stage S MAX , available number of local bus BN, available number of each type of processors, and available number of each type of ASIC are 3, 3, 1, and 1, respectively. In addition, there are three discrete voltage levels for each type of processor, but only one voltage level for each ASIC. To compare with other optimal approaches, we simplify our ILP formulation to approximate the conventional optimal approaches and compare their design quality in terms of total area and total power consumption. Three approximate approaches (‘-dvs’, ‘-pipe’, and ‘-local’) are adopted in this experiment. The minus symbol means that the corresponding design step is removed from our ILP model. That is, ‘-dvs’ represents that DVS doesn’t be performed by removing the two lower voltage levels for all processors. The ‘-pipe’ is the approach without pipelined scheduling by setting the available pipeline stage number S MAX to 1. The ‘-local’ approach gives the limitation that all data are only transferred through the system bus and the number of available local bus BN is set to 0.
Fig. 10 presents a chip photograph of this paper and its summary. The technology used is a 0.18-μm complementary metal–oxide–semiconductor (CMOS) with a supply voltage of 1.8 V. The BIST circuit area is 15% of the SSCG, which includes the DFFs and phase-shift detector of the MPD but does not include the accumulator or other digital circuits. The BIST is operated at 20 MHz, which is the same as the reference clock frequency. The SSCG is a fractional-N PLL with a ten-phase 1.2-GHz VCO, a third-order loop filter, and a MASH-111 SDM to meet the SATA-III specification. In total, 10 4 data outputs by the phase-shift detector are recorded by a logic analyzer.
Power consumption is an important design issue of current em- bedded systems. It has been shown that the instruction cache accounts for a signiﬁcant portion of the power dissipation of the whole chip. Data caches also consume a signiﬁcant portion of total processor power formultimedia applications because they are data intensive. In this paper, we propose two mecha- nisms to reduce dynamic power consumption for both instruc- tion and data caches. The HotSpot cache adds a small cache between the CPU and L1 instruction. It identiﬁes frequently accessed instructions dynamically and stores them in the L0 cache. The software-controlled cache architecture improves the energy efﬁciency of the data cache by allocating data types in an application to different cache regions. On each access, only the allocated cache regions need to be activated. We ﬁnd that on the average, the HotSpot cache and software-controlled cache can achieve 52% and 40% energy reduction on instruc- tion and data caches, respectively. Both schemes incur little performance degradation.
Index Terms— Correlator, delay-locked loop, direct sequence spreadspectrum, tracking error variance.
I. S YSTEM D ESCRIPTION AND S IGNAL M ODEL
I N THIS letter, we present a code tracking receiver with less complexity, by employing a differentially coherent tech- nique originally proposed for pseudonoise (PN) acquisition receiver . The proposed differentially coherent delay-locked loop (DCDLL) scheme is shown in Fig. 1. The received signal is first filtered by front-end band-pass filter (BPF) and the bandwidth of BPF is . is set to be chip rate ( , where is the chip duration). Then this proposed DCDLL scheme processes the received signal using a differential decoder with a delay of -chip duration in the delay path. The decoder output is then correlated with the difference of the advanced (early) and retarded (late) versions of the local PN code to produce an error signal. After the error signal is filtered by a low-pass filter (LPF), then it drives the voltage-controlled clock (VCC) through the loop filter and corrects the code phase error of the local PN code generator. In this proposed system, the bandwidth of LPF, denoted as , is set to be the system data rate ( , where is the data bit duration). The processing gain of this direct-sequence spread-spectrum (DS/SS) system is thus given by or . Usually, if the system is applied in ranging, and