Opportunistic interference alignment (OIA) is known to achieve the optimal degrees-of-freedom (DoF) in the interfering multiple-access channel (IMAC) with independent and identically distributed (i.i.d.) Rayleigh fading, provided that a certain user scaling condition is satisfied. We analyze the performance of OIA in a poor scattering K-cell single-input multiple-output IMAC, where there exist finite paths between the transmitter and receiver sides. Under the feasible model, we derive a new fundamental user scaling law, required to achieve a target DoF, which generalizes the existing achievability result shown for the i.i.d. Rayleigh fading case. Our main result indicates that KS DoF is achievable if the number of per-cell mobile stations (MSs) scales at least as SNR(K-1)min(L,S), where L denotes the number of paths and S denotes the number of simultaneously transmitting MSs per cell. To verify our achievability result for finite system parameters, computer simulations are performed along with comparison to the i.i.d. Rayleigh channel case. The amount of leakage of interference is numerically evaluated and is shown to be consistent with our theoretical result. The achievable sum-rates are also evaluated.

A variety of methods for color constancy have been proposed in order to come up with the ability of the human visual system which recognizes the inherent colors of objects under different illuminants. The methods based on the low-level statistics are widely used due to their low computational complexity and satisfactory results with adequate parameters. However, none of these methods utilize color channel correlation explicitly to improve the color constancy. In this paper, we propose a novel color constancy method using the color correlation. From a lot of observations, we find tendencies according to saturation values and finally incorporate a saturation weighting function into the existing methods. Experiments are performed on two widely used datasets and the results demonstrate that the proposed method improves the color constancy with a simple and effective manner.

In this paper, we investigate a joint scheduling and power control for a multi-cell cognitive radio network (CRN) in coexistence with a cellular primary radio network (PRN). The PRN uses a set of licensed orthogonal channels for transmission. This set of channels is also accessed in an opportunistic manner by a set of cognitive radio base stations (CR-BSs) to support secondary users (SUs). The problem is formulated to maximize the spectrum utilization of SUs without causing excessive interference to active primary users (PUs) of the PRN. In addition, all the served SUs must meet a certain QoS. A centralized solution for joint scheduling and power control is derived to give the global accessing decision for all unserved SUs. With the assumption that the knowledge of all subscribers is available, the coordinator of the CRN can use the joint scheduling and power control algorithm to maximize the spectrum utilization of served SUs by solving a mixed-integer linear programming (MILP) with an NP-hard complexity. To avoid the NP-hard complexity, we propose a sub-optimal heuristic algorithm that can be obtained at a much lower complexity based on the coloring interference graph among unserved SUs effected by served SUs and active PUs. Simulation results show that the proposed algorithm can yield the performance close to the optimal solution's.

In this paper, we investigate a joint scheduling and power control for an infrastructure-based cognitive radio network (CRN) in coexistence with a cellular primary radio network (PRN). The PRN uses a set of licensed nonoverlapping orthogonal frequency channels for transmission. This set of channels is also accessed in an opportunistic manner by a set of cognitive radio base stations (CR-BSs) to support secondary users (SUs). The problem is formulated to maximize the spectrum utilization of SUs without causing excessive interference to active primary users (PUs) of the PRN. In addition, all the serviced SUs must meet a certain Quality of Service (QoS), such as satisfying a predefined signal to interference noise ratio (SINR). A centralized solution for joint scheduling and power control is derived to make the global accessing decision for all unserved SUs. With the assumption that the knowledge of all subscribers is available, a coordinator of the CRN can use the joint scheduling and power control algorithm to maximize the spectrum utilization of serviced SUs by solving a mixed-integer linear programming (MILP) with an NP-hard complexity. To avoid the NP-hard complexity, we propose a suboptimal heuristic greedy algorithm that can be obtained at a much lower complexity based on the coloring interference graph among unserved SUs effected by serviced SUs and active PUs. Its superior performance over the existing algorithms is demonstrated through simulations.

In this paper, we propose dynamic spectrum sharing methods for femto base stations (BSs) operated within a macro BS's spectrum area. As the spectrum sharing methods, we provide traffic pattern based resource hopping schemes of the femto BSs. Traffic pattern information which reflects channel activities for each subband used by the macro BS is utilized by the femto BSs to select proper resources. We evaluate our proposed method compared with pure random hopping method in terms of resource hit probabilities. Also we evaluate our proposed method combined with direct interference measurement results, compared with pure interference measurement based random hopping method. Since traffic pattern information reflects future channel activities, simulation results show that the proposed methods can outperform pure random hopping and conventional methods which utilize past information such as interference measurement results only.

In this paper, we investigate a joint scheduling and power control for an infrastructure-based cognitive radio network (CRN) in coexistence with a cellular primary radio network (PRN). The PRN uses a set of licensed orthogonal channels for transmission. This set of channels is also accessed in an opportunistic manner by a set of cognitive radio base stations (CR-BSs) to support secondary users (SUs). The problem is formulated to maximize the spectrum utilization of SUs without causing excessive interference to active primary users (PUs) of the PRN. In addition, all the served SUs must meet a certain QoS. A centralized solution for joint scheduling and power control is derived to give the global accessing decision for all unserved SUs. With the assumption that the knowledge of all subscribers is available, the coordinator of the CRN can use the joint scheduling and power control algorithm to maximize the spectrum utilization of served SUs by solving a mixed-integer linear programming (MILP) with an NP-hard complexity. To avoid the NP-hard complexity, we propose a sub-optimal heuristic greedy algorithm that can be obtained at a much lower complexity based on the coloring interference graph among unserved SUs effected by served SUs and active PUs. Its superior performance oveover other related algorithms is demonstrated through simulations.

In the multicarrier transmissions such as orthogonal frequency division multiplexing (OFDM), one of the main problem is the high peak-to-average power ratio (PAPR) of transmission signals due to the superposition of many subcarriers. In this paper, we propose an approach for PAPR reduction using tone reservation (TR). The proposed approach guarantees a certain bound of PAPR reduction gain by using only one reserved subcarrier symbol. The bound depends on the ratio of the first and second peak magnitude. The proposed approach can be applied to not only one reserved subcarrier but also multiple reserved subcarriers. In the latter case, two methods are presented. To get more PAPR reduction gain, the proposed approach can be modified by adjusting the appropriate threshold.

Multi-hop wireless mesh networks (M-WMNs) have recently emerged as an alternative network infrastructure. For M-WMNs to serve as a network infrastructure, the protocol designs for M-WMNs should target a high network capacity. An efficient backoff algorithm is required to achieve a high network capacity. Most of existing works focused on backoff algorithms for infrastructure environment based on IEEE 802.11 distributed coordination function (DCF) or did not consider the features of M-WMN. In this paper, we propose an adaptive backoff algorithm based on the number of contention nodes for an IEEE 802.11 DCF-based M-WMN. The proposed backoff algorithm is founded on cross-layer cooperation between the routing and the medium access control (MAC) layers. The number of contention nodes of a node is obtained from the routing layer, and then this information is used to adaptively adjust the value of the minimum contention window (CWmin) in the backoff mechanism of the MAC protocol. To evaluate the performance of the proposed backoff algorithm and existing ones, computer simulations were performed. The simulation results show that the existing algorithms do not work well in M-WMN and the proposed backoff algorithm outperforms existing ones.

A Femto base station (BS) is connected to operator's network using a residential DSL or cable broadband connection and provides service in the licensed spectrum. When a BS operates with time division duplex mode (TDD), synchronization of the transmission time at the BS is important to avoid inter-slot interference with adjacent BSs. Because femto BSs are generally located indoor, they may not receive the global positioning system (GPS) signal which is used to synchronize the wireless cellular network. We propose a multi-hop based network synchronization scheme for femtocell systems. In the femtocell system, synchronization can be achieved by aligning its frame start point with the neighbor cell's preamble signal hop-by-hop. We evaluate the performance of the multi-hop synchronization and compare it to the single-hop synchronization.

In the orthogonal frequency division multiplexing (OFDM) based cellular wireless communication systems such as the Mobile WiMAX system, the communication between a mobile station (MS) and a base station (BS) starts with synchronization that includes time and frequency synchronization, and cell search. During the initial synchronization process in the Mobile WiMAX system, the cell search and the integral carrier frequency offset (CFO) estimation require computational complexity that consumes much of computational resources in downlink synchronization block. Therefore in this paper, we propose an efficient method to reduce the computational complexity of the synchronization process without performance degradation. The simulation result shows that the proposed method reduces computational complexity up to 1/16 compared with a conventional method.

The band adaptive modulation and coding (band AMC) scheme is used in orthogonal frequency division multiple access (OFDMA) systems to increase mobile communication system throughput, however, an amount of feedback overhead degrades uplink system performance. In this paper, we propose an adaptive feedback method that can significantly reduce the amount of channel feedback information when a femto base station (BS) and mobile stations (MS) communicate in the band AMC mode. Since the channel variation of the MS in the femto cell is relatively small, instead of informing the femto BS of the channel quality information (CQI) periodically, MSs only send CQI feedback to the femto BS when channel condition changes above a threshold. Simulation results show that the proposed scheme can reduce feedback overhead up to 70% without resulting in system throughput degradation compared to the conventional best-M feedback scheme.

An accurate model of IEEE 802.11e EDCA for a performance analysis that captures all major QoS features of the standard is introduced in this paper. Using the model, we derive the saturation throughput, and validate the accuracy of the proposed model through comparisons with other models via simulations.

In this paper, we give an analytical justification how to control the QoS parameters using a throughput approximation model for IEEE 802.11e EDCA. Using the model, we propose a new parameter control algorithm which successfully achieves the dynamic parameter assignment under desired throughput ratio constraints.

In S-MAC-based sensor networks, border nodes consume more energy since they follow multiple listen and sleep schedules. Therefore they switch into the listen state frequently and reduce the network lifetime. This paper proposes a simple but powerful algorithm, termed the schedule unifying algorithm (SUA), to minimize energy consumption of border nodes by unifying multiple listen and sleep schedules into a single unified schedule. The simulation results show that SUA incorporated SMAC- based nodes consume less energy, thereby extending the network lifetime approximately 2 times more.

Although each application service has its own quality of service (QoS) requirements, the resource allocation for multiclass services has not been studied mainly in multiuser orthogonal frequency division multiplexing (OFDM) systems. In this paper, we investigate the resource allocation in an effort to satisfy the QoS of multiclass services. A power minimization problem is examined for the various data rates and bit-error rate (BER) traffic. Lagrangian relaxation and a subcarrier timesharing factor are used to find an optimal solution. We propose an iterative algorithm to find optimal power consumption and a low complex suboptimal algorithm for practical implementation. The main contribution of this paper is that we apply the optimal solution obtained from the power minimization problem into the proposed algorithms. We execute extensive simulations under various multiclass service scenarios. These scenarios provide a more complete assessment of the proposed algorithms. Simulations results show that the proposed algorithms significantly outperform in total transmit power consumption compared to existing resource allocation algorithms.

For the next generation wireless communication systems, inter-cell interference (ICI) is the main source of performance degradation in mobile stations, especially, cell-edge mobile stations. This paper presents a novel orthogonal resource allocation algorithm (ORAA) for our proposed system architecture, termed the OFDMA-based cellular wireless system architecture using relays (OCWSAR). By using relays to cover the cell-edge areas and a robust algorithm with priority resource allocation, the ICI of the system is significantly reduced, while the sector frequency reuse factor is maintained at 1. A system-level simulator is developed in a multi-cell/multi-user environment, where the effect of the proposed ORAA can be evaluated precisely in the presence of inter-cell interference (ICI). Furthermore, the simulation results assure that the new system architecture with the associated algorithm yield higher SINR and data throughput for cell-edge users.

In this paper, we propose an improved orthogonal resource allocation algorithm (iORAA), which is designed to incorporate with the OFDMA-based cellular wireless system architecture using relays (OCWSAR) in order to reduce the inter-cell interference (ICI) significantly for mobile stations (MSs), especially, cell-edge MSs while maintaining the sector frequency reuse factor at 1. Thanks to that, the cell data throughput is also improved. The performance of the proposed algorithm is evaluated in a multicell/multi-user environment, where the effect of the iORAA can be evaluated precisely in the presence of ICI. Numerical results show that the new system with the associated algorithm yield high SINR distribution and data throughput for cell-edge MSs over the conventional algorithms.

Wireless mesh networks (WMNs) are one of the key features of beyond 3G system because of their flexible and low-cost deployment. IEEE 802.16 mesh mode has recently emerged as an alternative protocol for establishing WMN. Because the WMN is used as wireless backbone network, the design of WMN protocol should target a high network throughput. Then, in IEEE 802.16 mesh standard, the performance improvement is limited by the holdoff algorithm in the bandwidth reservation procedure. In this paper, we propose an adaptive holdoff algorithm for IEEE 802.16 mesh mode with coordinated distributed scheduling. The proposed holdoff algorithm improves network throughput by adaptively adjusting the transmission holdoff exponent according to current node state. In addition, the proposed holdoff algorithm maintains the backward compatibility with the IEEE 802.16 mesh standard and does not require any information from other layers. Simulation results show that the proposed holdoff algorithm performs better than existing holdoff algorithm in terms of total network throughput.

An efficient hardware simulator for the design of a WCDMA interference cancellation repeater (ICR) is presented. The ICR, which is composed of a digital interference canceller and analog parts, is a time-varying system that reacts to a time varying input signal. The design of the digital interference canceller is time-consuming due to the need for many iteration cycles of debugging, tuning, and performance verification. Furthermore, its performance, which is represented by the error vector magnitude (EVM) and the output power spectrum, can be verified only after it is implemented on a digital board, integrated into the repeater hardware with analog parts, and hooked up with expensive test equipment. Hence, a special computerized simulator is needed for fast and efficient verification of the interference canceller's design. The simulator consists of a signal generator, which emulates a base station, the ICR with an interference canceller, a feedback channel that models wireless fading channels, and a receiver that measures the EVM of the ICR. The accuracy of the simulator is verified by showing that the EVM and the output power spectrum of a reference interference canceller measured on this simulator match well within an error range of less than 8% with the corresponding values measured on the repeater hardware. The usefulness of the simulator is also confirmed by simulating the EVM performance of the ICR under various multipath fading channel conditions.

In this paper, we introduce a new scheme to improve the performance of the terrestrial-digital multimedia broadcasting (T-DMB) system, which exploits the properties of the Reed-Solomon (RS) decoder to correct more errors by predicting as many erasures as possible in each RS packet in advance before decoding in the RS decoder. The concept depth of prediction (DoP) used for predicting erasures is introduced and derived. Simulation results show that with optimal DoP values, the new scheme yields performance gain of around 1dB at BER of 10-8resulting in the extended coverage area of the T-DMB system.

In this paper, we propose two radio resource control (RRC) schemes where both base station (BS) and relay station (RS) perform dynamic resource allocation. The target system is a two-hop mobile communication system employing OFDMA technique. These schemes operate in such a distributed manner that both BS and RS control the mobile stations directly connected to each of them. However, the mobile station (MS) connected to RS is influenced by BS-RS and RS-RS links as well in addition to RS-MS link. To solve this problem, a BS-RS link control mechanism based on the information from RS should be supplemented into the system. The proposed mechanism, a semi-distributed method, is that BS allocates resources to the directly connected mobile stations and relay stations, first. Then RS allocates resources to the connected mobile stations in two ways; RS uses the assigned subcarriers from BS or all system subcarriers independently. The former method is called the `separate and sequential allocation (SSA)' and the latter the `separate and reuse allocation (SRA)'. The SSA method yields a superior quality without regard to the location of RS, but the SRA method reveals performance results depending on the location of RS. Computer simulations show that the two proposed methods have almost the same performance as the existing centralized method but lower complexity and higher flexibility compared to the centralized method.

A selective hybrid automatic repeat request (HARQ) scheme operating based on channel conditions for high speed packet data transmission systems is presented. This scheme exploits the channel condition identified by the channel quality indicator (CQI) to determine whether the retransmitted packet is directly decoded without soft combining or it is soft combined with the original packet and decoded. Since the channel condition is already measured and mapped to the CQI number by the receiver to report to the base-station for adaptive modulation and coding, additional increase of computational complexity at the receiver is not required for this scheme. The proposed selective HARQ scheme shows about two to three times performance improvement in terms of throughput because it takes the quality of retransmitted packets into account based on channel conditions before performing soft combining with the original packet with error.

A mobile router, which is in an ad-hoc network is not only a general ad-hoc network entity, but also an anchor node, which takes role of an access point to make a flat network environment into a hierarchical one. Thus, a mobile router must have both ad-hoc routing protocol and a mechanism to communicate with the infrastructure. The role of a mobile router is similar to the one of a mesh router in a wireless mesh network. From the point of view of an infrastructure, a mobile network represented by mobile router is identified as a node. Even when the mobile network moves while changing its point of attachment to the infrastructure, the mobile router should continuously transmit packets to the nodes in the network. But, the existing tunneling protocol derived from IP-in-IP encapsulation of Mobile IPv4 to send packets to the nodes of the mobile network causes an additive delay and an IP header overhead. In this paper, we propose two tunneling protocols and two location registration methods. We compare the performance of the proposed schemes against legacy ones. Our simulation results indicate that the new schemes improve network performance with respect of delay, link recovery time and packet delivery rate

To support very high data rate services that require higher system capacity, the high speed downlink packet access (HSDPA) was proposed in the UMTS standard. One of key techniques supporting the HSDPA services is the adaptive modulation and coding (AMC) in which the modulation scheme and the coding rate are adaptively changed to match the current channel condition reported by the user equipment (UE). Therefore, the mapping between the channel quality indicator (CQI) and signal to interference ratio (SIR) is closely related to the accuracy of AMC and the performance of HSDPA. This paper proposes a novel SIR to CQI mapping method that satisfies the 3GPP requirements. In order to verify the performance of the proposed mapping method, we implemented the link-level simulator which is composed of all the physical layer blocks depicted in the 3GPP standard. With the proposed mapping method, we show that UE can report the exact channel condition and the system can yield performance exceeding the requirements in the 3GPP technical specification.

This letter presents two adaptive interpolation methods based on applying an inverse gradient to conventional bilinear and bicubic interpolation. In simulations, the proposed methods exhibited a better performance than conventional bilinear and bicubic methods, particularly in the edge regions. In addition, the proposed methods can be used irrespective of the magnification factor (MF) and easily implemented due to their simple structure.

In spite of using IEEE 802.11 as medium access control (MAC), most of the ad-hoc routing protocols do not consider contention time that occurs in the medium reservation procedure. Large, contention times can be more critical than hop counts in determining the end-to-end delay. In ad-hoc networks, traffic concentration on some mobile nodes leads to long queuing delays, low packet delivery ratios, and inefficient power consumption. In this paper, we propose an ad-hoc routing protocol with minimum contention time and load balancing (MCL). MCL has two main characteristics. Firstly, MCL selects a route with minimum contention among many possible routes between source and destination in the route selection procedure; secondly, intermediate nodes do not reply to RREQs in the route discovery procedure. These characteristics reduce contention time in medium reservation procedure and distribute traffic throughout the network. We have compared the proposed MCL and the ad-hoc on-demand distance vector (AODV) routing protocol. Simulation results show that MCL outperforms AODV in terms of packet delivery ratio, average end-to-end delay, and normalized routing overhead. In situation where several mobile nodes send data to a fixed node such as an access point or a server, our proposed routing protocol MCL delivers a high performance gain as a result of load balancing.

In this paper, a frequency selective NEXT cancellation algorithm for xDSL transmission is proposed. NEXT is cancelled by subtracting estimated noise from the received signal. An adaptive filter is used to estimate the band limited interference noise from the interference data. This algorithm is applied for a DSL application where asymmetric and symmetric services coexist. It showed about 40% improvement of transmission speed.

We propose an approximation of the line spectral frequencies (LSFs) to the LPC cepstral coefficients (LPCCs). A direct relationship between LPCC and LSF is derived, and new parameters, named pseudo-cepstral coefficients, are obtained by reducing some factors in the relationship. We also address the statistical properties of the pseudo-cepstral coefficients and show their useful application to speech recognition.

In digital communication networks, a speech recognition system extracts feature parameters after reconstructing speech signals. In this paper, we consider a useful approach of incorporating speech coding parameters into a speech recognizer. Most speech coders employ line spectrum pairs (LSPs) to represent spectral parameters. We introduce weighted distance measures to improve the recognition performance of an LSP-based speech recognizer. Experiments on speaker-independent connected-digit recognition showed that weighted distance measures provide better recognition accuracy than unweighted distance measures do. Compared with a conventional method employing mel-frequency cepstral coefficients, the proposed method achieved higher performance in terms of a recognition accuracy.

We propose an adaptive fixed code-excited linear prediction (AF-CELP) speech coder operating at 4 kbps. By exploiting the fact that a fixed codebook contribution to the speech signal is also periodic as the corresponding adaptive codebook contribution, the adaptive fixed codebook model efficiently represents excitation signals. In order to overcome the quality degradation caused by the coarse quantization of excitation, a paired pulse algebraic codebook structure is also applied to the excitation model. Additionally, a pitch prefiltering, a noise spreading, and a harmonic enhancement technique are adopted in the decoding process. The spectrogram reading and informal listening tests proved that the AF-CELP reproduces high quality speech.

An interlacing property of the line spectrum pair frequency (LSF) is proved on the basis of the logarithmic spectral difference function defined by the autoregressive models of successive orders. The property that the LSFs of an order are interlaced with those of lower order, provides a tight bound on the formant frequency region.

This paper proposes a linear predictive (LP) analysis method where sample autocorrelations are estimated from the spectral envelope of a speech signal on the basis of the spectral autocorrelation. The spectral autocorrelation is defined as discrete quantities of speech spectrum with spectral resolution identical to the discrete Fourier transform (DFT) used to obtain the speech spectrum. From analytical and empirical derivation of its properties, we can estimate the fundamental frequency and the maximally correlated frequency for voiced and unvoiced speech, respectively, and then obtain the spectral envelope by sampling at a rate of the estimated frequency. A frequency normalization can be applied to the estimated spectral envelope because the number of samples of the spectral envelope usually differs from frame to frame. The spectral envelope is warped into the mel-frequency scale and the inverse DFT is applied to extract the estimate of sample autocorrelations. From the result of LP analysis on the sample autocorrelations, we finally obtain the spectral envelope cepstral coefficients (SECC). Hidden Markov model (HMM) recognition experiments show that SECC significantly improves the performance of a recognizer at low signal-to-noise ratios (SNRs) over several other representations.

The authors present a vector quantization (VQ) codebook design algorithm for recognition systems based on the phoneme-level hidden Markov model (HMM). In order to improve the discrimination ability of HMMs, the VQ codewords are extracted from the HMM state segments of the recognition units. An iterative VQ codebook design method based on the LVQ2 or MKM algorithm is proposed for the unified estimation of VQ codebooks and HMM parameters. Although the codebooks designed by the proposed algorithm showed larger quantization distortions for input speech spectrum than the conventional codebooks, they always improved the recognition accuracy of the recognition systems as compared to the conventional phoneme-independent or phoneme-dependent codebooks.>

Invariant acoustic-phonetic features conveyed simultaneously by the stable and the transient portion of the speech signal are investigated quantitatively using a relational cue. To represent the relational cue effectively, a peak-weighted binary spectrum is proposed. Prevocalic nasal recognition results show speaker independence and improved performance in comparison with dynamic time-warping methods.>

The performance of speech waveform coders that incorporate a noise spectral shaping filter is discussed. The coders studied are adaptive pulse code modulation (APCM), adaptive differential PCM (ADPCM), and adaptive delta modulation (ADM). For APCM and ADPCM, the noise shaping filter is designed to minimize the C-message weighted quantization noise power. As for ADM, the noise shaping filter is used to move some portion of in-band noise outside the signal band. Simulation results show that the performance improvement of these waveform coders with spectral shaping is about 0.5-3 dB over the systems without noise shaping. Although this improvement is relatively small, the waveform coders with noise spectral shaping yield subjectively more pleasing and intelligible sound than those without it.

In this paper, implementation of a compact and efficient multirate speech digitizer with variable transmission rates of 2.4, 4.8, 9.6, and 14.96 kbits/s is presented. The multirate algorithm has been made based on the residual-excited linear prediction (RELP) vocoder with a transmission rate of 9.6 kbits/s. The residual encoder employed in the RELP vocoder uses hybrid companding delta modulation (HCDM). This HCDM is also used as a 14.96 kbit/s coder. If the residual in the RELP system is down-sampled before encoding, a 4.8 kbit/s coder can be realized. If the residual encoder is not used, a 2.4 kbit/s linear predictive coder (LPC) can be realized by incorporating a pitch extractor. In the 4.8 and 9.6 kbit/s coders the pitch-implanted residual excitation method has been used to generate the excitation signal to the synthesis filter. The multirate speech digitizer algorithm has been implemented using 2900 series bit-slice microprocessors. The external memory is composed of 2K RAM's and 2K ROM's. The system design is a two-bus structure with a 204 ns cycle time. With efficient hardware and software design, the multirate speech digitizer requires almost the same hardware complexity as compared with the conventional 2.4 kblt/s LPC vocoder.

We present a reasonably complete account of an improved adaptive delta modulation (ADM) system called hybrid companding delta modulation (HCDM). The HCDM system that is far superior to continuously variable slope DM (CVSD) or constant factor DM (CFDM) is advantageous, particularly for speech coding. It employs both syllabic and instantaneous companding schemes. Performance analysis of the system has been done and verified by computer simulation. In getting the mathematical formula for HCDM granular noise, a new method based on amplitude distribution is proposed. Optimization of the system parameter values by simulation is also discussed. In addition, an efficient method of hardware implementation is considered.

New analytical formulas for quantization noise of two widely known adaptive delta modulation (ADM) systems, continuously variable slope delta modulation (CVSD) and Jayant's firstorder constant factor delta modulation (CFDM), have been obtained. These formulas are derived based upon the previously obtained results for linear delta modulation (LDM) and adaptive differential pulse code modulation (ADPCM). To verify the results, computer simulation has been done using band-limited Gaussian signals. Close agreements between the formulas and the simulation results have been obtained in the wide range of the input signal level.

A performance comparison of three representative ADM systems has been made by computer simulation using real speech. The three systems studied are continuously variable slope delta modulation (CVSD), Jayant's constant factor delta modulation (CFDM), and a modified version of Un and Magill's hybrid companding delta modulation (HCDM). Among the three systems, HCDM yields the best performance in signal-to-quantization noise ratio (SQNR) and dynamic range regardless of the channel bit error rate. Comparing CVSD and CFDM in an ideal channel, the dynamic range of the latter is significantly wider than that of the former, although their peak SQNR's are almost the same. In a noisy channel, CFDM degrades more rapidly than the other two as the bit error rate increases. In the channel with an error rate above 10-3, the use of CFDM appears to be impractical when the bit rate is below 16 khits/s. However, intelligible speech transmission is possible with HCDM or CVSD even at the error rate of as high as 10-1.