The inflection S-shaped software reliability growth model (SRGM) proposed by Ohba (1984) is one of the well- known SRGMs. This paper deals with the optimal software release problem with regard to the expected software cost under this model based on the Bayesian approach. To reflect the effect of the learning experience for the updated software system, we consider several improvement factors to adjust the values of parameters characterizing the inflection S-shaped SRGM. Appropriate prior distributions are assumed for such factors and the expected total software cost is formulated. The optimal release time is shown to be finite and uniquely determined. Because of the flexibility of prior distributions, the proposed Bayesian methods may be applied in many different situations. Numerical results are presented on the basis of the real data.

Boundary detection is one of the most studied problems in computer vision. It is the foundation of contour grouping, and initially affects the performance of grouping algorithms. In this paper we propose a novel boundary detection algorithm for contour grouping, which is a selective attention guided coarse-to-fine scale pyramid model. Our algorithm evaluates each edge instead of each pixel location, which is different from others and suitable for contour grouping. Selective attention focuses on the whole saliency objects instead of local details, and gives global spatial prior for boundary existence of objects. The evolving process of edges through the coarsest scale to the finest scale reflects the importance and energy of edges. The combination of these two cues produces the most saliency boundaries. We show applications for boundary detection on natural images. We also test our approach on the Berkeley dataset and use it for contour grouping. The results obtained are pretty good.

This letter analyzes quantum-based scheduling of real-time tasks when each task is allowed to have a different quantum size. It is shown that generalized quantum-based scheduling dominates preemption threshold scheduling in the sense that if tasks are schedulable by preemption threshold scheduling then the tasks must be schedulable by generalized quantum-based scheduling, but the converse does not hold. To determine the schedulability of tasks in quantum-based scheduling, a method to calculate the worst case response time is also presented.

A QoS support technique for easily minimizing delay in multihop wireless networks is proposed. Using a priority queue operation that reduces delays overall, the proposed technique, Reduced Congestion Queuing (RCQ), solves problems peculiar to multihops. By adding RCQ to a multihop system, base station or access point density and cost can be more effectively curtailed than by simply applying multihops to a cellular network or wireless LAN because RCQ expands the multihop service area. Due to its simplicity, the proposed technique can be used in a wide range of applications, including VoIP.

We propose a medium access control (MAC) protocol for real-time applications in one-hop ad-hoc wireless networks. It is a distributed mechanism that takes account of priority and has a bounded packet delay. Nodes use energy signals to contend for the right to access the channel. Nodes, which have a packet to transmit, send energy signals or listen to the channel based on their binary frame. The node that has sent energy signals and has not heard any energy signals wins the right to access the channel. We use two schemes to determine the binary frame: at the beginning of a session, a node determines it based on its priority level and a random number; after successful transmission, based on a count of successful packet transmissions. With the first scheme, in order to reduce contention losses, the nodes that had won the right to access the channel but failed in transmission have priority over the other nodes. With the second scheme, the node that has the largest count, the one that has been waiting the longest, can send a packet without risking collision. The protocol provides higher probability of successful transmission and a limit on maximum packet delay. An analysis of the protocol provides conditions for the protocol to be stable. We evaluate the performance of the proposed protocol using simulations of a network with a mixed population of data and real-time nodes, whose source is constant bit rate (CBR) and a two state Markov on/off process.

Based on the expectation-maximization (EM) algorithm, an iterative time-domain channel estimation approach capable of using a priori information is proposed for orthogonal frequency division multiplexing (OFDM) systems in this letter: it outperforms its noniterative counterpart in terms of estimation accuracy as well as bit error rate (BER) performance. Numerical simulations demonstrate that an SNR gain of 1 dB at BER=10-4 with only one iteration and estimation mean square error (MSE) which nearly coincides with the Cramer-Rao bound (CRB) in the low SNR region can be obtained, thanks to the efficient use of a priori information.

In this letter, we propose effective feature vectors to improve the performance of voice activity detection (VAD) employing a support vector machine (SVM), which is known to incorporate an optimized nonlinear decision over two different classes. To extract the effective feature vectors, we present a novel scheme that combines the a posteriori SNR, a priori SNR, and predicted SNR, widely adopted in conventional statistical model-based VAD.

In this paper, a channel allocation scheme is studied for overlay wireless networks to optimize connection-level QoS. The contributions of our work are threefold. First, a channel allocation strategy using both horizontal channel borrowing and vertical traffic overflowing (HCB-VTO) is presented and analyzed. When all the channels in a given macro-cell are used, high-mobility real-time handoff requests can borrow channels from adjacent homogeneous cells. In case that the borrowing requests fail, handoff requests may also be overflowed to heterogeneous cells, if possible. Second, high-mobility real-time service is prioritized by allowing it to pre-empt channels currently used by other services. And third, to meet the high QoS requirements of some services and increase the utilization of radio resources, certain services can be transformed between real-time services and non-real-time services as necessary. Simulation results demonstrate that the proposed schemes can improve system performance.

We examined how spatial disparity between the auditory and visual stimuli modulated the audio-visual (A-V) prior entry effect. Spatial and temporal proximity of multisensory stimuli are crucial factors for multisensory perception in most cases (e.g. [1],[2]). However our previous research[3],[4] suggested that this well-accepted hypothesis was not applicable to the A-V prior entry effect. In order to examine the effect of the spatial disparity on the A-V prior entry effect, six loudspeakers and two light emitting diodes (LEDs) were used as stimuli. The loudspeakers were located at 10, 25, and 90 degrees from the midline of the participants to both right and left sides. A preceding sound was presented from one of these six loudspeakers. After the preceding sound, two visual targets were presented successively at a short interval and participants judged which visual target was presented first. Two colour changeable ('red' or 'green') LEDs were used for the visual targets and participants judged the order of visual targets by their colour not by their side in order to avoid the response bias as much as possible. The visual targets were situated at 10 degrees or 25 degrees from the participants' midline to both right and left in the Experiment 1. Results showed a biased judgment that the visual target at the sound presented side was presented first. The amplitude of the A-V prior entry effect was greater when the preceding sound source was more apart from the midline of participants. This effect of spatial separation indicated that the clarity of either right or left side of the preceding sound enhanced the amplitude of the A-V prior entry effect (Experiment 2). These results challenge the belief that the spatial proximity of multisensory stimuli is a crucial factor for multisensory perception.

This paper considers the problem of contention resolution algorithm for multi-class with quality of service (QoS) constrained for wireless communication. Five different channel reservation schemes are proposed, namely, UNI+MLA, UNI+DS, UNI+DS+MLA, Partial UNI and Partial UNI+MLA schemes for multimedia traffic, all are extensions of our recently proposed UNI scheme for single-class traffic. The goal is to achieve the highest system performance and enable each traffic type to meet its QoS requirements. We evaluate the performance of each scheme by mathematical analysis. The numerical results show that our proposed schemes are effective in enabling each traffic type to achieve the best successful rate possible in this kind of environment. Finally when comparing between our proposed schemes and conventional technique in terms of both throughput performance and QoS requirements it is found that the UNI+MLA, UNI+DS+MLA and Partial UNI+MLA schemes are relatively efficient and suitable for practical applications.

Rapid advances in wireless sensor networks require routing protocols which can accommodate new types of power source and data of differing priorities. We describe a QoS-aware geographic routing scheme based on a solar-cell energy model. It exploits an algorithm (APOLLO) that periodically and locally determines the topological knowledge range (KR) of each node, based on an estimated energy budget for the following period which includes the current energy, the predicted energy consumption, and the energy expected from the solar cell. A second algorithm (PISA) runs on each node and uses its knowledge range to determine a route which meets the objectives of each priority level in terms of path delay, energy consumption and reliability. These algorithms maximize scalability and minimize memory requirements by employing a localized routing method which only uses geographic information about the host node and its adjacent neighbors. Simulation results confirm that APOLLO can determine an appropriate KR for each node and that PISA can meet the objectives of each priority level effectively.

The Ubiquitous sensor network (USN) node is required to operate for several months with limited system resources such as memory and power. The typical USN node is in the active state for less than 1% of its several month lifetime and waits in the inactive state for the remaining 99% of its lifetime. This paper suggests a power adjustment dual priority scheduler (PA-DPS) that offers low power consumption while meeting the USN requirements by estimating power consumption in the USN node. PA-DPS has been designed based on the event-driven approach and the dual-priority scheduling structure, which has been conventionally suggested in the real-time system field. From experimental results, PA-DPS reduced the inactive mode current up to 40% under the 1% duty cycle.

With the rapid increase of link speed in recent years, packet sampling has become a very attractive and scalable means in collecting flow statistics; however, it also makes inferring original flow characteristics much more difficult. In this paper, we develop techniques and schemes to identify flows with a very large number of packets (also known as heavy-hitter flows) from sampled flow statistics. Our approach follows a two-stage strategy: We first parametrically estimate the original flow length distribution from sampled flows. We then identify heavy-hitter flows with Bayes' theorem, where the flow length distribution estimated at the first stage is used as an a priori distribution. Our approach is validated and evaluated with publicly available packet traces. We show that our approach provides a very flexible framework in striking an appropriate balance between false positives and false negatives when sampling frequency is given.

In this letter, we propose PQ-MAC: a priority-based medium access control (MAC) protocol for providing quality of service (QoS) in wireless sensor networks (WSNs) which minimizes the energy consumption with traffic-based sleep-wakeup scheduling and supports QoS using differentiating channel access policy, packet scheduling, and queue management. The PQ-MAC utilizes the advantages of time division multiple access (TDMA) and slotted carrier-sense multiple access (CSMA). The proposed protocol is an energy-efficient, priority-based, and QoS compatible MAC protocol. It consists of multi-level queue management, sleep-wakeup scheduling, and an ordered contention period (CP) scheme. It also guarantees time-bounded delivery of QoS packets. Performance evaluation is conducted between PQ-MAC and S-MAC with respect to three performance metrics: energy consumption, throughput, and average latency. The simulation results show that the performance of PQ-MAC is better than that of S-MAC.

This paper presents a novel approach to single channel speech enhancement in noisy environments. Widely adopted noise reduction techniques based on the spectral subtraction are generally expressed as a spectral gain depending on the signal-to-noise ratio (SNR) [1]-[4]. As the estimation method of the SNR, the well-known decision-directed (DD) estimator of Ephraim and Malah efficiently is known to reduces musical noise in noise frames, but the a priori SNR, which is a crucial parameter of the spectral gain, follows the a posteriori SNR with a delay of one frame in speech frames [5]. Therefore, the noise suppression gain using the delayed a priori SNR, which is estimated by the DD algorithm matches the previous frame rather than the current one, so after noise suppression, this degrades the performance of a noise reduction during abrupt transient parts. To overcome this artifact, we propose a computationally simple but effective speech enhancement technique based on the sigmoid type function to adaptively determine the weighting factor of the DD algorithm. Actually, the proposed approach avoids the delay problem of the a priori SNR while maintaining the advantage of the DD algorithm. The performance of the proposed enhancement algorithm is evaluated by the objective and subjective test under various environments and yields better results compared with the conventional DD scheme based approach.

In this paper, we present a novel method to incorporate metadata into data mining. The method has many advantages. It can be completed automatically and is independent of a specific database. Firstly, we convert metadata into ontology. Then input a rule set to a reasoner, which supports rule-based inference over the ontology model. The outputs of the reasoner describe the prior knowledge in metadata. Finally, incorporate the prior knowledge into data mining.

We propose a simple and scalable scheduler called Elapsing Time-based Priority (ETP) scheduler for the optical burst switch (OBS) control channel; it eliminates the requirements of global synchronization and core-state information. This scheduler reduces edge-to-edge delays of burst control packets and total latencies of data bursts.

We investigate the issues involved in designing a packet scheduler for the proportional delay differentiation (PDD) model in differentiated services (DiffServ) networks. The PDD model controls the average waiting time of each class such that the average waiting time is proportional to its corresponding delay differentiation parameter. This paper proposes a novel packet scheduler for PDD referred to as the longest waiting time first (LWTF). By adding certain conditions, we found that the LWTF scheduler can be reduced to a known packet scheduler-priority queue with linear priorities (PQ-LP). The properties and behaviors of LWTF can be predicted from the analysis of PQ-LP. The simulation results in comparison with other PDD algorithms have also revealed that LWTF provides no worse level of service quality in long timescales and affords more accurate and robust control over the delay ratio in short timescales.

This letter proves the finish time predictability of EDZL (Earliest Deadline Zero Laxity) scheduling algorithm for multiprocessor real-time systems, which is a variant of EDF. Based on the results, it also shows that EDZL can successfully schedule any periodic task set if its total utilization is not greater than (m+1)/2, where m is the number of processors.

We consider the reliability-based heuristic search methods for maximum likelihood decoding, which generate test error patterns (or, equivalently, candidate codewords) according to their heuristic values. Some studies have proposed methods for reducing the space complexity of these algorithms, which is crucially large for long block codes at medium to low signal to noise ratios of the channel. In this paper, we propose a new method for reducing the time complexity of generating candidate codewords by storing some already generated candidate codewords. Simulation results show that the increase of memory size is small.