Search

The performance of an OpenCL program is strongly influenced by both hardware and software attributes. To achieve superior performance, developers may leverage automatic performance tuning techniques to determine the optimal parameters on the target device. Although existing approaches have shown promising tuning results in their target scenarios, other requirements such as efficiency, portability, and usability should also be considered because of the rapid growth of heterogeneous computing applications and platforms. In this paper, we re-examine the workgroup size tuning problem and propose a novel approach to meet the aforementioned requirements. We abstract the architectural details into a set of hardware parameters so that the proposed approach can be applied without the presence of target devices, which makes it more accessible to developers. The proposed approach is evaluated on 20 OpenCL kernels and six devices, including both CPUs and GPUs. Experimental results demonstrate that, with negligible overhead, our approach filters out 88.6 percent of the possible workgroup sizes on average. Among all the workgroup size candidates, the best- and worst-performing candidates can achieve average performance of 95.5 and 92.1 percent, respectively, compared with the optimal workgroup size.

Studies on object rearrangement planning typically consider known objects. Some learning-based methods can predict the movement of an unknown object after single-step interaction, but require intermediate targets, which are generated manually, to achieve the rearrangement task. In this work, we propose a framework for unknown object rearrangement. Our system first models an object through a small-amount of identification actions and adjust the model parameters during task execution. We implement the proposed framework based on a low-cost tabletop robot (under 180 USD) to demonstrate the advantages of using a physics engine to assist action prediction. Experimental results reveal that after running our adaptive learning procedure, the robot can successfully arrange a novel object using an average of five discrete pushes on our tabletop environment and satisfy a precise 3.5 cm translation and 5° rotation criterion.

Structured-light and stereo cameras, which are widely used to construct point clouds for robotic applications, have different limitations on estimating depth values. Structured-light cameras fail in black, transparent, and reflective objects, which influence the light path; stereo cameras fail in texture-less objects. In this work, we propose a depth fusion model that complements these two types of methods to generate high-quality point clouds for short-range robotic applications. The model first determines the fusion weights from the two input depth images and then refines the fused depth using color features. We construct a dataset containing the aforementioned challenging objects and report the performance of our proposed model. The results reveal that our method reduces the average L1 distance on depth prediction by 75% and 52% compared with the original depth output of the structured-light camera and the stereo model, respectively. A noticeable improvement on the Iterative Closest Point (ICP) algorithm can be achieved by using the refined depth images output from our method.

There is growing interest in object detection in advanced driver assistance systems and autonomous robots and vehicles. To enable such innovative systems, we need faster object detection. In this work, we investigate the trade-off between accuracy and speed with domain-specific approximations, i.e. category-aware image size scaling and proposals scaling, for two state-of-the-art deep learning-based object detection meta-architectures. We study the effectiveness of applying approximation both statically and dynamically to understand the potential and the applicability of them. By conducting experiments on the ImageNet VID dataset, we show that domain-specific approximation has great potential to improve the speed of the system without deteriorating the accuracy of object detectors, i.e. up to 7.5x speedup for dynamic domain-specific approximation. To this end, we present our insights toward harvesting domain-specific approximation as well as devise a proof-of-concept runtime, AutoFocus, that exploits dynamic domain-specific approximation., 6 pages, 6 figures. Published in IEEE Micro, vol. 38, no. 1, pp. 31-40, January/February 2018

This paper presents an object-centric method for efficiently performing two types of challenging pick-and-place tasks, namely sequential pick and place and object sorting. We propose multiclass dense object nets (MCDONs) for learning object-centric dense descriptors that maintain not only intra-class variations but also inter-class separation. Intra-class consistency is also inherently learned and is useful for our pick-and-place tasks. All the tasks only require a single demonstration from users, which can then be generalized to all class instances. A dataset containing eight classes and a total of 52 objects was provided in this study. We obtained a task success rate of 93.33% on a five-block stacking task and 97.41% on a three-class object sorting task.

Distributed renewable energy has emerged as a promising resource because of its environmental friendliness and economic considerations. However, most renewable energy sources are unreliable and may require considerable effort to be efficiently utilized in a computing center for providing services. In this paper, we exploit distributed renewable energy (e.g., solar energy) and peer-to-peer (P2P) technologies to aggregate distributed computing resources to provide an infrastructure called solar-based energy-efficient distributed server (SEEDS) farm for distributed computing and distributed storage. Energy-efficient devices (e.g., embedded devices) powered by solar energy form a P2P computing system to provide their computing resources to end-users. Specifically, this paper uses a Web-based service as a case study. A group of solar-powered embedded devices acts as a front-end cooperative caching system for Web servers. Web objects may be accessed through the distributed caching system without going through servers, and thus we can reduce brown energy consumption of the servers. This paper also develops an analytical model to evaluate the total energy consumption of Web-based services with and without SEEDS. Theoretical and simulation results show that the SEEDS system can support services and achieve significant improvements in energy efficiency by aggregating distributed energy resources.

Machine-type communication (MTC) is a crucial service for next-generation cellular networks. Mass access to the network by MTC devices may result in the overload of radio access networks (RANs) and degrade the service quality of human-to-human communication. Group paging is one of the mechanisms proposed to alleviate the RAN-overload problem. This paper presents an analytical model based on a recursive contending-users estimation (RCE) method proposed in the literature to derive the performance metrics of collision probability, access success probability, average access delay, statistics of preamble transmissions, statistics of access delay, and utilization of random-access opportunities (RAOs) for group paging with various combinations of group sizes and reserved radio resources in a paging access interval. The optimal group size and required RAOs are subsequently derived based on the given target access success probability. Numerical results demonstrate that the proposed model can accurately estimate the performance of group paging.

Co-channel interference seriously influences the throughput of a wireless mesh network. This study proposes an end-to-end channel allocation scheme EECAS that extends the radio-frequency-slot method to minimize co-channel interference. The EECAS first separates the transmission and reception of packets into two channels. This scheme can then classify the state of each radio-frequency-slot as transmitting, receiving, interfered, free, or parity. A node that initiates a communication session with a quality of service requirement can propagate a channel allocation request along the communication path to the destination. By checking the channel state, the EECAS can determine feasible radio-frequency-slot allocations for the end-to-end path. The simulation results in this study demonstrate that the proposed approach performs well in intra-mesh and inter-mesh communications, and it outperforms previous channel allocation schemes in end-to-end throughput. Copyright © 2013 John Wiley & Sons, Ltd.

Recently, relay technology has been adopted to enhance the coverage and performance of wireless networks such as Worldwide Interoperability for Microwave Access and Long-Term Evolution Advanced (LTE-A). However, using relays to forward packets may induce more packet losses than traditional single-hop wireless networks because transmissions are conducted through multiple radio links. When there are lost packets, relay stations (RSs) decide whether to retransmit these packets with automatic repeat request (ARQ) strategies. We observe that an improper ARQ strategy increases latency, blocked packets, and workloads on the multihop relay network. This paper proposes a new relay ARQ (RARQ) scheme, providing efficient acknowledgement to reduce packet latency and the number of blocked packets with small workloads. We also propose an analytic model to evaluate the performance of the proposed mechanism. Simulation results have validated the proposed model and demonstrated that our ARQ scheme outperforms conventional approaches.

Energy efficiency has become one of the most important design issues for embedded systems. To examine the power consumption of an embedded system, an energy profiling tool is highly demanded. Although a number of energy profiling tools have been proposed, they are not directly applicable to the embedded processors with power management functions that are widely utilized in battery-operated embedded systems to reduce power consumption. Hence, this study presents a high-level energy profiling tool, called SEProf, that estimates the energy consumption of an embedded system running multithread software and a multitasking operating system (OS) that supports power management functions. This study implements the proposed SEProf in Linux 2.6.19 and evaluates its performance on an ARM11 MPCore processor. Experimental results demonstrate that the proposed tool can provide accurate energy profiling results with a low profiling overhead.

This paper presents a combinatorial model and approximation formulas to estimate the average number of successful and collided users for a one-shot random access in finite-user multichannel slotted ALOHA systems. The proposed model and approximation can be used to evaluate the performance of group paging for machine-type communication (MTC) in 3GPP LTE. Numerical results demonstrate the applicable range of the approximation formulas and the accuracy of the derived performance metrics.

We propose a rank-based network coding scheme for distributing content in vehicular ad-hoc networks, in which vehicles adaptively broadcast packets based on the content reception status of their neighbors. In contrast to the previous work, CodeOn, vehicles in the proposed method only exchange content reception information in the service channel. The packet injection probability in the proposed method is proportional to the estimated innovation that a vehicle can provide to its neighbors. We also give an analytical model to analyze the system throughput when vehicles apply rank-based network coding to broadcast data. Simulation results demonstrate that the proposed method improves the throughput by more than 32% while maintaining lower end-to-end delay and protocol overhead than the previous work.

Decentralized traffic information systems realize real-time traffic information services without the need for a server infrastructure. However, existing systems rely on either the vehicular ad hoc network or application-layer peer-to-peer protocols over a broadband wireless network suffering from low lookup success rate, high lookup latency, and maintenance overhead of the P2P network. This article proposes a two-tier VANET/P2P architecture that exploits both VANET and P2P technology. In the low tier, vehicles form a VANET via intervehicle communication to exchange traffic information. On top of the VANET, a portion of the vehicles further establish a P2P overlay through a broadband wireless infrastructure to mitigate the disconnectivity problems of the VANET. Simulation results demonstrate that the two-tier architecture can significantly improve the lookup success rate compared to the single-tier VANET systems while reducing lookup latency and maintenance overhead compared to the single-tier P2P systems.

In recent years, IEEE 802.11 wireless local area networks (WLANs) have been widely deployed, and more and more mobile devices have built-in WLAN interfaces. However, WLAN employs the carrier sense multiple access with collision avoidance (CSMA/CA) medium access control (MAC) protocol, which consumes a significant portion of the energy resources of a mobile device. Hence, minimizing the energy consumption of the WLAN interface in mobile devices has recently attracted considerable interest in both academia and industry. This article provides a survey and an experimental study of the energy consumption issues and energy-efficient technologies of the MAC protocol in IEEE 802.11 WLAN.

While mesh networking is gaining momentum with widespread application, we are concerned with fast handoff in a secure mesh environment. To this end, this paper presents a means in the context of IEEE 802.11s of letting a mesh portal act as an IEEE 802.1X authenticator to reduce costly IEEE 802.1X authentication processes during handoff. Our approach is developed for alignment with IEEE 802.11s and 802.11i, keeping protocols at the station side operable with no changes. As another strength, our design applies to generic multihop wireless networks. Both analytical and simulation modeling are conducted to evaluate our scheme as well. Performance results show that our approach reduces handoff delay by up to 268% or achieves comparable performance resulting from the counterpart IEEE 802.11i scheme with high likelihood of 70%-85% successful preauthentication. Moreover, our performance analysis suggests an optimal number of access points managed by one mesh portal in a network. Qualitative and quantitative discussions indicate that our approach is applicable in pragmatic settings.

Wireless mesh networking (WMN) is regarded as a low-cost technology for rapid wireless network deployment. However, in a single-channel WMN, the overlapped transmission ranges between relaying mesh points could introduce serious interference. Using multiple radios over multiple channels can decrease the interference and improve the capacity of a WMN, but it increases the cost of a mesh point. One possible solution to balance between interference, hence performance, and cost is utilizing fewer radios that switch among multiple channels instead of using per-channel radios. In this paper, we propose a channel-switching method, called the traffic-aware switching scheme (TRASS), for a mesh point with a limited number of radios. TRASS utilizes the existing IEEE 802.11 mechanisms, i.e., hybrid-coordination-function-controlled channel access and power saving, to avoid packet loss during channel switching. A TRASS mesh point monitors the occupied channel time to schedule radios that switch among channels. The implemented TRASS demonstrates 75% throughput improvement by (2, 1), i.e., two-channel single-radio, over (1, 1). (3, 2) and (3, 1) achieve 69.8% and 39.7%, respectively, of the throughput of (3, 3) in the simulated TRASS.

One of the most important issues in offering real-time communication services in a mobile environment is support for seamless handover between base stations during a communication session. In Mobile WiMAX, a mobile station may have to perform a scan and association process before handover, but unfortunately, this scan and association process introduces service disruption. In this letter, we investigate the scan and association latency in Mobile WiMAX, and evaluate its performance through analytic models and simulations.

Wireless mesh networking, as a low-cost and reliable technology for rapid network deployment, has attracted considerable attention from academia and standardization in the industry. The IEEE 802.11s standard defines a wireless LAN mesh using the IEEE 802.11 medium access control and physical layers, and is one of the most active standards with increasing commercial opportunities. This study presents the design and development of a WLAN mesh system conforming to the latest IEEE 802.11s draft amendment. Without costly hardware modifications, the proposed solution is a pure software extension for commercial off-the-shelf WLAN chipsets. This study constructs an experimental testbed, and evaluates issues such as the transmission reliability of mesh broadcast-type control messages and multichannel transmissions. Experimental results demonstrate that the delivery of mesh broadcast-type control messages, such as routing construction frames, using the multiple acknowledged unicast scheme improves mesh stability from an 86 to a 98 percent success ratio in a 16-node grid. Transmitting packets using a single radio interface switching between multiple channels reduces inter-flow interference and doubles the throughput in our testbed.

WLAN has been widely deployed over public and private areas in recent years and has become one of the most popular access technologies for mobile Internet services. Handoffs between WLAN access points (APs) that introduce packet loss and delay during a network session is one of the critical issues for real-time communication services. Unfortunately, most of the previous studies in reducing handoff latency and packet loss in WLANs rely on WLAN infrastructure upgrades, and those solutions suffer from deployment problems in the well-established WLAN hotspots. In this work, a pure station (STA)-side approach which only requires the firmware or software upgrade on WLAN STAs without the enhancements of the WLAN standards and infrastructures is presented. The proposed mechanism developed from a time division duplex concept maintains both connections with the serving and target AP simultaneously using two different medium access control (MAC) addresses. Thus, an STA can perform WLAN association, authentication, security key handshake procedures with the target AP, or further acquires an IP address in a new subnet while transmitting and receiving real-time packets through the serving AP at the same time. Simulation results demonstrate that seamless handoffs for real-time communications in secured WLANs can be easily achieved by the proposed mechanism.

The integration of cellular and VoIP over WLAN (VoWLAN) systems recently has attracted considerable interest from both academia and industry. A cellular/VoWLAN dual-mode system enables users to access a low-cost VoIP service in a WLAN hotspot and switch to a wide-area cellular system without WLANs. Unfortunately, cellular/VoWLAN dual-mode mobiles suffer the power consumption problem that becomes one of the major concerns for commercial deployment of the dual-mode service. In this study, we present a novel power saving mechanism, called PIANO (paging via another radio), for the integration of heterogeneous wireless networks, and further apply the proposed methods to implement a cellular/VoWLAN dual-mode system. Based on the proposed mechanisms, a dual-mode mobile can completely switch off its WLAN interface, only leaving the cellular interface awake to listen to paging messages. When a mobile receives a paging message from its cellular interface, it wakes up the WLAN interface and responds to connection requests via WLAN networks. Therefore, a dual-mode mobile reduces the power consumption by turning off the WLAN interface during idle, and can also receive VoWLAN services. Measurement results based on the prototype system demonstrate that the proposed methods significantly extend the standby hours of a dual-mode mobile.

The Embedded Software (ESW) consortium under the Very-Large-Scale Integration (VLSI) Circuits and Systems Education Program which is supervised by the Ministry of Education (MOE) of Taiwan was initiated in 2004 to develop and promote embedded software education. One of the major missions for this consortium is to develop and deploy embedded software curricula which are not well established in traditional computer science (CS) and electrical engineering (EE) education programs in Taiwan. Therefore, the ESW consortium has spent three years in developing embedded software curricula with total twelve new courses. In this paper, our strategies, implementation and experiences for developing and deploying embedded software curricula are presented.

Traditional Computer Science curricula focus on the training for logic reasoning and programming skills. System integration is often not covered in computer science curricula. As the embedded platforms migrate from 8-bit microprocessors to 32-bit microprocessors, the engineers require different skills to design modern embedded systems. The Computer Science faculties at several universities in Taiwan have collaborated to design a new course to meet such needs. In this course, we focus on the capability of integrating hardware and software in networked embedded systems. In particular, the students learn the skills for embedded real-time programming, multi-thread programming, low-power network protocols, security, and wireless networks protocol. In this article, we report the design rationale and current status of this course.

Embedded graphic processing unit (GPU) is an indispensable component in enabling real-time rendering and graphic applications on mobile devices. However, embedded GPU consumes a considerable energy [1] which is critical for battery-operated devices. To understand the energy consumption of a graphic application, conventional approaches suggested the energy model based on hardware performance counters. However, those low-level energy models are mainly derived from GPUs of desktop computers, and they cannot be applied to embedded GPUs directly and the low-level models are less intuitive from a programmer's point of view. In this study, we consider a high-level energy consumption model for embedded graphic processors, and then we can estimate energy consumption of a graphic application based on graphic attributes of a scene. We conduct a number of experiments on real platforms to validate the proposed model. Our experimental results demonstrate that an average energy estimation error rate of 7.30% can be achieved.

Current heterogeneous wireless networks often overlap because of their complementary characteristics and the large deployment of various wireless access technologies. Mobile devices, which are equipped with multiple wireless interfaces, called multiple radio access technologies (multi-RATs) mobile stations (MSs), are also becoming increasingly popular. Therefore, common radio resource management (CRRM) has been proposed to coordinate heterogeneous radio resource allocations and improve the resource utilization of heterogeneous wireless networks. However, CRRM is an NP-hard problem, and low-complexity approaches for dynamic resource management are in high demand. In this paper, the resource request and allocation between multi-RAT MSs and heterogeneous wireless networks are modeled as a directed graph. Thus, the problem of searching for feasible radio resource allocations is simplified to finding trees in a directed graph. Based on the proposed model, a heuristic scheme can be used to find a feasible solution efficiently, and dynamically adjust the workload of heterogeneous BSs to accommodate new requests. Experimental results demonstrate that the heuristic scheme can reduce the request reject rate by 10%-55% compared with conventional approaches.

A near video-on-demand system is one of the most economic and practical applications to provide high quality digital video services through a network in terms of cost, storage and network capacity requirements. In this paper, an optical storage server with a novel data placement and retrieval schemes is proposed for a near video-on-demand system. The proposed data placement scheme on the optical discs fully exploits the periodical broadcast characteristic of an NVOD service to optimize the disc bandwidth utilization. By further applying our retrieval schedule strategy, the buffer requirement of a system can be significantly reduced. In that way, the maximal number of supported streams with the minimal buffer requirement can be easily achieved, The storage system can be applied not only to CD-ROMs storing MPEG-1, but also to Digital Video Disks (DVDs) storing MPEG-2 in the near future.

Previous studies on video storage servers focused on improving the disk throughput and reducing the server buffer size. However, the initial delay of a new request, one of the most important quality of service (QoS) parameters from the users' point of view, is almost neglected while designing a storage subsystem or evaluating its performance. For different types of video-on-demand (VOD) services such as interactive video game, digital library, or movie-on-demand system, the initial delay can vary from 0.5 s to 5 min. This criterion brings some impacts on designing a storage server for a particular VOD application. In this paper, we investigate the storage server design and the performance evaluation of VOD systems with different initial delay guarantees. We propose a new performance model on evaluating the efficiency of a video storage server so that a cost-effective configuration can be easily obtained under a specified initial delay bound.

In this paper, a novel initial videos allocation scheme and a dynamic load balancing strategy are proposed for a distributed video storage server in order to increase the availability and reduce operation cost. The initial allocation scheme determines the allocation of video replicas on the servers to achieve static load balance and to obtain a configuration for efficient dynamic load adjustment. From the simulation results, the proposed load shifting algorithm can reduce up to 50% request fail rate if compared with the same initialization algorithm without load shifting. The proposed initial allocation with load shifting also reduces 25% to 60% request fail rate from the least load first initial allocation scheme with load shifting, 5% to 10% request fail rate and 5% to 25% the number of shifting steps from the DASD dancing method. Moreover, a prototype is implemented on Windows NT to examine the correctness and practicability of the proposed schemes.

To provide high accessibility of continuous-media (CM) data, CM servers generally stripe data across multiple disks. Currently, the most widely used striping scheme for CM data is round-robin permutation (RRP). Unfortunately, when RRP is applied to variable-bit-rate (VBR) CM data, load imbalance across multiple disks occurs, thereby reducing overall system performance. In this paper, the performance of a VBR CM server with RRP is analyzed. In addition, we propose an efficient striping scheme called constant time permutation (CTP), which takes the VBR characteristic into account and obtains a more balanced load than RRP. Analytic models of both RRP and CTP are presented, and the models are verified via trace-driven simulations. Analysis and simulation results show that CTP can substantially increase the number of clients supported, though it might introduce a few seconds/minutes of initial delay.

Recently, zoning technique has been widely applied to disks to increase their capacities. Under the technique, an interesting feature of a disk is that there are a number of various bandwidths on it. Herein, a novel data layout scheme called cluster-pairing on efficiently exploiting this feature for continuous media (CM) servers is proposed. We first applied track-pairing method between a pair of homogeneous disks, and then partitioned each disk into a same number of clusters to facilitate the retrieval of region-based data placement. The proposed method can take the advantages of track-pairing and region-based data placement schemes to fully utilize the various bandwidths from zoned-disks and meanwhile reduce the seek time overhead. According to the simulation results, the disk throughput after applying our approach can be promoted by 35% to 65% than that of the traditional data striping strategies, 10% to 30% than that of region-based data placement method, and about 10% than that of track-pairing scheme. And the wasted storage space is less than 1% which is negligible in terms of the improvement of the disk throughput.

An efficient storage server with a novel data placement scheme for a near video-on-demand (NVOD) system is presented. The proposed data placement scheme fully exploits the periodical broadcast characteristic of a NVOD server to reduce the disk seek overhead. After employing our method, the disk throughput can be promoted to a great extent and the buffer requirement is much reduced. According to the simulation results, the total buffer requirement of a NVOD server can be reduced by around 75% of the previous approaches.

Near video-on-demand (near-VoD) system, which broadcasts popular movies on a number of channels according to a predetermined program schedule, is regarded as one of the practical applications through a CATV network. A broadcasting video storage server is proposed to support this application with flexible configuration and management functions, low system construction cost, and high efficiency. A generic file layout and a retrieval scheme are presented to accommodate variant types of videos with single, multiple, constant, and variable bit rates. Simulation results show that our broadcasting storage server can reduce by 10-20% the system construction cost than the previous approaches.

Of late, a zoning technique has been applied on modern disks to increase their capacities. As a result, a disk is partitioned into several zones with various sectors/track ratios. Another effect of disk zoning is that disks have higher transfer rates (bandwidths) on outer zones than on inner zones due to the disks spinning at a constant angular velocity. A compensatory region-based data placement is proposed to efficiently utilize the disk bandwidths of zoned-disks. Due to the same reading time of a data block, more data can be read on the outer zones to compensate for the insufficient read on the inner zones, which can significantly improve the disk throughput. From the simulation results and the performance evaluation of the prototype, the disk throughput can be promoted by 25%-35% compared with that of the original region-based data placement scheme. In addition, a frame converting strategy is designed to reconstruct the video data. With the frame converting technique, the proposed scheme can facilitate the interactive operations of video playout for video servers.

Proprietary protocols on the Internet often pose challenges for users and analysts to determine how secure the applications are or how easy private information can be accessed. This work presents a tool that processes packets generated by Skype and analyzes its proprietary implementation. The goal of the exercise is to determine how one may reverse engineer a proprietary protocol and obtain private information. The results demonstrate that a buddy-list of a target user can be obtained through the developed Skype Protocol Analyzer.

Random access channels (RACHs) in cellular networks are normally designed for Poisson-distributed arrivals with a constant rate. Unexpected bursty arrivals may result in severely collisions in RACHs and thus, degrade users' service qualities. This paper presents an analytical model to investigate the transient behavior of the RACHs with bursty arrivals generated in a specific time interval in OFDMA wireless networks. The proposed model considers the effect of new arrivals and the real-world implementation constraints of the OFDMA random access procedure such as periodic access characteristic, uniform random backoff policy, and power-ramping effect. The average number of contending and success mobile stations (MSs) in each random-access slot are then derived using the proposed model. The accuracy of the proposed model was verified through computer simulations and the results show the accuracy of the analysis.

Conventional user behavior detection relies on a large amount of sensors and expensive monitoring devices. Moreover, the systems are usually intrusive and may suffer from deployment problems. In this paper, we design and implement an energy management system (EMS) consisting of a non-intrusive load monitoring (NILM) meter, gateway, server and mobile device. The NILM meter provides a non-intrusive and low-cost solution to recognize the states of appliances and to disaggregate the energy consumption of appliances in a house/building. Based on the proposed EMS, we further implement a data mining scheme to detect users' behaviors based on the usage patterns of appliances. A prototype system verifies our design concept and the simulation results show that the detection accuracy of users' behaviors is more than 80% for most of the activities.

I/O operations consume a significant portion of energy of an embedded system. To profile the energy consumption of I/O requests issued by a specific application helps developers to understand the energy consumption of the software and to further optimize the energy efficiency of the designs. However, most of the existing energy profiling tools concentrate on the energy consumption analyses of processors and memory, and provide limited supports to associate the energy consumption of an I/O device with high-level I/O function calls. In this paper, we propose and implement a generic framework, called e nergy pro filing m odule ( EPROM ), which can estimate the energy consumption of I/O requests in application processes. The experimental results based on two typical I/O devices, i.e. WLAN and TFT-LCD, demonstrate that our proposed framework can provide accurate estimates on the energy consumption of I/O function calls and the errors between the estimation and measurement results are below 4%.

The rapid growth in the capability of modern FPGA devices allows developers to build a complete system on a single chip. These types of FPGA-based SoC (System-On-a-Chip) can normally achieve reduced system power, cost and size, and at the same time offer users a great deal of flexibility. The development of such SoCs normally starts from using a hardware / software co-design methodology in order to partition system tasks into computation-intensive and flexibility-demanding parts. Then, dedicated hardware and software will be implemented to realize these two parts. This paper presents an example which demonstrates the result of applying the hardware / software co-design methodology, a power efficient and performance reliable system architecture for realizing audio delay effects. Compared to similar implementations, our system architecture can save 40% of dynamic power consumption while offering the same data throughput and user flexibility.

In this paper, we present an architecture and implementation of a measurement-based energy profiling tool for mobile devices, called PowerMemo (power meter for mobile). The tool composes of a software event profiler and power measurement hardware to analyze process-level and function-level power consumption of mobile applications on Linux operating system and Dalvik virtual machine. Wireless signal attenuators and RF-shielded chambers are further integrated with the tool so that developers are able to emulate a real-life mobility scenario that a mobile device may encounter. The proposed tool overcomes the issue for profiling asynchronous I/Os and can correlate energy consumption of I/O events with software activities. This tool gives developers a broader view of software energy consumption in a mobile environment so that the developers can optimize the energy efficiency of their mobile applications.

Home and building automation system is proposed not only for improving user comfort and safety but also for energy management of a house and building. However, conventional home and building automation system may not know detailed information of appliances and appliance states in a house and building so that some advanced services might not be able to provide. In this study, we integrate nonintrusive load monitoring (NILM) meters with a gateway so that the gateway can understand the power consumptions of appliances in a house and building and provide detailed information of appliances and their states. We design and implement the proposed gateway based on Open Service Gateway Initiative (OSGi) framework with the integration of NILM smart meters, sensors and actuators. Based on this infrastructure, we then present a data mining and event association algorithm to establish the relationships between appliances, sensors and actuators. The association and relationship information can be used to identify appropriate smart home and building services to specific houses and buildings, to locate appliances in a particular room and area, and to help an NILM system to collect the signatures of unknown appliances. Our prototype system demonstrates that the functionalities of a gateway and its services for a smart home and building can be significantly improved by integrating NILM meters.

A nonintrusive load monitoring (NILM) system analyzes the changes of voltage and current on an electrical circuit and can disaggregate the usage and power consumption of appliances on the same circuit loop according to different power consumption signatures of the appliances. An NILM system is regarded as one of the most important features for energy monitoring and management in a smart home and building. Advances in computing, communication and storage technologies further make an NILM system possible to accommodate a huge amount of power consumption signatures of appliances, and to detect appliances and/or appliance states through database search techniques. In this study, we propose a low-cost NILM system for a smart home and building to facilitate end users contributing the power consumption signatures of their appliances to a datacenter through confirmations or feedbacks of the NILM search. Based on those user contributed datasets, a new NILM search algorithm is then presented in this paper. To evaluate and compare the performance and accuracy of NILM search results by employing the proposed and existing approaches, an evaluation test-bed is established. The experimental results demonstrate that the proposed scheme outperforms the existing algorithms and can achieve 85% to 96% accuracy of appliance state detections.

Intelligent Transportation Systems (ITSs) have attracted much attention recently. One of the most important services for ITSs is to provide real-time traffic information service. To reduce the deployment cost and achieve high service availability, decentralized approaches such as vehicular ad hoc network (VANET)-based or infrastructure-based peer-to-peer (P2P) traffic information systems were proposed. To further improve the success rate and lookup latency of traffic information lookups, a new two-tier traffic information system which integrates low-tier VANET and high-tier P2P overlay has been developed. However, conventional two-tier traffic information system may introduce extra routing overheads since the information lookups are broadcasted over two tier networks and introduce redundant lookup messages. This paper proposes an adaptive routing mechanism in the two-tier traffic information system to improve the efficiency of traffic information lookup. Simulation results demonstrate that the proposed adaptive scheme reduces the lookup latency and overhead while achieving the same lookup success rate compared with the conventional approach.

To support delay-sensitive applications on IP over wireless LAN (WLAN), both layer-2 handoff (L2H) and layer-3 handoff (L3H) must be conducted efficiently. Prior studies toward a fast L2H/L3H assume simple networking environments where Extended Service Set (ESS) exactly matches subnets. This paper identifies performance issues associated with inter-ESS L2Hs and L3Hs under mismatching ESS-subnet configurations, and analyzes actual performance impact on existing systems. Experimental results show that inter-ESS L2Hs are much time expensive than intra-ESS L2Hs while mismatching ESS-subnet settings lead to either time-consuming L3H detections or redundant L3H executions. We also discuss possible remedies for this problem.

Broadband wireless communication has brought users a number of multimedia services for years. As wireless broadband markets mature, system operators face new problems. First, system capacity is bounded by finite radio resources. Second, service providers increasing spend on network deployments and services provisions. With limited radio resource and increased costs, operators have no choice but to force users to pay higher rates to preserve the same service quality than they had several years ago. The expectation is that these problems are solved by introducing relay station (RS) into traditional wireless systems. The relaying technology has proven to be a feasible technology to expand system capacity and reduce deployment costs simultaneously. One of the most popular technologies is cooperative communications that boost network throughput significantly by improving the radio quality for transmission and reception (Nosratinia et al., 2004); (Mohr, 2005); (Doppler et al., 2007). Numerical results (Soldani & Dixit, 2008) showed the cost saving benefits when conventional communication networks adopt relay functionality. From this, a commercial cellular system can save more than 56 % on capital expenditures when RSs are deployed. RS is a wireless communication station that provides relay services for receiving and forwarding radio signals between two stations. In a cellular system, RS is in charge of receiving, decoding, and forwarding data between the Base Station (BS) and the Subscriber Stations (SSs). During the relaying, RS provides additional features to assist data transmissions. For example, (Tao et al., 2007) proposed a novel data forwarding scheme in this environment. The novel data forwarding scheme improved the transmission efficiency by approximately 66 %. Although data relay shows superior performances in throughput enhancements and cost savings. Multi-hop relaying introduces new issues that impact overall performance. To indicate a routing path in multi-hop circumstances, additional control information should be introduced. The routing indication design for multiple radio links introduces new overhead and may impact transmission efficiency during relay. Imprecise indications may waste RS computation and storage since it does not know what it shall relay and where to forward. Moreover, there may be a case that data destined to others are received by the RS that is not in charge. Imprecise relay indication will not only increase overheads but also lead RS to use its processor and storage inefficiently. In other words, data forwarding with ambiguous control information would increase RS complexity and buffer storage unnecessarily. During