System architecture is presented to realize distribution-service systems using unmanned aerial vehicles (multicopters). This architecture is characterized by deployment of power-supply stations and direct-route-based distribution. Planning of direct-route-based distribution is discussed, and uniform distribution job assignment among multiple multicopters is given on the basis of square-area modeling. Discussions are made on such issues as power- supply station deployment, power-supply coordination, comparison of multicopter and airship as the vehicle that performs distribution service, and distribution job-assignment adjustment for biased destination distribution cases. The numerical evaluation shows cases where the multicopter-based delivery is superior to the truck-based delivery in terms of service completion time.

An emergency node-combined wireless network that uses balloons, electric vehicles, and unmanned aerial vehicles (UAVs) was proposed. We discuss the construction and specific issues of this network. The communication area is evaluated by changing the node elevation. In this network, the elevations of balloon and UAV nodes vary, and the communication area changes with elevation. In addition, the network may be temporarily disconnected by changes of the communication area. To solve this problem, we employ a delay-tolerant network technique. We evaluate the number of nodes and node density according to the desired application and regional characteristics.

A distributed and expansible architecture for providing message communication service to nodes, which are typically shelters, in a wide disaster-affected area is presented. The large disaster-affected area is divided into a number of sub-areas, to each of which a pair of electric and unmanned aerial vehicles (EVs and UAVs) is assigned. Message files are sent and received through a wireless link between each node and the UAV flying over the node. They are also carried by this UAV (end-to-end-transfer) or carried/relayed by multiple UAVs (relay transfer) from/to the center station in the area. For simple square area modeling, sub-area formation methods such as time-uniform, space-uniform, and hybrid-uniform divisions are established for end-to-end transfer. The relay can be performed synchronously or asynchronously in relay transfer. Several observations are made through numerical examples.

This paper describes a network that was constructed in the disaster area of the Great East Japan Earthquake. The practical experience gained from the endeavor made us aware that for the provision of a serviceable emergency network of this kind in a disaster area, a significant decrease in network construction time is needed. In this paper, we summarize the benefits of using electric vehicles (EVs) in the context of near-future disaster recovery efforts. In particular, we determine the benefits of using EVs in vehicular networking. Unlike gasoline-powered conventional vehicles, an EV has a large-capacity battery that can be used to power communication nodes. Multiple stationary EVs, widely distributed in a disaster area, can be connected through wireless systems to form an emergency network. Each EV can be equipped with a communication station whose antenna is positioned as high as 9 m above its body with the help of an elastic pole. We present experimental results that show that an antenna at this height can extend the transmission range significantly, by taking advantage of a greater probability of line-of-sight radio propagation and a decrease in radio wave reflection from the ground. Finally, we present a prototype of an EV with functionalities that can be of tremendous utility in disaster recovery operations.

In this study, we assume that a number of electric vehicles (EVs), each with its respective partner unmanned electric helicopter (EH), are cooperatively engaged in wide-area disaster surveillance and data delivery to the center node. We present a simple model for area zoning assuming a wide square area and one-dimensional EV placement. Specifically, each EV-EH pair is in charge of a non-overlapping sub-area surveillance, and the surveillance data are temporarily stored in the EV. The partner EH is used to transfer the surveillance data, which are obtained in the sub-area in addition to those transferred from the immediate downstream EV, to the upstream EV via carry and forward. We present two principles for time-efficient surveillance and data transfer, that is, task-balanced zoning and synchronous data handover. Numerical examples are provided to show that the task-balanced zoning yields a slightly lower cycle time than does the uniform zoning, and synchronous handover results in significantly shorter data delivery times than does asynchronous handover.

The Great East Japan Earthquake and the resulting tsunami caused severe disruptions in the functioning of telecommunication systems as well as power supply systems in the wide area of East Japan. Those in the disaster area had no way or difficulties to communicate after the occurrence of the earthquakes for a few days, weeks, or even a few months. While telecom's disaster recovery was performed on the best-effort basis with limited resources, it is necessary to explore a new approach where communication quality during disaster recovery should be guaranteed in some form, meeting communication needs for those in the disaster area. For this goal, two key concepts, that is, the shelter communication service and trans-locatable design are proposed. These concepts were applied to the disaster recovery in Higashi-Matsushima City on the occasion of the great East Earthquake. The lessons learned through this experience are presented.

Network coding has been studied as a solution to mitigate traffic concentration at a bottleneck node in wireless multihop networks. When network coding is applied to a unicast flow, the decoding nodes overhear packet transmissions, receive all packets which are headed for not only themselves but also the other nodes, and store them. The reception of undesired packets is, however, unreliable, and this unreliability causes the decoding failure for network coding. In this case, the encoding node has to retransmit the packet somehow. In this paper, we discuss an ARQ scheme using a negative acknowledgement (NACK). By simple model simulation, our proposed scheme can improve overflow rate, compared with no network coding. Furthermore, a realistic protocol for our proposed scheme is developed, and its simulation results show that our proposed scheme is very effective to mitigate the degradation at the bottleneck node.

IEEE802.11a/b/g Wireless LAN (WLAN) supports multiple transmission rates for packet transmission. In a mesh network backbone using WLAN, it is desirable to select and use an optimum transmission rate in order to achieve high throughput. Transmission rate can be set fixedly or adaptively. The former is further classified to link-based and node-based. In this paper, three transmission rate selection methods, link-based fixed optimum rate transmission, node-based fixed rate transmission and adaptive rate transmission are experimentally evaluated. It is shown that the link-based fixed optimum rate transmission method outperforms the other methods in the average throughput.

In the proposed bird monitoring system, a bird-tag, equipped with GPS, is attached on the bird body. Base stations are geographically distributed in a wide interest area to receive the GPS data sets from bird-tags using low power wireless transmission system. A data center collects the GPS data from the base stations to track the position of each bird. Simple evaluation shows that the life of the bird-tag can be nearly one year by using the proposed power saving methods. Simple analysis shows that the real-time bird tracking performance is greatly improved by the proposed position estimation method using the concept of the monitoring coverage of each base station. The real-world testbed of the proposed monitoring system is described. The preliminary experimental results show that the monitoring coverage of each base station is as far as 7 km.

IEEE 802.11 wireless interface supports multiple transmission rates. It is an important problem how to select the transmission rate so that the throughput is maximized. In a wireless mesh network (WMN), nodes are stationary placed in general. The characteristic of each link will not change significantly over relatively long time. It is effective to assign fixed transmission rate which achieves the best transmission performance at each link while link quality does not change significantly. There are two of for fixed transmission rate assignment schemes. One type assigns a transmission rate at each link (link-fixed), and the other assigns it at each node (node-fixed). In this paper, we propose a novel node-fixed transmission rate assignment scheme. Our proposed scheme tries to increase throughput on the condition that the network connectivity of all nodes in the WMN is kept. We evaluate the proposed scheme by experimental measurement using our testbed. Compared with the scheme that the same transmission rate is assigned to all nodes, our proposed scheme can improve the minimum throughput although the average throughput of both schemes is almost the same.

A wireless mesh network (WMN) is one of important applications for mobile ad hoc networks (MANETs). Most researches about WMN employ IEEE 802.11 wireless LAN in wireless links. IEEE 802.11 supports multiple channels. Performance improvement strategies by using multiple interfaces and multiple channels were discussed. In this paper, we propose to apply a 3-sector antenna system to multi-interface and multi-channel WMN. In cellular networks, the 3-sector antenna system is widely used for enhancement of user capacity. Further improvement will be obtained if the 3-sector antenna system is applied to WMN. WMN has different features from cellular networks. In particular, the hidden node and exposed node problems are very important in WMN. The 3-sector antenna system is compared with the omni-directional antenna systems in terms of contention and collision rates. As a result, both rates of the 3-sector antenna system is smaller than or equal to those of the omni-directional antenna systems. The throughput is also analyzed with the consideration of hidden and exposed node problems. From the numerical results of the throughput analysis, the 3-sector antenna system achieves much better throughput performance than those of the omni-directional antenna systems.

Wireless Mesh Networks (WMN) is an essential technology based on mobile ad hoc networks (MANETs). A three sector antenna system applied to WMN was proposed in early 2010, and system performance was analyzed. In this paper, we introduce a real-world three sector antenna system set up in the current WMN test bed in Niigata University, which employs IEEE 802.11 wireless LAN and supports multi-interface and multi-channel. Channel assignment is performed to the system, and two network patterns are acquired. A series of real-world single-flow and multi-flow throughput measuring experiments are performed to the two network patterns using three sector antenna system. The performance and characteristics of the system are discussed. The test bed node has the exchange function of three sector antenna and omni directional antenna. We discuss a preliminary comparison between the two antenna system. And the three sector antenna system achieves much better throughput performance than the omni directional antenna system on the basis of the one-hop throughput measuring experiment.

We propose a novel communication system for evacuees of shelters, termed "Shelter Communication System (SCS)". SCS is composed of a computer (termed Shelter Server) connected to the Internet and a set of personal computers (termed shelter PCs), one in each shelter, connected to Shelter Server through a temporal link and the Internet. SCS provides a message communication service between shelters as well as between shelters and outside the disaster area. Neither computer terminals nor cellular phones are required for those in the shelters. They write messages on specially-designed message sheets by hand. They receive messages on the sheets of paper printed by Shelter PCs. A prototype of SCS has been developed. A simple evaluation shows that SCS could provides service for 15,000 shelters and 750,000 people there, using 10 earth stations and wireless mesh networks.

An inter-vehicle communication system, called vehicular ad hoc network (VANET), is one of the most promising applications of a mobile ad hoc network (MANET). One of the important issues in VANET is scalability. Compared with topology-based routing protocols, a routing protocol using position information may alleviate the issues of scalability or control overhead. Car navigation systems using a global positioning system (GPS) are widespread, and position-based routing is a useful approach in VANETs. A next-hop forwarding method is a type of position-based routing protocol, wherein a selection scheme for the next-hop node is very important. In this paper, we propose a novel selection scheme for the next-hop node in the next-hop forwarding method in VANET. In the proposed scheme, a new link metric, expected progress distance, is introduced in order to consider both forwarding distance and transmission quality of the wireless link. We evaluate the performance of our proposed scheme, and show that it can achieve much higher throughput and a better packet delivery ratio than existing conventional schemes.

One of the important issues in Mobile Ad Hoc NETworks (MANETs) is assignment of Internet Protocol (IP) addresses to nodes. Several address assignment schemes are discussed in Autoconf working group in Internet Engineering Task Force (IETF). These schemes are based on autonomously decentralized manner, and it may cause duplication of assigned IP addresses. The duplication of IP addresses disturbs route computation because routing protocols assume uniqueness of IP addresses. Therefore, a Duplicate Address Detection (DAD) scheme is needed. Weak DAD uses a unique key for the detection of duplicated IP addresses. This key is advertised together with the IP address in control messages for routing. By finding the conflict of keys, duplicated IP addresses are detected. This scheme can use a simple rule, but it requires large overhead because key size should be large enough to satisfy the uniqueness. In this paper, we propose a partial-key DAD for the reduction of control overhead. In the partial-key DAD, the original key is divided into several partial keys, and one of them is attached to each control message. We evaluate the performance of the partial-key DAD, and show that it can reduce the control overhead although the resolution time of duplicated IP addresses increases slightly.

One of the solutions to mitigate degradation due to traffic concentration on certain nodes is network coding in ad hoc networks. When network coding is applied to unicast traffic, it is necessary to satisfy two condition; (a) an encoding node has to know which packets the neighbor node stores, and (b) decoding nodes have to hold all native packets encoded in the received packet except for the desired native packet. Because of these conditions, some retransmission schemes can be considered. In our study, we propose three retransmission schemes for network coding in ad hoc networks, and compare the performance. From the result, it is found that the scheme in which decoding nodes that fail in decoding the received packet and return NACK to the encoding node decreases the overflow rate the most.

We study the connectivity in vehicular ad-hoc networks, where the motion of vehicles is constrained on a lattice-shaped road network. First, we theoretically investigate the connectivity under the Poisson-positioning assumption, where vehicles are positioned according to a Poisson process on each road at any arbitrary instants. We find that the Poisson-positioning assumption allows the existence of the finite critical-vehicle density; that is, if (and only if) the density of vehicles is greater than the finite critical density, then there exists a large (theoretically infinite) cluster of vehicles and an arbitrary pair of vehicles in the set is connected in single or multiple hops. We obtain an analytical expression for the critical density as a function of the transmission range of each vehicle and the distance between intersections. Next, we consider the connectivity under more realistic movement patterns of vehicles where the Poisson-positioning assumption does not hold. We numerically find that, even in non-Poisson-positioning cases, there exists the critical vehicle density. The critical density in non-Poisson-positioning cases is, however, larger than the one under the Poisson-positioning assumption. We also gain some insight on the efficiency of roadside-relay-station deployment to provide better connectivity between vehicles.

In a sensor network, a lot of distributed sensor nodes observe a certain area or environment, and transmit the observed data to a sink. The observed data sequences may have space- and/or time-dependent correlation. This correlation is regarded as redundant information and can be used for error correction by the joint decoding. In this paper, we propose the joint decoding scheme using the time-dependent correlation between the continuous data sequences generated by a sensor node. We also propose a power control scheme using the time- dependent correlation for saving energy consumption. The packet error rate and the energy consumption ratio are evaluated and the effect of our proposed schemes are clarified.

OLSRv2 is being standardized in the IETF MANET (Mobile Ad hoc Network) working group. OLSRv2 uses the common packet format, that is also being standardized in the MANET WG. It is necessary to validate the OLSRv2 protocol specification by implementing OLSRv2. In addition, OLSRv2 has a great capability to reduce message size compared to OLSR. This is due to address compression. The change of control message size directly affects the total network performance, but how much it affects the various parameters are dependent on the network scale, number of nodes, and address compression implementation. In this context, we have implemented the OLSRv2 protocol, which can run as a real-time protocol or in the QualNet simulator. In addition, we added link disconnection feedback function from MAC layer to OLSRv2 modifying wireless LAN device driver. In this paper our OLSRv2 implementation detail is introduced, and OLSRv2 performance is compared with OLSR by using both a 50 node testbed built in Niigata university and also in the Qualnet simulator. In addition, IP address compression effect is analyzed in the simulator, and the performance of OLSRv2 with link layer disconnection feedback is evaluated in the testbed.

VANET is an application of a mobile ad hoc network to inter-vehicle communication. Each vehicle is aware of its position information by GPS or other methods. Position-based routing is a useful approach in VANET. The Position-based routing protocol can be classified roughly into a next-hop forwarding method and a directed flooding method. We evaluate performance of both methods by analysis and simulation and compare them in this paper.

A wireless mesh network has much attention for various applications such as home networks or hot spot services, and its standardization is now progressing at the IEEE 802.11s task group. One of the most important issues of the wireless mesh network is how to handle the association information of huge number of legacy user stations (STAs). In this paper, we propose a low-overhead handling scheme of the STA association information in order to solve this issue. In the proposed scheme, we introduce an on-demand approach to handle the STA association information. We also evaluate the overhead of the proposed scheme, and clarify its superiority.

In the multihop cellular networks, multiple routes from a user station to a base station can be established because of the flexibility of network topology. In order to reduce packet errors of wireless transmissions, we proposed the multi-route coding scheme, in which packets are encoded and divided into subpackets, and they are transmitted on several routes from the user station to the base station. In order to improve the performance of the multi-route coding, a route selection scheme is very important issue. In this paper, we propose the route selection scheme for the multi-route coding in multihop cellular networks. The proposed scheme is based on the bit error rate of wireless links. We also apply the proposed route selection scheme to a hybrid routing protocol and evaluate its performance.

In sensor networks, because of densely deployment of sensor nodes, sensing area of each sensor node is overlapped with each other and its observed data sequence is correlated. The joint decoding scheme that uses the correlation between the observed data sequences can reduce bit errors in the wireless channel. In this paper, we extend the joint decoding to the case of more than two-hop transmission. For such extension, selection of fusion scheme is very important issue. If the global fusion scheme is employed and the observed data sequence of each sensor node is transmitted on an independent route, route diversity effect could be obtained by using the correlation between observed data sequences. We compare the bit error rate performance of local and global fusion schemes and clarify the route diversity effect

For reduction in transmission errors on wireless links, we proposed the multi-route coding on multiple routes for wireless multi-hop networks. In the proposed coding, a source node encodes a packet, divides it into sub packets, and transmits them on multiple routes. A destination node combines and decodes the arriving sub packets to get coding and diversity gain. In this paper, we propose a hybrid ARQ scheme for the multiroute coding. In a hybrid ARQ, important information is firstly transmitted, and redundant information is retransmitted later. On the other hand, in the multiroute coding, all sub packets should have the same importance because of efficient diversity combining. We investigate the trade-off between hybrid ARQ and multi-route coding

In the ad hoc communication system "SKYMESH", an urgent communication network backbone is built 50-100 meters high over ground by using balloons, taking advantage of good line-of-sight, low interference and long transmission range. SKYMESH is used as an emergency communication network when a large-scale natural disaster occurs. SKYMESH can be used for collecting information on disaster areas for rescue recovery and survey purposes. In this paper, an outline of SKYMESH and the result of the experiment performed at Niigata University campus in November 2005 are presented.

It is generally believed that IntServ architecture is more suitable for strict QoS guarantees than DiffServ architecture because the former has per-flow-base traffic handling mechanisms but the latter does not. This general belief, however, has not been fully verified yet. In this paper, we theoretically and numerically compare IntServ and DiffServ architectures in terms of x-percentile end-to-end queueing delay as well as the maximum end-to-end queueing delay. We find that, in some network topology, DiffServ architecture makes the maximum queueing delay smaller than IntServ architecture does. We also find that in most cases x-percentile queueing delay in DiffServ-based networks is much smaller than that in IntServ-based networks

We propose a new technique for measuring the bottleneck bandwidth of a network path between end hosts. Our proposal is based on the packet-pair technique and uses a simple statistical procedure for filtering out the cross traffic effect. We explain the basic mechanism of our proposal and its implementation in two different types of tools, single end-point and double end-point tools. The former runs in one end-point of the path, assuming that the other end-point or intermediate routers should reply to each received packet using ICMP (Internet control message protocol) packets. The latter works under the cooperation of both source and destination hosts. We have extensively conducted several experiments on nationwide and international networks to show that our proposal outperforms existing tools, such as Pathchar and Nettimer, in terms of accuracy and robustness of measurement.

We argue the end-to-end measurement based admission control (EMBAC) in VoIP networks. We present a new probability-based EMBAC scheme and show that its performance is close to the ideal method using virtual trunk-based admission control. We then present a QoS allocation approach and the highest load link-based method for selecting an admission threshold and dimensioning link capacities. It is shown that the QoS allocation approach guarantees the end-to-end target in terms of blocking probability and average packet loss rate and the highest load link-based method is cost-effective.

We propose an estimator of the effective bandwidth, called Gaussian estimator using buffer measurement, for superposition of sources in IP networks. Our estimator does not rely on a specific traffic model and works only based on a small set of measurements of the workload in the buffer of a router. We conducted simulation experiments by using real traces, which were collected on a LAN at Chiba University and collected at an US-Japan link of WIDE backbone, and show that our proposed estimator performs better than other existing estimators that rely only on the measurement of traffic amount from sources.

In this paper, we propose a dynamic routing with endpoint measurement based admission control (EMBAC) for VoIP networks with mesh topology. The basic idea of this admission control is to use probe flows to measure two routes, which are direct route and one of the two-link routes between each VoIP call's originating-terminating node pair. Based on the probe flow's QoS evaluation, the node decides a route for each call. We compare performance of direct routing, basic dynamic routing and advanced dynamic routing by using simulation. It is shown that an advanced dynamic routing with EMBAC works better for minimizing call blocking probability and packet loss rate than direct routing.

This paper proposes enhanced end-to-end measurement based admission control mechanisms for VoIP networks. In this architecture, active measurement is used for call admission control to maintain packet loss rate for voice flows below a pre-determined value under various network conditions. This paper presents several new concepts, including constant probing instead of conventional on-demand probing and automatic adjustment of admission threshold based on feedback control. Simulation is used to evaluate the network performance such as blocking probabilities and packet loss rate. It is shown that the proposed admission control mechanisms work well to control packet loss rate under various network conditions.

This paper presents a QoS management architecture for VoIP networks. Voice quality, specifically MOS (mean opinion score), is continuously and automatically monitored by using an objective voice quality evaluation tool for periodically generated test calls carrying artificial voice signals. Edge-to-edge measurement based admission control is used to maintain packet loss rates for both directional voice flows below a predetermined value under various traffic conditions. A network design method is given to meet target blocking probability and packet loss rate. Simulation is used to evaluate network performance such as blocking probabilities and packet loss rates for the network with edge-to-edge measurement based admission control. It is shown that admission control works well to control packet loss rate under various network conditions.

We compared several server algorithms and found that performance of server selection can be improved using QoS statistics (average of past weeks QoS values) and real time QoS probing. Specifically, a set of candidate servers are selected based on QoS statistics for a given time section of the day and day of the week. A target server is, then, selected among the pre-selected candidate servers based on real time QoS measurement using echoping.

The conventional go-back-NARQ is modified so as to be applicable to point-to-multipoint communication by using a control block that indicates the occurrence of retransmission to all the receivers. This paper also presents a "tandem error control scheme," in which the up-link and the down-links use separate go-back-NARQ's. The throughput efficiency of the tandem go-back-NARQ is analyzed. Specifically, the mathematical formula for the throughput efficiency is obtained for cases in which there is a single receiver. The throughput efficiencies of these schemes are also evaluated by simulation.

This paper discusses adaptive capacity allocation for high-usage groups (direct routes) in telephone networks so as to best maintain performance for each hour's traffic load. We present two algorithms for this adaptive allocation scheme. The first, called a capacity allocation algorithm, determines capacity allocation in a manner that minimizes the maximum value of point-to-point blocking probabilities for each hour's traffic load. The second, called a frame assignment algorithm, determines the channel connection pattern with a minimum of changes from the old pattern, for a new capacity allocation.

This paper presents the concept of dividing a local office service area under the present telephone network into blocks of switching areas to make up a local network with small, distributed switching nodes because switching costs will become much cheaper relative to cable laying costs due to LSI technology progress. In this paper, cable laying cost and switching cost are taken as the two main factors in determining future local network configurations. First, the cable laying cost is formulated using cable duct and circuit length. Next, the switching system cost is formulated based on the relation between capability and cost of components such as gates and memories. Finally, local telephone network cost is calculated, using the cost formulas obtained and data available at present. The conclusion is that from the cost-effective point of view, conventional local network office service areas should be divided into 9 to 16 blocks.

DSI (digital speech interpolation) is a technique for increasing the utilization of telephone transmission facilities by allocating transmission channels during intervals of active speech (talkspurts) only. The use of DSI for tandem connections, where DSI links are connected in cascade via circuit switching, poses a problem, however, since the voice-quality degradation that is inevitable in each DSI link can be accumulated to an unacceptable level as the number of DSI links in an end-to-end call connection increases. To cope with this problem, we propose the use of a priority channel assignment technique. An example of such a technique is one that provides preference in channel assignment at each DSI process in a network to calls that have passed through at least one prior DSI link. We evaluate the performance of priority assignment schemes for TASI-type DSI analytically and show that the overall voice-quality degradation of a multi-DSI link call can be improved significantly through the use of priority assignment.

This note indicates an observation on the relationship between the independent works of Hellerman on a measure of computational work and of Cook and Flynn on the average minimum cost of logical networks.