Index tuning is a challenging task aiming to improve query performance by selecting the most effective indexes for a database and a workload. Existing automatic index tuning methods typically rely on "what-if tools" to evaluate the benefit of an index configuration, which is costly and sometimes inaccurate. In this paper, we propose RIBE, a novel method that effectively eliminates redundant queries from the workload and harnesses statistical information of query plans to enable fast and accurate estimation of the benefit of an index configuration. With RIBE, a considerable portion of what-if calls can be skipped, thereby reducing index tuning time and increasing estimation accuracy. At the heart of RIBE is a deep learning model based on attention mechanism that predicts the impact of indexes on queries. A practical advantage of RIBE is that it achieves both improved accuracy of benefit estimation and time savings without making any changes to DBMS implementation and index configuration enumeration algorithms. Our evaluation shows that RIBE can achieve competitive tuning results and 1--2 orders of magnitude faster performance compared with the tuning method based on the full workload, and RIBE also attains higher tuning quality and comparable efficiency against the tuning methods based on the state-of-the-art workload compression methods.

In this paper, we propose an online scheduling system for tours by one-way electric vehicle (EV) sharing. The system takes reservations of tours with their multiple destinations, time windows and stay time for visiting each destination. It then tells the user the sequence and scheduling of visits to each destination, and which specific EVs to take. The system tells the user to change to another EV when the battery runs out. This enables users to move long distances without concerns about remaining battery. The system also calculates a schedule for moving the EVs to rebalance their distribution. In order to verify the effectiveness of our system, we evaluated it with simulations with 500 to 700 users. Our results showed that the proposed system improved the acceptance rate of reservations while reducing the relocation count(the number of EVs to be moved rebalance the distribution) by up to 15%.

Image-based road network extraction approaches encounter a number of problems related to the lack of prior road data. In this paper, we present a road extraction approach using OpenStreetMap road vectors to accurately and robustly extract complete road information. Due to the persistent misalignment between image and map data, map registration is carried out first to move road vectors to corresponding image road centerlines. Junction templates derived from road vectors are matched with image-derived binary road mask and curvilinear response image to correct junction points. The non-junction point matching approach then effectively aligns the vector road network based on image curvilinear structures, within which width and orientation estimations are also embedded and can be obtained to recover the piecewise width of road segments. Road label mask is finally created with full road knowledge - centerline, width, connectivity, and topology. Our approach was tested on complex scenes and its effectiveness could be verified.

To optimize overlay networks on the Internet, it is useful to estimate the qualities of each overlay link. Existing tools for measuring link qualities send probe packets from one end of the network to the other, but this requires the execution of measurement software at both end nodes. Besides, these tools need to send many probe packets to measure quality accurately, and this can disrupt other communication. In this paper, we propose a technique based on supervised learning for estimating link quality based on the quality observed for other similar links. Since our technique does not require sending probe packets, we can quickly estimate the qualities of many links without disrupting other communication or wasting processing time at many nodes. Through experiments on PlanetLab, the proposed method showed better performance on path latency estimation than the estimating results based on path length.

Finding a vacant parking space in a large crowded parking facility takes long time. In this paper, we propose a navigation method that minimizes the parking time based on collected real-time positional information of cars. In the proposed method, a central server in the parking facility collects the information and estimates the occupancy of each parking zone. Then, the server broadcasts the occupancy data to the cars in the parking facility. Each car then computes a parking route with the shortest expected parking waiting time and shows it to the driver. We conducted simulation-based evaluations of the proposed method using a realistic model based on trace data taken from a real parking facility. We confirmed that the proposed method reduced parking waiting time by 20%-70% even with low system penetration.

When disasters occur, a common characteristic is the partial or complete failure of the telecommunications infrastructure so the usual means of communication are often not available. However, accurate and timely information of the disaster area is important. In this paper, a data collection method from an area of interest (AoI) within the disaster zone is proposed that uses the mobile phones of the people to serve as sensing nodes. To achieve maximum AoI coverage while minimizing delay, we propose a DTN-based data aggregation method. Mobile phone users create messages containing disaster-related information and merge them with their respective coverage areas into a new message with the merged coverage to reduce message size and to minimize the overall message collection delay. However, simply merging messages may result in duplicate counting thus, to prevent this, a Bloom filter is constructed for each aggregated message. Also, to reduce further the message delivery time, the expected time of a node to reach its destination is introduced as a routing metric. Through computer simulation with a real geographical map, we confirmed that the proposed method achieved a smaller delay in message delivery than epidemic routing.

In this paper, aiming to reduce cellular network load, we propose a cooperative download method by which user terminals exchange fragments of a content called chunks through short-range wireless communication like Bluetooth or Wi-Fi with other user terminals who want the same content. In order to reduce the cellular network load, it is necessary to receive as many chunks as possible via short-range wireless communication. Our basic idea is to predict contact time when other user terminals get within a range of short-range wireless communication and to obtain chunks efficiently from other user terminals. To predict the contact time, each terminal periodically registers information of its location and retaining and requiring chunks with a server in the fixed network. The server calculates for each terminal the probability and time to contact other terminals so that the terminal can determine its communication schedule with those meeting terminals to obtain chunks efficiently. Through computer simulations with measured energy consumption parameters, we confirmed that terminals using our method obtained from other terminals 50% more chunks with 20 % less energy consumption than the method of greedily collecting chunks by always turning on the wireless device.

In this paper, we propose a method for improving the playback quality of one segment terrestrial television broadcasting (called 1seg, hereafter) at mobile terminals staying in a weak 1seg radio reception area (called WRA terminals) by wireless multi-hop video streaming from terminals in strong 1seg radio area (called SRA terminals). First, we formulate the problem to derive the set of video forwarding paths that maximize the number of WRA terminals to be “relieved” by receiving the forwarded video, under the constraints on the wireless bandwidth and the maximum number of hops. Since this problem is NP-hard and deriving the optimal solution in real-time is infeasible, we propose a greedy algorithm that each WRA terminal find a video forwarding path from a SRA terminal receiving the required 1seg channel based on the information periodically exchanged among neighboring terminals. In order to increase the number of relieved terminals, we introduce a technique to let a SRA terminal change its watching 1seg channel, forward the video to a WRA terminal, and ask for the relief for itself. Through computer simulations, we confirme that the proposed algorithm with 2-hop relief paths improved the video playback quality in more than 60% of the total WRA staying time.

For accident prevention at intersections, it is useful for drivers to grasp the position of vehicles in blind spots. This can be achieved without infrastructure if some vehicles passing near the intersection capture and share live video of the intersection through inter-vehicle communications. However, such video streaming requires a congestion control mechanism. In this paper, aiming to let a driver grasp the situation at an intersection, we propose a method to select vehicles that send a video in order to generate a live bird's-eye-view video of the intersection. In our method, each vehicle at an intersection exchanges information with others, such as the sub-areas of the intersection it captures, the quality of its video, and its position and speed. Based on the exchanged information, each vehicle autonomously judges whether it should send its video or not. Through simulation with a QualNet simulator, we confirm that our method achieves a good video arrival rate and video quality sufficient for practical use.

In vehicular ad hoc networks (VANETs), delivering messages to a specific location is difficult due to the high mobility of vehicles. In this paper, we propose a method for efficient message delivery in VANETs utilizing the route information in car navigation systems. In the proposed method, each car periodically exchanges the information on its current position and scheduled route in the car navigation system with neighboring cars within radio range. By referring to the exchanged information, each car forwards messages to the neighboring car that will most closely approach the destination. Through simulations, we confirm that the proposed method achieves a better delivery rate with low bandwidth usage than a geocast-based method and epidemic routing.

In this paper, we present a QoS routing protocol called GVGrid for multi-hop mobile ad hoc networks constructed by vehicles, i.e., vehicular ad hoc networks (VANETs). GVGrid constructs a route on demand from a source (a fixed node or a base station) to vehicles that reside in or drive through a specified geographic region. The goal of GVGrid is to maintain a high quality route, i.e. a robust route for the vehicles movement. Such a route can be used for high quality communication and data transmission between roadsides and vehicles, or between vehicles. The experimental results have shown that GVGrid could provide routes with longer lifetime, compared with an existing routing protocol for VANETs

In this paper, we evaluate the performance of real-time data transmission by our routing protocol GVGrid. In Ref. [5], we have presented GVGrid, a position-based routing protocol for multi-hop mobile ad hoc networks constructed by vehicles. GVGrid constructs a route on demand from a source (a fixed node or a station) to vehicles that exist in a destination region. The experimental results have shown that GVGrid could achieve high quality data transmission in a region of a medium and high density of vehicles.