論文

In this paper, we propose the UAV path planning method in urban environments with dynamic risk-map generation by vehicle and pedestrian perception. In urban environments, it is the challenging task for Unmanned Aerial Vehicles (UAVs) to generate the safety flight path such that they will not collide with the ground vehicles and pedestrians in case of accidental fall. To mitigate the collision risk, the proposed method dynamically generates the risk-map, where collision risks within specific areas are mapped based on mobility of ground vehicles and pedestrians. To perceive the mobility of them, the proposed method utilizes Vehicle-to-Everything (V2X) messages that include position, velocity, and acceleration of the ground vehicles and pedestrians. Based on the proposed risk-map, UAVs seek their flight paths that minimize the collision risk by Rapidly-exploring Random Tree Star (RRT*) algorithm. Simulation result shows that the flight path of the proposed method decreases the collision risk compared with the shortest path and previous safety path in the situation that there are 200 vehicles and 400 pedestrians in 1250 m×1050m city.

Reducing traffic accidents and congestion remains a major challenge, especially at intersections which often cause traffic bottlenecks. Connected and Automated Vehicles (CAVs) are expected to improve traffic safety and efficiency through cooperative driving based on communication with other vehicles and infrastructure. While various cooperative driving methods for unsignalized intersections have been proposed, most of these methods assume that all vehicles in the system are CAVs. However, it is also necessary to consider mixed traffic environments involving both CAVs and human-driven vehicles since full CAV market penetration is not expected for several decades. There are several challenges in considering human-driven vehicles such as speed deviation and acceleration specifications. In this paper, we propose a cooperative driving method in mixed traffic at unsignalized intersections based on Virtual Platoon. We determine minimum following distances based on vehicle types and human-driven vehicles’ target speed which is indicated to the drivers. Simulation results using a traffic simulator demonstrate that the proposed method successfully avoids collisions in mixed traffic and reduces average travel time by up to 18.7 seconds compared to signal control.

Simultaneous Localization and Mapping (SLAM) is a technique that estimates a self-position while simultaneously constructing a map of the surrounding environment. In SLAM, errors in the estimated self-position can significantly affect overall performance. Global Bundle Adjustment (GBA) is a widely used optimization process intended to correct such errors and improve estimation accuracy. While GBA often enhances trajectory accuracy, in some cases it can introduce substantial degradation. To address this problem, we propose Motion Check (MC), a method that compares the estimated trajectories pre- and post-GBA. When a significant discrepancy in trajectory shape is observed, the method judges that GBA has caused a deterioration in accuracy. By identifying points where accuracy degrades due to GBA and applying appropriate corrective actions, Motion Check effectively suppresses such deterioration. In evaluations using the EuRoC dataset, Motion Check demonstrated its effectiveness by reducing the maximum trajectory error by up to 81.52 cm and the root mean square error by up to 3.40 cm.

In vehicular networks, vehicles periodically exchange messages that include their positions and IDs. On the other hand, adversaries can violate the vehicles' location privacy by obtaining the vehicles' movement history from the messages. In recent studies, mix-zone-based schemes protect the vehicles' location privacy. In the schemes, vehicles utilize pseudonyms instead of their IDs in encrypted messages, and therefore adversaries cannot track them by their IDs. However, since the messages are encrypted by public keys of Road Side Units (RSUs), the vehicles can change their pseudonyms only around the RSUs. In this paper, we propose a Dynamic Mix-zone scheme considering Communication delay for Location Privacy (DMCLP). The DMCLP enables vehicles to change their pseudonyms even if there are no RSUs around them by removing necessity of RSUs from the mix-zone-based schemes. To remove the necessity, the DMCLP adopts both Identity-Based Signature (IBS) and proposed two-stage mix-zones. The IBS enables vehicles to certify the correctness of public keys by themselves. Therefore, it removes the necessity to tell the correctness to trusted authorities through RSUs. Furthermore, since RSUs guarantee connectivity of vehicular networks, influence of communication delay on pseudonym change becomes a problem without RSUs. The proposed two-stage mix-zones mitigate the influence of communication delay by removing vehicles that will cause communication delay from pseudonym change processes. Simulation results show that the proposed scheme improves update frequency of pseudonyms by 50 at most in 1000 seconds compared with the previous mix-zone-based scheme.

As the Internet of Vehicles (IoV) develops, vehicular applications become more complex. The existing offloading methods in the vehicular networks predict the time it takes to process a task and offload it to the computation resources that minimize the delay in execution. However, these methods do not adequately consider other vehicles on the road or the time constraints of the applications. Therefore, we propose a dynamic task offloading method with a centralized controller to improve the task success rate considering all vehicles on the road. We formulate the objective function which maximizes the success rate as a multiple knapsack problem by using an approximation algorithm. In the proposed method, the centralized controller gathers the information on vehicles and edge servers, and based on the algorithm it determines the edge server that maximizes the task success rate in the entire area. The algorithm calculates vehicle priority depending on the task success status of vehicles and determines the edge server based on that priority. In this paper, we conducted the experiments in which we implemented the prototype of the proposed method, and evaluated its performance. The experiments showed that the offloading method improved the task success rate by up to 38% compared to the existing methods.

Connected and Automated Vehicles (CAVs) have potential to improve the traffic efficiency near the highway on-ramps by sharing their merging times through wireless communication. For the traffic efficiency and vehicle safety, it has been the technical challenge to determine the merging times without collision of vehicles merging from different roads. Especially in the mixed traffic of CAVs and human-driven vehicles (hereinafter called non-CAVs), CAVs cannot detect all the non-CAVs even by their sensors. Therefore, it is necessary to determine CAVs' merging times for preventing the undetected non-CAVs from colliding with other vehicles. In this paper, we propose a group-based CAVs control method considering undetected non-CAVs in mixed traffic. The main idea of the proposed method is that it indirectly controls the undetected non-CAVs by CAVs. Specifically, the method forms vehicle groups each of which consists of a leading CAV and following non-CAVs. The method assigns merging times to the leading CAVs such that different groups do not merge onto the main road simultaneously. As a result, the undetected non-CAVs in a group do not collide with vehicles in the other groups. Simulation results show that the proposed method mitigates stop-and-go driving near the merging zone and improves outflow traffic.

Intraoperative hypotension may occur during surgery, depending on the patient's condition and the amount of anesthetics administered. Intraoperative hypotension causes adverse events such as myocardial infarction, acute kidney injury, and stroke, which may affect patient prognosis and should be avoided. However, it is challenging to predict intraoperative hypotension before surgery. Therefore, the authors propose a system that predicts the occurrence of intraoperative hypotension with higher accuracy than existing studies to help preventing intraoperative hypotension. Specifically, the occurrence of intraoperative hypotension is predicted by using information about personal lifestyle and medical history obtained from the pre-anesthesia questionnaire in addition to the patients' baseline characteristics such as age and gender that have already been used. The authors constructed prediction model, that predicted intraoperative hypotension with an F-measure of 0.724.

In this paper, we utilize multihop communication in sidelink Cellular Vehicle-to-Everything (C-V2X) communication for extending the perceptible range of surrounding environments. The challenging task of the sidelink multihop communication is to maintain periodically reserved wireless resources. Since the multihop communication generates aperiodic traffic, it has the potential to break the reserved resources in C-V2X. The proposed method filters the packets that will break the reserved resources. Furthermore, to avoid degrading the perception quality due to the packet filtering, the proposed method prompts vehicles to forward packets when there is enough room in their resources for forwarding the packets. In the simulation results, we confirm that the proposed method extends the perceptible range compared with single-hop communication and multihop communication with several ETSI parameters.

In this paper, we propose a Q-learning-based spatial reuse method considering throughput fairness in Wireless LANs (WLANs). In Spatial Reuse (SR) methods, wireless nodes try to use wireless resources efficiently by controlling both the Transmission Power (TP) and Carrier Sense Threshold (CST). When wireless nodes are densely deployed, the SR methods have difficulty to achieve both the high aggregate throughput and throughput fairness because the mutual interference among the wireless nodes becomes severe. The proposed method removes the difficulty by utilizing Q-learning where wireless nodes can learn the adequate CST and TP by themselves. The proposed method motivates nodes to use wireless resources actively by rewards, while it suppresses nodes with high throughput using the resources by negative rewards. As a result, the wireless resources are distributed among nodes with low throughput, and the proposed method achieves both the high aggregate throughput and throughput fairness. Simulation results show that the proposed method improves the aggregate throughput with keeping throughput fairness.

The European Telecommunication Standards Insti-Tute (ETSI) defined message including information of surrounding objects in V2P (Vehicle to Pedestrian) for traffic safety such as Cooperative Awareness Message (CAM), Vulnerable road user Awareness Message (VAM) and Collective Perception Message (CPM) In urban environment, convergence is occurred because large number of vehicles and pedestrians send messages. Some works suggested methods of reduction traffic of messages which are clustering or changing traffic rate. In clustering methods, cluster participation conditions are strict for pedestrian in the environment.The methods of changing traffic rate don't consider the are which require high awareness rate to avoid traffic accidents. In this paper, we propose the new threshold to relax joining conditions and the method to determinate VAM generation time in areas designated by vehicles. We compare our threshold with conventional method's one to show improving awareness rate. And we evaluate proposed methods to improve awareness rate in the specific area. In the result, the proposed method increased the object recognition rate, and showed a higher object recognition rate than the conventional method.

Connected and Automated Vehicles (CAVs) utilize Vehicle-To-Everything (V2X) communication to exchange messages and share information with surrounding vehicles. Collective Perception Messages (CPMs) contain object information detected by the onboard sensors. Through the exchange with nearby CAVs, CAVs can perceive the surrounding objects. Higher vehicle traffic densities, however, can lead to a shortage of communication resources and degradation of communication quality. This, as a result, leads to a decreased object recognition frequency. To mitigate such degradation, it is necessary to reduce the traffic volume of CPMs. Therefore, it is important to decrease the message size of CPMs. Various existing studies aimed at it have reduced the redundancy of CPMs. Most of the studies, however, do not prioritize the inclusion of high-Accident risk objects in CPMs, which leads to a higher Age of Information (AoI) for such objects. In this paper, we propose a risk and redundancy-based object selection method, called RRS, which selects objects for CPMs based on the accident risks for surrounding CAVs and the redundancy of CPMs in C-V2X. In this proposal, we define an accident avoidance acceleration a as an indicator of the accident risks. Based on a for surrounding CAVs and the number of CPM receptions indicating the redundancy, CAV s calculate the priority of objects for CPMs. CAV s select a fixed number of objects based on the priority for CPMs. We evaluate RRS through simulation experiments in an intersection scenario. The results show that RRS decreases the AoI for objects with high accident risks and the number of the highest risks.

Reducing traffic accidents and congestion remains a significant challenge, especially at intersections crucial for safety and efficiency. Virtual Platooning has been proposed to improve traffic efficiency. However, existing research has focused on Connected Autonomous Vehicles (CAVs), without taking into account human-driven vehicles, which lead to issues such as delay in reaction to stimulus onset and difficulty in controlling acceleration. In this paper, we propose method accounts for Connected Human-driven Vehicles (CHVs) and address these challenges through control interval adjustments, a target speed determination, and larger minimum following distances. Simulation using a traffic simulator confirms that our approach improves traffic safety and efficiency.

The Collective Perception Service (CPS) enables Connected and Automated Vehicles (CAVs) to distribute data detected by their onboard sensors. Higher vehicle traffic densities, however, can cause communication resource shortages and reduced communication quality. To address this issue, minimizing the traffic volume of Collective Perception Messages (CPMs) becomes necessary. The ETSI standard approach prioritizes objects with high mobility when including them in CPMs. This approach, however, may result in multiple vehicles transmitting data about the same highly dynamic objects simultaneously, which leads to inefficient use of communication bandwidth. Furthermore, only relying on mobility parameters that surpass a threshold does not account for accident risks that depend on the positional and directional relationships between objects and CAVs. This paper proposes a risk and redundancy-based object selection method, RRS, which selects the objects for CPMs based on accident risk and redundancy in C-V2X. We introduced accident avoidance acceleration as a metric to assess accident risk, and redundancy as the number of CPM receptions for an object. Using these two indicators, CAVs calculate the priority of objects for inclusion in CPMs. We evaluate RRS through simulation in a traffic intersection scenario. The results demonstrate that RRS effectively reduces CPM traffic volumes while maintaining object recognition capabilities, even in environments with high CAV densities.

Vehicular networks exchange a wide range of data to improve safety and comfort on the road. For data security and other reasons, some data should only be accessed by authorized vehicles. Existing secure data delivery schemes utilize authentication and data encryption to realize secure data access control. However, many of these schemes require multiple round-trip times (RTTs) of communication to execute these processes, meaning that the overhead, such as authentication and key setup, consumes valuable network resources. In this paper, we propose a self-managed authentication-based data delivery protocol, SMAD, using elliptic curve cryptography, TESLA protocol, Pedersen commitment, and Schnorr signature. SMAD reduces the number of RTTs by introducing a self-managed authentication. SMAD also uses Diffie-Hellman (DH) key exchange in an unconventional hybrid manner, combining the ephemeral-static and ephemeral-ephemeral modes to achieve both security and good efficiency of data delivery. The security analysis shows that the proposed SMAD satisfies the requirements for secure data delivery in vehicular networks. Moreover, the performance analysis shows that SMAD has better data acquisition time performance than existing schemes.

© 2018 IEEE. This paper proposes a method that predicts customer value by focusing on purchasing behavior. The method generates a relevance model for purchase days and amount in each period between customer value and purchasing histories beforehand based on a consumer panel survey. The authors adopt the random forest method to generate the prediction model. The proposed method facilitates the provisioning of smart customer management to each customer according to level such as suggesting products or services.

© 2019 IEEE. Mobile devices such as smartphones have various restrictions in terms of resources such as batteries and processors. The load reduction on the mobile devices is possible by offloading, migrating application to the resource-rich environment such as cloud servers. However, further study is required on multi-tier environment which contains multiple resources composed of edge servers and cloud servers. This paper proposes an offloading system based on estimated response time (ERT) in multi-tier environment which contains multiple resources. The proposed offloading system calculates ERT based on the state of network and computation resources of each execution resource. Offloading is executed on execution resources with the minimum ERT. We implemented the prototype of the proposed offloading system and evaluated the performance.

Recently, connected and automated vehicles have potential to mitigate traffic congestion and improve traffic efficiency at a highway on-ramp. Especially, vehicle to infrastructure communication based scheduling schemes are paid attention because a central control server can assign merging timing to automated vehicles to improve traffic efficiency on whole roads. Although several schemes calculate merging timing based on their criteria, these schemes cause traffic congestion when inflow traffic becomes high. In order to solve this problem, it is necessary to schedule merging timing of vehicles such that density of vehicles decreases on whole roads. In this paper, we propose a scheduling scheme for cooperative merging at a highway on-ramp with maximizing average speed of automated vehicles. The main idea of the proposed scheme is to utilize the traffic model that density and speed of vehicles have a negative correlation. By utilizing the traffic model, the proposed scheme preferentially assigns merging timing to vehicles which are on a congested road by maximizing average speed of all vehicles on whole roads. Simulation result shows that the proposed scheme decreases density of vehicles on whole roads even if inflow traffic becomes high. Furthermore, the proposed scheme improves outflow traffic from a highway on-ramp by 400 (veh/h) at most in comparison with related schemes.

© 2020 IEEE. In disaster situation, collecting information about victims and disaster area quickly is required. However, it is expected that the communication environment is not fully prepared because communication facilities may be damaged. Mobile Edge Computing (MEC) is attracting attention, in which servers are distributed to the vicinity of end users. Thus, dynamic task offloading adapting to changes of network resources is useful in unstable network. This paper proposes a dynamic task offload system based on the state of network. The proposed offload system consists of two mechanisms. First mechanism is Software Defined Mobile Edge Computing (SDMEC) architecture. Second mechanism is dynamic task placement based on the state of network resources. We formulated an optimization problem to minimize the total response time while application execution. In the proposed offload system, multiple tasks are placed dynamically depending on the load of servers and network resources to minimize response time of application. We implemented the prototype of the proposed offload system and evaluated its performance.

© 2020 IEEE. Mobile devices have certain restrictions such as processing power and network bandwidth. Edge computing can reduce load and delay by migrating the parts of application processing from mobile devices to edge servers. Edge computing is a method that extends cloud computing to the edge of network. Conventional schemes of face recognition based on edge computing can reduce the amount of network transmission data between the edge servers and the cloud servers by preprocessing data on the edge servers. However, the schemes execute each task on a certain resource, thus application response time may increase according to the conditions of computational resources and network. Hence, a task placement system considering the conditions of computational resources and network is necessary. This paper proposes a task placement system for face recognition applications in edge computing. The proposed system calculates estimated response time of possible task placements based on the task placement decision formula. The formula parameterizes power and CPU utilization of computational resources and delay and bandwidth of network. The task placement that minimizes estimated response time is determined. We implemented a prototype of the proposed task placement system and evaluated its performance through experiment.

© 2019 IPSJ. Dynamic Route Guidance (DRG) systems determine the traveling routes of vehicles and solve traffic problems in urban areas. Ad-hoc network composed of vehicles, called VANET is being conducted and VANET-based DRG systems are proposed. In VANET-based DRG systems, a dynamic topology on VANET is one of the biggest challenges. Packet loss prevents decision of proper traveling routes. Therefore, this paper adapts the ICN concept to VANET, called IC-VANET. In IC-VANET, each vehicle acquires traffic information by multi-hop request / response type communication, and it caches contents. Then, we propose a traffic information collection and routing protocols using Virtual Nodes (VNRP) where packets are relayed via Virtual Nodes preferentially for improving data acquisition rate and communication overhead. Virtual Node (VN) is a logical node that is maintained at a specific area considering the vehicle density by inter-vehicle communication with surrounding vehicles, and it collects traffic information. We compared the proposal method with the conventional VANET method by simulation and evaluated the performance.