Wataru Uemura

A tight curve warning system using FSK visible light and road-to-vehicle communication

Rolling accidents often happen on the tight curves, so road-to-vehicle communication is required. In this paper we apply visible light communication to road-to-vehicle communication and propose a tight curve warning system. We use the hierarchical coding of FSK because a receiver can receive both low frequency data and high frequency data when it is close to the sender, and it can receive only low frequency data when it is far away from the sender. We made a 1/32 scale prototype model and the experimental results show its ability to distinguish the communication distance from the frequency. The car on our prototype system put on its brakes 2 m from an LED using 16.6 kHz.

Development of a surveillance cameras system in ad hoc network

This paper presents a surveillance cameras system that consists of only nodes without a server or a station using ad hoc network technology. Each node shares the picture that another node takes, in order to have the robustness against car thieves.

Information delivery tactile pavings using visible light communication

In Japan, there are tactile pavings for visually impaired persons walk support. Tactile pavings tell existence of a way and escalator. Tactile pavings are tool to tell. And, it is one of the auxiliary systems of navigation. Military GPS (global positioning system) became public welfare apparatus. And it is convenience for visually impaired persons. So we propose a navigation system for visually impaired persons. The precision of GPS is about 10m. The precision of IPS (Indoor positioning system) using technology of wireless LAN (A-GPS: assisted GPS) is about 5m. Similarly the precision of visible light communication using lighted is about 5m. This precision is not good for visually impaired persons walking on tactile pavings. We propose the method of sending information using visible light communications from tactile pavings, not from the roof. Our proposal is good precision, because distance of between transmitter and receiver is short. Visible light communication transmits information by flashing light. It can transmit to pinpoint area, because of the linearity of light. However received signals from the neighboring tactile pavings interfere each other. Avoiding interference requires the multiple access method. So we use CDMA (code division multiple access). And we made a prototype and confirmed good precision.

A broadcasting method based on RTS/CTS for an ad-hoc network

An ad-hoc network, which does not require a base station such as an access point, has a hidden node problem because nodes communicating in the same time cannot know each others communication status. In other hand, in the wireless infrastructure network the access point manages the communication with the node using RTS (Request to Send) and CTS (Clear to Send). However, in the broadcasting, we cannot use RTS/CTS method because the number of target nodes is not 1 but more than 2. In this paper, we propose a broadcasting method based on RTS/CTS to avoid this problem. Sender node selects the target node. And they communicate with each other. Neighbouring nodes can listen to these packets. Then sender node can broadcast its information to multi nodes. We use wireless nodes specified by IEEE 802.15.4 which is the physical layers specification. And our experiment results show the effective transmission.

About Q-values of Monte Carlo method

Profit sharing method is one of the reinforcement learning methods. Profit sharing can work well on the partially observable Markov decision processes (POMDPs). Because it is the typical non-bootstrap method, and itpsilas Q-value is usually handled accumulative. Profit sharing, however, does not work well on the probabilistic state transition. This paper we propose the novel learning method which can work well on the probabilistic state transition. It is similar to the Monte Carlo method. So we discuss about Q-values of our proposed method. In the environment with deterministic state transitions, we show the same performance both the conventional profit sharing and proposed method. And show the good performance of proposed method against the conventional profit sharing.

About distributing rewards to a rule with probabilistic state transition

Profit Sharing is one of the reinforcement learning methods. On Profit Sharing, an agent as a learner distributes rewards to rules selected by the agent after reaching a goal state. If there is a non-deterministic state transition rule, for example a probabilistic one, an agent must consider the estimate value of its rule with the probabilistic state transition. Conventional Profit Sharing does not consider the probabilistic state transition because it distributes same rewards even if the state transition probability is 10%, 1%, and so on. In this paper, we propose the novel Profit Sharing method which considers the probabilistic state transition. In the environment with deterministic state transitions, we show the same performance both the conventional Profit Sharing and proposed Profit Sharing. And show the good performance of proposed Profit Sharing against the conventional Profit Sharing.

A proposal for surveillance cameras system by Ad Hoc Network

We propose a new surveillance cameras system. This system uses nodes with cameras and without a camera. When taking a picture, it is transmitted to each node and stored by each node. If the crime occurs in the parking lot, it is useful for the calculation of the criminal by using the taken image. The newly feature of the system that we propose is to transmit the taken image to other nodes and to share image data with each node. Our monitoring system is different from a conventional monitoring system in the point that the file server is not needed. We construct a simple system.

The coded hierarchical modulation with amplitude for estimating the position in visible light communications

In a museum, a visitor can know introductions of the exhibition gallery with various methods to satisfy a visitors interest. There is a method using a portable terminal in order to guide. Furthermore, the visitor can know the local introduction of a gallery, if gallery guidance takes in position estimate and calculates a works sequence number from the visitors present location. Therefore, we pay attention to the method using VLC (visible light communication) as the high-precision position estimating method (its precision is about several meters) because the communication range of VLC is equal to the visible light irradiation range. Using VLC the portable terminal cannot distinguish from works, when works are crowded. The distance between a transmitter and a receiver can be estimated by received signal strength in VLC. However the visitor needs to change the threshold value of a receiver for every place. So it will become a visitors effort. Then we have to adjust the brightness of lighting equipment in order to change the receiving range if the threshold value keeps the constant. So this conventional method is not useful for the lighting equipment of an exhibition work. On the other hand there is the coded hierarchical modulation that receives the different information according to the distance between a transmitter and a receiver. This modulation needs two dimensions of a camera in order to receive a signal. It uses a high frequency in order to change information according to distance. However it is hard to use this method because this method need high-speed camera. So in order to avoid this problem we propose a method which uses the elements of photo detector in order to estimate the distance in one dimension. Then our method uses the amplitude. The receivable distances are changed according to receive intensity if amplitude takes two or more values. We call this method the coded hierarchical modulation with amplitude. We show the transmitter and a receiver. We show the relationship between the distance and error rates in order to show a change of the distance and received information in the coded hierarchical modulation with amplitude.

An ad-hoc network routing protocol for a disaster scene

In recent years, ad-hoc network technology consisting of not access points and base stations but of wireless nodes has gained attention because the disaster scene requires quickly construction of the network. Usually the aim of routing protocol in ad-hoc network is to provide the stable network service. In order to maintain the network, node must communicate with each other in exchanging their routing tables. So the control packet does not include the information of an application layer. We propose the novel routing protocol for a disaster scene which includes the application information in the control packet. By using our proposed protocol, nodes can construct the routing table and send the nodes information to neighboring nodes at the same time. Then we investigate the tradeoff between relay hops and delay times. We show the experimental results for the real world.

A novel ad-hoc network routing protocol with an acceleration sensor

On an ad-hoc network which is consisted of movable nodes, the network topology is dynamic because each node is adaptive. So we must maintain the routing table in order to send a packet. In proactive routing protocols, each node exchanges their routing information in order to maintain the routing tables. In this type of protocol, we can send a packet quickly, but a lot of control packets are needed. So the battery life is short. On the other hand, in reactive routing protocols, a node starts making the routing table after receiving a packet sending request. So this protocol takes a lot of seconds to send a packet. In this paper, we focus on the cause that changes the routing table. Because we should update the routing table only when the node moves, we propose a novel routing protocol which has the acceleration information from the sensor. And a node exchanges the routing table only if it moves. We apply it to the surveillance camera system at a parking lot.