Eiji Teramoto

NETSTREAM: traffic simulator for evaluating traffic information systems

A traffic simulator NETSTREAM (NETwork Simulator for TRaffic Efficiency And Mobility) has been developed for evaluating traffic information systems such as dynamic route guidance systems. In this paper, a traffic flow model for evaluating traffic information systems in large-scale networks is proposed. The block density method which is one of macroscopic models is improved so that each vehicle is distinguished and lane configurations are taken into consideration. Traffic now can be calculated at high speed in consideration of route choice of each driver using the proposed model. Reduction of average travel time is evaluated in case that the dynamic route guidance system (DRGS) is installed in a whole provincial city using NETSTREAM. In the situation that nearly 66000 vehicles are generated in the network including over 3000 links, the travel times of all the vehicles are calculated in a short time.

Trajectory estimation improvement based on time-series constraint of GPS Doppler and INS in urban areas

With the development of various driver assistance systems using vehicle position, improvement of positioning accuracy is desired. GPS, widely used as a positioning system, can provide accuracy of a few meters in suburban areas. However, in urban areas, this accuracy can be degraded to ten meters or more because of the reflection and blocking of GPS signals by tall buildings. In order to maintain accurate positioning in such an environment, it is important to develop a method of trajectory estimation which is robust against the surrounding buildings. In this paper, the heading estimation which is the most influential factor in trajectory accuracy is improved by tight coupling of time-series GPS-Doppler and INS. As a result, the accuracy of trajectory estimation is improved in urban areas.

Proposal for a new localisation method using tightly coupled integration based on a precise estimation of trajectory from GPS Doppler

Driver assistance systems require the function of precise localisation. However, the accuracy of standard global positioning system (GPS) is inadequate for this. In this paper, a new method for precise localisation using the tightly coupled integration of inertial navigation system (INS) and raw GPS data is investigated. GPS Doppler corresponds to the relative velocity between the satellite and the receiver. Our new method enables us to make a robust and accurate estimation of the vehicle trajectory based on INS and GPS Doppler. Furthermore, the total optimisation by combining GPS pseudorange data with the accurate trajectory enables us to estimate the position of the vehicle precisely.

M-epoch ambiguity resolution technique for single frequency receivers with INS aid

Accurate positioning for land vehicles is required to meet the requirements of new safety systems. Though dual frequency RTK has sufficient accuracy for these needs, the cost makes it difficult for this to become common in consumer vehicles. The objective of this study is to realize commercially feasible RTK using practical single frequency receivers with the support of vehicle embedded gyros and odometers. The information of vehicle trajectory limits the ambiguity search space and improves the performance of single frequency RTK. A driving experiment showed that this approach has advantages both in a suburban area and urban canyon, compared to the traditional RTK.