Toichi Sawada

FRAMEWORK OF TAILORMADE DRIVING SUPPORT SYSTEMS AND NEURAL NETWORK DRIVER MODEL

Drivers are positioned as the nucleus of driving helped by driving support systems of ITS. An automatic driving system in the future may release drivers partially from driving but will never release them completely. This is because automobiles are a door-to-door means of transport, and the concept of an automobile is a driver controlled vehicle system in essence. Therefore, it is desirable for driving support systems of automobiles to have a reasonable interface with a focus placed on personal characteristics of drivers. Today, various systems aimed at pre-crash safety of drivers and reduction in driving loads are being made fit for practical application. These systems have excellent mechanical functions but systems are not good enough to fit driving feelings. This is because the driver models of these systems are the same, though each driver has different driving characteristics. Nowadays, tailormade medical treatment is receiving much attention in the field of medical care. It is also desirable for driving support systems to reflect the driving characteristics of individuals as much as possible, begin monitoring the driver when a driver starts driving and calculates the driver model, and supports them with a model that makes the driver feel quite normal. That is the construction of Tailormade Driving Support Systems (TDSS). This research proposes a concept and a framework of TDSS, and presents a driver model that uses a neural network to build the system. As for the feasibility of this system, the research selects braking as a typical constituent element, and illustrates and reviews the results of experiments and simulations.

Measuring readiness potential in driving simulator toward investigation of driver's cognitive process

regression, called smooth sparse regression, which has a spatio-temporal “smoothing” prior that encourages weights that are close in time and space to have similar values. This causes the classifier to select spatio-temporally continuous groups of features, whereas standard sparse regression classifiers often select a scattered collection of independent features. We applied the proposed method to both simulation data and real MEG data from two separate experiments. We found that the method consistently increases classification accuracy, and produces weight vectors that are more meaningful from a neuroscientific perspective. Research fund: JST PRESTO.

BASIC STUDY ON TAILORMADE BRAKING SUPPORT SYSTEM

It is desirable for driving support systems to improve the safety of driver vehicle systems, and at the same time to have a performance that does not make individual drivers feel uncomfortable. Since human beings have various control characteristics, any system that supports driving under fixed conditions without taking such characteristics into consideration cannot be a driving support system in the true sense. The authors believe that only those systems that reflect the characteristics of individual drivers improve safety and pave the way for their widespread use, and proposal Tailormade Driving Support System (TDSS) in IATSS RESEARCH Vol. 28 No. 1. This TDSS is composed of three systems that support braking, steering and accelerating, and it gives assistance fitted to individual drivers with a driver model that uses a neural network. This research reviewed the construction of models of a Tailormade Braking Support System (TBSS) for braking to stop vehicles and the evaluation of drivers. As a result, the following conclusions were drawn. (1) Braking factors were found to change in the period from the start of braking to stopping; (2) Changes in braking factors can be logically incorporated into the control elements of braking support system; (3) Readymade Driver Model is effective as a model to be incorporated into the base system of TBSS; (4) Tailormade Driver Model built on Neural Network is effective as a main model to construct TBSS; (5) As for TBSS, both subjective and objective ratings on the timing and magnitude of braking are favorable, and its safety and sense of security are improved.