This paper proposes an automatic driving system for lane change maneuver on highway. The automatic driving system is based on a reference driver model which employs on-board vehicle external sensing signals and estimates the initiation timing and the magnitude of longitudinal and lateral accelerations as reference signals. The patterns of reference accelerations are determined by modeling the expert drivers' experimental maneuvers. The proposed automatic driving system is tested on a driving simulator (DS) of Tokyo University of Agriculture and Technology (TUAT) by applying control systems which convert the reference accelerations to vehicle control input such as the accelerator pedal displacement and the steering wheel angle. The results by the automatic driving show good consistency with the results obtained by the expert driver experiments. The effectiveness of the proposed automatic driving system is verified by changing the velocity difference between the lanes to show that the system can handle various lane change situations.

抄録全体を表示

Directional response and directional stability of vehicle combinations are the main research topics in vehicle safety. This paper uses simplified mathematical models to analyze the steering sensitivity in lane-change maneuvers of multi-articulated vehicles with multiple axles. It also calculates the frequency response, pole, and zero in the transfer functions of each variable for two types of tractor and double-trailer combinations having multiple axles. The following conclusions are drawn. (1) Steering sensitivity of multi-articulated vehicles with multiple axles in lane-change maneuvers is composed of steering sensitivity in steady-state turns and the parameters of lane-change maneuvers. (2) The effect of the number of trailer axles on the damping ratio of zero is larger than that of the other parameters for multiple axles. (3) A large steering angle is necessary to change running lanes for multi-articulated vehicles with multiple axles that have low steering sensitivity in lane-change maneuvers; therefore, the rise time is short. Moreover, if the vehicle combination has a low damping ratio of oscillatory modes, the settling time is long, and the overshoot and maximum deviation are large.

抄録全体を表示

In this study, an optimization method is developed to reduce the ground noise impact of helicopter approach flights, and flight experiments are conducted to confirm the effectiveness of this optimization. First, we build a noise model of the JAXA (Japan Aerospace Exploration Agency) experimental helicopter based on measurement results. Then, we define optimal control problems to minimize the noise levels measured at points on the ground surface. The main objective of this paper is to propose a real-time flight trajectory optimization method for solving the optimization problems using actual flight conditions with an on-board computer. To validate the effect of the proposed method, some numerical simulations are conducted under real flight experiment scenarios. The obtained optimal solutions are characterized by steep flight path angles so as to avoid the flight conditions that generate large noise, avoidance of the measurement points, and short flight times; these optimal flights are different from conventional flight trajectories. Finally, the validity of the proposed optimization method is examined by means of flight experiments with the experimental JAXA helicopter. The actual flight experiments result in noise reduction, and prove the effectiveness of the optimization.

抄録全体を表示

Various types of low-speed vehicles have been developed for welfare and ecological purposes, some of which are equipped with autonomous locomotion functions. Although a number of technologies for safe autonomous locomotion have been developed, few researches have been conducted into technologies that provide the riders of such vehicles with a sense of security. Therefore, this research aims to develop a Human Machine Interface that provides a sense of security. Based on discussions held on information that would be useful for riders of intelligent low-speed vehicles, we constructed two prototype haptic interfaces termed as "Information Steering Systems," which provide advance notification of upcoming vehicle locomotion and alert notifications for surrounding pedestrians. We determined the characteristic parameters of the proposed system by conducting experiments using the first prototype interface with a powered wheelchair driving simulator. The operation of the proposed system was verified in two driving environments using the second prototype interface and a prototype intelligent powered wheelchair.

抄録全体を表示

Although electric-powered mobility devices with autonomous locomotion functions are considered helpful in assisting the elderly with decreased abilities to go on outings, technologies for safe locomotion in an outdoor environment with pedestrians have not yet been researched sufficiently. This research focuses on the autonomous locomotion on sidewalks where pedestrians coexist. An interactive locomotion strategy involving velocity control is also discussed. Taking into consideration the characteristics of sidewalks, a velocity control method that considers the interaction prediction of surrounding pedestrians was proposed and implemented into our prototype intelligent powered wheelchair. Through an experiment on a sidewalk test course, the operations of the proposed system were verified.

抄録全体を表示