Abstract: Based on activity method, this paper focuses on morning commuting behavior when commuters travel with autonomous vehicles (AVs). Firstly, a net utility function of commuters is constructed by the activity utility of commuters at home, in car and at workplace, and the disutility of travel time cost and that of schedule delay cost. Then, this net utility function is applied to build an equilibrium model. Finally, under the assumption of constant marginal activity utility, the properties of equilibrium are analyzed. The results show that, in autonomous driving, the starting and ending time of morning peak and the number of commuters who arrive early and late at workplace are the same as those in manual driving. In automatic driving, however, the departure rate of arriving early at workplace is higher than that of manual driving, while the departure rate of arriving late is just the opposite. In addition, compared with manual driving, the departure time of arriving at workplace on time is earlier and the number of people queuing at the bottleneck is larger in automatic driving. However, the net utility of commuters and the total net utility of system in automatic driving are greater than those in manual driving.
Abstract: Autonomous mobile robots can be found in a wide
field of applications. Their types range from household robots over
workshop robots to autonomous cars and many more. All of them
undergo a number of testing steps during development, production
and maintenance. This paper describes an approach to improve
testing of robot behavior. It was inspired by the RoboCup @work
competition that itself reflects a robotics benchmark for industrial
robotics. There, scaled down versions of mobile industrial robots
have to navigate through a workshop-like environment or operation
area and have to perform tasks of manipulating and transporting
work pieces. This paper will introduce an approach of automated
vision-based testing of the behavior of the so called youBot robot,
which is the most widely used robot platform in the RoboCup
@work competition. The proposed system allows automated testing
of multiple tries of the robot to perform a specific missions and
it allows for the flexibility of the robot, e.g. selecting different
paths between two tasks within a mission. The approach is based
on a multi-camera setup using, off the shelf cameras and optical
markers. It has been applied for test-driven development (TDD) and
maintenance-like verification of the robot behavior and performance.