Design of an intelligent outdoor mobile robot with autonomous road-crossing function for urban environments

Abstract

This paper describes the total system design and implementation of an outdoor mobile robot with an autonomous road-crossing function intended for urban environments. The developed robot is capable of navigating along sidewalks while simultaneously detecting pedestrian push-button boxes. If a button box is detected, the robot can autonomously traverse up to it with high precision and activate the push button with a finger. Next, it approaches the crossing zone, detects the zebra stripes and perceives the trajectory required to cross. Upon detecting a green signal, the robot then finally crosses the road to reach the opposite side of the road and continue to its intended destination. We tested the robot in a real and unmodified outdoor environment and obtained promising results. More application-oriented, the novelty of this work is that not only is the robot capable of crossing roads autonomously, it can also autonomously activate pedestrian push buttons in a real world outdoor environment; there are not many robots that can perform such an action in an outdoor environment.

AutOTranS: an Autonomous Open World Transportation System

“…The research into these robots has been growing [1] but there are still significant challenges handling the uncertain character of many outdoor operating environments. Some recent work does utilize mobile manipulators in outdoors scenarios: using pedestrian cross-walks and traffic lights [2], moving around a campus to get you a coffee [3] or working throughout a solar plant [4]. However, none of these systems can handle outdoor pick and place tasks involving novel objects.…”

“…1 Dept. of Physics, System Engineering and Signal Theory, University of Alicante, Alicante, Spain (email: brayan.impata@ua.es) 2 Dept. of Electrical and Computer Engineering, Northeastern University, Boston, Massachusetts, USA (email: shah.vi@husky.neu.edu, ha.singh@northeastern.edu) 3 College of Computer and Information Science, Northeastern University, Boston, Massachusetts, USA (email: rplatt@ccs.neu.edu) the picking that does not make any assumptions about the objects.…”