Bonaventure Undergraduate Robotics Laboratory

PeopleBot entering an office supervised by John Postl '11

The Bonaventure Robotics Laboratory supports the integration of robotics into the undergraduate computer science curriculum and permits experimentation with behavior control algorithms. The lab is located in room 103 of the William F. Walsh Science Center. The director of the lab is Dr. Robert Harlan.

The lab has an Adept MobileRobots PeopleBot robot, acquired through a 2007 grant from the George I. Alden Trust. The robot supports the Department's Inside the Mind of a Robot project, which aims to develop a software system that enables the robot to perform tasks autonomously in a human environment. The guiding idea behind the project is that the robot will act in and reason about the environment in the same way that humans do, thereby enabling the robot to communicate its reasoning to humans. The first part of the project, enabling the robot to find its way to an office in Walsh using a topological map, was completed in April, 2011, and is discussed in the lab's Technical Report 10.

The robotics lab supports the Department's Robotics and Computer Vision course, CS 342,and original research under faculty supervision by its undergraduates. Eight undergraduates have served as authors and co-authors of the labs technical reports listed below.

About the Robotics Lab

The lab was established in 1997 to support a robotics course. We purchased four "MIT-style" autonomous robots robots based upon the Handy Board micro-controller and Legos. Students experimented with basic behavior control algorithms. The capstone project was Cabbie, an autonomous robot that could navigate between locations in an abstract downtown Manhattan.

Robert Harlan and David Levine wrote a successful National Science Foundation CCLI-AI grant application to expand the lab and to integrate it into the undergraduate curriculum. The grant, awarded in May of 2000, enabled the purchase of the Khepera miniature robots. This platform was selected because it permits the development of large programs that run on a host computer and communicate with the robot via a tether. Each robotic workstation has a PC running Linux, a Khepera robot and a four-foot square environment within which to experiment with control algorithms. The lab also had turrets to support object manipulation and computer vision which can be added to the Khepera platform. It also has a larger Koala robot to permit real-world experimentation.

The department has developed an object-oriented interface for the Khepera robots in C++, which it makes available for non-commercial use free of charge through the GNU General Public License.

The Khepera robots were retired in 2010.

The four-foot tall Adept MobileRobotics PeopleBot robot discussed above was purchased throught the George I. Alden grant in 2007. One or two upper division undergraduates work on the Inside the Mind of a Robot project. The coding of programs controlling the robot’s behavior, planning and reasoning have been and will continue to be developed by computer science undergraduates. “It will provide them with a platform for developing real-time, mission-critical software,” Harlan said, adding that students in other disciplines will be able to design and conduct experiments on how humans interact with the robot.

The PeopleBot will enable Dr. Anne Foerst, a theologian, a member of the Computer Science Department and a co-author of the Alden grant with Harlan, to continue her experimentation with human-robot interaction begun at the Artificial Intelligence Laboratory at the Massachusetts Institute of Technology. Foerst, co-author of the grant and an internationally known expert on human-robot interaction, is the author of God in the Machine, a book that examines what robots can teach us about being human. 
    

Technical Reports and Associated Software of the Laboratory

1. Parsing as Search: An Easy-to-Understand RTN Interpreter
   Bonaventure Undergraduate Robotics Laboratory Technical Report 1, September, 1993
   Robert Harlan and David Patrone '93
   Originally published SIGCSE Bulletin, 25, 3, September 1993
   
   This paper discusses an RTN interpreter developed in LISP for testing the conformity of commands given in English to a grammar.
   
   Software

2. Parsing as Search 2: An Easy-to-Understand ATN Interpreter
   Bonaventure Bonaventure Undergraduate Robotics Laboratory Technical Report 2, September, 1996
   Robert Harlan, David Patrone '93 and Scott Alexander '96
   
   An extension of technical report 1, this paper discusses an semantic grammar capable of interpreting commands given in English and issuing commands in LISP to a simulated robot problem-solver.
   
   Software

3. BARD: The Bonaventure Autonomous Robot Delivery System

   Bonaventure Bonaventure Undergraduate Robotics Laboratory Technical Report 3, April 2000

   Shelly McClarigan '00
   
   BARD uses a topological map of an office complex model to plan and carry out delivery tasks. It was developed using a Khepera robot.
    

4. The kRobot Class Interface for the Khepera Robot
   Bonaventure Bonaventure Undergraduate Robotics Laboratory Technical Report 4, July 2000
   Robert M. Harlan, David B. Levine, Shelly McClarigan '00
   Originally published in the Proceedings of the 32^nd SIGCSE Technical Symposium on Computer Science Education SIGCSE Bulletin vol. 33(1), 2001, ACM Press, New York, NY, pp. 105 – 109.
   
   The kRobot Interface for the Khepera and Koala Robots, version 3.2

5. Descriptions of Projects at the Bonaventure Robotics Laboratory
   Bonaventure Bonaventure Undergraduate Robotics Laboratory Technical Report 5, March 2001
   Robert M. Harlan
   Originally presented at AAAI Robotics and Education Conference, Stanford University, March 2001
   
   This paper discusses both lab assignments and student work on autonomous robot navigation developed at St. Bonaventure.
   
    
6. Mapper II: A Robust Mapper of Unknown Environments
   Bonaventure Bonaventure Undergraduate Robotics Laboratory Technical Report 6, July, 2001
   Robert M. Harlan, Michael Neel '02 and Brian Zimmel '02
   
   This paper discusses an autonomous robot map generator that uses its wheel encoders to generate a metric map an unknown, rectilinear environment such as a model office building. The robot uses the rectilinear features of the environment to maintain its localization, generating an accurate map that can then be used for autonomous navigation by the robot.
   
   Mapper I
   Mapper II
   
    
7. Scout: A Robot Capable of Generating a Map of an Indoor Space and Using the Map for Navigation
   Bonaventure Bonaventure Undergraduate Robotics Laboratory Technical Report 7, July, 2002
   Robert M. Harlan, Matthew Cowles '03 and Jamie Casilio '03
   
   This report discusses an autonomous robot explorer that is capable of generating a topological map of its environment and then using the map to navigate. Localization of the robot is enhanced by recording information about the features of the environment that connect the landmarks or distinctive, identifiable features of the environment.
   
   The report is currently being revised and thus is not available.
    
8. AI and Robotics Labs
   Bonaventure Bonaventure Undergraduate Robotics Laboratory Technical Report 8, March 2004
   Robert M. Harlan
   Originally published in Accessible Hands-on Artificial Intelligence and Robotics Education. AAAI Press. Menlo Park, CA. 2004
   
   The paper discusses six labs used in the Robotics and Computer Vision course to introduce applied artificial intelligence techniques.
    
9. Creating Emergent Behavior
   Bonaventure Undergraduate Robotics Laboratory Technical Report 9, August 2004
   Robert M. Harlan and Shelly McClarigan
   Originally published in 36^th SIGCSE Technical Symposium on Computer Science Education. February, 2005.

   This paper presents two labs that enable undergraduate students to create emergent behavior, behavior that is not programmed by the roboticist but that emerges from the interaction of lower-level behaviors and the environment.
   
   Emergent Behavior Labs
   Video

10. Inside the Mind of a Robot
    Bonaventure Undergraduate Robotics Laboratory Technical Report 10, April, 2011
    John Postl
    The goal of the “Inside the Mind of a Robot” project is to enable a life-sized, PeopleBot robot to carry out navigation tasks in St. Bonaventure University’s Walsh Science Center. This paper discusses the behavior control program that enables the robot to navigate to an office assigned by the user autonomously from a fixed position in the Center's lobby and to return to the starting point. Low-level behaviors, such as detecting doors and aligning with walls, are sequenced to enable the robot to complete more complex tasks such as entering offices. At the planning level, the robot uses a map to plan a route from its current position to a goal. The map is used to guide the execution of its plan.