100 IEEE TRANSACTIONS ON EDUCATION, VOL. 51, NO. 1, FEBRUARY 2008 Hybrid Educational Strategy for a Laboratory Course on Cognitive Robotics Catalin Buiu Abstract—This paper presents an innovative cognitive robotics laboratory course based on the latest developments in mobile robotics, communication networks, software agents, and educa- tional technologies. The hybrid educational strategy combines simulation and experiments on real robots, as well as in-class experiments and remote control over the Internet, while the stu- dents are assisted by human tutors and artificial software agents. Evaluation data demonstrate the validity and attractiveness of this approach. Index Terms—Education, educational technology, mobile robots, software agents. I. INTRODUCTION C OUNTRIES, institutions, and individuals are facing sig- nificant challenges regarding the educational system. The need for a flexible learning system, tailored to students’ edu- cational and social backgrounds, is much more acute in tran- sitional economies (such as that of Romania) where financial resources are scarce and where the lack of practical, experi- mental, engineering education is common in the Romanian uni- versities. Remote experimentation has attracted a growing re- search interest, which has generated numerous applications. A distributed laboratory for remote operations and robotics re- search that uses an object-oriented protocol (TelRIP) and the In- ternet is described in [1]. Most of the remote experiments avail- able now are Web-based. For example, the distance learning application presented in [2] allows a remote user to conduct experiments in the Control Engineering Lab at Oregon State University, Corvallis. The Swiss Federal Institute, Lausanne, Switzerland, has also demonstrated a number of robots and other devices connected to the Internet. A LabView-based solution to distance learning that allows the real-time control of elec- trical drives and robotic arms is presented in [3]. Many other researchers have followed a similar path. For example, another solution described in [4] allows students to control remotely a pilot-scale DC servomotor, while [5] presents the remote fuzzy control of a pedagogical power plant. An Internet-based moni- toring application of distributed control systems (DCSs) is pre- sented in [6], along with the hardware and software design con- siderations that enable it to access the process variables of the DCSs via a Web browser. An educational application of remote, Manuscript received December 21, 2005; revised November 2, 2006, Feb- ruary 16, 2007, and June 19, 2007. This work was supported by the European Commission under the Leonardo da Vinci Program (Contract N/04/B/PP 165.011 AutoTech) and by Microsoft Romania’s educational program. The author is with the Department of Automatic Control and Systems En- gineering, University Politehnica of Bucharest, Bucharest 060042, Romania (e-mail: cbuiu@ics.pub.ro). Digital Object Identifier 10.1109/TE.2007.906605 Internet-based, level control is presented in [7]. LABNET is an- other Internet-based remote laboratory for control engineering education developed at University Europea, Madrid, Spain. At present, this remote laboratory integrates three basic physical systems: level control, temperature control and a ship stabilizing system [8]. This transition to remote labs has also been recognized by laboratory equipment manufacturers, who are developing new ways to connect their equipment to the Internet. However, most educational institutions have not yet been able to take full ad- vantage of this technology. Automation Technicians Vocational Training Repository (AutoTech) is a Leonardo da Vinci pilot project currently funded by the European Commission (2004–2007) [9]. The main goal of the project is to develop a set of motivating and innovative Internet-based vocational training packages for automation technicians. Students at vocational training schools and students in higher education institutions are secondary target groups of this project. The learning material is developed by AutoTech partners from four countries (Norway, Spain, Germany, Romania) and comprises traditional learning com- ponents such as theory, presentations, exercises, quizzes, and visual aids (pictures, videos, interactive animations, virtual reality). The main innovations in the project stem from 1) learning material, which includes interactive and dynamic simulations, games, and competitions that are industrially relevant, innovative, and motivating; 2) the use of recent developments in remote experimentation which integrate online interactive remote experimentation with real physical laboratory equipment (including robots). Robotics is a favorite application area for remote and vir- tual experimentation over the Internet. Remote experimentation refers to the control of real manipulators or mobile robots, while virtual experimentation refers to the control of virtual models of the robots. The application presented in [10] is an example of virtual experimentation, where virtual reality modeling lan- guage (VRML) models of industrial robots are used for remote control. A direct control architecture for Internet-based personal robots that is unaffected by the inherent time delay over the net is described in [11], while [12] discusses vehicle teleoperation and several characteristics that distinguish this from remote con- trol. In [13], the authors present the remote operation of an ABB IRB 1400 S4 robot, which is intended to be used in a remote lab- oratory at the University of Leon, Spain. Most of the remote robotics applications reported in the lit- erature so far have concentrated on low-level servo motor and joint control of robots. The applications presented in this paper concentrate on higher levels of control that are associated with cognitive robotics. Rather than simply moving a robotic arm, a 0018-9359/$25.00 © 2008 IEEE