Robotics and Virtual Worlds: An Experiential Learning Lab Barbara Caci1, Antonella D’Amico1, and Giuseppe Chiazzese2 1
Dipartimento di Psicologia, Università degli Studi di Palermo, Viale delle Scienze, Ed. 15. 90128 Palermo - Italy 2 Istituto per le Tecnologie Didattiche di Palermo, Consiglio Nazionale delle Ricerche, Via Ugo La Malfa, 153, 90146 Palermo - Italy {barbara.caci,antonella.damico}@unipa.it,
[email protected]
Abstract. Aim of the study was to investigate the cognitive processes involved and stimulated by educational robotics (LEGO® robots and Kodu Game Lab) in lower secondary school students. Results showed that LEGO® and KGL artifacts involve specific cognitive and academic skills. In particular the use of LEGO® is related to deductive reasoning, speed of processing visual targets, reading comprehension and geometrical problem solving; the use of KGL is related to visual-spatial working memory, updating skills and reading comprehension. Both technologies, moreover, are effective in the improvement of visualspatial working memory. Implications for Human-Robot Interaction and BICA challenge are discussed. Keywords: Educational robotics, cognitive skills, academic performance, human-robot interaction.
1
Introduction
Educational robotics is one of the most challenging research area in the domain of Human-Robot Interaction (HRI), and is eligible as one of the best practical domain for the BICA challenge [1; 2]. In the constructivist framework [3], robotic and virtual interfaces are considered powerful tools for learning concepts about Mathematics, Computer programming, and Physics [4], for improving visual-constructive abilities, reasoning and problem-solving skills [5; 6] and for enhancing narrative and paradigmatic thinking [7]. Both LEGO® robotic kits and KGL allow children to build robots or agents able to act goal-oriented behaviors, such as to direct themselves toward a light, avoid obstacles or move inside a maze, and so on. Using LEGO® robotic kits, children build the body of the robot assembling motors and sensors (e.g., touch, ultrasonic, or light) with the brain of robot, a programmable microcontroller-based brick (NTX). Then, they may program the mind of the robot using an object-based interface inspired to Logo, and based on if-then rules. However, LEGO® kits are limited in the kind of behaviors that subjects can design and program. A. Chella et al. (Eds.): Biologically Inspired Cognitive Architectures 2012, AISC 196, pp. 83–87. springerlink.com © Springer-Verlag Berlin Heidelberg 2013
84
B. Caci, A. D’Amico, and G. Chiazzese
A more extensive programming practice is offered by the recent Kodu Game Lab (KGL) [8], a 3D virtual stage for the creation of virtual worlds, in which children can define more complex behaviors, movements and interactions with characters and objects using similar programming rules [9]. Using KGL children may design virtual environments, defining terrain scenarios and adding mountain, volcanoes, lakes, and so on. Then, they may enrich the environment by adding funny characters (e.g. Kodu, cycle, boat, fish), environmental elements (trees, apples, rocks clouds), paths (streets, enclosures, walls) and city elements (houses, factories, huts, castles) and, finally, they may assign specific behaviors to some of the elements, generating an interactive virtual world. Although the literature about KGL is quite recent, first experiments on use of Kodu have shown that its visual programming language is particularly easy to use also for novice students [10], compared to textual language models used by Alice [11], Greenfoot [11] and Scratch [13]. Starting from this theoretical framework, an experiential learning laboratory (32hours) involving children was designed with a twofold goal: to study some of the cognitive and academic abilities involved in building and programming robots/agents; to measure the effectiveness of the laboratory in the enhancement of the same cognitive and academic skills.
2
Method
2.1
Participants
The study involved 59 students of 11 years of age, attending an Italian Secondary School, that were casually assigned to experimental (F =14, M=22) and control condition (F=15, M=18). Children of the experimental group (EG) followed the laboratory described below; children of the control group (CG) followed the regular school activities. 2.2
Materials and Procedures
A pre-post test design was adopted. During the pre-post test phases the cognitive abilities and academic performances of EG and CG were measured using: ─ eight syllogistic and conditional reasoning tasks, drawn from the Automated System for the Evaluation of Operator Intelligence [14]. ─ the PML working memory battery [15], aimed at measuring phonological and visual-spatial working memory, executive functions (shifting, updating, inhibition), rate of access to long term memory and speed of processing. ─ the MT Reading Comprehension tasks [16] requiring children to read a narrative and an informative passage and to answer to 15 multiple-choice questions for each passage. ─ two arithmetical problem solving tasks and one geometrical problem solving task.
Robotics and Virtual Worlds: An Experiential Learning Lab
85
Moreover, during laboratory the programming skills of children were assessed using two questionnaires (LEGO® -Q and KGL-Q) respectively aimed at examining the acquisition of concepts about fundamentals of LEGO® and KGL interfaces, and their programming rules. 2.3
The Laboratory
The laboratory consisted of eight four–hours sessions, performed during the curricular school time, and described below: 1. Familiarization with Robots – movies describing different kind of fantasy and real robot (e.g. Transformers, AIBO, RunBot) were presented to children, ad used as a frame to start a circle time discussion about scientific concepts related to Biology, Physics and Mathematics (e.g., the notion of brain, mind, sensory-motor apparatus, velocity, light spectrum). 2. Construction of LEGO® robots – In order to let children familiarize with LEGO® robotic kit, and to build a small-mobile robot following the instruction provided by the manual. 3. Programming training with LEGO® and creation of the arena – The programming interface was introduced to children. Then, they created a narrative scenario (e.g. a script) for the robot behavior and built a physical environment (i.e. the arena) using pasteboard, colors, modeling paste and other materials. 4. Programming session with LEGO® and verification - Children realized the programming algorithms for adapting the robot to the created environment. Finally, they completed the LEGO®-Q questionnaire about the acquisition of LEGO® programming skills. 5. Familiarization with KGL environment – Children acquired the functionality of the Xbox controller for interacting with KGL interface and explored some illustrative virtual worlds. Next, they designed and realized their own environment, choosing terrains, mountains, lakes, and specific characters. 6. Familiarization with tile-based visual programming of KGL - Children were trained to program the virtual world using appropriate rules (a rule in KGL is composed by a when-do tile). 7. Construction of the virtual world and programming of virtual agents- Children reproduced the narrative scenario previously realized with LEGO® (session 3), enriching it with all the elements and additional features available in KGL. 8. Programming session with KGL and verification – Children were involved in collaborative programming sessions, and were then presented with the KLG-Q questionnaire aimed at verifying the acquisition of KLG programming skills.
3
Results
A correlation analysis was performed with the aim to explore the relationships among the performance in cognitive or academic tests used in the pretest phase, and the results obtained by children in LEGO®-Q and KLG-Q. Results at Pearson’s
86
B. Caci, A. D’Amico, and G. Chiazzese
product-moment correlation tests showed that deductive reasoning skills (r=.72; p
