Evaluating a Mouse-Based and a Tangible Interface Used for Operator Intervention on Two Autonomous Robots

Abstract

This paper presents an experiment where participants used two different interfaces of an application designed for controlling the movements of two vineyard spraying agricultural robots. The goal of the application was to evaluate users’ efficiency and perceived satisfaction while controlling the robots’ movements using a mouse-based and a tangible interface. In a controlled experiment conducted at the Software Quality and Human-Computer Interaction Laboratory of the Computer Engineering and Informatics Department of Patras University 32 participants, experienced in using mouse-based interfaces, used both interfaces. During the experiment, the users interacted with a simulation of the robot movements, without moving and colliding the actual robots, while the robot speed and behavior were based on actual data from our field experiments. The experiment followed a within-group design and for the tangible interface, the Wizard of Oz method was used. For data collection, eye-tracking glasses were used to capture the interaction with the tangible interface combined with recording devices. The hypotheses tested included efficiency measured in time to complete a task, accuracy measured in unsprayed or double-sprayed areas and robot collisions, and perceived user satisfaction measured with a short custom questionnaire. Results indicated that no statistically significant differences were found related to interface efficiency and accuracy, but users liked using the tangible version of the application the most.