Recent studies regarding human and robotic behaviours are focused either to detect their differences in online video games, or to define criteria with which explorations processes, namely incremental processes that discover the features of initially unknown (or partially known) environments. In video games, design of maps, or generally speaking environments where players are placed, is performed in order to lead them to reach specific locations on maps. Using core concepts of the autonomous mobile robotics field, like path planning and decision making, and of video game design, like collision detection, we created a framework capable of defining an autonomous exploring behaviour to bring a robotic agent to explore an unknown environment, with the goal to complete the task simulating the behaviour of a human agent. To do this, the framework allows to perform a sequence of tasks, as an experiment, all regarding explorations with a low visibility condition, for human players, saving information about locations visited, defining a trajectory. The experiment culminates with a survey, to obtain other information about the human agent and his, or her, impressions about the explorations performed, like the percentage of environment explored. The robotic exploration is performed in the framework by perceiving the environment through its sensors, merging the information gathered in a virtual map of the environment, and choosing a location to reach according to a decision making process modeled after the data collected from humans. Moreover, the framework calculates the distance between two trajectories, one robotic and another one human, using this value to make a comparison between them, as well as to verify the goodness of the robotic trajectories obtained.
Studi recenti riguardanti i comportamenti umani e robotici si sono focalizzati sia sulla rilevazione delle loro differenze nei videogiochi online, sia sulla individuazione dei criteri con i quali vengono definiti i processi esplorativi, ovvero processi incrementali che scoprono le caratteristiche di ambienti inizialmente sconosciuti (o parzialmente noti). Nei videogiochi, viene svolta la progettazione di mappe o, più in generale, di ambienti nei quali i giocatori sono collocati, per portarli a raggiungere specifiche posizioni. Utilizzando concetti chiave del campo della robotica mobile autonoma, come la pianificazione dei percorsi e il decision making, e dello sviluppo dei videogiochi, come il collision detection, abbiamo creato un framework in grado di definire un comportamento autonomo in ambito esplorativo per portare un agente robotico a esplorare un ambiente sconosciuto, con l'obiettivo di completare l'attività, simulando il comportamento di un agente umano. Per fare questo, il framework propone una sequenza di compiti, sotto forma di esperimento, tutti riguardanti l'esplorazione di ambienti in una condizione di scarsa visibilità, per agenti umani, salvando le informazioni sui luoghi visitati. L'esperimento termina con un sondaggio, per ottenere ulteriori informazioni sull'agente umano e sue impressioni sul come ha effettuato le esplorazioni. L'esplorazione robotica viene eseguita nel framework percependo l'ambiente attraverso i sensori, integrando le informazioni raccolte in una mappa virtuale dell'ambiente e scegliendo una posizione da raggiungere in base alle informazioni estratte dai dati raccolti dai giocatori umani. Inoltre, il framework utilizza diverse funzioni matematiche per calcolare la distanza tra due traiettorie, una robotica e l'altra umana, per confrontarle.
Simulating human behaviour in environment exploration in video games
LAZZARETTI, SIMONE;ZHAN, YUAN
2017/2018
Abstract
Recent studies regarding human and robotic behaviours are focused either to detect their differences in online video games, or to define criteria with which explorations processes, namely incremental processes that discover the features of initially unknown (or partially known) environments. In video games, design of maps, or generally speaking environments where players are placed, is performed in order to lead them to reach specific locations on maps. Using core concepts of the autonomous mobile robotics field, like path planning and decision making, and of video game design, like collision detection, we created a framework capable of defining an autonomous exploring behaviour to bring a robotic agent to explore an unknown environment, with the goal to complete the task simulating the behaviour of a human agent. To do this, the framework allows to perform a sequence of tasks, as an experiment, all regarding explorations with a low visibility condition, for human players, saving information about locations visited, defining a trajectory. The experiment culminates with a survey, to obtain other information about the human agent and his, or her, impressions about the explorations performed, like the percentage of environment explored. The robotic exploration is performed in the framework by perceiving the environment through its sensors, merging the information gathered in a virtual map of the environment, and choosing a location to reach according to a decision making process modeled after the data collected from humans. Moreover, the framework calculates the distance between two trajectories, one robotic and another one human, using this value to make a comparison between them, as well as to verify the goodness of the robotic trajectories obtained.File | Dimensione | Formato | |
---|---|---|---|
Sim Human Behaviour in. Robots.pdf
accessibile in internet per tutti
Descrizione: Testo della tesi
Dimensione
2.49 MB
Formato
Adobe PDF
|
2.49 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/148001