| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 11.47 MB | Adobe PDF | |||
| 190.85 KB | Adobe PDF | |||
| 85.57 KB | Adobe PDF | |||
| 113.73 KB | Adobe PDF | |||
| 2.3 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
Atualmente o desenvolvimento de hardware vem superando as necessidades dos videojogos,
com isso os videojogos atuais conseguem ter um aspecto visual cinematográfico, mas não tem
evoluído muito além de oferecem suporte da grandes resoluções e taxas de atualização acima
dos 60Hz, chegando ao ponto de novos lançamentos simplesmente não apresentarem muitas
novidades ao público.
O sucesso é, de maneira geral, o objetivo das empresas ao desenvolverem um videojogo. Alguns
estudos como o de de Byl (2015) mostram que o sucesso dos videojogos está relacionado com
as emoções individuais. Baseando-se nessas informações foram definidos dois problemas de
estudo que guiaram essa pesquisa. O primeiro, onde se observou que os videojogos não possuem
uma componente automática que os ajudem a manter os jogadores motivados e envolvidos
emocionalmente e o segundo em que foi analisado que os videojogos não são capazes de fazer
um balanceamento personalizado, de forma a apresentarem uma experiencia única para cada
jogador.
Essa motivação levou a buscar uma maneira de inovar os videojogos utilizando o estado emocional dos jogadores para criar alterações nas mecânicas de jogo e na forma como a narrativa
é apresentada. Para isso, nessa dissertação é apresentado um modelo de como a utilização do
biofeedback pode ser aproveitada para a criação de videojogos afetivos, onde através da leitura
de sensores fisiológicos, o jogo seja capaz de detectar a emoção do jogador e poder ser capaz
de se adaptar para alterar o estado emocional do jogador e mantê-lo em um nível de imersão e
entretenimento maior.
Para a realização dos experimentos foi utilizado uma versão de código livre do Pac-Man, um
videojogo transgeracional e com uma mecânica simples. Foram realizadas algumas alterações
no código para que fosse possível realizar a coleta dos dados do jogo e durante o desenvolvimento
desse estudo foram realizadas duas etapas de experimentos.
Os experimentos consistiram em voluntários jogarem o Pac-Man e durante o tempo em que
jogavam, seus dados fisiológicos eram coletados através de biossensores e ao mesmo tempo
também eram coletados os dados das variáveis do jogo.
Na primeira etapa dos experimentos o objetivo foi produzir material suficiente para que fosse
possível observar os padrões básicos de comportamento nas curvas dos biossensores, para isso
foram alteradas algumas variáveis de jogo manualmente entre uma fase e outra para que fosse
possível fazer uma analise desses dados e então buscar padrões de comportamento. Enquanto na segunda etapa dos experimentos foi buscando aplicar o conhecimento adquirido
com as analises, utilizando o biofeedback para auxiliar no balanceamento do jogo. Nessa etapa
o jogador jogava a primeira fase configurada manualmente e as próximas eram geradas por um
sistema de predição de níveis, que fazia uma análise do desempenho do jogador e comparava
com os dados fisiológicos e então gerava a configuração da próxima fase do jogo.
A utilização dos dados de desempenho do jogador durante a fase jogada sendo comparados com
os dados fisiológicos ajudou a resolver um dos atuais problemas relacionados à interpretação
das emoções. Com essa informação é possível entender o que causou um disparo emocional
no jogador, isso porque mesmo conseguindo identificar um padrão nos dados dos biossensores
e associa-los a uma emoção, se não for possível saber o que a causou, pode não ajudar no
balanceamento do videojogo. Isso porque se ao analisar os dados dos biossensores e a emoção
extraída dos dados foi medo, primeiro deve-se olhar para o contexto do jogo, pois se a pessoa
está jogando um videojogo de terror, é esperado que ela sinta medo, porque é isso que vai
ajudar a mantê-la motivada a jogar, mas se ao invés disso apenas for levado em consideração as
escalas de emoções, vai ser encontrado que medo é uma emoção negativa e que isso deve ser
alterado.
Assim, o que foi observado ao longo dessa pesquisa é que é possível aumentar o envolvimento
e a atenção dos jogadores se o jogo se adaptar ao estado emocional da pessoa e a principal
constatação foi que além de buscar padrões apenas nos dados dos biossensores é necessário
buscar o contexto da emoção encontrada nos dados do videojogo. Mas o que deve ser observado
é a necessidade de pesquisas em diversas áreas, principalmente em inteligência artificial para
auxiliar no reconhecimento de emoções e no aprimoramento dos biossensores para evitar a
saturação dos dados.
Nowadays, the development of hardware is exceeding the need of videogames, and with that, the videogames could technically achieve a visual aspect similar to the visual achieved in movies. However, this feature is not achieving this evolution yet, and videogames are only limited to offer higher resolutions and refreshment rates above the 60Hz, with few surprises to the public in new releases. Success is, in general, the goal of companies to develop a video game. Some studies like de de Byl (2015) show that the success of video games is related to individual emotions. Based on this information, two study problems were defined that guided this research. The first, where it was observed that videogames do not have an automatic component that helps them keep players motivated and emotionally involved and the second where it was analyzed that videogames are not capable of custom balancing to present an experience unique to each player. This motivation led to finding a way to innovate videogames by using the emotional state of the players to create changes in the game mechanics and the way the narrative is presented. For this, this dissertation presents a model of how the use of biofeedback can be used for the creation of affective video games, where through the reading of physiological sensors, the game is able to detect the player’s emotion and be able to adapt to change the emotional state of the player and keep them at a higher level of immersion and entertainment. For the experiments was used an open-source version of Pac-Man, and a cross-generation video game with simple mechanics. Some changes were made to the code so that it was possible to collect the game data and during the development of this study two stages of experiments were performed. The experiments consisted of volunteers playing Pac-Man and during the time they played, their physiological data were collected through biosensors and at the same time the data of the game variables were also collected. In the first stage of the experiments, the objective was to produce enough material to observe the basic behavioural patterns in the biosensor curves. To do so, some game variables were manually changed between phases, so that it was possible to make an analysis of these data. so look for patterns of behaviour. While in the second stage of the experiments was trying to apply the knowledge acquired with the analysis, using biofeedback to help balance the game. At this stage, the player played the first manually configured phase and the next were generated by a level prediction system, which made an analysis of the player’s performance and compared with the physiological data and then generated the configuration of the next phase of the game. Using player performance data during the play phase compared to physiological data has helped to solve one of the current problems related to the interpretation of emotions. With this information, it was possible to understand what caused an emotional trigger in the player, because even if can identify a pattern in the biosensors data and associate them with an emotion if it is not possible to know what caused it, may not help in balancing the video game. This is because if when analyzing biosensor data and the emotion extracted from the data was fear, one must first look at the context of the game, because if one is playing a horror video game, one is expected to be afraid, because that is which will help keep her motivated to play, but if only the scales of emotions are taken into account, it will be found that fear is a negative emotion and that it must be changed. Thus, what was observed throughout this research is that it is possible to increase the involvement and attention of players if the game adapts to the emotional state of the person and the main finding was that in addition to seeking patterns only in the biosensors data is necessary to seek the context of the emotion found in the video game data. But what should be noted is the need for research in various areas, especially artificial intelligence to aid in the recognition of emotions and enhancement of biosensors to avoid data saturation.
Nowadays, the development of hardware is exceeding the need of videogames, and with that, the videogames could technically achieve a visual aspect similar to the visual achieved in movies. However, this feature is not achieving this evolution yet, and videogames are only limited to offer higher resolutions and refreshment rates above the 60Hz, with few surprises to the public in new releases. Success is, in general, the goal of companies to develop a video game. Some studies like de de Byl (2015) show that the success of video games is related to individual emotions. Based on this information, two study problems were defined that guided this research. The first, where it was observed that videogames do not have an automatic component that helps them keep players motivated and emotionally involved and the second where it was analyzed that videogames are not capable of custom balancing to present an experience unique to each player. This motivation led to finding a way to innovate videogames by using the emotional state of the players to create changes in the game mechanics and the way the narrative is presented. For this, this dissertation presents a model of how the use of biofeedback can be used for the creation of affective video games, where through the reading of physiological sensors, the game is able to detect the player’s emotion and be able to adapt to change the emotional state of the player and keep them at a higher level of immersion and entertainment. For the experiments was used an open-source version of Pac-Man, and a cross-generation video game with simple mechanics. Some changes were made to the code so that it was possible to collect the game data and during the development of this study two stages of experiments were performed. The experiments consisted of volunteers playing Pac-Man and during the time they played, their physiological data were collected through biosensors and at the same time the data of the game variables were also collected. In the first stage of the experiments, the objective was to produce enough material to observe the basic behavioural patterns in the biosensor curves. To do so, some game variables were manually changed between phases, so that it was possible to make an analysis of these data. so look for patterns of behaviour. While in the second stage of the experiments was trying to apply the knowledge acquired with the analysis, using biofeedback to help balance the game. At this stage, the player played the first manually configured phase and the next were generated by a level prediction system, which made an analysis of the player’s performance and compared with the physiological data and then generated the configuration of the next phase of the game. Using player performance data during the play phase compared to physiological data has helped to solve one of the current problems related to the interpretation of emotions. With this information, it was possible to understand what caused an emotional trigger in the player, because even if can identify a pattern in the biosensors data and associate them with an emotion if it is not possible to know what caused it, may not help in balancing the video game. This is because if when analyzing biosensor data and the emotion extracted from the data was fear, one must first look at the context of the game, because if one is playing a horror video game, one is expected to be afraid, because that is which will help keep her motivated to play, but if only the scales of emotions are taken into account, it will be found that fear is a negative emotion and that it must be changed. Thus, what was observed throughout this research is that it is possible to increase the involvement and attention of players if the game adapts to the emotional state of the person and the main finding was that in addition to seeking patterns only in the biosensors data is necessary to seek the context of the emotion found in the video game data. But what should be noted is the need for research in various areas, especially artificial intelligence to aid in the recognition of emotions and enhancement of biosensors to avoid data saturation.
Description
Keywords
Biofeedback Detecçãodeemoção Gamedesign Evolutivo Videojogosadaptativos Videojogosafetivos