Friendly lords implementing an artificial intelligence social architecture in Mount& Blade II: Bannerlord

Detalhes bibliográficos
Autor(a) principal: Ricardo, David Cristóvão
Data de Publicação: 2021
Tipo de documento: Dissertação
Idioma: eng
Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
Texto Completo: http://hdl.handle.net/10400.1/18128
Resumo: Despite living in a golden age for video games where there is an increase in the development and popularity of new technologies, such as Virtual Reality and Augmented Reality, and constant improvements being made on immersion, either through rich storytelling, high graphical fidelity or even gameplay itself, one area that is generally lacking on innovation is the social interaction of Non-Playable Characters. The credibility of these virtual characters (usually called NPCs) requires that they have human characteristics, such as emotions and the ability to think and make decisions depending on their own will. One of the most important characteristics is the ability to socialize and interact with each other. The main objective of this work was to implement a Social Architecture of Artificial Intel ligence in the game Mount&Blade II: Bannerlord to enhance the credibility of NPCs, making them socially more active and interesting, and improving the User Experience. In the game in which this study takes place (respectively Mount&Blade II: Bannerlord), although there are plenty of NPCs in the world, these are extremely limited, predictable and rarely demonstrate social behaviors. It also addresses some socio-emotional architectures such as CiF and FAtiMA. While FA tiMA is more focused on the generation of emotions, and how those emotions may affect the behaviour of characters, CiF has an explicit representation of social relations between NPCs and how the can influence behaviour. Due to this, the system architecture used as the basis for our model is the CiF architecture, also used before in some popular games like Skyrim and Conan Exiles. It is a system designed to be able to generate social behavior for the social agents, in this case the virtual characters. Instead of looking static and appearing to be clones of each other, the NPCs will appear more natural, making them more interesting and believable from the point of view of social interaction. One of the major differences between the CiF model and the developed model, named Comme il Faut - Bannerlord (CiF-BL), was, respectively, an addition of a centralized component, which will manage about which NPCs will be able to engage in interactions, how many Social Interactions will be able to take place at the same time, when these are started and finished From CiF-BL’s perspective, the player is just another character, and just as NPCs will want to interact with each other, they will also want to interact with the player. To do this work, it was necessary to adapt the game’s existing dialog system to make CiF BL’s changes possible. Adding the options that allow the player to interact socially with an NPC and vice-versa. It was also necessary to adapt and implement another game system, responsible for signaling and informing the most relevant locations and characters. This adaptation was responsible for making it possible to visualize the interactions between the different NPCs. Thus, during the time that the user is present in the virtual world, the virtual characters will no longer have a totally irrelevant and figurative role in the game, and will have a greater social participation. With the goal of appearing more believable, natural, and, to look more like "living beings". The developed model was validated and evaluated through user tests. Unfortunately, due to COVID-19 restrictions that occurred during the execution of this thesis, it was not possible to conduct a controlled in-person evaluation (which would be the ideal way to evaluate our work). Instead, a mod was produced and released online to the Bannerlords player and modding community. The only difference between the mod and the base version of the game was the CIF-BL model used to control the NPCs. The mod was called "Friendly Lords" and released in 18th of August 2021, both in NexusMods and in ModDB. The mod was announced on social networks, such as Reddit, in groups that were related to the game. Participants were voluntarily invited to complete an anonymous questionnaire regarding their gaming experience, and the answers were collected. From the analysis of these same answers, very favorable results were obtained for the objectives we defined. However, overall, the feedback from players and the community has been very positive about the mod and its modifications to the game. Some examples are that it looks promising, that they want active development, and that are grateful to help bring new life to the game. People give suggestions and constructive feedback to implement in the mod and help get a better experience. Meanwhile, some players have shown more interest and offered to collaborate directly in the development, either for more dialogue options between the NPCs, or to translate the mod respectively into Turkish which is the native language of the studio currently developing the game. In short, using the social architecture implemented in the game, the Credibility of NPCs and the Social Presence improved by more than 30%, having successfully achieved the goals. This document describes in more detail the process of researching, implementing and testing the model. To improve this work and the Artificial Intelligence social architecture, it would be necessary to at least add more personality traits and a greater number of different social interactions, for the NPCs to have a greater diversity in terms of their social behavior.
id RCAP_996f3182863c7ae9678560f851106e1c
oai_identifier_str oai:sapientia.ualg.pt:10400.1/18128
network_acronym_str RCAP
network_name_str Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
repository_id_str 7160
spelling Friendly lords implementing an artificial intelligence social architecture in Mount& Blade II: BannerlordArquitectura socialPersonagem Não-JogávelInteração socialInteligência artificialDomínio/Área Científica::Engenharia e Tecnologia::Outras Engenharias e TecnologiasDespite living in a golden age for video games where there is an increase in the development and popularity of new technologies, such as Virtual Reality and Augmented Reality, and constant improvements being made on immersion, either through rich storytelling, high graphical fidelity or even gameplay itself, one area that is generally lacking on innovation is the social interaction of Non-Playable Characters. The credibility of these virtual characters (usually called NPCs) requires that they have human characteristics, such as emotions and the ability to think and make decisions depending on their own will. One of the most important characteristics is the ability to socialize and interact with each other. The main objective of this work was to implement a Social Architecture of Artificial Intel ligence in the game Mount&Blade II: Bannerlord to enhance the credibility of NPCs, making them socially more active and interesting, and improving the User Experience. In the game in which this study takes place (respectively Mount&Blade II: Bannerlord), although there are plenty of NPCs in the world, these are extremely limited, predictable and rarely demonstrate social behaviors. It also addresses some socio-emotional architectures such as CiF and FAtiMA. While FA tiMA is more focused on the generation of emotions, and how those emotions may affect the behaviour of characters, CiF has an explicit representation of social relations between NPCs and how the can influence behaviour. Due to this, the system architecture used as the basis for our model is the CiF architecture, also used before in some popular games like Skyrim and Conan Exiles. It is a system designed to be able to generate social behavior for the social agents, in this case the virtual characters. Instead of looking static and appearing to be clones of each other, the NPCs will appear more natural, making them more interesting and believable from the point of view of social interaction. One of the major differences between the CiF model and the developed model, named Comme il Faut - Bannerlord (CiF-BL), was, respectively, an addition of a centralized component, which will manage about which NPCs will be able to engage in interactions, how many Social Interactions will be able to take place at the same time, when these are started and finished From CiF-BL’s perspective, the player is just another character, and just as NPCs will want to interact with each other, they will also want to interact with the player. To do this work, it was necessary to adapt the game’s existing dialog system to make CiF BL’s changes possible. Adding the options that allow the player to interact socially with an NPC and vice-versa. It was also necessary to adapt and implement another game system, responsible for signaling and informing the most relevant locations and characters. This adaptation was responsible for making it possible to visualize the interactions between the different NPCs. Thus, during the time that the user is present in the virtual world, the virtual characters will no longer have a totally irrelevant and figurative role in the game, and will have a greater social participation. With the goal of appearing more believable, natural, and, to look more like "living beings". The developed model was validated and evaluated through user tests. Unfortunately, due to COVID-19 restrictions that occurred during the execution of this thesis, it was not possible to conduct a controlled in-person evaluation (which would be the ideal way to evaluate our work). Instead, a mod was produced and released online to the Bannerlords player and modding community. The only difference between the mod and the base version of the game was the CIF-BL model used to control the NPCs. The mod was called "Friendly Lords" and released in 18th of August 2021, both in NexusMods and in ModDB. The mod was announced on social networks, such as Reddit, in groups that were related to the game. Participants were voluntarily invited to complete an anonymous questionnaire regarding their gaming experience, and the answers were collected. From the analysis of these same answers, very favorable results were obtained for the objectives we defined. However, overall, the feedback from players and the community has been very positive about the mod and its modifications to the game. Some examples are that it looks promising, that they want active development, and that are grateful to help bring new life to the game. People give suggestions and constructive feedback to implement in the mod and help get a better experience. Meanwhile, some players have shown more interest and offered to collaborate directly in the development, either for more dialogue options between the NPCs, or to translate the mod respectively into Turkish which is the native language of the studio currently developing the game. In short, using the social architecture implemented in the game, the Credibility of NPCs and the Social Presence improved by more than 30%, having successfully achieved the goals. This document describes in more detail the process of researching, implementing and testing the model. To improve this work and the Artificial Intelligence social architecture, it would be necessary to at least add more personality traits and a greater number of different social interactions, for the NPCs to have a greater diversity in terms of their social behavior.Apesar de vivermos numa era dourada para os videojogos onde há um aumento do desen volvimento e da popularidade de novas tecnologias, como a Realidade Virtual e a Realidade Aumentada, e constantes melhorias a serem feitas na imersão, seja através de ricas narrativas, fidelidade gráfica elevada ou até mesmo gameplay em si, uma área que geralmente não é alvo de inovação é a interação social entre as Personagens Não Jogáveis. A credibilidade dessas personagens virtuais (normalmente chamadas de NPCs) exige que estas tenham características humanas, como emoções e a capacidade de pensar e tomar decisões dependendo da sua própria vontade. Uma das características mais importante é a capacidade de socializar e interagir uns com os outros. O objectivo principal deste trabalho é implementar uma Arquitectura Social de Inteligência Artificial no jogo Mount&Blade II: Bannerlord para incrementar a Credibilidade dos NPCs, tornando-os socialmente mais ativos e interessantes, e melhorar a Experiência do Utilizador. No jogo em que se realiza este estudo (respetivamente Mount&Blade II: Bannerlord), embora haja uma abundância de NPCs no mundo, estes são extremamente limitados, previsíveis e raramente demonstram comportamentos sociais. Este trabalho aborda diferentes arquitecturas socio-emocionais, como Comme il Faut (CiF) e FAtiMA. Enquanto FAtiMA está mais centrado na geração de emoções, e como essas emoções podem afectar o comportamento das personagens, CiF tem uma representação explícita das relações sociais entre os NPCs e de que forma o comportamento pode influenciar. Devido a isto, a arquitetura do sistema usada como base para o nosso modelo é a arquitetura CiF, também já usada antes em alguns jogos populares como Skyrim e Conan Exiles. É um sistema concebido para ser capaz de gerar comportamentos sociais para os agentes sociais, neste caso, os personagens virtuais. Em vez de parecerem estáticos e aparentarem serem clones uns dos outros, os NPCs vão parecer mais naturais, tornando-os mais interessantes e credíveis do ponto de vista da interação social. Uma das maiores diferenças entre o modelo CiF e o modelo final desenvolvido, nomeado de Comme il Faut - Bannerlord (CiF-BL), foi, respectivamente, uma adição de um componente centralizado, que irá fazer a gestão acerca de quais os NPCs que estarão aptos para se envolverem em interações, quantas Interações Sociais serão capazes de se realizar ao mesmo tempo, quando estas são iniciadas e terminadas. Da perspectiva do CiF-BL, o jogador é apenas uma outra personagem e, assim como os NPCs vão querer interagir uns com os outros, eles também vão querer interagir com o jogador. Para realizar este trabalho, foi necessário adaptar o atual sistema de diálogo já existente do jogo para possibilitar as alterações do CiF-BL. Adicionando assim as opções que permitem ao jogador interagir socialmente com um NPC e vice-versa. Foi necessária também uma adaptação e a implementação de um outro sistema do jogo, responsável por assinalar e informar as localizações e os personagens mais relevantes. Esta adaptação foi responsável para tornar possível a visualização das interações entre os diferentes NPCs. Assim, durante o tempo em que o utilizador estiver presente no mundo virtual, as personagens virtuais irão deixar de ter um papel totalmente irrelevante e figurativo no jogo, e, terão uma maior participação social. Com o objectivo de parecerem mais credíveis, naturais, e, de se assemelharem mais com “seres vivos”. Para a validação e a avaliação do modelo, este foi sujeito a testes de utilizador. Infelizmente, devido às restrições COVID-19 que ocorreram durante a execução desta tese, não foi possível realizar uma avaliação controlada presencialmente (que seria a forma ideal de avaliar o nosso trabalho). Em vez disso, um mod foi produzido e lançado online para os jogadores do jogo Bannerlord e para a comunidade modding. A única diferença entre o mod e a versão base do jogo foi o modelo CIF-BL usado para controlar e melhorar os NPCs. O mod foi chamado de "Friendly Lords" e lançado em 18 de agosto de 2021, tanto no site NexusMods como no ModDB. O mod foi anunciado nas redes sociais, como o Reddit, em grupos que estavam relacionados com o jogo. Os participantes foram convidados voluntariamente a preencher um questionário anónimo relativamente à sua experiência de jogo, sendo feita a recolha das respostas. A partir da análise dessas mesmas respostas, foram obtidos resultados muito favoráveis para os objectivos que definimos. Entretanto, no geral, o feedback dos jogadores e da comunidade foi muito positivo sobre o mod e sobre as suas modificações no jogo. Alguns exemplos são que parece promissor, que querem um desenvolvimento ativo, e que estão gratos por ajudar a trazer uma nova vida ao jogo. As pessoas dão sugestões e feedback construtivo para implementar no mod e ajudar a obter uma melhor experiência. Entretanto, alguns jogadores mostraram mais interesse e disponibilizaram se para colaborar diretamente no desenvolvimento, quer seja para mais opções de diálogos entre os NPCs, quer seja para a tradução do mod, respectivamente para Turco que é a língua materna do estúdio que atualmente desenvolve o jogo. Resumindo, usando a arquitetura social implementada no jogo, a Credibilidade dos NPCs e a qualidade do comportamento social melhorou mais de 30%, tendo atingido com sucesso os objetivos. Este documento descreve mais em pormenor o processo de pesquisa, de implemen tação e de teste do modelo. Para melhorar este trabalho e a sua arquitetura social de Inteligência Artificial, seria necessário, pelo menos, adicionar mais traços de personalidade e um maior número de diferentes interações sociais, para os NPCs terem uma maior diversidade quanto ao seu comportamento social.Dias, João Miguel de Sousa de AssisSapientiaRicardo, David Cristóvão2022-07-28T13:22:30Z2021-12-202021-12-20T00:00:00Zinfo:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/masterThesisapplication/pdfhttp://hdl.handle.net/10400.1/18128TID:202911071enginfo:eu-repo/semantics/openAccessreponame:Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)instname:Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informaçãoinstacron:RCAAP2023-07-24T10:30:22Zoai:sapientia.ualg.pt:10400.1/18128Portal AgregadorONGhttps://www.rcaap.pt/oai/openaireopendoar:71602024-03-19T20:07:55.591378Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) - Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informaçãofalse
dc.title.none.fl_str_mv Friendly lords implementing an artificial intelligence social architecture in Mount& Blade II: Bannerlord
title Friendly lords implementing an artificial intelligence social architecture in Mount& Blade II: Bannerlord
spellingShingle Friendly lords implementing an artificial intelligence social architecture in Mount& Blade II: Bannerlord
Ricardo, David Cristóvão
Arquitectura social
Personagem Não-Jogável
Interação social
Inteligência artificial
Domínio/Área Científica::Engenharia e Tecnologia::Outras Engenharias e Tecnologias
title_short Friendly lords implementing an artificial intelligence social architecture in Mount& Blade II: Bannerlord
title_full Friendly lords implementing an artificial intelligence social architecture in Mount& Blade II: Bannerlord
title_fullStr Friendly lords implementing an artificial intelligence social architecture in Mount& Blade II: Bannerlord
title_full_unstemmed Friendly lords implementing an artificial intelligence social architecture in Mount& Blade II: Bannerlord
title_sort Friendly lords implementing an artificial intelligence social architecture in Mount& Blade II: Bannerlord
author Ricardo, David Cristóvão
author_facet Ricardo, David Cristóvão
author_role author
dc.contributor.none.fl_str_mv Dias, João Miguel de Sousa de Assis
Sapientia
dc.contributor.author.fl_str_mv Ricardo, David Cristóvão
dc.subject.por.fl_str_mv Arquitectura social
Personagem Não-Jogável
Interação social
Inteligência artificial
Domínio/Área Científica::Engenharia e Tecnologia::Outras Engenharias e Tecnologias
topic Arquitectura social
Personagem Não-Jogável
Interação social
Inteligência artificial
Domínio/Área Científica::Engenharia e Tecnologia::Outras Engenharias e Tecnologias
description Despite living in a golden age for video games where there is an increase in the development and popularity of new technologies, such as Virtual Reality and Augmented Reality, and constant improvements being made on immersion, either through rich storytelling, high graphical fidelity or even gameplay itself, one area that is generally lacking on innovation is the social interaction of Non-Playable Characters. The credibility of these virtual characters (usually called NPCs) requires that they have human characteristics, such as emotions and the ability to think and make decisions depending on their own will. One of the most important characteristics is the ability to socialize and interact with each other. The main objective of this work was to implement a Social Architecture of Artificial Intel ligence in the game Mount&Blade II: Bannerlord to enhance the credibility of NPCs, making them socially more active and interesting, and improving the User Experience. In the game in which this study takes place (respectively Mount&Blade II: Bannerlord), although there are plenty of NPCs in the world, these are extremely limited, predictable and rarely demonstrate social behaviors. It also addresses some socio-emotional architectures such as CiF and FAtiMA. While FA tiMA is more focused on the generation of emotions, and how those emotions may affect the behaviour of characters, CiF has an explicit representation of social relations between NPCs and how the can influence behaviour. Due to this, the system architecture used as the basis for our model is the CiF architecture, also used before in some popular games like Skyrim and Conan Exiles. It is a system designed to be able to generate social behavior for the social agents, in this case the virtual characters. Instead of looking static and appearing to be clones of each other, the NPCs will appear more natural, making them more interesting and believable from the point of view of social interaction. One of the major differences between the CiF model and the developed model, named Comme il Faut - Bannerlord (CiF-BL), was, respectively, an addition of a centralized component, which will manage about which NPCs will be able to engage in interactions, how many Social Interactions will be able to take place at the same time, when these are started and finished From CiF-BL’s perspective, the player is just another character, and just as NPCs will want to interact with each other, they will also want to interact with the player. To do this work, it was necessary to adapt the game’s existing dialog system to make CiF BL’s changes possible. Adding the options that allow the player to interact socially with an NPC and vice-versa. It was also necessary to adapt and implement another game system, responsible for signaling and informing the most relevant locations and characters. This adaptation was responsible for making it possible to visualize the interactions between the different NPCs. Thus, during the time that the user is present in the virtual world, the virtual characters will no longer have a totally irrelevant and figurative role in the game, and will have a greater social participation. With the goal of appearing more believable, natural, and, to look more like "living beings". The developed model was validated and evaluated through user tests. Unfortunately, due to COVID-19 restrictions that occurred during the execution of this thesis, it was not possible to conduct a controlled in-person evaluation (which would be the ideal way to evaluate our work). Instead, a mod was produced and released online to the Bannerlords player and modding community. The only difference between the mod and the base version of the game was the CIF-BL model used to control the NPCs. The mod was called "Friendly Lords" and released in 18th of August 2021, both in NexusMods and in ModDB. The mod was announced on social networks, such as Reddit, in groups that were related to the game. Participants were voluntarily invited to complete an anonymous questionnaire regarding their gaming experience, and the answers were collected. From the analysis of these same answers, very favorable results were obtained for the objectives we defined. However, overall, the feedback from players and the community has been very positive about the mod and its modifications to the game. Some examples are that it looks promising, that they want active development, and that are grateful to help bring new life to the game. People give suggestions and constructive feedback to implement in the mod and help get a better experience. Meanwhile, some players have shown more interest and offered to collaborate directly in the development, either for more dialogue options between the NPCs, or to translate the mod respectively into Turkish which is the native language of the studio currently developing the game. In short, using the social architecture implemented in the game, the Credibility of NPCs and the Social Presence improved by more than 30%, having successfully achieved the goals. This document describes in more detail the process of researching, implementing and testing the model. To improve this work and the Artificial Intelligence social architecture, it would be necessary to at least add more personality traits and a greater number of different social interactions, for the NPCs to have a greater diversity in terms of their social behavior.
publishDate 2021
dc.date.none.fl_str_mv 2021-12-20
2021-12-20T00:00:00Z
2022-07-28T13:22:30Z
dc.type.status.fl_str_mv info:eu-repo/semantics/publishedVersion
dc.type.driver.fl_str_mv info:eu-repo/semantics/masterThesis
format masterThesis
status_str publishedVersion
dc.identifier.uri.fl_str_mv http://hdl.handle.net/10400.1/18128
TID:202911071
url http://hdl.handle.net/10400.1/18128
identifier_str_mv TID:202911071
dc.language.iso.fl_str_mv eng
language eng
dc.rights.driver.fl_str_mv info:eu-repo/semantics/openAccess
eu_rights_str_mv openAccess
dc.format.none.fl_str_mv application/pdf
dc.source.none.fl_str_mv reponame:Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
instname:Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação
instacron:RCAAP
instname_str Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação
instacron_str RCAAP
institution RCAAP
reponame_str Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
collection Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
repository.name.fl_str_mv Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) - Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação
repository.mail.fl_str_mv
_version_ 1799133324493455360