AI behavior in equal intersections : modification of Simulandia traffic AI
Lahdenranta, Karl (2021)
Lahdenranta, Karl
2021
All rights reserved. This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2021061816349
https://urn.fi/URN:NBN:fi:amk-2021061816349
Tiivistelmä
When creating a simulation of any sort, one of the most important aspects is immersion and when the player is put into a completely lifeless environment it is rarely very immersive. To solve these issues developers cannot simply plant people, animals, etc. to our simulation, these objects also need to interact with the player and the interactions should feel real. This is where we need to implement AI as part of our simulation.
AIs are usually fairly complex because a simple AI is often not enough to create an illusion of interacting with a living being. This complexity can, unfortunately, become an issue when there is a need to alter existing AI behavior.
This thesis will discuss methods for incorporating new behavior for existing AI, specifically iTS Intelligent Traffic System (Unity asset) used by TTS Työtehoseura ry’s Simulandia project (made by Turku University of Applied Sciences and ADE Oy) and the new behavior being comprehension of equal junctions.
Although iTS has many build-in options for modifying AI behavior, it has no understanding of equal intersections. This will increase the complexity of our problem, as our task changes from simply modifying existing behavior to creating a completely new one and incorporating it as a functioning part of the AI. The AI consists of multiple fairly large scripts, which also brings a problem of accidentally creating new bugs and unwanted behavior when editing said files. For this reason, the thesis will partly cover how the AI works currently and lists some possible issues that may arise when trying to modify it in a certain way.
This thesis presents three possible methods on how to implement equal intersections to the driving environment and discusses strengths, weaknesses, and optimal integration environments for each method. All methods are then compared on which one would be the best choise to implement to the current version of Simulandia driving environments and what possible aspects these methods have for further development.
AIs are usually fairly complex because a simple AI is often not enough to create an illusion of interacting with a living being. This complexity can, unfortunately, become an issue when there is a need to alter existing AI behavior.
This thesis will discuss methods for incorporating new behavior for existing AI, specifically iTS Intelligent Traffic System (Unity asset) used by TTS Työtehoseura ry’s Simulandia project (made by Turku University of Applied Sciences and ADE Oy) and the new behavior being comprehension of equal junctions.
Although iTS has many build-in options for modifying AI behavior, it has no understanding of equal intersections. This will increase the complexity of our problem, as our task changes from simply modifying existing behavior to creating a completely new one and incorporating it as a functioning part of the AI. The AI consists of multiple fairly large scripts, which also brings a problem of accidentally creating new bugs and unwanted behavior when editing said files. For this reason, the thesis will partly cover how the AI works currently and lists some possible issues that may arise when trying to modify it in a certain way.
This thesis presents three possible methods on how to implement equal intersections to the driving environment and discusses strengths, weaknesses, and optimal integration environments for each method. All methods are then compared on which one would be the best choise to implement to the current version of Simulandia driving environments and what possible aspects these methods have for further development.