Name:Tianyu HU
Position:Junior Gameplay Programmer
Company:UBISOFT


I worked in a start-up company as a gameplay programmer creating arcade games for 2 years. And then I went to University of Utah to study in a master degree program Entertainment Arts & Engineering for 2 years. Now I have been working in Ubisoft Shanghai FARCRY wildlife team as a junior gameplay programmer for the last year.


Far Cry 5 is a very large open world game, featuring around forty animals, more than fifteen thousand animations, and complex, emergent Artificial Intelligence. It would not be practical for traditional Quality Assurance teams to manually test all of the possible interactions between all of these characters and game world. Furthermore, Ubisoft is a global company with many studios around the world collaborating together on the same game, so efficient and reliable code development was critical to the project’s success. The Shanghai studio was responsible for all of the animals in Far Cry 5, so we needed a way to deliver all of the animation, design, and programming to all of the other studios across many different time-zones, so that they would all be able to smoothly continue their own work with the latest output from Shanghai.

 

A specialist team of tools and automation programmers was established in Shanghai to develop a suite of automated tools, processes, and dedicated servers to help achieve this. The team’s objectives were to reduce the burden on manual human testers, provide high cost efficiency, detect problems within hours, and so allow the animation & design teams to rapidly iterate their work, helping to ensure the highest quality of animation and artificial intelligence behaviours.

 

This presentation will introduce some of the approaches used by the team, including:

Developing a suite of test programs to validate the technical compliance of the animations.

Designing small AI test maps to validate the motion and combat behaviour for each of the individual animals.

Automated video recording of interactions between animals, for the traditional QA team members to review when necessary without having to run the tests again.

Basic statistical data analyses of large numbers of AI interactions, to validate that the animals’ behaviours matched the designers’ intentions.

Summary email reports to provide management with a quick overview of the day-to-day issues, and help predict when trend of when all problems will finally be resolved.

 

Audience members will gain an insight into the challenges faced by the team, solutions implemented, and ideas for future development on one of Ubisoft’s best-selling games of all time.


View