The Evac Test

With Australia having its worst bushfire season to date last summer (2019), the dangers of climate change are becoming more real.

Project overview

To create a fun yet informative experience that placed users in a real-life situation to better equip them for a bushfire evacuation.

Tools Used

Team

My Role

Timeline

Blender

Unity

Procreate

Adobe After Effects

Adobe Premiere

Christian Cerda

Alannah Frankel

User researcher

Interface designer

CAD modeler

Lead animator

C# programmer

8 weeks

The project began by researching into aspects of climate change that my team would find interest in and due to the huge impact of the Black Summer fires that occurred in early 2019, we chose to focus on bushfire preparation and evacuation in the hopes to have an impact on future bushfire disasters that may happen.

We began by conducting background research into bushfires, by looking into current solutions that were related to emergency preparation and prevention as well as teaching people about bushfires.

During our research, we discovered a lack of approachable information for anyone over the age of 12, as most documents were lengthy documents.

We then researched the current techniques that Australia uses in order to fight the fires. These include:

Volunteers
Water and fire retardant
Earth boundaries
Controlled burns

However, despite these techniques, changes in Australia's climate mean being aware of and prepared for bushfires is crucial now more than ever.During the Black Summer, Australia’s worst bushfire season on record, 2,439 homes were lost.

It then became imperative that families weren't informed and knowledgeable about proper evacuation procedures.
This became the basis for our virtual reality prototype app.

We then chose to develop our app into the form of a game so that users could implicitly learn about evacuation procedures, and base it around the emergency kit that the County Fire Authority (CFA) recommends for one's bushfire plan.

In order to flesh out our idea, semi-structured interviews were then conducted in order to understand peoples' preparation and understanding of the evacuation process and what to include in an emergency kit.

Despite their value however, our user research showed most people did not have an evacuation pack ready or know what items to include if facing a bushfire threat.

This led to the virtual reality game concept becoming a scavenger-based game that puts players in a situation where they need to evacuate.

Users must race against a clock to move around their homes and grab everything they deem necessary. With total freedom of the items they can choose, the users’ raw knowledge will be tested.

The game will place the user in the shoes of someone who needs to evacuate. The home users will find themselves in will contain common household objects to build upon this idea.

Through the use of interviews, non-essential objects people would like to bring were identified and will be placed throughout the house. This will force users to think about their items in real life and if they would take them.

Before beginning to iterate, we identified the constraints in which we would have to design for, as we were limited to using a Google Cardboard as our Virtual Reality device due to its affordability, and therefore its potential to reach more people without them needing to purchase expensive VR headsets.

Therefore our game would have to function solely on either gaze used to control the player or a touch interaction for the screen.

We then began to iterate upon our concept by creating storyboards that would work as a rough outline of the game's stages and giving us a base to work from.
The storyboards would also assist us in mapping out each room of the house and where each item would be located.

When analysing the storyboards, this led us to consider an addition of features.
These included:

How users will travel between rooms.
A possibility we chose was a dwelling gaze in which users would have to look directly at a door for a limited time to be teleported to the next area. However, due to users having a limited time to explore, this would have cut precious seconds that could be necessary for users when playing.

Therefore we simply went with touch interaction when users were focused on the door.

A scoring system

This was so users would have a goal to seek after while playing and adding making the game enjoyable for them.

Points would only be rewarded if users would collect an item listed in the essential kit.

User positioning

Placing users in the centre of rooms in view of each item that could be collected, thus removing the need for them to move around the 3D space of the room.

To better understand the user and identify possible problems that they may experience, a persona-based walkthrough, user journey map, and bodystorming was done, which led to us further fixing issues that could be faced by players.

persona-1_orig

persona_orig

persona-1_orig

1/2

The main improvements include:

The home screen welcome sign appearing too similar to a button despite being unclickable, thus breaking the principle of affordance.

The addition of hover effects so that users could identify buttons that can be clicked.

Adding a "How to Play" tab so that players can learn and prepare before jumping into the game.

We then created final mock-ups for interfaces that would be required and what they should include, as well as conducted A/B testing to determine how to structure the gameplay around collecting objects within the game.

The debate was around adding a glow around objects that are interactable, as light is a good way to guide the user towards interactive objects.
Despite this, we did not want to take away from the freedom that users had and felt that a glow would act more like a direction for users rather than a suggestion.

To test this, we created two mock-up images (one with and one without a glow around interactable objects) and tested 7 participants. They were given the direction “You are in a VR game about bushfires and you need to evacuate. You are told to click the items you want to grab before leaving. This is the screen you see, which items do you click?”.

4 participants were given the image with the glow and 3 were not, in order to find if those without could identify the interactable objects without direction, or if they would attempt to grab unclickable objects.

Our findings showed that all users who were shown the first test image (without the glow) identified the clickable items, however, 1 user thought the television was obtainable. They also did not always choose every object, often choosing just 2 or 3 items. This showed they were exercising their freedom of choice.

Test participants who were shown test image 2 (with the glow) showed a variety of different answers. Some immediately asked if they were only allowed to select from the glowing objects, some said they would grab everything glowing, some asked if the essential objects were the glowing ones.

One user asked if they could grab the television, even though it wasn’t glowing, pushing against our theory that light would guide them.

Another user was confused with the task and did not know they had any freedom. They stated they would grab everything glowing because that is what the game is telling them.

Our findings led us to conclude that glowing objects were not an accurate way to represent interactable objects and give users freedom of choice.

One user asked if they could grab the television, even though it wasn’t glowing, pushing against our theory that light would guide them.

Another user was confused with the task and did not know they had any freedom. They stated they would grab everything glowing because that is what the game is telling them.

Our findings led us to conclude that glowing objects were not an accurate way to represent interactable objects and give users freedom of choice.

With tutor think-alouds, interviews, storyboards, personas, bodystorming, a persona-based walkthrough, and A/B testing we identified 5 key findings:

1. The design must give users freedom.
2. The design needs to resemble aspects of the users’ real life.
3. The design must be easy to understand and use.
4. The design must be immersive.
5. The design must be enjoyable.

Once we had identified our key findings and conducted our testing, we began to develop our game using Blender to create our models for our game, and Unity to control the environment and construct the game. Glitch (an online browser used to develop web apps would be used to host our game so it could be accessed online).

We began by constructing a list of all objects that would be interactable within the game, and dividing them amongst our team, using a low poly style to make the environment appear friendlier due to the possible confronting nature of the situation the game is based on.

In total, 20+ interactable models were made, along with numerous models being made to build the house environment.

Once the models were completed, they were imported into Unity. Here we then set up how the game would operate.

My roles here during the development of the game included learning C# programming in order to control the changing of levels, changing between panels on the main and end menus, controlling how buttons interacted and control of the sound.

In order to improve the atmosphere, we added sounds and manipulated the soundtrack which would speed up as the player's time would run out, thus creating a sense of urgency.

Once all the models had been made and placed, all the coding and interactions were programmed and music was added, we exported our game and imported it into Glitch so that we could test on other desktops.

We also created a promo video (can be found in the overview), highlighting the significance of the issue and why our game was an appropriate solution to this problem.

Follow this link to try our game out yourself.
https://a3-maincam.glitch.me/BushfireGameV4MainCam/

There are numerous aspects we would have liked to add with more time. The first features we wanted to add were to give the user more control.

These included the ability to put items back after adding them to the inventory and an options/settings menu to allow users to turn off background music and game sounds.

Visibility was another principle that could have been improved with the use of labels in the inventory, an error message for uncollectible items and labels on the doors.

Enforcing the educational element of the game was another section we would have loved to improve. This would have been achieved with the use of an end panel that lists which items the user missed and a link to the Rural Fire Service (RFS) so that individuals could be prompted to find out more.