top of page
Search

#3 Tour.Go: Monuments and AR Interactions

  • Oliver
  • Feb 1, 2024
  • 4 min read

Welcome to another blog discussing progress for Tour.Go!


In the last few weeks, I have been focusing on AR, and more specifically, object manipulation. I have also started testing the app, along with making improvements for an easier experience.


AR Object Manipulation

One of the most significant aspects when developing AR experiences is how the user can control the environment. This point introduces many challenges that have to be considered. For example, how are objects moved in the 3D space, what are the optimal actions the user can perform to scale or rotate an object? Manipulating objects is a core component in this app, due to the complications presented when interacting with monuments in the real world. The user could be a distance away from a large scale monument, or a certain target could be inaccessible. As such, I have made ray cast interactions with no limit distance. Although, this may represent problems later with a host of monuments, but this can be refined and tested.



The video above shows how objects can be 'grabbed', moved, rotated and scaled. The user can select their desired tool, and then move their finger across the screen both in a horizontal and vertical direction depending on the tool. The move tool uses a vertical delta input to push the object away, and draw the object in. Moreover, the rotation tool can be used by sliding horizontally across the screen for orientation in both directions on the world Y axis. Finally, the scale tool requires up and down input, scaling the object respectively.


Blender Add-on for Landmark 3D Assets

One of the challenges for creating this app involves viewing monuments in 3D. I had planned to find an asset collection of monuments that might be available, but these were either too expensive or low quality. As such, I found a good solution using real world data.



Using a blender addon, any landmark with defining features and that is not too complex in structure can be imported into Unity. I am still configuring the models and defining the pipeline to ensure a consistent 3D viewing experience, but using this tool is a significant step in making one of the most engaging features in the app a reality.


Image Via Web Request

Unity has extensive functionality for web requests, and this was essential for a variety of images  for each monument. Using the landmark’s Wikipedia image, in the public domain, it was possible to download the image and display live during the app. The main reason for this approach was so that the app did not need to be manually updated for new content. Using the Firebase real-time database module, it is possible to edit the JSON file and add a new entry with a URL stored. The same could be applied for 3D assets, as some monuments will have a model that the user can inspect using the object manipulation tools. Accessing 3D models from a URL might be a consideration in the future.



Nearby Monument Menu

One problem I encountered when testing the app, was that the waypoint was too small and difficult to find. I had tested the app in a room-sized environment, and this had differences in the outside world. For a solution, I created a menu that can display the nearby monuments in the area, and gives indication for the distance between the player and each waypoint. Moving away from a waypoint will result in the distance value increasing, giving good indication to the user. 



The next step as part of this menu, is creating a way for the user to ‘lock on’ to a monument, as such, it will allow for the distance value to be locked to the screen for a better viewing experience. The user will be able to disable the current targeted monument by pressing the menu button once again.


One point that I have found during testing, is that is can sometimes be difficult to find the monument if you are too close or too far away, as the user is required to tap on the screen. As mentioned in a prior blog, a ray is cast to detect if the user taps onto a waypoint object. Creating a pop-up on the HUD (heads-up-display) might be a better solution. I will investigate further when refining the app over the next few weeks.


The nearby monument menu was useful for testing, as it was clear the distance between each monument was incorrect. I found that the calculation for positioning each waypoint to its corresponding geo-location was too inaccurate. The Unity world coordinates measure by 1 unit equals 1 real world metre. However, this assumption does not mean an easy implementation for high accuracy. This is due to the current implementation treating the earth as a sphere, as it is not and is instead an ellipsoid. Implementing a new calculation should solve the issue. However, the core functionality is working appropriately.


Conclusion

I have successfully produced an effective MVP (minimal viable product), as the current version encompasses all the central functionality for the app. Over the next few weeks I will be focusing on gamification elements. More specifically, in the next stage, I will focus on the badge collection system, saving user data and creating the archive for discoveries. I will also focus on improving location accuracy, including how the user can more easily interact with real-world monuments.


I am currently conducting a feasibility survey to gain opinions that will influence the alpha version of the app. Any entries are appreciated: https://forms.gle/E45SVV15oaqhVbw88


References




bottom of page