New zombie model & AI implementation preparations

This week’s focus was on finishing the first working version of the new runtime navigation mesh baking as well as researching and planning the implementation of the new zombies. The new navigation mesh baking system is now fully implemented and the framerate is also very stable on my machine. There are a few problems to fix with the connection of two navmeshes but that is something I will be focusing on at a later time. For the new zombie implementation I did start by replacing the old low poly model with the new model(s) I purchased on the Unreal marketplace. The replacement went pretty smoothly. Since I would like to have a fairly large amount of zombies close to the player, the performance of zombies is crucial. Unfortunately the approach I took when implementing the zombies last year is not efficient enough to achieve this. Because of that I did quite a bit of research and planning on how to approach the new implementation of zombies and AI in general. I believe I have come up with a better approach that should in theory be significantly more efficient. I tested it with a very limited test setup which yielded promising results. It allowed me to have about a hundred or so zombies within 100m radius around the player. That amount should be enough to meet my goals for zombie density.

While doing my research I came across a Unity asset for character customization called UMA. I looked into that asset several years ago but didn’t consider it an option for the game at that time. After having spent 2 days of researching and testing the asset I would like to try and implement the zombie visuals using said asset. One major reason is that it also supports Unity’s addressables system, the system I built my modding system on. This support and the fact that it’s free (meaning modders can download that asset too) should enable modding capabilities for zombie visuals by allowing modders for example to provide zombie clothing via mods. If it works out the zombie variants would be more or less infinite thanks to the capabilities the asset provides. Varied zombie visuals are super important for my game because I want the player to be able to loot anything the zombie is wearing. For immersion and player decision making I want to have the clothing on the zombie representing the actual loot when the zombie is dead. Since the zombies are such a vital aspect of the game and require a whole lot of time to implement fully I will be doing the implementation in fairly small steps. This allows me to test the performance frequently and adjust quickly if necessary. Attempting the implementation with the asset I mentioned is not without risk but I believe that taking that risk is well worth it if that means having tons of customization options available if it does work out.

Notable tasks this week:

• Finished first implementation for runtime navigation mesh baking

• Implemented new zombie model

• Stress tested old zombie implementation which turned out to not be efficient enough

• :Researched and tested UMA asset, an asset providing massive character customization capabilities

• Researched and started planning new zombie implementing

Wishlist on Steam

Join the Discord

That’s all I got this week. I hope you are all doing well.

Stay safe out there and I’ll see you next week.