Boids FishTank: Predator–Prey Flocking Simulation (Orca vs Tuna)

Overview

Boids FishTank: Predator–Prey Flocking Simulation (Orca vs Tuna) is a real-time aquatic ecosystem built in Unity HDRP. It extends the classic Boids flocking algorithm into a predator–prey scenario where tuna school together and orca pods coordinate hunts using role-based behaviors.

The goal of the project is to explore how far flocking models can be pushed toward more cinematic, natural-looking, and fun-to-watch behavior—almost like a digital fish tank powered by AI.

---

Features

• Prey Agents (Tuna) – Classic Boids flocking (separation, alignment, cohesion) with obstacle and predator avoidance.
• Predator Agents (Orcas) – Extended, role-based behaviors (leader / flanker / striker / support) with targeting, pursuit, interception, flanking and attack.
• Predator–Prey Interaction – Tuna schools tighten, scatter and reform in response to orca pressure and strikes.
• Obstacle & Boundary Avoidance – Both species steer around tank walls and terrain using lightweight spherecast-based avoidance.
• Real-Time UI Tuning – In-game panels let you adjust flocking weights, speeds, radii and predator parameters while the simulation is running.
• HDRP Environment – Built in Unity HDRP using the Water/Ocean and Terrain systems for reflections, depth and atmosphere.
• Cinemachine Cameras – Free-roam camera plus follow/overview rigs to watch the school or focus on individual orcas.
• Underwater / Above-Water Audio Blend – Audio mix shifts smoothly as the camera moves below or above the water surface.

---

Controls

• Camera Move: W A S D
• Camera Look: Mouse
• Zoom / FOV: Mouse scroll
• Boids (Tuna) Parameters: F1 – open/close tuning panel
• Orca Pod Parameters: F2 – open/close predator tuning panel
• Audio Mute / Volume Toggle: F3 (and additional audio controls in the small on-screen panel)

Note: This is an interactive simulation / sandbox, not a traditional game with win/lose conditions. The fun is in watching the behaviors emerge as you tweak the parameters.

---

Performance Notes

This project focuses on behavior and visual quality rather than extreme agent counts. The simulation uses Unity Jobs + Burst and a spatial hash grid to accelerate neighbor queries, but performance can still drop at very high fish counts, especially with HDRP water, terrain, and post-processing enabled. If your frame rate is low, try reducing the number of boids in the UI panels.

---

Planned / Possible Future Extensions

• More species and behavior variations (e.g., different fish types and predator styles).
• More customizable “behavior presets” similar to a particle system for easy authoring.
• Additional environment types and lighting setups.
• Further optimization for higher agent counts.

Thanks for checking out the project! Feedback on flocking behavior, predator–prey dynamics, or performance is very welcome.

---

Credits

3D models used in this project are licensed under the Sketchfab Standard / CC Attribution licenses. Proper attribution to the original creators is provided below:

• Lowpoly Coral Pack by assetfactory – https://sketchfab.com/3d-models/lowpoly-coral-pack-29d5be0e220f48818346cabfa065e887
• Seaweed by OwenCalingasan – https://sketchfab.com/3d-models/seaweed-dec5b256a37f40acb63b0bf30b45d756
• Soft Coral Set by Kanna-Nakajima – https://sketchfab.com/3d-models/soft-coral-set-256355f15fcb4095af17b75ae572bff0
• Coral by Sandra_s – https://sketchfab.com/3d-models/coral-c26e47859f0945d69a4e2944ee80b995
• Tuna Fish by GoldenZtuff – https://sketchfab.com/3d-models/tuna-fish-642c6515d893474a8fc8f129491efcac

Assets are used under their respective Sketchfab Standard / CC Attribution licenses. No redistribution of original source files is included.