Project Orvus
Technologies Used
Game Info
Gallery
Orvus Overview Video
Orvus is a vertical slice PC shooter demo built by a solo developer for Game Studio in Unreal Engine 5.8, utilizing a hybrid C++ and Blueprint architecture to balance performance with rapid iteration. The project is built around Motion Matching to achieve incredibly fluid character movement, cover transitions, and shooter mechanics.
A major focus of Orvus is its Advanced AI & Cover System. The AI utilizes Behavior Trees (BT) and EQS to dynamically assess the battlefield. Instead of standing in the open, AI actively seek out specific crouch or stand cover points. They use a custom peek-and-fire mode with Line of Sight overrides to pop out, suppress the player, and return to safety. Target prioritization relies on a Threat Matrix differentiating between the Player and AI Companions, while a Token-based "Rage Mode" allows enemies to aggressively push.
The Combat & Damage System is fully data-driven. Weapons are configured entirely through Data Assets (combining animations, SFX, VFX, and impact data), making scaling the arsenal effortless. The damage system features localized hit-zones (headshot multipliers), explosive damage scaling, and "fair-play" player damage reduction. Combat feel is enhanced by robust horizontal and vertical Aim Assist, and a dual-reaction hit system combining animation impacts with physical impulses, culminating in full ragdoll physics on death.
The game also features a custom Stage/Mission Controller to manage enemy wave spawning and win/lose conditions, alongside a diegetic 3D Projected UI for seamless weapon pickup and drop interactions.
Gameplay
Core Mechanics
Game Modes
Difficulty
Key Features
Fluid character movement and shooter mechanics driven entirely by Motion Matching
Advanced AI targeting via Threat Matrix handling Player vs. Companion prioritization
Dynamic AI Combat Radii that shrinks and expands based on the number of alive enemies in the encounter
Token-based "Rage Mode" allowing a limited number of enemies to aggressively push the player's cover
Intelligent Cover System supporting crouch/stand cover variations and peek-and-fire modes
Data-driven combat framework using Weapon Data Assets for instant configuration of VFX/SFX and logic
Advanced Damage System with headshot multipliers, explosive damage, and player-advantaged scaling
Dual-layer Hit Reactions combining physical impact impulses with animation-based flinches and ragdolling
Diegetic 3D projected UI for seamless weapon interaction and inventory management
Custom Stage Controller for orchestrating enemy spawns, missions, and win/loss states
Robust Aim Assist system providing both horizontal and vertical tracking
Architectural Implementation
Data-Driven Combat Framework
Weapons and projectiles are abstracted away from hardcoded logic. Designers create a WeaponData asset linking the mesh, fire rates, Niagara VFX, Cascade impacts, and SFX. Attaching this data to the base weapon class instantly yields a fully functional, replicated weapon. Aim assist mathematics handle both horizontal friction and vertical magnetism.
Advanced Cover & AI Logic
The CoverComponent evaluates the geometry to classify cover as crouch or stand capable. When an AI snaps to a cover point, C++ overrides standard LineOfSight checks, allowing the Behavior Tree to execute smooth peek-and-fire states. Aggro logic uses event-driven delegates to lock Threat Matrix targets without heavy O(N) polling.
Physics-Integrated Damage System
Damage routing calculates bone-specific hit results to apply precise headshot multipliers, while simultaneously reducing damage taken by the player for a "fair advantage" feel. On impact, characters receive a physical impulse blended with an animation hit-react, gracefully transitioning into full physics ragdoll upon death.
Mission Control & UI Projection
A centralized Mission Controller dictates level pacing, managing triggered enemy spawns and evaluating score-based win/loss states. World interaction utilizes a 3D projected widget system, allowing the player to seamlessly pick up and drop weapons via UI elements rendered in world space.
Performance Benchmarks
Motion Matching queries evaluate efficiently without degrading frame time during heavy combat
AI Threat Matrix and Cover EQS queries process without frame drops during large firefights
Data Assets load weapon configurations asynchronously, keeping memory footprint low