Artificial Intelligence in Unreal Engine 5 (Marco Secchi)（Z-Library）

【One-Line Pitch】 A hands-on guide to building game AI in Unreal Engine 5, taking you from basic concepts like nav meshes and behavior trees to advanced systems such as Mass and Smart Objects using both Blueprints and C++. Ideal for Unreal developers with some engine and C++ experience who want to make NPCs and enemies behave intelligently. 【Book Arc】 - **Opening (~0%–12%)**: Sets expectations and prerequisites — who the book is for, required hardware/software (UE 5.4+, Visual Studio or Rider), and the assumption of prior Unreal and C++ familiarity. - **Early (~12%–35%)**: Introduces AI in games broadly (pathfinding, rule-based systems, FSMs, behavior trees, machine learning) and maps the Unreal AI framework — Gameplay Framework, Navigation, Behavior Trees, Mass Entity, State Tree, Smart Objects, EQS, and AI Perception. - **Middle (~35%–55%)**: Deep dives into the Navigation System (nav mesh setup, modifiers, link proxies, optimization) and Behavior Trees (creation, configuration, extension), reinforced by practical project templates like the Unreal Agility Arena. - **Late (~55%–75%)**: Covers the Perception System and Environment Query System (EQS) — generators, contexts, tests — and how to wire them into behavior trees for responsive agents. - **Ending (~75%–100%)**: Explores advanced topics: hierarchical state machines with State Trees, data-oriented calculations with Mass, interactable elements via Smart Objects, plus an appendix on C++ in Unreal Engine. 【Key Takeaways】 - **AI in games is a toolbox, not a single technique** (Early): The book surveys pathfinding, rule-based systems, FSMs, behavior trees, machine learning, reinforcement learning, and generative AI, helping you pick the right approach for a given behavior. - **The Navigation System is the backbone of agent movement** (Middle): Nav mesh generation, modifiers, link proxies, and avoidance are covered through concrete project steps, not just theory. - **Behavior Trees are the primary authoring tool for AI logic** (Middle): You learn to create, configure, and extend them, including integrating perception and EQS queries. - **Perception and EQS make agents aware and adaptive** (Late): The Perception System handles stimuli and debugging, while EQS lets agents query the environment for smart decision-making. - **Advanced systems target scale and complexity** (Ending): State Trees offer hierarchical state machines, Mass enables data-oriented calculations for crowds, and Smart Objects support interactable world elements. - **Blueprints and C++ are used side by side** (throughout): The book assumes both, with C++ appearing in AI controllers and custom logic, and an appendix covering C++ fundamentals in Unreal. - **Practical gyms reinforce every concept** (throughout): Each chapter builds small test levels ("gyms") to validate navigation, perception, EQS, and state tree behavior before moving on. 【Reading Tips】 - **Skim the opening if you already know Unreal**: The first chapters cover prerequisites and AI taxonomy; experienced developers can jump to Part 2 for navigation. - **Deep-read the navigation and behavior tree chapters**: These are the foundation for everything later; the project templates are worth following step by step. - **Treat the gyms as mandatory practice**: The book's value comes from building and testing; skipping the hands-on parts will leave gaps in understanding. - **Use the appendix as a C++ refresher**: If your C++ is rusty, review it before tackling chapters that mix Blueprints with C++ AI controllers. - **Check UE version compatibility**: The book targets UE 5.4+, but later versions should work; watch for minor API changes in advanced systems like Mass and State Trees. 【Coverage Limits】 This guide is based on stratified excerpts covering the table of contents, preface, and early chapters; detailed content of later chapters (e.g., specific Mass processors, Smart Object definitions) is not fully represented. The excerpts do not cover every code example or project detail.