Mathworks UAV Toolbox User’s Guide (MathWorks) (z-library.sk, 1lib.sk, z-lib.sk)

【One-Line Pitch】 A hands-on reference for engineers who already know MATLAB/Simulink and want to model, simulate, and visualize UAV behavior—from flight-log playback and waypoint tuning to Unreal Engine 3D scenes—without building a hardware rig first. Best for practitioners who learn by running examples rather than reading theory. 【Book Arc】 - **Opening (~0%–10%)**: Example-driven entry point. Flight-log import and playback (MAVLink TLOG, ULOG), flight-instrument gauge visualization, and custom plots establish the "read a log, see the flight" workflow before any simulation. - **Early (~10%–30%)**: Deepens log analysis—signal mapping tables, unit/frame conventions, extracting attitude/position/velocity/airspeed, and building custom signal mappings and FFT-style plots. Solves the "what do these fields mean and how do I trust them" problem. - **Middle (~30%–55%)**: Moves into control and model fidelity. Waypoint follower tuning for fixed-wing UAVs, heading control, and approximating a high-fidelity plant with the UAV Guidance Model block—including a low-fidelity-first tuning strategy and RRT motion planning. - **Late (~55%–75%)**: Transitions to 3D simulation with Unreal Engine: how the environment communicates with Simulink, block execution order, and the three coordinate systems (Earth-fixed, body, Unreal world). Solves the "why doesn't my model line up in the scene" problem. - **Ending (~75%–100%)**: Practical scene work—sensor selection, simple flight-and-sensor scenarios, depth/semantic-segmentation visualization, installing the scene-customization support package, editing scenes in Unreal Editor, and packaging custom scenes into executables. 【Book Arc note】The excerpts are heavily weighted toward the examples chapter and the 3D simulation guide; the arc above reflects that emphasis. 【Key Takeaways】 - **Logs are the front door to UAV analysis** (Opening): MAVLink and ULOG files are imported, mapped to named signals with units and reference frames, and replayed over UDP at a controlled rate—mirroring real hardware publishing. This is the fastest way to understand a flight after the fact. - **Signal mapping is where correctness lives** (Early): The `flightLogSignalMapping` tables expose signal names, fields, and units; mismatched units or frames silently corrupt gauge output, so inspecting `info` before extracting is essential, not optional. - **Custom analysis is a first-class workflow** (Early): You can register new signals via function handles (e.g., flagging accelerations above a threshold) and add custom plots, including FFT/periodogram views—turning the toolbox into a lightweight flight-data analysis environment. - **Fidelity is a dial, not a fixed cost** (Middle): The UAV Guidance Model block approximates a high-fidelity fixed-wing plant; tuning gains on the cheap model first (critically damped second-order PD, settling-time-based) then validating on the high-fidelity variant is the recommended loop—and it runs both directions. - **Waypoint following decomposes into understandable pieces** (Middle): A waypoint follower computes desired heading from pose, lookahead distance, and waypoint list; a proportional heading controller regulates roll under coordinated flight; visualization approximates pitch with flight-path angle in windless conditions. - **Planned paths and flown paths diverge for real reasons** (Middle): 3D Dubins paths assume instantaneous roll changes, while actual dynamics lag—so small tracking errors are expected, not bugs. Knowing this prevents over-tuning. - **3D simulation hinges on coordinate discipline** (Late): Earth-fixed, body, and Unreal world frames coexist; understanding communication with the simulation environment and block execution order is prerequisite to debugging any scene. - **Scenes are customizable end-to-end** (Ending): From installing the support package through reparenting actor blueprints in Unreal Editor, running the simulation, and packaging scenes into executables—plus applying semantic segmentation labels to custom scenes for perception work. 【Reading Tips】 - **Skim the examples chapter first, then return**: Read the opening examples for workflow shape, but treat the long signal-mapping tables as reference—look them up when a specific field or unit confuses you. - **Deep-read the fidelity and coordinate-system sections**: The low/high-fidelity tuning loop and the three-frame coordinate discussion are the conceptual load-bearing walls; everything else is procedure. - **Run, don't just read**: These are executable examples with sliders, `sim` calls, and build logs. The value is in observing behavior changes (lookahead distance, heading gain) rather than memorizing parameters. - **Watch for version and environment constraints**: Requirements, hardware minimums, and limitations for Unreal Engine simulation are stated explicitly—check them before investing time in scene customization. - **Take away the loop, not the numbers**: The transferable skill is "tune cheap, validate expensive" and "inspect units before plotting," not the specific gain values shown. 【Coverage Limits】 This guide is based on stratified excerpts concentrated in the examples chapter and the 3D simulation guide; the excerpts do not cover the full reference material, so block-by-block API details and any chapters beyond these sections are not summarized here.