SEAM Studio
Guide 05 · SEAM Studio docs

Trajectories and UAVs: moving receivers, playback, and POV views

A static TX→RX solve tells you one snapshot; a trajectory tells you how the link evolves as a receiver flies, drives, or walks through the scene. This guide covers the two ways to move a UE, UAV actors with attached antennas, the exact semantics of trajectory playback in the viewer, and the per-entity POV inset. Everything here works with the Mock backend alone — no GPU or Sionna RT installation required.

1. Two ways to move a UE

ApproachWhereWhat moves
Device trajectoryResults mode → UE trajectory panelAn existing RX device is swept along a start→end line or drawn waypoints; one solve per step.
Actor flight pathSelect an actor → inspector TRAJECTORY sectionThe actor (e.g. a UAV) follows its own authored waypoints; an attached RX rides along and flies the path.

Both feed the same trajectory engine: the result is a time-series of samples (RSS / path gain / SINR / RMS delay spread / path count per step) plus per-step live rays you can play back in the viewer.

2. Device trajectory — the UE trajectory panel

  1. Switch to Results mode and find the UE trajectory panel (dockable, like every panel — move it with ◧/◨/⧉).
  2. Define the route. Three options:
    • ⌖ Pick start → end in viewport — click two points in the 3D view: first the start, then the end (Esc cancels). A dashed yellow preview line shows the segment. Start at RX seeds the Start / End fields from the current first RX position instead.
    • Draw route (Esc finishes) — click waypoints one by one in the 3D view; Esc finishes the route (at least 2 points). With more than one RX in the scene, a selector chooses which RX the drawn route belongs to, so you can author one route per UE and they all move together.
    • ⤓ Import JSON — load waypoints from a JSON file. Cartesian x/y/z or geographic lat/lon points are auto-detected (geographic needs the scene's geodetic anchor), and underground points are flagged with a ⚠ warning on the route row. Files with per-waypoint antenna orientations show an oriented chip. Format details: ../point_import.md.
  3. (Optional, outdoor) Check Follow terrain to drape each waypoint onto the surface below it plus the UE height offset — keeps the UE at a constant height above sloped ground. Leave it off indoors.
  4. Set Num points (2–200 solve steps; every route is resampled to this step count) and dt (seconds per step).
  5. Press Simulate trajectory. Each step is a full TX→RX solve at the moved position; with multiple TX the per-step SINR and interference appear too. With routes for several UEs, Include fixed UEs also solves every un-routed RX at its fixed position each step.

The result strip shows the UE id, a moving UE chip, the sample count and backend, and a ✕ Remove button that clears the marker/trail/ray overlay immediately.

3. UAV actors

Create and pose

In the scene tree's ACTORS row, press +UAV (next to +Car, +Human, +Custom). The UAV spawns hovering above the scene center. Select it and the inspector shows Actor · <id> with a POSE & SIZE section: Name, X / Y / Z (m), Yaw (°), Length / Width / Height (m), then Apply. Positions are Z-up ENU meters (X east · Y north · Z up); yaw rotates about +Z. You can also drag the actor with the viewport gizmo.

The UAV renders as a procedural quadrotor (fuselage, four rotor arms, skids, gimbal camera). In large outdoor scenes the visual model is scaled up to device-marker visibility — a real 0.6 m drone would be sub-pixel at campus scale — while the RF projection keeps the true size_m box.

UAV actor selected: translation gizmo, yellow waypoint line over the campus, POV inset uav_001 → TX tx_001, and the inspector with POSE & SIZE, ATTACHED DEVICES, and TRAJECTORY waypoint rows
A selected uav_001: gizmo + yellow flight path in the viewport, the live POV inset (uav_001 → TX tx_001), and the actor inspector with + Attach RX here / + Attach TX here and the waypoint table.

Attached devices — how an actor becomes an RF endpoint

Actors are scatterers; rays terminate at antennas. To make a UAV an RF endpoint, open ATTACHED DEVICES and press + Attach RX here (or + Attach TX here for a UAV relay/base station). This creates the antenna on top of the actor and checks it in the attached-device list. Attached devices ride along when the actor moves, so paths always solve from the actor's current position.

Flight path — the TRAJECTORY section

Check Define waypoints in the TRAJECTORY section to give the actor a flight path. Each waypoint row has:

Below the table: + Waypoint appends the actor's position, ⌖ Pick waypoint appends a clicked point from the viewport, and Record current pos captures the pose you just dragged to. dt (s), Speed (m/s) and Mode (once / loop / pingpong) control the animation, which plays in Results → Scenario playback (Simulate scenario animates every actor along its own waypoints).

Pacing: dt or Speed. By default the path is paced by dt: every segment takes dt seconds, however long it is. Waypoints placed far apart then mean a very fast actor. A drone with waypoints 50 m apart at dt = 0.25 s flies at 200 m/s, and its Doppler follows that speed. Below the Speed field the inspector shows the speed this implies (implied N m/s from dt, plus the peak when legs differ; amber when the peak is over 3× the kind's typical speed). Enter a Speed to move at that constant pace instead: each segment then takes its length / speed, so unequal legs take proportionally longer. The placeholder is a typical speed for the kind (car 10, human 1.4, UAV 10, custom 5 m/s). It is only a suggestion, and the field stays blank (dt pacing) until you type a value. While a speed is set, dt is dimmed and unused but kept, and clearing the speed returns to dt pacing. The same pacing drives scenario frames, the sensing target's t = 0 velocity, datasets that sample the actor (actor_id), the AODT scatterer route, and ⚡ Simulate paths along trajectory. Stored as trajectory.speed_m_s (see ../scene_format.md).

A trajectory also gives the actor a velocity for Doppler (Sionna backend). Simulate scenario applies each frame's velocity. Every other paths solve (Simulate paths, channel analysis, datasets, UE trajectories including ⚡ Simulate paths along trajectory, GT playback) leaves the actor at its authored pose but moves it at its trajectory velocity at t = 0. A path that reflects off the actor therefore carries its Doppler even when the TX and RX are static, and a device attached to the actor moves with it. The exception is a dataset that samples the actor's own path: that actor stays at rest (see §4 of sensing.md).

That velocity is the trajectory tangent × speed. At an exact waypoint time it is the outgoing leg's velocity, at the end of a once trajectory it is 0 (the actor stops), and at a pingpong turnaround it is the reversed leg. Before v0.1.14 those frames reported the average of the two legs (7.07 m/s at a 90° turn at 10 m/s) or half the speed, so Doppler and velocity_true at waypoint frames changed in v0.1.14.

⚡ Simulate paths along trajectory

The ⚡ Simulate paths along trajectory button (enabled with ≥ 2 waypoints and an attached RX) sends the actor's waypoints straight to the trajectory engine: one TX→RX solve at each step along the flight path, with per-step rays. The attached RX is what flies the path — without one the button's tooltip tells you to attach an RX first. Results land in the UE trajectory panel exactly like a device trajectory.

4. Trajectory playback — exact semantics

A trajectory result adds a playback transport under the UE trajectory panel: ▶ / ⏸, a frame slider with a frame/total counter, ⟳ repeat, and a speed select (0.5×–4×). Multi-UE runs add per-UE scrub bars and a KPI UE selector. The viewer behavior is deliberately strict:

Playback engaged mid-flight: magenta per-frame rays converging on the moving rx_001 marker while the static TX and scene stay intact
Playback engaged mid-flight — the per-frame rays converge on the moving rx_001 marker; the static TX/RX markers stay exactly at their scene positions.

5. The entity POV inset

Click any TX, RX, or actor and a live inset opens in the top-right of the viewport: it renders that entity's point of view toward its link partner. The header reads e.g. TX tx_001 → with a partner selector (a TX looks at the first RX by default; everything else looks at the first TX). During trajectory or scenario playback the inset tracks the live rendered pose. The source's own airframe is hidden in the POV pass, so a UAV's camera is not blocked by its own rotors. The camera button saves the POV view as a full-resolution PNG; × closes the inset (it reopens on the next selection).

TX selected: POV inset TX tx_001 → RX rx_001 with the green ray toward the RX, and the device inspector showing ANTENNA ORIENTATION with Yaw/Pitch/Roll, Look at…, and Aim
With tx_001 selected, the inset shows TX tx_001 → RX rx_001 along the green LOS ray. The inspector's ANTENNA ORIENTATION section (Yaw / Pitch / Roll, Look at…, Aim) points the boresight at another device.

Related: the device inspector's ANTENNA ORIENTATION section sets Yaw (°) / Pitch (°) / Roll (°) directly, or pick a target under Look at… and press Aim to point the boresight at it — handy for aiming a directional TX at the UAV you just flew.

6. Where the numbers go

The playback strip shows per-frame KPIs (t, position, RSS, Path gain, SINR/SNR, RMS delay, Paths). For the full time-series, open the Metrics dashboard panel: once a trajectory result exists it adds paper-style charts — Trajectory: power vs time (RSS/SINR, derived RSRP), RMS delay spread vs time, path count vs time, plus Doppler and serving-cell/handover views when available — each with PNG/SVG/CSV export. Trajectory samples also feed ML datasets (trajectory sampling mode, see ../ml_datasets.md) and the RFData export (trajectory.csv).

PreviousSimulating: paths, radio maps, beamforming, and channel analysis NextML datasets and exports