Skip to content

Functionality

Lift & Drone ingests ArduPilot Dataflash .bin logs, runs a single processing pipeline, and presents results in a Streamlit UI (app.py). A legacy Plotly Dash app (drone_dashboard.py) reuses the same plot builders for an all-in-one layout and optional HTML export.

End-to-end pipeline

When you upload a log, the app parses the binary file and runs:

  1. parse_log — extract typed message streams (GPS, IMU, BAT, etc.).
  2. run_ekf_on_log — optional EKF fusion for derived motion (when the pipeline produces data).
  3. compute_metrics — mission metrics (distance, speeds, altitude, battery, warnings, etc.).

Orchestration is centralized in service/orchestrator.py; the Streamlit loader calls process_log_file() so views do not import parser/metrics modules directly.

Log parsing

  • Format: Dataflash binary via pymavlink.
  • Message types used across analytics and UI include: ATT, GPS, BAT, VIBE, RCOU, RCIN, IMU, MODE, ERR, MSG, EV, BARO (and related fields inside those streams).
  • Time: Microsecond timestamps are normalized to seconds relative to mission start for plotting and metrics.

Metrics and warnings

Mission metrics (distance, duration, GPS speeds, IMU acceleration and integrated speed, altitude bands, battery current and energy, EKF peaks when available, gyro-extremes notice, etc.) are computed in service/metrics/metrics.py. The Mission Summary stat cards surface the main values; an expander lists metric notices (distance, altitude gain, battery, gyro) when warnings exist.

For a field-level description of how distance, speeds, IMU, and battery numbers are defined, see the Analytics & Metrics section in the project README.

Streamlit application (app.py)

  • Upload one or two .bin files; with two logs, radio buttons choose which file drives the dashboard.
  • Color trajectory by: speed, altitude, or elapsed time (affects 2D map and 3D trajectory coloring).
  • Quick stats: duration, distance, max speed, energy (when available).
  • Firmware line when a version string is found in MSG records.

Mission Summary

A grid of stat cards (duration, distance, speed breakdown, altitude ASL and above takeoff, max acceleration, energy, average satellites, etc.) with optional per-metric warnings.

Real-world GPS trajectory (2D)

Interactive Plotly map with OpenStreetMap tiles, trajectory coloring from sidebar settings, takeoff/landing context, and mission bounds details: reverse-geocoded takeoff/landing hints (Nominatim), coordinates, and valid GPS points vs total with average HDOP/satellites for filtered points.

AI Flight Analysis

Google Gemini post-flight text based on the metrics dict only (not the raw log): Short / Detailed reports or a Custom question, with results cached per log and mode. Configure GEMINI_API_KEY in .streamlit/secrets.toml, or paste a key in the sidebar expander for the current session.

For model id, prompts, what data is sent to the API, and privacy notes, see AI flight analysis (Gemini).

Telemetry panels (toolbar)

Switchable Plotly views (pills or radio, depending on Streamlit version):

Panel Content
Battery Voltage and current over time
Vibration & Motors Vibration axes and motor PWM outputs
Attitude Actual vs desired attitude
Events Mode changes, errors, and events timeline
3D Trajectory Local E–N–U path with coloring tied to the same speed / altitude / time choice as the map

Panel builders live in drone_dashboard.py and are imported by views/telemetry.py.

Legacy Dash dashboard (drone_dashboard.py)

  • CLI: python drone_dashboard.py path/to/log.bin — serves a multi-panel layout (typically port 8050).
  • Static export: --export report.html writes a standalone HTML file.
  • Includes a GPS health style panel in the classic layout that is not wired into the Streamlit toolbar (Streamlit focuses the map + filtered GPS stats instead).

Theming and assets

The Streamlit app uses project CSS under ui/ and .streamlit/config.toml for a dark, dashboard-style appearance.