type
status
date
slug
summary
tags
category
icon
password
这里写文章的前言:
一个简单的开头,简述这篇文章讨论的问题、目标、人物、背景是什么?并简述你给出的答案。
可以说说你的故事:阻碍、努力、结果成果,意外与转折。
- The updated PPT;
 
- Three video clips recording of the current experiment result.
 
- The issues and challenges to be discussed.
 
Here’s a slide‑by‑slide outline you can drop into your PPT to walk your boss through this week’s code work. Feel free to adapt or reorder as you like:
Slide 1: Title & Context
- Project: Tello Autonomous Orbit & Precision Landing
 
- Date: Week of July 29, 2025
 
- Presenter: Qiyue Chen (Embedded AI Intern, CTI One)
 
Slide 2: Objectives
- Build a reliable orbiting routine around a mission pad
 
- Layered precision‑landing capability down to 10 cm
 
- Comprehensive flight‑data logging for post‑flight analysis
 
- Real‑time visual overlay of telemetry
 
Slide 3: High‑Level Architecture
main.pyorchestrates mission phases
drone_interface.pyabstracts Tello SDK calls & state reader
pad_tracker.pyhandles AprilTag pad detection & centering
pid_controller.pycomputes velocity commands from errors
precision_landing.pydrives the multi‑layer descent logic
logger.pycaptures timestamped CSV logs
video_overlay.pyrenders pad/metrics overlay on live feed
Slide 4: PID‑Based Orbit Control
- PIDController class with tunable Kp/Ki/Kd parameters
 
- In 
orbit_phase(), read pad offset → feed into PID → issue velocity commands 
- Dynamically adjust threshold based on pad loss timeout (configurable)
 
- Benefit: smooth, stable circling even with intermittent pad visibility
 
Slide 5: Multi‑Layer Precision Landing
- Layer definitions from 
config.LAND_ALTITUDES(e.g. 120 cm → 50 cm → 20 cm) 
- Alignment loop at each layer:
 - Run up to 
LAND_ALIGNMENT_ITERATIONS, check centering error ≤LAND_CONV_RADIUS - Fine‑tune X/Y before descending
 
- Descent control:
 - Fixed step 
LAND_DESCENT_DISTANCEper iteration - Delay 
LAND_DESCENT_DELAYfor stability 
- Benefit: avoids drift, ensures pad‑centered touchdown
 
Slide 6: Flight Data Logging
FlightLoggerclass writes CSV with fields:
time, padx, pady, error_x, error_y, vx, vy, vz, h, tof, camstart()/log()/log_step()/close()API
- Logs on every control loop iteration
 
- Benefit: detailed post‑flight charts, error‑trend analysis, regression testing
 
Slide 7: Video Overlay & Telemetry
video_overlay.pyruns in its own thread
- Draws pad bounding box, centering error, current altitude & velocities
 
- Provides live feedback to operator during test flights
 
- Benefit: immediate visual confirmation of control logic
 
Slide 8: Robust Error Handling & Cleanup
- Exception traps around orbit & landing phases
 
emergency_land()on any unhandled error
- Always call 
logger.close(),stop_state_reader(),cleanup_sockets() 
- Ensures no dangling threads or lost logs
 
Slide 9: Key Challenges & Solutions
- NameErrors in logger/precision code → unified single 
FlightLoggerinstance passed through modules 
- Method mismatches (e.g. missing 
start_time, undefined variables) → aligned function signatures 
- Pad detection noise → dynamic thresholds and retry loops
 
Slide 10: Next Steps
- PID tuning with live flight data (adjust Kp/Ki/Kd per altitude layer)
 
- Simulation harness to automate orbit & landing tests
 
- Sensor fusion (optical‑flow + ToF) for wind‑disturbance compensation
 
- Dashboard for post‑flight CSV charting & performance KPIs
 
Slide 11: Repo & Resources
- Config docs: covers all 
config.pyparameters 
- Test logs: under 
log/flight_log_*.csv 
Slide 12: Questions & Feedback
- Open floor for any clarifications or feature requests
 
Feel free to copy these bullets directly into your slides. Good luck with your presentation!
- Start Mission
 
- Initialize & Takeoff
 
- Start Video Streaming Thread
 - [Side branch/annotation]
 
OpenCV real-time overlay
YOLACT Edge (AI detection) integration — In Progress
- Move to Orbit Start Position
 
- Begin Orbit (PID Control)
 - (loop until orbit complete)
 
- Start Layered Precision Landing
 
- For each layer:
 - Descend to Target Altitude
 - Is Pad Detected?
 - Yes → Begin Alignment & Landing
 - No → Auto-Search Pattern
 - Is Pad Detected?
 - Yes → Begin Alignment & Landing
 - No → Manual Keyboard Control
 - Descend to Next Layer (repeat)
 
(WASD/arrows, system monitors for pad, press ‘L’ to resume)
- Land
 

- Author:Jeremy Chen
 - URL:https://www.qychen.com/article/24248e57-eb37-802e-8c97-e0dd727fd84f
 - Copyright:All articles in this blog, except for special statements, adopt BY-NC-SA agreement. Please indicate the source!
 
