We make modern drone applications easier to build with an autopilot ecosystem made for integrations and AI models Contact: support@droneforge.us
El Segundo, CA
- Tweets 489
- Following 0
- Followers 8,200
- Likes 1,130
Update 2.7.0 out now!
⭐️ Enjoy our new Replayable routes feature. Fly it once and run it again for continuous monitoring
⚙️ Watch your drone follow its path with lower error
📦 Easily search for available features in our features tab!
Download today at thedroneforge.com
2
2
21
1,433
DRONEFORGE retweeted
DRONEFORGE IS HIRING
We just 10x'd autonomous drone deployments, and we're planning to do it again
We're hiring two very important positions on-site in El Segundo
> Entry-level Flight Controls / GNC (directly own and improve the control stack of our real-world deployments)
> Hardware Design and Test Engineer (advance and develop our products that are being used by hundreds of customers)
Help us build America's largest commercial autonomous drone fleet. DM me or go to our company page to learn more :)
thedroneforge.com/company
9
32
299
14,825
[7.18.2026] Researchers at the University of Konstanz made Visual SLAM run up to 9x faster on embedded hardware without CUDA
GLidE-SLAM uses the GPU for fast, photometric pose tracking between frames
It then returns to feature based SLAM only when it needs to extend the map or regain stability
The result is far less compute, with nearly the same trajectory accuracy
Read the paper here: arxiv.org/pdf/2607.16897
2
3
35
1,991
Interesting with deploying slam to low-cost systems? Check out Nimbus -- the quickest way to enable autonomy on ready-to-fly FPV drones!
thedroneforge.com
4
431
Update 2.6.0 out now!
New:
Turn any FPV drone into a cinematic drone. Autonomous tracking helps you fly handsfree
Improved:
Application performance improved. BMI270 IMU supported.
3
19
2,519
The code for self-correcting drones just dropped btw
github.com/uzh-rpg/LAFR
2.10.2026 -- Researchers at RPG UZH demonstrated a drone system that learns its own dynamics and improves its controller WHILE flying
This is its flow:
1. fly a trajectory
2. learn the unmodeled forces, delays, and disturbances (online residual dynamics)
3. re-optimize the policy from the latest state (RASH-BPTT)
4. increase speed until tracking error approaches the safety boundary (adaptive temporal scaling)
It went from 2.0 m/s to 7.3 m/s in roughly 100 seconds of flight
Read more about the paper here:
arxiv.org/pdf/2602.10111v1
2
28
4,119
2.10.2026 -- Researchers at RPG UZH demonstrated a drone system that learns its own dynamics and improves its controller WHILE flying
This is its flow:
1. fly a trajectory
2. learn the unmodeled forces, delays, and disturbances (online residual dynamics)
3. re-optimize the policy from the latest state (RASH-BPTT)
4. increase speed until tracking error approaches the safety boundary (adaptive temporal scaling)
It went from 2.0 m/s to 7.3 m/s in roughly 100 seconds of flight
Read more about the paper here:
arxiv.org/pdf/2602.10111v1
2
5
65
11,853
Update v2.5.0 out now!
What's new:
> Friendlier and easier to use UI
> Customization for the text-to-flight feature
What's improved:
> Improved flight performance
> General app performance
Get started today at thedroneforge.com
1
4
20
4,361
We are back in stock! Get text-to-flight autonomous drones today!
thedroneforge.com
4
277