
Delft Dryad
Drone Swarms for Extreme Sports Cinematography
We are developing an autonomous swarm of 3–5 drones made to deliver multi-angle, cinematic coverage for extreme sports like downhill biking and freeride skiing. By replacing static cameras / manual operators with onboard Edge AI and LiDAR obstacle avoidance, we aim to automate the filming process.
Status
Overview
We are actively building our first prototype to achieve autonomous human-tracking flight. Currently we have all the necessary components needed for it and are integrating all of them together.
Progress
We have successfully managed to run Edge AI human recognition directly on our main MCU ( STM32N6). As well as built a functioning drone which is currently in the process of being fine-tuned for the first flight.
Looking ahead
For our future goals regarding obstacle avoidance we have selected our ToF LiDAR sensor and are actively developing our custom obstacle avoidance algorithm.
Context
The problem
After reviewing countless extreme sports clips, we noticed one thing they all had in common: creators constantly saying, “It was way steeper than it looks” or “The camera didn’t do it justice.” Currently the only solution is to tag someone along with lot of drone piloting experience to capture these unique and unrepeatable moments.
The solution
We want to change that. By developing an autonomous drone swarm capable of multi-angle tracking, athletes can choose the best perspective for every second of their stunt. Capturing the line exactly as it felt in reality, without needing a professional crew.


Timeline
Project Initialization
Apr
2026
Dryad is officially formed with a core two-person team to brainstorm and establish the foundation for the upcoming drone.
Architecture & Team Expansion
Jun
2026
Decided on the technical approach and the team expanded to 5 members to account for all the new tasks.
Hardware Assembly & Integration
Sep
2026
Managed to secure all primary hardware components and began end-to-end software and system integration for the initial prototype.
Liftoff
Nov
2026
Targeted to begin with the first flight testing. With the goal of stable autonomous human-tracking with a single drone (excluding obstacle avoidance).
Dryad Alpha
Jan
2027
Deployment of the first complete system prototype with full feature integration (tracking, edge AI, and initial obstacle avoidance) for a proof of concept testing.
Dryad v1
Jun
2027
Release of the first production-grade single drone, fully optimized with edge-case handling and refined autonomous navigation.
Dryad v2 (Hardware Miniaturization)
Dec
2027
Hardware architecture scaled down and compressed into a compact form factor designed specifically for multi-agent swarming.
Dryad Swarm
Jun
2028
Full commercial-ready deployment of an autonomous drone swarm capable of multi-angle athlete tracking with high-reliability collision avoidance.
Project Initialization
Apr
2026
Dryad is officially formed with a core two-person team to brainstorm and establish the foundation for the upcoming drone.
Architecture & Team Expansion
Jun
2026
Decided on the technical approach and the team expanded to 5 members to account for all the new tasks.
Hardware Assembly & Integration
Sep
2026
Managed to secure all primary hardware components and began end-to-end software and system integration for the initial prototype.
Liftoff
Nov
2026
Targeted to begin with the first flight testing. With the goal of stable autonomous human-tracking with a single drone (excluding obstacle avoidance).
Dryad Alpha
Jan
2027
Deployment of the first complete system prototype with full feature integration (tracking, edge AI, and initial obstacle avoidance) for a proof of concept testing.
Dryad v1
Jun
2027
Release of the first production-grade single drone, fully optimized with edge-case handling and refined autonomous navigation.
Dryad v2 (Hardware Miniaturization)
Dec
2027
Hardware architecture scaled down and compressed into a compact form factor designed specifically for multi-agent swarming.
Dryad Swarm
Jun
2028
Full commercial-ready deployment of an autonomous drone swarm capable of multi-angle athlete tracking with high-reliability collision avoidance.
Open Roles
Application Details
We are currently not recruiting new members. However as we will transition from our prototype new positions will open. We’ll post all updates on our website and announce new roles during project evenings. So if you would like to contribute stay tuned.
Application
The Team
Embectrical

Tomáš Kováč
Embectrical and Team Lead
Department

Name
Role
Department

Name
Role
Software

Mateusz Marciniak
Software Lead
Department

Andrei Stan
Software Engineer
Department

Fabio Cantini
Software Engineer
Hardware

Franciszek Adamski
Hardware Lead
Department

Ondrej Mlcoch
Hardware Engineer



