Skip to content

Hardware

Overview

A drone is a computer system with a collection of subsystems and the ability to issue commands in order to change the global state (eg move). We can issue commands to the drone by manipulating a control interface which then converts our commands into radio signals (eg a radio controller), which a receiver on the plane then relays to our flight controller. Alternatively, we can utilize a computer to generate the same commands and define flight patterns via software. When controlling the drone manually, radio dropouts or loss of signal can occur. Likewise, when flying pre-programmed missions we also need the ability to perform manual overrides, for example if an aircraft is in the vicinity. For both scenarios, we incorporate mission planning to have pre-defined fallbacks and safety protocols. Then we can have the drone execute a plan/return to a known good point (RTL) in case of loss of control or an emergency.

Broadly a drone has:

Flight control system

  • Flight controller -- your drone's brain, which contains an IMU (Inertial Measurement Unit), which estimates what direction is up/the current state. This allows the flight controller to compute our goal state and by extension what thrust the motors need to provide to maintain flight. A brief overview of power levels: F405 (SpeedyBee/Matek/clones) --> F7 --> H7 (OrangeCube) --> M7 (Auterion/CUAV/ARK).
  • Flight controller interface -- a board which allows the FC to communicate with hardware and other sensors. On low-end devices this will often be integrated.
  • Receiver -- this relays messages from a transmitter (radio controller) to pilot the drone.
  • Temperature compensation -- More advanced flight controllers can regulate temperature in order to improve/stabilize their IMU performance.
  • DroneCAN -- this is a common interface/protocol (CAN) for sensors and other devices to communicate.
  • Compass -- Most flight controllers will have one built in, to provide a frame of reference using the local magnetic field. More advanced designs often combine the results of multiple compasses to reduce noise and normalise the results. Remember to calibrate!
  • Barometer -- this provides a backup altitude reference while in flight. Many flight controllers will have one built in.
  • GPS -- this provides the primary navigation data, often via a u-blox chip. GNSS (integrating results from other satellite systems) improves our redundancy and reduces variance. RTK is an advanced form of GPS which provides even higher resolution, but requires custom equipment.

Power/flight system

  • Battery -- A large 6S LiPo provides a decent amount of power (~1/2 hour) in the field.
  • Electronic Speed Controller -- this receives signals from the flight controller to regulate voltage states (eg control the motors). DShot is a popular protocol.
  • Power management unit -- this tracks voltage/amperage to estimate battery health and by extension remaining flight time.
  • Motors / propellers / frame -- Your choice here. Carbon fiber provides a lot of rigidity but figuring out how to position elements so that the centers of gravity and propulsion are aligned is an important goal when designing airframes. Minimizing radio interference and electrical noise is an important design consideration as well.

Communication/control

Laptop/tablet to run your control layer + chat with the backend.

  • Transmitter -- for manual control of a drone in the field via radio.
  • Telemetry radio -- SiK (915Hz) is a common starting point, but more powerful radios (RFD900x) and repeaters can dramatically improve the range and the amount of data that can be transmitted to monitor the drone's state.
  • Network -- we can interface with the drone over a local (WiFi) or remote (cell) network.
  • USB -- you will probably have to connect the flight controller to your computer (eg run Ardupilot in serial mode) at some point.

Sensors/payloads

Potential options:

  • Companion computer -- to run your control layer locally and relay telemetry data/state.
  • Cameras -- Sony ILX-LR1 is camera built specifically for mapping with a manual shutter, 1/8000 shutter speed, a 61MP sensor and an ISO range of 100-32000.
  • Infrared sensors -- Teledyne makes sensors used in many popular drones.
  • Lidar -- By observing patterned laser light across a space, we can calculate the orientation and distance of any reflections. This is valuable for 3D imagery and terrain detection.
  • Radar -- We can embed low-cost millimeter wave radar and use this to estimate distance. This is often used to estimate distance to the ground or detect collisions with objects in our flight path.
  • Optical flow -- this employs computer vision to estimate velocity by computing the amount of pixel change and by extension distance using a pinhole camera.
  • Gimbal -- Gremsy is a popular choice for tweaking camera orientation mid-flight.

NDAA

As of 2024, NDAA hardware is required for drone operations in the United States in certain locations. BlueUAS is a program to certify hardware and simplify the procurement process. In late 2025, the FCC increased the restrictions in conjunction with the 2025 NDAA act, banning non-NDAA hardware going forward. In early 2026, the FCC clarified that custom drones with BlueUAS components are permitted as long as domestic components exceed 65 percent of the total cost of the finished product.

  • Cube Orange+ -- We build software to communicate with OrangeCube flight controllers (Ardupilot/PX4) using standard telemetry protocols (Mavlink2) via an onboard companion computer (Raspberry Pi w/ 4G uplink) running a custom linux distribution. This allows us to plan, distribute and coordinate missions, then calibrate and capture sensor data locally in order to synchronize with the backend post-flight (w/ WiFi). From there, we can generate a map and apply custom AI models for advanced analytics.

Companies in this space:

  • 3DR -- US made drone and radio hardware.
  • NWBlue -- NDAA drone hardware vendor.
  • Lumenier -- NDAA drone hardware vendor.
  • ModalAI -- PX4 developer kits, using Qualcomm processors with support for running TFLite models.
  • Wingtra -- Advanced VTOL surveying drone platform.
  • Ebee-X -- Advanced surveying drone platform.
  • GreenSight -- Agriculture drone platform.
  • AeroCogito -- Affordable flight controllers (H7).
  • Auterion -- Advanced flight controllers with integrated cellular networking and Jetson (NVidia) integration.
  • ARK Electronics -- Advanced flight controllers (M7) with Raspberry Pi/Jetson integrations.
  • Tycho Voyager -- A low-SWaP autonomy stack for navigation.