All projects

Project 09 Simulation · Physics

Drone Physics Lab

A drone physics sandbox: change mass, thrust and wind, fail a motor mid-flight, then fly it with your phone.

Role
Simulator & physics
Timeline
2026
Stack
Physics engine · WebSockets · PID control
Status
Live
Drone Physics Lab: a physics parameter sidebar next to a 3D drone flying over hills and lakes

Overview

Drone Physics Lab is a drone physics testing simulator. It lets you manipulate real-life physics, components and the drone’s balance, and simulate critical situations like motor failures, so pilots can train effectively.

When it runs locally, your phone becomes the remote: connect through its browser and fly with dual virtual joysticks.

The problem

Motor failures and bad balance are exactly the situations pilots need practice for, and exactly the ones you can’t safely rehearse with a real drone.

How it works

  • Tinkerable physics. Mass, inertia, drag, thrust coefficients, and environmental wind and gravity are all exposed as parameters.
  • Failure scenarios. Trigger motor failures to see how the drone reacts, and practise the recovery.
  • Control modes. Fly a quadcopter or a hexacopter in Direct mode (raw motor percentages), PID Stabilized, from the keyboard, or from the phone remote, with every PID gain live on a slider.
  • Flight challenges. Hover Test, Ring Course, Precision Landing and Altitude Ladder, watched from an orbit, chase or FPV camera.
  • Phone as remote. Running locally, the simulator accepts a phone connection through its browser, with dual virtual joysticks sending input over WebSockets.

Key features

  • 01

    Motor failures

    Fail a motor mid-flight and train the recovery.

  • 02

    Tunable physics

    Mass, inertia, drag and thrust coefficients.

  • 03

    Environment

    Wind with gusts, gravity down to the Moon’s, and ground effect.

  • 04

    Direct & PID

    Direct, PID, keyboard or phone, with live gains.

  • 05

    Phone remote

    Dual virtual joysticks in the mobile browser.

  • 06

    Runs online

    Try it at drone.sriram4n.com.

Challenges & what I learned

Every exposed parameter is another way for the model to go unstable, so good defaults matter as much as flexibility. Tuning PID gains by hand is also one of the best ways to really understand them.

What’s next

More failure scenarios, such as sensor faults and battery sag, and recorded runs for comparing pilots.