
World's most advanced autonomous vehicle software
Showing 5 of 36 projects. Click any project card for scope, mentors, and proposal studio.
Mentors: Student: Omkar Sarkar
ArduPilot flight logs are currently analyzed manually, making it difficult to understand why crashes or unexpected behavior occur. Users have to inspect multiple signals across time, which requires experience and is not reliable for most people. A Python tool that runs locally and takes a .BIN log as input. The data is aligned across time and processed to extract signals such as attitude error, battery behavior, GPS error and other signals. Physics-based checks detect sustained issues based on flight behavior, while a Hidden Markov Model observes patterns over time to filter out short disturbances. Results are treated as reliable when both agree over a consistent duration. An LLM based agent coordinates the workflow by running the analysis steps and combining their outputs into a clear explanation, while all calculations follow rule-based logic. The analyser adapts to different firmware versions by reading parameters and detecting configuration related issues by comparing them with expected ranges from the ArduPilot codebase. It outputs what happened, when it happened, and why, along with linked telemetry evidence and possible fixes. Deliverables include a complete Copter focused diagnostic pipeline, ML integration, dynamic parameter lookup, structured output generation, and a packaged command line tool with documentation.
Mentors: Student: Neeta Misericordia
ArduPilot flies over a million vehicles but has never controlled a humanoid, I intend to change this. I've spent the past several weeks building an 8-DOF humanoid in Gazebo Harmonic with ArduPilot SITL already connected, with the EKF3 reading attitude at 50Hz, joints responding, and the robot standing under real physics. During GSoC I'll extend it into a real platform with stable ZMP-based walking, active arms with reach and grasp, crouch and stand functions, MAVLink navigation commands, and a PR to ArduPilot main. By the end of summer ArduPilot will be able to command a humanoid to walk somewhere and interact with objects, using the same battle-tested infrastructure that already runs drones and rovers.
Mentors: Student: Kwaku Richter
This project aims to add a practical mesh networking framework to ArduPilot so multiple vehicles can continue sharing telemetry and coordination data even when direct links to the ground station are lost. The proposed solution is a new ArduPilot subsystem, AP_SwarmMesh, which builds standard MAVLink peer streams, wraps them in a lightweight peer routing header for multi-hop forwarding, and exchanges them over a dedicated peer link using both a "Full" and "Lite" backend. The main deliverables are the AP_SwarmMesh library, peer packet framing and forwarding logic, bounded onboard peer state tracking, SITL validation, and hardware demonstrations showing resilient multi-vehicle communication across both backend types.
Mentors: Student: Elton Satiyev
This project makes drone control easier by using a web-based AI assistant. Pilots can type or speak commands like “Arm the drone” or “Fly 10 meters north,” and the tool converts them into MAVLink messages for ArduPilot vehicles. It reimplements MAVProxy’s AI chat module as a modern WebTool, integrating the OpenAI API with JavaScript and WebSocket for Mission Planner connection. Development focused on core actions such as arming, takeoff, directional flight, and switching modes.
Mentors: Student: Bhajneet Singh Bedi
The project aims to make an electro-mechanical model of a motor which would simulate a motor in gazebo. The purpose is to enhance the motor testing behaviour before real world deployment. It also converts the RC command to motor joint control command more correctly and accuratly.