Self-healing aerial mesh for rescue operations.

SkyMesh is a self-healing offline mesh that uses UAV relays to reconnect rescue teams across terrain, distance, and damaged infrastructure.

Read Full Proposal

Mission Focus

Problem

Terrain can split responders into disconnected mesh groups.

Goal

Reconnect field teams and restore push-to-talk voice using UAV relays in a secure offline mesh.

Proof

Show the system failing before the UAV relay, then recovering it.

System Architecture

Command, team nodes, and UAV relay operate in one offline network.

Mesh Backbone

Decentralized routing algorithm (Babel) and 2.4 GHz Wi-Fi with fixed node roles.

Push-To-Talk Voice

Local Murmur + Mumble with no internet dependency.

Telemetry

Geo Location, link quality, and network fails are tracked locally monitored in Command dashboard

ML-Driven Mesh Repair

Multi-agent reinforcement learning guides UAV relays to reconnect isolated teams.

Operational Workflow

0. Pre-Plan

Estimate initial UAV hover point and relay count.

1. Bring Up

Start command services and deploy ground nodes.

2. Detect

Confirm partition with route, ping, and PTT failure.

3. Launch UAV

UAV relay joins mesh and begins autonomous monitoring.

4. Recover

Autonomous reposition restores route and cross-team PTT.

Blog Posts

Core Team and Advisors

Vicente Ferreira

    Vicente Ferreira

    Field Hardware and Deployment Lead | 110420

    vicente.frazao.ferreira@tecnico.ulisboa.pt

    Pedro Gonçalves

      Pedro Gonçalves

      Networking and Integration Lead | 87559

      pedrorenato@tecnico.ulisboa.pt

      Leonardo Catarino

        Leonardo Catarino

        UAV Payload and Flight Operations Lead | 106778

        leonardo.catarino@tecnico.ulisboa.pt

        Eduardo Curião

          Eduardo Curião

          Power and Autonomy Lead | 109709

          eduardo.curiao@tecnico.ulisboa.pt

          Dylan Félix

            Dylan Félix

            Monitoring and Guidance Software Lead | 110749

            dylan.m.felix@tecnico.ulisboa.pt

            Henrique Ferreira

              Henrique Ferreira

              Voice and User Workflow Lead | 110385

              henriquesantosferreira40@tecnico.ulisboa.pt

              Prof. Ayman M. Radwan

                Prof. Ayman M. Radwan

                Scientific Advisor
                Mariana Santana

                  Mariana Santana

                  Scientific Co-advisor