Pedro Trindade
Portrait of Pedro Trindade

Pedro Trindade

GNC & Autonomy Engineer

UAS Swarm AI Engineer at TEKEVER

Lisbon, Portugal

I build autonomy for unmanned aircraft at TEKEVER, working on how several aircraft plan and fly together. My work runs across the autonomy stack, from system architecture and software design through to flight tests on real platforms.

Before that, I was a researcher at ISR, where I did my PhD on distributed control: how a team of vehicles holds a formation using only what each one knows about its neighbours. I also built the quadrotors and the indoor flight arena to test that work.

What I enjoy most is taking an idea from the maths on paper to an aircraft that flies it, and seeing it hold up outside the lab. Outside work, I still build things, from small drones to a home server.

Interests

  • Drones and aircraft, from the bench to the air
  • Swarm autonomy
  • Guidance, navigation and control
  • Autonomy software architecture
  • Taking research out of the lab
  • AI-assisted engineering
  • How good engineering teams work

Skills

Robotics & control
ROS2PX4ArduPilotSITLJSBSimGazebo
Programming languages
C/C++PythonMATLAB/Simulink
Tooling
DockerDevContainersLinuxGitCI/CDAgentic coding (Claude Code)
Hands-on
CAD3D printingCNCcompositessoldering

Experience

Industry & Research

TEKEVER

Lisbon · 2025 – now

  1. UAS Swarm AI Engineer

    Apr 2025 – now

    • Working across TEKEVER's autonomy stack, contributing to architecture definition, mission-level autonomy, and integration with ground control and flight systems for testing on real platforms.
    • Contributed to the company's first swarming initiative from the start, helping define requirements, shape the system architecture, and build foundational autonomy modules, including path planning and ROS2 message translation.
    • Designed and implemented distributed cooperative formation flight, enabling UAVs to execute planned routes synchronously.
    • Drove the migration of a vision-based GNSS-denied navigation project to ROS2. Prepared the base repository and CI/CD pipeline, defined message interfaces, built tools to convert legacy logs into replayable rosbags, and standardised development with Docker and DevContainers.
    • Supported team growth by running technical interviews and providing input into hiring decisions, and helped onboard new engineers by sharing system and operational context.

Institute for Systems and Robotics

Lisbon · 2018 – 2025

  1. Researcher, Distributed Formation Control

    Feb 2019 – Mar 2025

    • Developed and analysed distributed consensus and formation control algorithms for multi-vehicle systems, from theory through to flight experiments on multirotors.
    • Designed, procured and assembled custom quadrotors to meet research requirements: hardware selection, PX4 integration, 3D printing, thrust testing, PID and EKF tuning.
    • Set up an indoor flight arena with a motion capture system, so students and researchers could test and validate algorithms on multirotors.
    • Prepared two interfaces for quadrotor control, one using Simulink and one using ROS and Python, for fast implementation and testing of control algorithms.
    • Published nine peer-reviewed papers, four in journals and five at international conferences.
  2. Researcher, Quadrotor Control and Integration

    Jul 2018 – Sep 2018 Short

    • Integrated the PX4 autopilot and onboard computers into DJI quadrotors, enabling execution of project software and real-time remote monitoring.
    • Tested and validated control algorithms for target tracking using a gimbal camera.

Education

Instituto Superior Técnico

Universidade de Lisboa · 2014 – 2026

  1. PhD in Electrical and Computer Engineering

    Feb 2021 – Jan 2026

    Advisors: Pedro Batista and Rita Cunha

    • Pass with Distinction and Honour
    • Coursework grade: 20/20
    • Thesis: Decentralized Consensus-Based Formation Control of Autonomous Vehicles, submitted March 2025 and defended January 2026.
    • Graduate school: EECI International Graduate School on Control, 2022, a one-week module at Paris-Saclay on equivariant systems and observer design for autonomous systems.
  2. MSc in Aerospace Engineering

    Sep 2017 – Nov 2019 2 thesis awards

  3. BSc in Engineering Sciences (Aerospace Engineering)

    Sep 2014 – Jul 2017

    • First cycle of the integrated master's in Aerospace Engineering.

Teaching & community

AeroCup Portugal

Student UAV competition · 2025 – now

  1. Organiser and coordinator

    May 2025 – now

    • Co-organising an international student UAV competition: event design, technical rules, logistics and operations.

Instituto Superior Técnico

Teaching · 2021 – 2022

  1. Laboratory teaching, Unmanned Aerial Vehicles

    Sep 2021 – Feb 2022 Short Teaching excellence

    • Taught the lab classes of Unmanned Aerial Vehicles for a semester: modelling, motion estimation and motion control of a small quadrotor.
    • Connected the lab's motion capture system to the students' Simulink kit, so their controllers could fly the quadrotor on precise position and attitude data.
    • Helped fix problems in the student kit: sensor message decoding, and a software conflict that crashed the lab computers.
    • Reviewed lab handouts and exam questions, and marked student reports.
    • Awarded the Certificate of Teaching Excellence 2021/2022, based on student surveys.
  2. Laboratory teaching, Control of Cyber-Physical Systems

    Sep 2021 – Nov 2021 Short

    • Taught the lab classes of Control of Cyber-Physical Systems for a quarter: sensor calibration, plant model identification and controller design.
    • Helped plan the classes and the assessment with the other lab staff, and built a working example to show students the goal of the lab.
    • Marked student reports, agreed the marks with the other lab staff, and supervised the first exam.

AeroTéc

Aerospace Students' Association · 2016 – 2019

  1. Board Member as Project Manager

    2018/2019

    • Communication between project teams, resource allocation, and keeping schedules aligned across concurrent projects.
  2. Vice President

    2017/2018

    • Elected to the board. Partner companies and organisations, events, fundraising and finances, and a revision of the statutes.
    • The association grew from around 35 to around 120 members.
  3. Team Leader of the Test Plane project

    2017/2018

    • Led a team through the design and construction of a UAV, while mentoring new members for the following year's Air Cargo Challenge.
  4. Collaborator

    2016/2017

    • Member of the Air Cargo Challenge 2017 competition team, working on a payload-optimised aircraft.
    • Helped organise Semana Aeroespacial, and presented introductory courses on aeromodelling and model rocketry.

Projects

Distributed formation control in flight

Consensus-based formation controllers from the PhD, flown on a team of multirotors in the lab's flight arena. The plots on the left follow the formation error and each vehicle's position as the team settles into formation.

Indoor flight arena

I set up the lab's flight arena with a motion capture system, so that students and researchers could test control algorithms on multirotors, safely and repeatably.

A small custom quadrotor on the arena floor, with motion-capture markers

Custom quadrotors

Research quadrotors built from the ground up: hardware selection, PX4 integration, 3D-printed parts, thrust testing, PID and EKF tuning.

Architecture of the prototyping platform: ground computer, PX4 autopilot, motion capture and ground control

Rapid prototyping for UAV testing

Two routes from algorithm to flight, through Simulink or through ROS and Python. Described in the ICARSC 2021 paper.

Student and personal projects

A fixed-wing aircraft dropping payloads over the field at AeroCup

AeroCup Portugal

Organiser · 2025 – now

Co-organising an international student UAV competition. Pictured: the UBI Aeronautics Team's aircraft dropping payloads at the last edition.

A small fixed-wing aircraft made of foam board faced with card

K-Line fixed-wing

Personal · 2026, paused

A fixed-wing in K-Line, foam board faced with card, built to learn to fly fixed-wing aircraft. Paused for lack of time.

A fixed-wing UAV under construction on a table

Fixed-wing UAV with friends

Personal · 2021

Built a fixed-wing UAV with friends.

One of the two quadrotors designed and built at home

Two quadrotors, designed and built

Personal · 2019 – 2020

Designed and built two quadrotors in my spare time, and used them to learn to fly.

A Pioneer robot carrying a laptop and ultra-wideband range sensors

FastSLAM with ultra-wideband ranging

Course project · Técnico, 2018

SLAM on a Pioneer robot from ultra-wideband range measurements only, in Python and ROS, in a team of four. Worked in 2D; in 3D, the landmarks' height was hard to observe.

The Test Plane project's UAV in the workshop

Test Plane

AeroTéc · 2017/2018

Led a team through the design and construction of a UAV, while mentoring new members for the following year's Air Cargo Challenge.

The Air Cargo Challenge 2017 team holding their aircraft, with the Portuguese flag

Air Cargo Challenge 2017

AeroTéc · Zagreb, 2017

Member of the competition team, working on a payload-optimised aircraft: CAD design, system integration and testing.

Publications

A network with optimised edge weights, and the agents agreeing faster than with a first guess

JournalEuropean Journal of Control · 2025

High-order consensus: Optimal controller design

How to choose the best coupling gains and network edge weights for consensus between agents modelled with any number of integrators. The design minimises an LQR-like cost averaged over the initial state, with a truncated Newton method made to fit the structure of the consensus problem.

Pedro Trindade, Pedro Batista, Rita Cunha

DOI

With a delay below the designed margin the agents agree; above it they drift apart

JournalAutomatica · 2025

Delay-margin design in consensus for high-order integrator agents

Consensus between agents modelled with any number of integrators, when every message arrives with a constant delay over a directed network. The paper gives a systematic way to design the coupling gains so that the team tolerates a chosen delay margin, which can be made as large as needed at the cost of slower convergence.

Pedro Trindade, Pedro Batista, Rita Cunha

DOI

Contours of the cost over the coupling gains and edge weights, with a Newton search stepping to the optimum

ConferenceIEEE Conference on Decision and Control (CDC) · 2024

Optimal Consensus for High-Order Integrator Agents

Optimal coupling gains and edge weights for a high-order consensus protocol on a given directed network. The parameters minimise an LQR cost averaged over the initial state, found with a truncated Newton method.

Pedro Trindade, Rita Cunha, Pedro Batista

DOI

Six quadrotors in a circular formation, keeping its shape while the whole team travels at constant speed

JournalControl Engineering Practice · 2023

Third-order consensus for robust distributed formation control of double integrator vehicles

A robust, distributed formation tracking controller for vehicles modelled as double integrators, such as multirotors. The paper gives exact bounds on the coupling gains of a third-order consensus protocol for any fixed network, uses it to reject constant disturbances, and validates it in multirotor experiments.

Pedro Trindade, Pedro Batista, Rita Cunha

DOI · Flight video

A network where every message is delayed by the same time, and the agents' states still settling to agreement

ConferenceIEEE Conference on Decision and Control (CDC) · 2023

First- and Second-Order Consensus with Constant Uniform Delays

First- and second-order consensus over a directed network when every message arrives with the same constant delay. The paper derives bounds on the coupling gains that still guarantee consensus: exact bounds for the first-order case, and sufficient bounds for the second-order case through a Padé approximation of the delay.

Pedro Trindade, Rita Cunha, Pedro Batista

DOI

The cost against the coupling gain for three cases: each curve is convex, with its best gain marked

ConferenceAmerican Control Conference (ACC) · 2022

Consensus: Optimal Coupling Gain Design

How to choose the coupling gain of a consensus protocol for single integrator agents. The paper proposes an LQR-based cost, shows that it is convex in the gain, gives its exact gradient, and derives a closed-form optimal gain under simplifying assumptions.

Pedro Trindade, Rita Cunha, Pedro Batista

DOI

Top view of the arena: a real and a simulated quadrotor flying the same Lissajous trajectory, tracked by motion capture

ConferenceCo-authorIEEE ICARSC · 2021 Highly Commended

Rapid Development and Prototyping Environment for Testing of Unmanned Aerial Vehicles

An indoor platform for developing and testing control and navigation for unmanned aerial vehicles: motion capture for ground truth, offboard computers, and several quadrotors. Software modules hide the low-level flight and communication details, and a simulator lets real and simulated quadrotors fly together.

Tiago Oliveira, Pedro Trindade, David Cabecinhas, Pedro Batista, Rita Cunha

DOI

Four quadrotors holding a square formation around a virtual leader despite a constant disturbance, where they would otherwise drift

ConferenceIFAC World Congress · 2020

Distributed Formation Control of Double-Integrator Vehicles with Disturbance Rejection

A robust, distributed formation tracking controller for double integrator vehicles, aimed at multirotors. A consensus protocol for triple integrators, with necessary and sufficient conditions for convergence, adds integral action that rejects constant disturbances. Tested in flight on two real quadrotors, flying with simulated ones.

Pedro Trindade, Rita Cunha, Pedro Batista

DOI · Flight video

Theses

Six quadrotors: four holding a square formation, and two orbiting around its edges

PhD thesisInstituto Superior Técnico · 2026

Decentralized Consensus-Based Formation Control of Autonomous Vehicles

How a team of vehicles can reach and hold a formation when each one only knows about its neighbours. The thesis covers consensus for agents described by any number of integrators, formation tracking that rejects constant disturbances, controllers designed to tolerate a chosen communication delay, optimal design of the controller gains, and flight experiments on multirotors.

Advisors: Pedro Batista and Rita Cunha
Pass with Distinction and Honour.

Full text · Experiments video

Awards

  1. 2021 / 22

    Certificate of Teaching Excellence, Instituto Superior Técnico, for the Unmanned Aerial Vehicles course.

  2. 2021

    Highly Commended, Industrial Robot journal (Emerald Publishing), for the ICARSC 2021 paper on a rapid prototyping environment for UAVs.

  3. 2020

    Best MSc Thesis in Robotics, Portuguese Society of Robotics.

  4. 2020

    Outstanding MSc Thesis Award, IEEE Portugal Section.

  5. 2018 / 19

    Certificate of Academic Excellence, Instituto Superior Técnico.

  6. 2017 / 18

    Certificate of Academic Merit, Instituto Superior Técnico.

  7. 2016 / 17

    Certificate of Academic Merit, Instituto Superior Técnico.

  8. 2015 / 16

    Certificate of Academic Merit, Instituto Superior Técnico.

  9. 2014 / 15

    Certificate of Academic Merit, Instituto Superior Técnico.