Lectures

Online Lectures

Please find the set off all lectures in our Youtube lectures channel. In the following there's the listing of current lectures.

Quaternion Lectures

• Quaternions Intro, Addition and Multiplication

• Quaternion Properties

• Rodrigues' Rotation Formula

• Quaternion Rotation Formula

• Quaternion Kinematics

ISAE-SUPAERO Lectures

• MIMO Control (Fall 2023, MSc)

Course Material in our intranet LMS page.

• Drone Guidance, Navigation and Control (Fall 2023, PhD)

Programme:

We will start by investigating (from a Control Systems point of view) Rigid-Body Dynamics, Stationary Aerodynamics (Buckingham-Pi-based description of forces and moments), and Flight Mechanics. Students will sketch a Flight Dynamics Simulator of their chosen Aircraft (fixed-wing or rotary-wing).

After manually flying their aircraft, the students will design autonomous guidance systems to follow autonomous missions (such as waypoint following or region surveillance). We will understand the different levels of control we expect from a typical drone, the sensors required to implement the corresponding laws under uncertainties and disturbances (e.g., cascaded PID, Total Energy Control Systems, L1 Guidance logic, Proportional Navigation), commonplace state observers, the bleeding-edge theoretical results encircling this technology (from a Controls perspective, i.e., stability analysis, convergence regions, convergence rates, controllability, observability), and current research areas.

Finally, we will match theory and practice through an in-depth study of the GNC laws implemented in the widely used Pixhawk PX4 open-source Autopilot project. For starters, we will quickly examine Embedded Real-Time concepts such as hardware abstraction, task scheduling, race conditions, semaphores, and interprocess communications in the context of the PX4 Project and corresponding Real-Time operating system (i.e., NuttX). If successful, the student will be able to read and modify the PX4 code by then. Afterward, we will explore Pixhawk's distributed-computing avionics architecture and analyze different Control task separation topologies regarding their loop delay margins and impact on achievable performances through Bode Integral arguments. We also explore PX4's nonlinear strategies to cope with actuator saturation (e.g., anti-windups, pseudo-inverse with output clipping, pseudo-inverse with sequential desaturation technique) and their relations to fundamental Control results in the literature.

Pré-requis:

Linear Algebra, Ordinary Differential Equations, Programming (MATLAB/Simulink and C/C++), Newton's Laws, and Control Systems at an Undergraduate Level.

Aerodynamics and Flight Mechanics at an Undergraduate Level are optional since the Lecturer will introduce these topics for fixed-wing and rotary-wing architectures at the beginning of the course.

IMPORTANT: To fully follow the course and assignments, a laptop equipped with Linux Ubuntu and MATLAB is mandatory (dual-boot or virtual machine). We will compile and install in Linux all other tools (i.e., Pixhawk-related software) during class. Linux Ubuntu is the de-facto system for developing open-source drones under PX4, and therefore is mandatory for this course. Prof. Lustosa's team can help you install Dual Boot during our first class if you are unable to do so, but please have a laptop with substantial free storage before subscribing to the course.

Equipe pédagogique:

Leandro Lustosa (ISAE-SUPAERO)

Méthode pédagogique:

20h Lectures (in-person) + 10h Assignments.