Robotics 2
Introduction
Prof. Alessandro De Luca
Robotics 2 ̶ 2025-26
n II semester February 24 ̶ May 27, 2026
n schedule Tuesday (8:00-10:00) - Wednesday (14:00-17:00), room B2
n master courses Artificial Intelligence and Robotics & Control Engineering
n credits 6 = 150h (1 ECTS = 25h of student work)
n regular lectures in the classroom (~60h)
n individual study and exercises (~90h)
n if needed, see the video lectures (recorded in 2019-20) on YouTube in the
Robotics 2 playlist of the Video DIAG – Sapienza channel
n G-group [Link]
active links below these!
Robotics 2 2
General information
n prerequisites
n Robotics 1 as a prerequisite (mandatory for the exam)
n aims
n advanced kinematics & dynamic analysis of robot manipulators
n design of feedback control laws for free motion and interaction tasks
n related courses
n Autonomous and Mobile Robotics 1st semester of year 2, 6 credits
n Elective in Robotics whole year 2, 12 credits (four modules)
or Control Problems in Robotics 6 credits (two out of four modules)
n Probabilistic Robotics 1st semester of year 2, 6 credits
n Medical Robotics 2nd semester of year 2, 6 credits
n research video channel [Link]/user/RoboticsLabSapienza
Robotics 2 3
The new textbook
B. Siciliano, L. Villani, G. Oriolo, A. De Luca: Foundations of Robotics,
Springer, 2025: [Link]
Advanced Textbooks in Control and Signal Processing Advanced Textbooks in Control and Signal Processing
published on
September 6, 2025
Siciliano · Villani
Oriolo · De Luca
Bruno Siciliano · Luigi Villani · Giuseppe Oriolo · Alessandro De Luca
Foundations of Robotics
Bruno Siciliano · Luigi Villani (eBook, Hardcover,
Giuseppe Oriolo · Alessandro De Luca
This textbook explores the foundational principles of robotics, focusing on its core
pillars: modeling, planning, and control. Balancing mathematical rigor and physical
intuition, a coherent formalism is established and used throughout the book. At the
same time, technological challenges and application-driven solutions are given appro-
priate consideration.
Softcover)
Foundations
With a general perspective that includes both fixed-base manipulators and mobile
12 chapters
robots, the book presents the essential tools for understanding key topics such as kin-
ematics, statics, trajectory planning, dynamics, and motion control. In its second part,
more advanced topics are addressed, including wheeled robots, visual control, motion
5 appendices
planning, force control, flexible robots and cooperative manipulation.
of Robotics
To support the learning process, appendices provide essential background material
on linear algebra, mechanics, differential geometry, control theory, and graph search
190+ problems
algorithms. The practical implementation of the methodologies is emphasized through-
out, with over 50 worked examples and case studies, many supported by simulations.
Additionally, more than 190 end-of-chapter problems are included, with a Solutions
Manual available for instructors adopting the book for their courses.
719 pages
Foundations of Robotics
Foundations of Robotics is designed for use as a textbook in both undergraduate and
graduate robotics courses within engineering programs, making it an ideal resource
for students and educators alike.
covering
n Robotics 1
n Robotics 2
n Autonomous and
Mobile Robotics
worldwide best seller since 2009
ISBN 978-3-031-85522-1
(with Italian, Chinese, Greek editions)
9 783031 855221
check the latest
Errata Corrige!
Robotics 1 4
Robotics
n algorithms for robotics*
n process inputs from sensors that provide noisy and/or partial data
n build geometric and physical models of the robot and the world
n from first principles and/or data-driven
n plan high- and low-level actions at different time horizons
n execute these actions on actuators with uncertainty/limited precision
n design & analysis of robot algorithms raise a unique combination of
questions from many fields
n control
control theory
theory
n computational geometry and topology
n geometric and
geometric and physical
physicalmodeling
modeling
n reasoning under uncertainty
n probabilistic algorithms and game theory
* = largely modified
n theoretical computer science from intro to WAFR 2016
Robotics 2 5
all on fixed-base
robot manipulators! Program - 1
n advanced kinematics Q: are redundant robots
“special” manipulators?
n kinematic calibration
n kinematic redundancy and related control methods
n dynamic modeling of manipulators
n direct and inverse dynamics
Q: why/when do we need
n Euler-Lagrange formulation dynamics for robot control?
n Newton-Euler formulation
n properties of the dynamic model
n identification of dynamic parameters
n inclusion of flexibility at the joints
n inclusion of geometric constraints
Robotics 2 6
Task-related redundancy
video of ABB robot in laser cutting
6-DOF robot for a 5-dimensional task
= 1 degree of kinematic redundancy
Robotics 2 7
Robot dynamics and control
video of WAM by Barrett Technology
…but things may go wrong anyway
@Ishikawa Lab, Tokyo University, 2012
Robotics 2 8
Robot dynamics and control
swing-up maneuver of a double inverted pendulum
video video
energy pumping, with PID control with model-based nonlinear control
Pendubot Quanser underactuated robot
(two links, only one motor at the base)
@DIAG Robotics Lab, 2005
Robotics 2 9
Robot dynamics and control
video of Atlas by Boston Dynamics, 2017
…but things may go wrong anyway
[Link]
Robotics 2 10
Program - 2
Q: why/when is a
feedback action needed?
n design of feedback control laws
n free motion tasks
n set-point regulation
n PD with gravity cancellation or compensation
e.g., on a n PID or saturated PID
soft robot
n iterative learning for gravity compensation
n regulation in the Cartesian/task space
torque input
n trajectory tracking commands
n feedback linearization and input-output decoupling
e.g., on a ▪ in the joint space
7R robot Q: why/when is kinematic
▪ in the Cartesian/task space
control not sufficient?
n passivity-based control
n adaptive (and robust) control
n on-line learning
Robotics 2 11
Iterative learning under gravity
continuum soft robots
§ hard to model: ∞-dimensional ⟹ PCC (= Piecewise Constant Curvature)
§ difficult estimation of the dynamic parameters
video
two-segment prototype @TU Delft I-RIM 2022 conference
Robotics 2 12
Feedback linearization and inverse dynamics
rigid multi-link robots
§ use a complete dynamic model, with feedback reaction to tracking errors
§ uncertainties handled by off-line identification, on-line adaptation, …
video
7R KUKA LWR4+ robot DEI UniPadova, I-RIM 2020 conference
Robotics 2 13
Position- vs. torque-controlled robots
𝑞" or
𝑦" 𝑞̇ ! 𝑖 𝑞̇ 𝑞 𝑦
task/trajectory high-level low-level actuation task
planner controller controller 𝜏 → robot ∫ kinematics
including feedforward
– e.g., 𝒒̇ ! (𝑡)
kinematic control sensing
… acceleration 𝑞̈ , mass, gravity, inertia, centrifugal forces, friction?
𝑞" or
𝑦" 𝑖! 𝜏! 𝑞̇ 𝑞 𝑦
task/trajectory torque task
planner controller
actuation robot ∫ kinematics
including feedforward
– e.g., 𝝉! (𝑡)
sensing
dynamic control
n both modes may be present even in the same robotic system
Robotics 2 14
Program - 3
n design of feedback control laws
n interaction tasks with the environment
in general,
n compliance/admittance control torque input
n impedance control commands
n hybrid force/velocity control
n image- and position-based visual servoing
n kinematic control treatment only Q: why multiple control
laws for handling the
n fault diagnosis interaction?
n detection and isolation of robot actuator faults
n extension to a class of sensor faults and to collision detection
n simulation tools
n Matlab/Simulink (including Robotics Toolbox)
n CoppeliaSim (formerly V-REP)
Robotics 2 15
Interacting with a rigid, irregular surface
video
more appropriate control law? what is the goal?
Robotics 2 16
Sneak preview of videos follows ...
n kinematic redundancy and related control methods
n robot dynamic modeling and identification
n motion control in the presence of joint flexibility
n interaction with the environment: force and motion control
Robotics 2 17
Kinematic/dynamic control and redundancy
SNS algorithm handles hard bounds on robot motion
KUKA LWR4+ robot video DIAG Sapienza/Stanford, IEEE ICRA 2012
Robotics 2 18
Kinematic control and redundancy
(standing) HRP-2 humanoid robot video @LAAS/CNRS Toulouse
HQP approach for multiple equality and inequality tasks with priorities
Robotics 2 19
Dynamic modeling and identification
4x faster
typical dynamic experiments
for model identification
data acquisition 2 videos ICRA 2014 @DIAG Robotics Lab model validation
for identification by torque prediction
KUKA LWR4+ dynamic
7R robot simulation with
with joint torque CoppeliaSim
sensing (was V-REP)
Robotics 2
video 20
Dynamic modeling and identification
e.g., linear parametrization of gravity term in robot dynamic model
𝑛=7 𝒈(𝒒) = 𝒀% (𝒒)𝝆% 𝑟) = 12
symbolic expressions of gravity-
𝝆" = related dynamic coefficients
numerical values
identified through
experiments
+" =
𝝆
gravity joint torques
prediction/evaluation on
new validation trajectory
Robotics 2 21
Dynamic modeling and identification
complete dynamic model estimation vs. joint torque sensor measurement
- friction meas
without the use of a joint friction model including an identified joint friction model
Robotics 2 22
Sensorless collision detection and isolation
using momentum observer (model-based)
video @DIAG Robotics Lab KUKA KR5 Sixx R650 robot
I-RIM 2021 conference
Robotics 2 23
Motion and interaction control
2 videos @DLR München
low-damped oscillations due to flexibility end-effector response to forces
of robot transmissions at the joints with impedance control
(use of Harmonic Drives) (selective behavior in different directions)
Robotics 2 24
Control of environment interaction
2 video clips from Springer Handbook of Robotics – Multimedia Extension
[Link]
surface contour following peg-in-hole insertion strategy
De Schutter et al @KU Leuven, Belgium (mid ‘90s)
Robotics 2 25
Physical human-robot interaction control
video ICRA 2015 @DIAG Robotics Lab
Robotics 2 26
Contacts
n student hours Tuesdays 12:00-13:30 (until early June 2026)
n in presence A-210, left wing, floor 2, DIAG
n via Zoom or G-Meet (see [Link]/deluca/[Link])
n send an email for other dates (check also My travel dates)
n communication mode
n use the G-group for questions and doubts: everyone would benefit!
n by mail (for personal issues) deluca@[Link]
n URL [Link]/deluca
n course material
n [Link]/deluca/rob2_en.php
n pdf of slides, link to video lectures, videos shown in class (zipped),
syllabus, written exams (most with solutions), …
Robotics 2 27
Exams and Master Theses
n type of exam
n midterm test qualifies for a final project (OR as part of the final exam)
n final written exam OR final project + report + oral presentation
n exam schedule for academic year 2025-26
n 2 sessions at the end of this semester
n between June 3 and July 30 to be published by April
n 1 session after the summer break on infostud & course web page
n between August 31 and September 22
n 2 sessions at the end of the first semester of next year
n January and February 2027
n sign in on infostud (code 1021883) up to one week before, one session
open at a time (OPIS questionnaire needed – to be filled in class!)
n 2 extra sessions only for students of previous years, part-time, etc.
n periods: March 16 to April 20 and October 7 to November 5, 2026
n theses samples at DIAG Robotics Lab [Link]/labrob
Robotics 2 28