Učni načrt predmeta

Predmet:
Humanoidna in servisna robotika
Course:
Humanoid and Service Robotics
Študijski program in stopnja /
Study programme and level		 Študijska smer /
Study field		 Letnik /
Academic year		 Semester /
Semester
Informacijske in komunikacijske tehnologije, 3. stopnja Inteligentni sistemi in robotika 1 1
Information and Communication Technologies, 3rd cycle Intelligent Systems and Robotics 1 1
Vrsta predmeta / Course type
Izbirni / Elective
Univerzitetna koda predmeta / University course code:
IKT3-623
Predavanja
Lectures		 Seminar
Seminar		 Vaje
Tutorial		 Klinične vaje
work		 Druge oblike
študija		 Samost. delo
Individ. work		 ECTS
30 30 30 210 10

*Navedena porazdelitev ur velja, če je vpisanih vsaj 15 študentov. Drugače se obseg izvedbe kontaktnih ur sorazmerno zmanjša in prenese v samostojno delo. / This distribution of hours is valid if at least 15 students are enrolled. Otherwise the contact hours are linearly reduced and transfered to individual work.

Nosilec predmeta / Course leader:
izr. prof. dr. Bojan Nemec
Sodelavci / Lecturers:
izr. prof. dr. Andrej Gams
Jeziki / Languages:
Predavanja / Lectures:
Slovenščina, angleščina / Slovenian, English
Vaje / Tutorial:
Pogoji za vključitev v delo oz. za opravljanje študijskih obveznosti:
Prerequisites:

Zaključena druga stopnja bolonjskega študija ali diploma univerzitetnega študijskega programa. Pri tem predmetu je potrebno predznanje matematike, fizike, znanje o sistemih vodenja in programiranja.

Priporočeni predmeti:
- inteligentni sistemi vodenja robotov
- robotski vid

Completed Bologna second-cycle study program or an equivalent pre-Bologna university study program. This course requires profound knowledge of mathematics, physics, theory of control systems and computer programming.

Recommended courses:
- Intelligent robot control
- Robot vision

Vsebina:
Content (Syllabus outline):

Osnovna zgradba in principi humanoidnih in servisnih robotov

- Strukture robotov, aktuacija, zaznavanje in krmilne arhitekture
- Značilnosti humanoidnih in servisnih robotov, ki delujejo v človekovem okolju

Predstavitev gibanja in robotskih nalog

- Parametrične predstavitve diskretnih in periodičnih gibanj
- Verjetnostne in dinamične predstavitve gibanja robotov
- Primitivi gibanja in njihove razširitve za interakcijo in podajno vedenje
- Predstavitev kompleksnih vedenj in veščin robotov

Učenje robotov

- Osnove spodbujevanega učenja in njihova povezava s sodobnim globokim spodbujevanim učenjem
- Učenje s posnemanjem in učenje iz demonstracij
- Globoko spodbujevano učenje za krmiljenje robotov in pridobivanje veščin
- Učenje v simulaciji in vzporedna učna okolja
- Prenos iz simulacije na realnega robota in prilagajanje
- Prenos znanja in posploševanje

Posploševanje in prilagajanje robotskih veščin in gibanj

- Posploševanje robotskih gibanj in veščin glede na nalogo in kontekst
- Učenje parametrov in metaparametrov politik
- Statistične in učne metode posploševanja
- Iterativno in optimizacijsko prilagajanje
- Prilagajanje naučenih veščin novim nalogam, okoljem in konfiguracijam robotov

Optimalno, prediktivno in podajno krmiljenje robotov

- Optimalno krmiljenje in linearno-kvadratne metode
- Razširitev optimalnega krmiljenja na nelinearne robotske sisteme
- Modelno prediktivno krmiljenje v robotiki
- Podajno krmiljenje in krmiljenje na osnovi sil
- Integracija modelnega krmiljenja in krmiljenja na osnovi učenja

Humanoidni in servisni roboti v človekovem okolju

- Sodelovanje med človekom in robotom
- Fizična interakcija človek–robot in robot–okolje
- Sinhronizacija, koordinacija in prilagajanje gibanja
- Varnost in robustnost v človekovem okolju

Večrobotski in dvoročni robotski sistemi

- Koordinacija več robotskih manipulatorjev
- Pristopi vodja–sledilec
- Simetrična in asimetrična razgradnja nalog
- Kooperativna manipulacija in dvoročne veščine

Zaznavanje in senzorski sistemi v servisni robotiki

- Robotski vid, globinski vid, merjenje razdalje in bližine ter drugi načini zaznavanja okolja
- Detekcija in lokalizacija robotov in objektov
- Združevanje senzorskih podatkov in multimodalno zaznavanje
- Zaznavanje za manipulacijo in interakcijo

Načrtovanje robotskih nalog in avtonomno izvajanje

- Klasično načrtovanje nalog in načrtovanje na osnovi učenja
- Visokonivojsko sklepanje in načrtovanje z uporabo splošno namembnih modelov (LLM, VLM)
- Programiranje robotov z naravnim jezikom
- Veliki jezikovni in vizualno-jezikovni modeli v robotiki
- Sestavljanje in izbira ponovno uporabnih robotskih veščin
- Agenti umetne inteligence za izvajanje in nadzor nalog
- Prilagajanje, obravnava izjem in okrevanje po napakah pri izvajanju
- Integracija zaznavanja, sklepanja, načrtovanja in krmiljenja robotov

Uporaba humanoidne in servisne robotike

- Praktični primeri učenja, načrtovanja, krmiljenja in interakcije pri humanoidnih in servisnih robotih
- Integracija zaznavanja, učenja, sklepanja, načrtovanja in krmiljenja v celovitih robotskih sistemih

Basic structure and principles of humanoid and service robots

- Robot structures, actuation, sensing, and control architectures
- Characteristics of humanoid and service robots operating in human environments

Robot motion and policy representation

- Parametric representations of discrete and periodic motions
- Probabilistic and dynamical representations of robot motion
- Motion primitives and their extensions for interaction and compliant behavior
- Representation of complex robot behaviors and skills

Robot learning

- Fundamentals of reinforcement learning and their relation to modern deep reinforcement learning
- Imitation and learning from demonstration
- Deep reinforcement learning for robot control and skill acquisition
- Simulation-based learning and parallel learning environments
- Sim-to-real transfer and adaptation
- Transfer learning and generalization

Generalization and adaptation of robot skills and motions

- Task- and context-dependent generalization of robot motions and skills
- Learning of policy parameters and meta-parameters
- Statistical and learning-based generalization methods
- Iterative and optimization-based adaptation
- Adaptation of learned skills to new tasks, environments, and robot configurations

Optimal, predictive, and compliant robot control

- Optimal control and linear quadratic methods
- Extension of optimal control to nonlinear robot systems
- Model predictive control in robotics
- Compliant and force-based robot control
- Integration of model-based and learning-based control

Humanoid and service robots in human environments

- Human–robot cooperation and collaboration
- Physical human–robot and robot–environment interaction
- Motion synchronization, coordination, and adaptation
- Safety and robustness in human environments

Multi-arm and bimanual robot systems

- Coordination of multiple manipulators
- Leader–follower approaches
- Symmetric and asymmetric task decomposition
- Cooperative manipulation and bimanual skills

Perception and sensory systems for service robotics

- Vision, RGB-D, range, proximity, and other environmental sensing
- Robot and object detection and localization
- Sensor fusion and multimodal perception
- Perception for manipulation and interaction

Robot task planning and autonomous execution

- Classical and learning-based task planning
- High-level reasoning and planning using foundation models (LLM, VLM)
- Natural-language-based robot programming
- Large language and vision-language models in robotics
- Composition and selection of reusable robot skills
- AI agents for task execution and supervision
- Adaptation, exception handling, and recovery from execution failures
- Integration of perception, reasoning, planning, and robot control

Applications of humanoid and service robotics

- Practical examples of learning, planning, control, and interaction in humanoid and service robots
- Integration of perception, learning, reasoning, planning, and control in complete robotic systems

Temeljna literatura in viri / Readings:

Izbrana poglavja iz naslednjih knjig: / Selected chapters from the following books:
- Siciliano, B., and Khatib, O. (eds.) Springer Handbook of Robotics, Springer-Verlag Berlin Heidelberg, 2016. ISBN 978-3-319-32552-1
- Corke, P. Field and Service Robotics, Springer, 2006. ISBN 10 3-540-33452-1
- Calinon, S. Robot Programming by Demonstration, EPFL Press 2009, ISBN-13: 978-1439808672
- Vadakkepat, P. and Goswami, A. (eds.) Humanoid Robotics: A Reference, Springer, 2017, ISBN 978-94-007-6045-5
- Haddadin, S.: Towards Safe Robots, Springer Berlin Heidelberg, 2014
- Kober, J. and Peters, J. Learning Motor Skills from Algorithms to Robot Experiments. Heidelberg: Springer-Verlag, 2014. ISBN 978-3-319-03193-4
- Sutton, R. S., Barto, A. G. Reinforcement Learning: An Introduction, 2nd ed., MIT Press, 2018.
- Nemec, B., and Ude, A. Robot skill acquisition by demonstration and explorative learning, In New Trends in Medical and Service Robotics, Springer 2014, ISBN 978-3-319-05431-8
- Calinon, S. A Tutorial on Task-Parameterized Movement Learning and Retrieval, Intelligent Service Robotics (Springer), 9:1, 1-29, 2016.
- Jaquier, Noémie, Michael C. Welle, Andrej Gams, Kunpeng Yao, Bernardo Fichera, Aude Billard, Aleš Ude, Tamim Asfour, and Danica Kragić. "Transfer Learning in Robotics: An Upcoming Breakthrough? A Review of Promises and Challenges." The International Journal of Robotics Research. 2024
- Hwangbo, J. et al. “Learning Agile and Dynamic Motor Skills for Legged Robots.” Science Robotics, 4(26), 2019.
- Driess, D. et al. “PaLM-E: An Embodied Multimodal Language Model.” ICML, 2023.
- O’Neill, A. et al. “Open X-Embodiment: Robotic Learning Datasets and RT-X Models.” ICRA, 2024.
- Kim, M. J. et al. “OpenVLA: An Open-Source Vision-Language-Action Model.” CoRL, 2024.

Cilji in kompetence:
Objectives and competences:

Cilj predmeta je osvojiti znanja iz osnov humanoidne in servisne robotike, vodenja, učenja ter uporabe humanoidnih in servisnih robotov. Poudarek je na sodobnih pristopih vključevanja robotskih mehanizmov v človekovo okolje.

Pridobljena znanja bodo omogočila študentom razumevanje principov gibanja in obvladovanje osnov sodobnih tehnologij s področja servisne robotike ter prenos teh tehnologij v prakso.

The objective of this course to obtain theoretical and practical knowledge of the basics of service and humanoid robotics, control, learning and applications of service and humanoid robots. The emphasis is on modern approaches of the integration of robot systems into human-like environments.

The obtained knowledge will allow the students to understand the basic principles of motion and handle modern technologies of service robotics and to apply these technologies into real practice.

Predvideni študijski rezultati:
Intendeded learning outcomes:

Študenti bodo z uspešno opravljenimi obveznostmi tega predmeta pridobili:
- razumevanje pomena in strukture humanoidnih in servisnih robotov;
- poznavanje vrste servisnih robotov, razlikovanje med servisnimi in humanoidnimi roboti, poznavanje njihovih značilnosti in tipičnih področji uporabe servisnih ter humanoidnih robotov,
- razumevanje sodobnih oblik zapisov trajektorij gibanja,
- razumevanje pomena ter principov avtonomne adaptacije gibanja robotov,
- razumevanje principov vodenja z uporabo generatorjev gibov oz. z optimizacijo,
- razumevanje osnov zapisa in izvajanja gibanja v latentnem prostoru,
- razumevanje sistemov navigacije, vodenja in učenja z demonstracijo,
- razumevanje pomena uporabe kompleksnih senzorskih sistemov v robotskih sistemih in razumevanje razlogov za uvajanje servisnih robotov ter razlogov za uvajanje humanoidnih robotov.

Students successfully completing this course will acquire:
- understanding of the structure and the aim of humanoid and service robots;
- knowledge of main characteristics of the various types of service robots and knowledge of the most common areas of applications for service robots and reasons for application of humanoid robots,
- understanding of contemporary form of encoding trajectories of motion
- understanding of principles of autonomous motion adaptation,
- understanding of control principles using motion primitives and optimization
- understand the basics of encoding and executing motion in latent spaces
- understanding of navigation, control and programming by demonstration principles,
- understanding of the importance of the complex sensory system in robotics, and knowledge of limitation and motivations for application of service and humanoid robots.

Metode poučevanja in učenja:
Learning and teaching methods:

Predavanja, seminar, konzultacije, individualno delo

Lectures, seminar, consultations, individual work

Načini ocenjevanja:
Delež v % / Weight in %
Assesment:
Ustni izpit
50 %
Oral exam
Seminarska naloga
25 %
Seminar work
Ustni zagovor
25 %
Oral defense
Reference nosilca / Lecturer's references:
1. SIMONIČ, Mihael, UDE, Aleš, NEMEC, Bojan. Hierarchical learning of robotic contact policies. Robotics and computer-integrated manufacturing. Apr. 2024, vol. 86, 1-12 str., ilustr. ISSN 1879-2537
2. KUSTER. Boris, UDE, Aleš, NEMEC, Bojan. Belief-driven tactile exploration for vision-free plug–socket insertion,” IEEE Robotics and Automation Letters, 2026,
3. NEMEC, Bojan, YASUDA, Kenichi, UDE, Aleš. A virtual mechanism approach for exploiting functional redundancy in finishing operations. IEEE transactions on automation science and engineering. [Print ed.]. 2021, vol. 18, no. 4, str. 2048-2060. ISSN 1545-5955
4. Jaquier, Noémie, Michael C. Welle, Andrej Gams, Kunpeng Yao, Bernardo Fichera, Aude Billard, Aleš Ude, Tamim Asfour, and Danica Kragić. "Transfer Learning in Robotics: An Upcoming Breakthrough? A Review of Promises and Challenges." arXiv preprint arXiv:2311.18044 (2023), to appear in The International Journal of Robotics Research.
5. GAMS, Andrej, PETRIČ, Tadej, NEMEC, Bojan, UDE, Aleš. Manipulation learning on humanoid robots. Current robotics reports. 2022, vol. 3, str. 97-109. ISSN 2662-4087