Topics in Autonomous Robotics

Topics in Autonomous Robotics

Students will be introduced to modern approaches in control and design of autonomous robots through lectures and a project. The course takes place on particular dates of the spring semester, typically half a day of lectures, see the schedule below.

There will be a choice of research projects for each topic. A project consists in experimental research or a critical review. Ideally, a research project could become a conference paper. Students choose one project and register on this website. Projects are be done with a lecturer from another lab than the host lab of the student.  Students will be assessed on the research projects: written report (80%) + oral presentation (20%). The grading of the research project is done by the corresponding professor. 

The written report will be 6-8 pages long, and formatted using the IEEE style files provided below and according to the guidelines of each project. All reports must be written in English. The oral presentation of the report will last 10 minutes, followed by 10-15 minutes of questions and answers.

Lecturers and Course dates in spring 2023. Please check dates and hours carefully as they change from lecture to lecture. Lectures will be given online using Zoom.

1. Tuesday March 28, 2023, 9:00 to 12:00 (CM09) – Reconfigurable robotics – Jamie Paik
2. Wednesday April 5, 2023, 9:00 to 12:00 (CM09) – Micro- and Nanorobotics – Selman Sakar
3. Tuesday April 18, 2023, 13:30 to 16:30 (CO120) – Visual perception for robotics – Amir Zamir
4. Tuesday April 25, 2023, 13:30 to 17:30 (CO120) – Soft electrically-driven actuators for robotics and haptics – Herb Shea
5. Tuesday May 2, 13:30 to 17:30 (CO120) – Deep learning for Autonomous Vehicles – Alexandre Alahi
6. Wednesday May 17, 2023, 9:00 to 12:00 (CM09) – Design and Control of Prosthetic Devices – Silvestro Micera
7. Tuesday May 23, 9:00 to 12:00 (CO121) – Locomotion control in swimming and legged biorobots – Auke Ijspeert
8. Tuesday May 30, 2023, 13:30 to 16:30 (C0123) – Robotics for Rehabilitation and Assistance – Mohamed Bouri
   

Deadline for project registration: June 30 2023. Note that the project can take place during the summer. You do not have to choose a project immediately, and can wait to see all lectures before signing up (Signup details sent by email on June 2nd).

Deadline for handing in project reports: To be defined with the host professor; by September 2023 at the latest.

Presentation Date:  To be defined with the host professor; by September 2023 at the latest.

Presentations should last for 10-15 minutes followed by 10 minutes question.

Abstract

A truly ubiquitous environment is where human-machine interactions are intuitive, reliable, and compatible. This requires an intelligent platform that is versatile and adaptable to evolving tasks and dynamic environments. While there are extensive efforts in addressing this challenge through massive data and learning algorithms, there is yet to be a cohesive solution to improve the actual physical interaction. Recent developments in soft robots with their unconventional material-based solutions and modular robots with a multitude of configurations propose possible avenues to extend the capacities of robotics. This talk will highlight the recent progress in soft-material robots and reconfigurable origami robots that aim at achieving comprehensive solutions toward diverse "softer" human-robot applications.

Topics that are covered are:

- Origami robot design platform

- Pneumatic actuators and soft sensors

- Reconfigurable robot applications   

Time

May 26^th 2021 @ 9h15 https://epfl.zoom.us/j/83551001695

Projects

 1 A life-scaled 3D origami platform for force-feedback manipulations: design, model, and prototype a motorized origami robot platform that can interact with an VR environment. You will base your transmission on the folding mechanism of Foldway but you are free to choose the application and the actuator. There are two parts to the project (ie need 2 members): a force controlled platform and its graphical representation of the force feedback (no need to be VR)

2 Literature review on multimodal force feedback: Force feedback interfaces and their technologies. 

• Quantify and tabulate technology types and their performance (i.e. number of modes, types of feedback, frequency, bandwidth, force, etc).
• Analyze pros/ cons (with quantity) of the used technology (i.e. new transmission, actuator, control, interface) for at least 2-3 design parameters (ie overall weight, volume, aspect ratio, power density, etc) 

When there are no direct examples of force feedback interface with a clear dataset beyond rotary/ linear motors, pick potential technologies for an active stiffness change (phase-changing materials, electrostatic sheets, vacuum-based flat pouches, etc) to project their performance for benchmark dimensions.  

Recordings

https://mediaspace.epfl.ch/channel/Reconfigurable+Robotics+Lab+-+RRL/49145

For questions, contact jamie.paik@epfl.ch with a subject heading ENG-615 Projects

Selman Sakar

Here is a pdf of the slides:  https://drive.google.com/file/d/13gwn-i048ZxYPl9CSgSU2soFhZaKXMNX/view?usp=share_link 

Tuesday April 25  2023 from 13:30 to 17:00. Soft electrically-driven actuators for robotics and haptics. Herb Shea, EPFL-LMTS

Tuesday May 2 from 13:30 to 17:30. Deep learning for Autonomous Vehicles. Alexandre Alahi.

Lecturers: Alexandre Alahi

Abstract: Deep learning is poised to reshape the future of mobility with Autonomous Vehicles (AVs) tackling the "last-mile" problem, i.e., how we move people and goods to their final destinations. Self-driving cars will reduce human error in driving and save millions of lives every year worldwide, reduce traffic and pollution, increase mobility for the elderly and those with disabilities. Similarly, new types of autonomous vehicles will share sidewalks to lower delivery costs or assist elderly/blind people in crowded social scenes.

Integration of these AVs into our society remains a grand challenge: they need to co-exist with humans in close proximity. There is an arms race of research groups from academia, startups, and large companies pushing the limits of deep learning to solve this challenge. 

Students will learn about the state-of-the-art deep learning methods to develop AVs. The course will focus on 3 pillars (the 3 P): 

1- Perceiving - detecting and classifying the activities of traffic participants; 
2- Predicting - forecasting their mobility;
3- Planning - a set of actions accordingly.

Projects:

1. Project 1: Implement an object detector for autonomous vehicles. 
   You will revisit our available neural network architecture for pose estimation to solve detection task 
2. Project 2: Implement a human trajectory forecasting model.
   You will participate to our Trajnet++ challenge by revisiting our existing models.

More details in the lecture. Contact: alexandre.alahi@epfl.ch.

Monday May 10 from 9:00 to 13:00. Robotics for Rehabilitation and Assistance.  Mohamed Bouri.

Rehabilitation Robotics (Part I)

• Actuation and sensing
• ‪Concepts and metrics
• Lower limb and Upper limb devices
• ‪Control strategies and therapies

Walk Assistive Robotics (Part II)

• Exoskeletons, definitions and review of existing devices
• ‪Control strategies
• Perspectives for daily living

Wednesday May 19 from 9:00 to 12:00. Neuro-Robotics. Silvestro Micera.

Main topics

1. Introduction to neuroprosthetics and bionics
   1. General definition
   2. Modularity as a key approach
2. Control of hand prosthesis
   1. Muscular control
   2. Neural control
   3. Shared control
   4. sensory feedback
3. Brain-machine interfaces
   1. Brain decoding
   2. sensory feedback

Please use the following zoom link for our class today:

https://epfl.zoom.us/j/2034249425?pwd=VU5Qa2hOMEtHSldmWmVQQUs2cTFLUT09

Possible activities:

1. Review on sensory feedback using implantable interfaces

2. Review on shared control for neuro-robotics integration

See you later

Silvestro

Tue May 23 from 9:00 to 12:00. CO 121  Investigating animal locomotion using biorobots and neuromechanical simulations. Auke Ijspeert.

 Thursday April 15 from 14:15 to 18:00. 

Lecturers: Dario Floreano and Omar Aloui

Zoom: https://epfl.zoom.us/j/85059432880

Contents:

- Introduction to tensegrity systems: history, definition, types and properties (Floreano)

- Tensegrity systems in art, architecture, engineering, biology (Floreano)

- Stability, equilibrium, and form finding (Aloui)

- Tensegrity robots for locomotion, manipulation, and biomimetism (Floreano)

- Manufacturing methods (Floreano)

- Modular tensegrity robots at LIS (Floreano)

Proposed projects:

1) A novel dual-cell tensegrity-based jumping mechanism

2) A dual tensegrity structure with bendable bars as a novel jumping tensegrity robot

(see attached description). If interested, please contact Omar.Aloui@epfl.ch

Monday April 8 2019: Small scall Robotics (Selman Sakar)

Monday April 8 2019: Mobile Robot Design (Francesco Mondada)

Structure:

• Existing Mobile robot products
• Robot design methodologies
• Case studies:
  • Khepera: Mechatronics >< Market
  • e-puck & Thymio II: Mechatronics >< Education
  • Leurre & ASSISIbf: Mechatronics >< Biology, methodology
  • S-bot & marXbot: Mechatronics >< Computer Science, approach
  • Robotic glasses & Ranger: Mechatronics >< user interaction
• Conclusion / assignements