Biorobotics Laboratory BioRob

Project Database

This page contains the database of possible research projects for master and bachelor students in the Biorobotics Laboratory (BioRob). Visiting students are also welcome to join BioRob, but it should be noted that no funding is offered for those projects. To enroll for a project, please directly contact one of the assistants (directly in his/her office, by phone or by mail). Spontaneous propositions for projects are also welcome, if they are related to the research topics of BioRob, see the BioRob Research pages and the results of previous student projects.

Amphibious robotics
Computational Neuroscience
Dynamical systems
Human-exoskeleton dynamics and control
Humanoid robotics
Miscellaneous
Mobile robotics
Modular robotics
Neuro-muscular modelling
Quadruped robotics


Amphibious robotics

609 – Refactoring code base for control of amphibious robots
Category:semester project, internship
Keywords:C++, Computer Science, Control, Electronics, Linux, Programming
Type:10% hardware, 90% software
Responsible: (MED 1 1215, phone: 32602)
Description:The EPFL Biorobotics Lab (BIOROB) operates several robots on a control framework developed over several years. The latest iteration involves an amphibious, sprawling-gait robot under and NCCR Robotics program to further the state of search and rescue robots. The aim of this project is to refactor the functioning code base to improve usability and readability. Specifically, we aim to separate the general control framework from specific instance of the running robot and extensively document classes using Doxygen comments. Good knowledge of UNIX-like environments (e.g. Linux) and good C++ skills are a requirement. Previous experience with mechatronics projects and/or robotics is a plus. Applicants able to provide examples of past programming projects are strongly preferred. Code base located at: https://gitlab.com/biorob-krock/Krock2_RobotController

Last edited: 15/08/2019
599 – Porting control architecture to ROS
Category:semester project, bachelor semester project, internship
Keywords:C++, Control, Linux, Programming, Robotics
Type:20% hardware, 80% software
Responsibles: (MED 1 1611, phone: 34793)
(MED 1 1215, phone: 32602)
Description:We are investigating highly articulated, amphibious robots within the scope of NCCR Robotics towards search and rescue scenarios. These robots are to be demonstrated cooperating with other legged and aerial platforms within the program. https://nccr-robotics.ch/ Our system is currently controlled through a C++ program running on an Odroid XU4, though most other systems under our NCCR project run within the Robot Operating System (ROS). To facilitate cooperation (e.g. sharing sensor data while mapping environments), we would like to port the control of the robot to this common, "middleware" platform. https://www.ros.org/ The project would consist of the initial setup and test of ROS, further documentation of the current C++ code base, and working up towards the control of a quadrupedal robot, such as k-rock. Ultimately, this would help more easily integrate additional actuation, sensors, and control. https://actu.epfl.ch/news/k-rock-meets-his-cousin/ Student should have experience coding in C++. Please contact Matt Estrada with a description of previous projects worked on.

Last edited: 05/06/2019
598 – Amphibious hardware design
Category:semester project, master project (full-time), internship
Keywords:Bio-inspiration, C++, Control, Electronics, Locomotion, Mechanical Construction, Programming, Robotics
Type:10% theory, 80% hardware, 10% software
Responsibles: (MED 1 1215, phone: 32602)
(MED 1 1215, phone: 32602)
Description:Aquatic, undulatory locomotion is investigated in the lab by highly-articulated, robotic systems that require actuation, electronics, and hence waterproofing, throughout the body of the robot. The ability to rapidly iterate hardware design is hindered by the overhead in designing these water-resilient systems. An alternative approach being used in the lab is to use a dry-suit, creating a layer between the water and electronics inside. However, this suit can hinder mobility of the robot and wear at contact points when traversing the ground. We are working with a dry suit for the K-Rock robot, using a similar skeleton as the one shown below, but a different (i.e. less-bio) suit. Desired refinements include a better process to put on the suit and reinforcing contact points with the ground. https://actu.epfl.ch/news/k-rock-meets-his-cousin/

Last edited: 03/06/2019

Human-exoskeleton dynamics and control

602 – Design (Mechanics, Electronics, Human-machine interface) and improvement of a hand exoskeleton
Category:semester project, master project (full-time), internship
Keywords:C++, Data Processing, Electronics, Embedded Systems, Mechanical Construction, Radio, Robotics
Type:100% other activities
Responsible: (MED 0 1023, phone: 34183)
Description:Impairments to hand sensorimotor functions are among the most common consequences of traumatic and neurological conditions affecting the central and peripheral nervous systems, such as strokes, spinal cord injuries, and cerebral palsies. Due to the fundamental role that our hands play during everyday living, deficits in arms and hand functioning are heavily disabling conditions that can drastically reduce the personal independence during daily life and the social participation of affected individuals.
At the CNBI and BioRob laboratories at the École Polytechnique Fédérale de Lausanne, we developed a novel hand exoskeleton with the main purpose of assisting independence and promoting intensive use during activities of daily living (ADL). The system was iteratively designed, developed and tested in collaboration with healthcare professionals and users with hand motor disabilities. Experimental results showed that the exoskeleton could help these users regain capabilities useful for independence in daily life that they had lost since their disabling accidents.
Short summary video
Journal paper
Tests with end-users
Images and short description

Within the framework of the hand exoskeleton, we have several opening for students semester and master projects:
(i) Industrial design
- Design for production
- Design for regulatory compliance
- Design for manufacturing and assembly
- Setup of supply chain

(ii) Electronics
- Microcontroller
- Radio communication
- Design for manufacturing and assembly
- Battery recharge

(iii) Human machine interface
- Electromyographic (EMG) control
- Software interfacing with commercial EMG device
- Development of subject-specific decoders

We’re interested in self-driven candidates, with background and experience in:
(i) Mechanical Design (e.g. Solidworks, Inventor, Catia, ...). Expertise in: 3D printing, Laser cutting, CNC, ...
(ii) Electronics design (e.g. Altium, Eagle, ...). Expertise in: breadboard prototyping, SMD soldering, ...
(iii) Software development (e.g. C, C++, Matlab, ...). Expertise in: software modeling (e.g. UML), signal processing, ...

Experience with any of the following is a plus:
- Design for production methods (DFMA)
- Worked with soft and wearable systems, fabrics and sewing
- Interest in assistive technologies and devices
- Passionate about product design

If you are interested in any of the previous topics, contact: luca.randazzo@epfl.ch

Last edited: 09/07/2019

Miscellaneous

601 – Multimodal controller design and characterization using somatosensory feedback for salamander locomotion
Category:semester project
Keywords:3D, Bio-inspiration, Control, Feedback, Locomotion, Programming, Simulator, Synchronization
Type:30% theory, 70% software
Responsible: (MED 1 1611, phone: 34793)
Description:A simulation model of the salamander is currently being developed for 3D simulations with the Bullet 3D physics engine. It has been implemented such that it could already simulate basic walking gaits using a network of oscillators based on CPGs. This project will consist in improving the walking gaits in simulation by upgrading the morphology and the oscillator networks. More precisely, the end goal of the project will be to explore how the combination of sensory feedback with CPG networks could improve walking in salamanders.

Last edited: 20/06/2019
597 – Building a 3D mouse treadmill for locomotion experiments
Category:semester project, master project (full-time)
Keywords:3D, Control, Electronics, Locomotion, Mechanical Construction, Quadruped Locomotion
Type:10% theory, 80% hardware, 10% software
Responsible: (MED 1 1611, phone: 36714)
Description:Mice have been a very crucial species in studying and understanding how mammalian locomotion operates. Current state of the art neuroscience research has provided insights into the neural circuits in the brain and spinal cord that are involved in turning behaviors during locomotion. In order to study the neural circuits while the mice are performing turning behaviors, it is essential that they are placed on a platform that allows them to not only walk in straight paths(like a conventional treadmill) but also allow them to turn while walking in place. At this moment one solution we consider is a large supported ball over/inside which the mouse walk and rotate. To this aim the goal of the project is to first explore the possibilities of a mechanism that can accomodate the above mentioned requirements. Then build a proof of concept system and test it with mice. A long term goal for the project would to integrate a Virtual Reality system with treadmill support. For this project we are looking for a student who has background/experience in building mechanical structures, mechanical design, mechatronic systems and may be some basic electronics.

Last edited: 08/05/2019
578 – Implementation of an optimization framework
Category:master project (full-time)
Keywords:C++, Computer Science, Linux, Programming
Type:100% software
Responsible: (MED 1 1025, phone: 36630)
Description:The EPFL biorobotics lab (BIOROB) has several servers that are used as computation nodes to run simulations. Currently, we use a custom optimization framework that allows users to easily distribute simulations (mainly running on Webots) on the nodes and to collect the results. The aim of this project is to (a) explore and document how to deploy SLURM (or a similar open-source well-maintained job scheduler) on such a setup, and to (b) implement an extensible optimization framework (in C/C++, with bindings at least for Python) capable of doing what our current setup can do, but on the top of SLURM. Good knowledge of UNIX-like environments (e.g. Linux) and good C++ skills are a requirement. Previous experience with job schedulers and/or optimization (e.g. particle swarm optimization) is a plus.

Last edited: 14/11/2018 (revalidated 09/08/2019)

Neuro-muscular modelling

608 – Matching healthy and pathological gaits with GA algorithm
Category:semester project
Keywords:Biped Locomotion, Computational Neuroscience, Feedback, Optimization, Programming, Reflexes
Type:30% theory, 70% software
Responsible: (MED 1 1611, phone: 36620)
Description:The EPFL Biorobotics laboratory is using neuromuscular simulation to better understand the mechanisms behind the neural control of human locomotion. These simulations have shown that a simple sensory driven controller is able to replicate healthy gaits stimulating reflexes based on stretch and force receptors and gained by specific constant parameters. This project aims to investigate whether our neuromuscular model is able to replicate gaits resulting from pathologies such as stroke, weakness and cerebral palsy matching joint trajectory obtained in simulations and experimental data using Genetic Algorithms. The code is implemented in Python. Past experience with Webots is a plus, but not compulsory.

Last edited: 18/07/2019
607 – Sensitivity analysis of muscle reflexes on multiple parameter dimension
Category:semester project
Keywords:Biped Locomotion, Feedback, Programming, Reflexes
Type:30% theory, 70% software
Responsible: (MED 1 1611, phone: 36620)
Description:The EPFL Biorobotics laboratory is using neuromuscular simulation to better understand the mechanisms behind the neural control of human locomotion. These simulations have shown that a simple sensory driven controller is able to replicate healthy gaits stimulating reflexes based on stretch and force receptors and gained by specific constant parameters. Yet, it is still not clear which parameter or combination of parameters influence most the gait patterns. Previously, an OFAT (one factor at time) study was conducted to investigate the influence of single parameters This project aims to perform a sensitivity analysis studying the changing effect of combinations of reflex parameters extracting data from multiple simulations in Webots. The code is implemented in Python. Past experience with Webots is a plus, but not compulsory.

Last edited: 18/07/2019
605 – Modulation of reflex parameters in human walking
Category:semester project, master project (full-time)
Keywords:Bio-inspiration, Biped Locomotion, Feedback, Optimization, Programming, Reflexes
Type:30% theory, 70% software
Responsible: (MED 1 1611, phone: 36620)
Description:The EPFL Biorobotics laboratory is using neuromuscular simulation to better understand the mechanisms behind the neural control of healthy and pathological human locomotion. These simulations have shown that a simple sensory driven controller is able to replicate healthy gaits stimulating reflexes based on stretch and force receptors and gained by specific constant parameters. These parameters are tuned by the reticulospinal tracts in order to achieve different objectives, ie. reaching a target speed, ground clearance, step length and cadence. This project aim to perform multiple optimizations on reflex parameters to reach target gait characteristics together with a study on the way these parameters change in function of the changing of the target. Computational programming skills with Python are required. Previous experience with Webots is a plus.

Last edited: 18/07/2019
606 – Modeling spasticity behavior in human pathological locomotion
Category:semester project
Keywords:Biped Locomotion, Feedback, Programming, Reflexes
Type:30% theory, 70% software
Responsible: (MED 1 1611, phone: 36620)
Description:The EPFL Biorobotics laboratory is using neuromuscular simulation to better understand the mechanisms behind the neural control of healthy and pathological human locomotion. These simulations have shown that a simple sensory driven controller is able to replicate healthy gaits stimulating reflexes based on stretch and force receptors and gained by specific constant parameters. These reflexes are modulated by descending pathways coming from the brain. However, brain and spinal injuries may result in an abnormal modulaton of reflexes leading to spasticity. This project aims to reproduce the spasticity behavior in neuromechanical simulations of pathological human walking. Computational programming skills with Python are required. Previous experience with Webots is a plus, but not compulsory.

Last edited: 18/07/2019
604 – Studying the role of CPGs and reflexes for Guinea Fowl locomotion using neuro-mechanical models
Category:master project (full-time), internship
Keywords:3D, Bio-inspiration, Biped Locomotion, Inverse Dynamics, Kinematics Model, Locomotion, Optimization, Programming, Simulator
Type:20% theory, 50% software, 30% other activities
Responsible: (MED 1 1611, phone: 36714)
Description:Ground birds such as the guinea fowl (GF) have become important bipedal animal models for understanding the biology of neuro-mechanical integration, and for bio-inspired design and control of bipedal robots [refs]. The goal of the project is to develop and study motor control strategies for bipedal locomotion in guinea fowl by combining experimental data with neuro-mechanical models. The project will be under the supervision of two labs, the Comparative Neuromechanics Lab at the University of California, Irvine, USA (Dr. Monica Daley) and BioRob at EPFL (Prof. Auke Ijspeert). The student will begin the project by becoming familiar with the existing literature on GF biomechanics, neuromuscular control and locomotion models []. The first aim of the project is to extend an existing OpenSim (Musculoskeletal Simulator) model of the GF to the simulation framework (FARMS) used in BioRob for closed loop neuro-mechanical simulations. The second aim of the project is to generate bipedal gaits using the GF model, validated using available experimental data. For this, the student will perform inverse kinematics and dynamics analysis of the model using available experimental data. Additionally, this process could be extended considering the muscles in the loop to predict muscle activations, which can be compared with experimentally measured muscle activation patterns measured using electromyography (EMG) during locomotion. Once the inverse dynamic validation step is completed, the student can then focus on developing simple controller models to produce the muscle activations necessary for stable locomotion. As a starting point, the student would use a combination of Central Pattern Generator (CPG’s) and simple reflex rules based on experimental evidence from existing literature. The student will use the GF neuromechanical model to reproduce terrain perturbation studies previously performed in experiments in Dr. Daley’s lab. The student can hypothesize multiple possible controller strategies that predict perturbation recovery dynamics. By comparing predicted perturbation recovery and stability between the model and experiment, we can rigorously tests hypotheses for the control strategies used by the guinea fowl, addressing the scientific question of the project. The resulting controllers could serve as bio-inpsired control for bipedal robots.
Work Load : Theoretical(20%), Programming(50%), Data Analysis(25%), Writing(5%)
Requirements : Good programming skills in C++ and Python. Experience in musculoskeletal modelling and neural modelling is beneficial. Experience in rigid body physics simulators such as OpenSim, Webots, Gazebo, Bullet is a bonus for the project.
REFERENCES :
Daley, M.A., Voloshina, A. and Biewener, A.A., 2009. The role of intrinsic muscle mechanics in the neuromuscular control of stable running in the guinea fowl. The Journal of physiology, 587(11), pp.2693-2707.
Daley, M.A. and Biewener, A.A., 2006. Running over rough terrain reveals limb control for intrinsic stability. Proceedings of the National Academy of Sciences, 103(42), pp.15681-15686.
Birn-Jeffery, A.V., Hubicki, C.M., Blum, Y., Renjewski, D., Hurst, J.W. and Daley, M.A., 2014. Don't break a leg: running birds from quail to ostrich prioritise leg safety and economy on uneven terrain. Journal of Experimental Biology, 217(21), pp.3786-3796.
Daley, M.A., 2018. Understanding the agility of running birds: sensorimotor and mechanical factors in avian bipedal locomotion. Integrative and comparative biology, 58(5), pp.884-893.
Reference Model:
OpenSim model of guinea fowl hindlimb (see http://locomotionlab.net/research/basic-research/)

Last edited: 17/07/2019
594 – Studying the role of different muscle models on the low level control of lower limb
Category:semester project
Keywords:Bio-inspiration, Biped Locomotion, Computational Neuroscience, Dynamics Model, Feedback, Locomotion, Muscle modeling, Programming, Simulator
Type:30% theory, 70% software
Responsibles: (MED 1 1024, phone: 37506)
(MED 1 1611, phone: 34793)
Description:Biological muscles have intrinsic stabilization features. Recent observation suggest that those features are not recapitulated in the most common muscle model used by neuro-muscular modelers, i.e. Hill Muscles. Building on the results of a previous master project, the current project aims at : - implementing a new muscle models incorporating a mechanism of force enhancement during eccentric contraction - testing the effect of this muscle model in the control of joint position on a simplistic one degree of freedom model - implementing the new muscle model in a neuromuscular model of human walking already available in the lab and study the effect of this new muscle model in terms of: - stabilization effect on gait - energy expenditure of the model The project will involve a little litterature review of new finding on muscle structure, programming of the new muscle models on two different models (a simple one segment model and a more complex 6 segment model of human walking).

Last edited: 22/02/2019

Mobile robotics

596 – Design of tugging tether for micro air vehicle manipulation
Category:semester project, master project (full-time)
Keywords:Electronics, Mechanical Construction, Motion Capture, Programming
Type:80% hardware, 20% software
Responsible: (MED 1 1215, phone: 32602)
Description:Micro air vehicles (MAVs) equipped with controllable attachment mechanisms have been shown to be maneuverable enough to navigate 3D settings and forceful enough to affect human-scale environments. Force instance, our previous investigation demonstrated two, 100 gram vehicles coordinating to open a door as a proof of concept [1]. The original investigation demonstrated forces up to 40x's the vehicle's weight using high-end hobby servos to generate sufficient forces. However, no mind was paid to the energetic efficiency and speed. A more thorough design of a tugging system would give the opportunity to better match motor capabilities to a chosen gripping mechanism to optimize for speed, incorporate simple sensing, and eventually better enable cooperation in manipulation. This project would define functional requirements and create a prototype for a 1 degree of freedom tugging system to be mounted to an off-the-shelf drone and simple attachment mechanism (e.g. magnets or hook). If time allows, three units will be constructed to demonstrate lifting and orienting an object. This would be a collaboration with Laboratory of Intelligent Systems. Relevant skills: mechanical design, motor analysis, mechatronics, microcontrollers, drones [1] Estrada, M. A., Mintchev, S., Christensen, D. L., Cutkosky, M. R., & Floreano, D. (2018). Forceful manipulation with micro air vehicles. Science Robotics, 3(23), eaau6903. Relevant press: https://actu.epfl.ch/news/small-flying-robots-able-to-pull-objects-up-to-4-2/ https://www.technologyreview.com/the-download/612344/wasp-inspired-robots-can-lift-40-times-their-own-weight-and-work-together-to/

Last edited: 24/05/2019
595 – Design and implementation of outdoor navigation algorithms
Category:master project (full-time)
Type:30% hardware, 70% software
Responsible: (GARAGE, phone: 078 843 99 35)
Description:Rovéo is an agile rover designed to easily overcome urban obstacles (see www.rovenso.com). The project will focus on designing, implementing and assessing algorithms for outdoor navigation (C++). A strong focus will also be put on the mechatronics aspects of outdoor navigation (choice of sensors and sensor fusion, waterproof protections, adaptation to existing mechanics and electronics). This is a challenging multi-disciplinary project for highly motivated students.

Last edited: 19/03/2019

15 projects found.