Ongoing Projects

Wearable haptic interface to control a drone

Hugo Kohli (CH)

Learning a teleoperation task normally requires a high cognitive effort as well as extensive training. In the frame of the Symbiotic Drone project, we are designing a new Human-Robot Interface (HRI) based on wearable technology in order to ease the process of learning a new interface as well as the dynamics of the controlled machine. For this project, we want to test how effectively a human is able to learn to use his own body instead of a predefined hardware interface (i.e. a joystick). In particular, a motion controller based on hand motion was implemented and used to control the position of the drone, following a path through obstacles in a simulated environment. At the same time, we implemented a haptic device (a glove) capable of transmitting relevant information about the navigation of the robot and the surrounding environment during flight. The glove shows promising capabilities for the teleoperation of a flying robot when limited or no visual feedback is available from the machine. A hardware platform was implemented for further testing on a real quadrotor. For this project, we want to extensively test the developed wearable interface and assess its potentialities for teleoperation of single and multiple quadrotors, assessing learning time, efficiency and cognitive effort needed, compared to standard hardware devices such as joysticks. Students interested in flying robotics, haptics, and wearable technology are encouraged to apply.

Type: Master project
Period: 16.09.2019 - 16.03.2020
Section(s): Robotics Microengineering
Type of work: 20% hardware, 40% software, 40% experimental
Requirements: Python, ROS C++
Subject(s): Wearable, telerobotics, haptics
Responsible(s): Matteo Macchini, Fabrizio Schiano

Development of an untethered soft robotic gripper

Maxime Boutot (Microengineering)

Soft materials are revolutionizing the field of robotics. Soft robots are highly resilient and safe for interacting with humans and natural elements. A possible strategy to design soft robots – currently pursued in our lab – is to use heterogeneous soft deformable modules that can be assembled into the robot morphology. In this proposed semester project, we seek to use the current kit of soft modules to develop a novel completely untethered soft robotic gripper, overcoming current state of the art solutions using external power supply, actuators or controllers. The student will first choose and fabricate the modules necessary to assemble the gripper, selecting them among the already designed modules. She/he will use lightweight materials, 3D printing technologies and off-the-shelf electronic components. Secondly, he/she will assemble and mechanically characterize the gripper. Finally he/she will develop a simple open loop Arduino code to conduct grasping tests. As a bonus objective some proof of concept sensing capabilities could be implemented.

Type: Semester project
Period: 15.09.2019 - 30.01.2020
Section(s): Robotics Microengineering Mechanical Engineering Materials Science and Engineering
Type of work: 10% theory, 80% hardware, 10% software
Requirements: CAD (preferably inventor or SolidWorks), good understanding of mechanisms and materials, Arduino programming, 3D printing experience
Subject(s): soft robotics, robotic grippers
Responsible(s): Davide Zappetti, William Stewart

Design, development, and testing of an adaptive drone test rig

Arthur Alain Bernard Gassner (Microengineering)

At the Laboratory of Intelligent Systems (LIS) at EPFL, we design drones and algorithms to make them fly.
Every engineer using drones knows how important is to have the right control parameters. The control parameters change according to several factors and especially due to the mechanical structure of the drone. The goal of this project is the identification of the relationship between the control parameters and the mechanical structure of the drone (a rotary-wing UAV). In order to achieve this goal the student’s role will be twofold: (a) design and development of a drone test rig able to adapt to different drone morphologies and (b) develop a software, based on ROS, Python and PX4, able to give attitude commands to the drone, analyze its response, and give guidance to the user on which parameters to change. The whole setup will be tested in our Motion Tracking Hall (also known as the Drone Dome). Depending on the quality of the results, we could publish a conference paper related to the topics tackled in this project.

Type: Semester project
Period: 15.09.2019 - 30.01.2020
Section(s): Robotics Microengineering
Type of work: 20% hardware, 50% software, 30% experimental
Requirements: Control - 3D printing basics - Python/C++ - Matlab - ROS knowledge is a plus - previous experience with drones is a plus
Subject(s): Control - Drones - Testing
Responsible(s): Fabrizio Schiano, Przemyslaw Kornatowski

Wearable telerobotics - study on different limb fitness

Jan Alexander Yamann Frogg (Microengineering)

Wearable interfaces based on body motion can make the teleoperation of a distal robot easier to learn for inexperienced users. When a person is asked to move freely following their natural intuition as if they were to control a flying robot (in this case a fixed-wing drone), the most common strategy is to mimic the robot's attitude with their torso. This is true if the point of view corresponds to the drone's frontal camera. This effect could be explained by the fact that the user 'embodies' in the robotic envelope transferring his body ownership into the machine. For this project, you will test the effectiveness of a system capable of creating a personalized human-robot interface based on body motion. Additionally, you will evaluate and quantify the variability of the interface when the person is constrained to use a defined body part instead of being free to use their whole body. Also, we want to understand if the instinctive choice of torso movements affects teleoperation performance. The project can be summarized in two main tasks : (i) you will design, prepare and run an experiment finalized at answering the above scientific questions. (ii) you will systematically analyze the obtained data on body motion and teleoperation performance.

Type: Semester project
Period: 15.09.2019 - 15.01.2020
Section(s): Robotics Microengineering School of computer and communication sciences
Type of work: 50% experimental, 50% data analysis
Requirements: Python, basics of statistics
Subject(s): Aerial robots, human-robot interfaces, body motion
Responsible(s): Matteo Macchini, Fabrizio Schiano

Development of Software Environment for Flapping Wing Test Rig

Charles Coster (Microengineering)

At the Laboratory of Intelligent Systems in collaboration with the Animal Flight Lab at Lund University, we have developed the mechanical model of an avian-inspired flapping wing. The next step is to actuate the motors to mimic the flapping behavior of birds. Therefore, the goal of this semester project is to (i) perform a literature survey on avian flapping wing behavior (ii) select a suitable microcontroller and implement a suitable control strategy, and (iii) develop a graphical user interface for a simple use of the flapping mechanism.

Type: Semester project
Period: 17.09.2019 - 18.01.2020
Section(s): Robotics Microengineering Electrical and Electronic Engineering Mechanical Engineering School of computer and communication sciences
Type of work: 10% literature, 70% software, 20% testing
Subject(s): Embedded programming, bioengineering
Responsible(s): Enrico Ajanic, Olexandr Gudozhnik

Design and Testing of Flapping Wing Test Rig

Paul Megevand (ME)

At the Laboratory of Intelligent Systems in collaboration with the Animal Flight Lab at Lund University, we have developed the mechanical model of an avian-inspired flapping wing. A first prototype was successfully developed in a previous project. The next step is to develop a second prototype and to test it in a wind tunnel environment. Hence, this semester project involves (i) a thorough literature survey, (ii) improvement of the current design followed by the manufacture of the system, and (iii) comprehensive testing in the wind tunnel.

Type: Semester project
Period: 17.09.2019 - 18.01.2020
Section(s): Robotics Microengineering Mechanical Engineering Materials Science and Engineering
Type of work: 10% theory, 50% hardware, 40% testing
Requirements: CAD
Subject(s): Mechanical design, bioengineering, aerodynamics
Responsible(s): Enrico Ajanic, William Stewart

Agile Flight Experimentation

Lucas Streit (Robotics)

LIS has recently developed a super-agile, avian-inspired unmanned aircraft. We are now beginning to explore research avenues that stem from the increased agility of the platform. This project will investigate the applications of increased agility starting with a dynamic model, moving to in-lab experiments, and finally outdoor flight tests. The student will have ample opportunity to demonstrate and learn a lot of hands-on skills in experiment design, UAV construction techniques, and their operation. The successful candidate will have experience in mechanics and electronics as well as good self-motivation.

Type: Semester project
Period: 15.09.2019 - 30.01.2020
Section(s): Robotics Mechanical Engineering
Type of work: 5% Theory, 5% Simulation 20% Electronics, 20% Hardware, 50% Experimental
Subject(s): Aerial Robots, Experimentation, Simulation, Mechanics
Responsible(s): William Stewart

Drone swarm simulator in Matlab

Andrea Giordano (Mechanical Engineering)

At the Laboratory of Intelligent Systems, we develop swarming algorithms for drones. The goal of this project is to complete the development of an existing drone swarm simulator written in Matlab. The first phase of the project will involve the conversion of functional code to classes, basing on the example of Randal Beard [1] available online. A uniform and appropriate selection of the variable/function names should make the code easy readable. State-of-the-art swarming algorithms are already implemented for you (Reynolds, Olfati Saber, Vicsek algorithms). However, proper integration, testing, and documentation are needed for the release. A challenge for most simulation environments is computational time. By comparing the current implementation with other available integration methods and by profiling with Matlab in-built toolboxes, you will characterize the running time of the software and identify strategies to improve it. A particular attention to the development of graphical function and a GUI is important to make the simulator easier to use. Finally, for increasing the usability of the code from a large community, you will produce example files that explain the main features of the simulator. Together with a schematic documentation, you will produce GIFs and plots that illustrates the results of your example files. Everything will be uploaded and documented on Github, following the standard website template. Previous experience with the cited software is required. [1]

Type: Semester project
Period: 01.09.2019 - 01.02.2020
Section(s): Robotics Microengineering Electrical and Electronic Engineering Mechanical Engineering School of computer and communication sciences
Type of work: 20% theory, 40% software, 40% testing
Requirements: Physical modelling and programming skills (Matlab, Simulink), knowledge on git
Subject(s): Swarm robotics, multi-agent control
Responsible(s): Enrica Soria, Fabrizio Schiano

Adaptive Human-Robot Interface for Aerial Robots

Xiao Zhou (Microengineering)

In the Laboratory of Intelligent Systems, we are working on the development of motion-based intuitive interfaces for the control of mobile robots. Such interfaces are based on the observation of people’s instinctive body motion while observing robot movements. Despite the high intuitiveness of motion-based interfaces, users would profit of an adaptive systems able to cope with their changes of preferences over time. For this reason, we are studying the design of an adaptive interface which learns online the user’s behavior during teleoperation and changes accordingly, to simplify the task. For this project, you will implement an autopilot for a fixed-wing drone simulator, which will play the role of a human operator, and a simulated interface (like a remote controller) that the autopilot will use. Later, you will implement a machine learning algorithm to study the performance of the autopilot during flight and adapt the interface to improve performance. Students interested in aerial robotics, human-robot interfaces and machine learning are encouraged to apply

Type: Semester project
Period: 16.09.2019 - 15.01.2020
Section(s): Robotics Microengineering Electrical and Electronic Engineering School of computer and communication sciences
Type of work: 30%+theory +50%+software +20%+testing
Requirements: python +basics+on+control+systems +basics+in+machine+learning+(possibly+reinforcement+learning)
Subject(s): aerial+robotics +drones +human-robot+interfaces +machine+learning
Responsible(s): Matteo Macchini, Davide Zambrano

Exploring parameters robustness in soft modular robots with Evolutionary Algorithms

Nathan Holzapfel (Robotics)

Soft modular robots are versatile systems that can be assembled into different task-specific morphologies to safely locomote and manipulate beside or cooperatively with humans or in un-constructed environments. In soft modular robots, the control, the mechanical properties and the morphology are intimately related to each other thus representing a complex design problem. For this, a, valuable solution is using Evolutionary Algorithms (EA) to explore the variety of, possible designs for proof-of-concept locomoting tensegrity modular robots. However, filling the gap between evolved robots and real platforms remains a challenge. Indeed, the real robots can have small differences in terms of morphologies, material properties and actuation. Such differences can affect the dynamics of the robot., The project aim is to theoretically investigate and explore new methodologies for evolving robust robot behaviors that can be resilient to small parameters change. The student will first experiment with EAs in an already available evolutionary robotics software framework (TenSoft) to analyze the most critical parameters in the evolved robots. For that, he/she will use the TenSoft simulation environment to perform multiple experiments to investigate whether the evolved behaviors are robust to parameters change (i.e. noise). Thereafter, he/she will explore EA solutions to evolve robust modular robots. Depending on the initial experimental results, this last phase could involve the co-evolution with Reinforcement Learning algorithms. The project goal is to quantitatively assess the robustness of the robot behavior and the tolerance within the simulated results.

Type: Semester project
Period: 24.09.2019 - 24.12.2019
Section(s): Robotics School of computer and communication sciences
Type of work: 50% theory, 50% software
Requirements: C++, python
Subject(s): robotics, evolutionary algorithms, physics simulations, reinforcement learning
Responsible(s): Davide Zambrano, Davide Zappetti

Wing folding mechanisms - design and manufacturing optimization

Simon Honigmann (Robotics)

The need to use winged drones in confined spaces presents many difficult challenges, particularly wingspan, and overall vehicle volume reduction. One method to address this is the ability to actively change morph the wings by folding them and thereby reducing the required operating space. With this project, the student will work on a previously developed concept of a foldable wing and will improve the design in terms of morphing repeatability, energy efficiency, component weight, and robustness to external forces. The student will design and manufacture a prototype that will be directly compared with the already developed prototype and provide useful insights in terms of design and manufacturing choices of the new design.

Type: Semester project
Period: 15.09.2019 - 30.01.2020
Section(s): Robotics
Type of work: 10%+Theory +10%+Software +80%+Hardware+
Requirements: Design +Manufacturing +Prototyping +CNC+Machining +3D+Printing
Subject(s): Robotics +Mechanics +Mechatronics +Design +Manufacturing+
Responsible(s): Charalampos Vourtsis, William Stewart

Fixed-Wing Perching

Luca Guarino (Robotics)

Perching today is studied predominantly with multirotor vehicles with few projects conducted using fixed-wing aircraft. The aim of this project is to implement perching on fixed-wing aircraft using technologies currently under development at the Laboratory of Intelligent Systems. The structure of the project follows. The student will begin by conducting a thorough study of the state-of-the-art on relevant topics. Once complete, the student will develop a dynamic model of fixed-wing perching. The output conditions of this model will be used to predict the airframe loading on the vehicle at impact. The student will use these results to design and build a mechanism for perching. Experiments will be conducted in the laboratory to characterize relevant parameters of the design. Finally, the student will integrate the hardware into a test vehicle conduct a flight test.

Type: Master project
Period: 01.10.2019 - 01.04.2020
Section(s): Robotics
Type of work: 33%+theory +33%+hardware +33%+experimentation
Subject(s): Mechanics +Dynamics +Materials +Soft+Robotics
Responsible(s): William Stewart, Egor Piskarev

Audio Instructions and Video Streaming for Delivery Drones

Vaibhav Gupta (Robotics)

At the Laboratory of Intelligent Systems (LIS) we are developing drones for last-cm delivery. These delivery drones are fully autonomous and can be monitored in real-time with the help of a web-application framework named Dronistics. To facilitate this operation of the drone and communication between a sender and recipient, and also to provide real-time instructions to the recipient, a camera, speaker and microphone will be deployed on the drone. The goal of this project is to extend the existing video-streaming solution in real drone and also to implement an algorithm to provide real-time instruction. The first goal of this project is to perform a comparative study of protocols, hardware, and software for this implementation. Then the existing streaming solution the algorithm should be adapted to run on a Linux-based external computer (Raspberry Pi Zero). The second goal is to develop the hardware and software to provide the pre-determined real-time instructions to the recipient. The third goal is to implement a safety system that prevents the arming of the propeller if the cage is not closed properly. Finally the implementation of the above work in a real delivery drone and test it in a real scenario. The sender should initiate/join the communication from a web-based interface and the person next to the drone should initiate the communication by clicking a button on the drone.

Type: Semester project
Period: 25.09.2019 - 25.03.2020
Section(s): Robotics
Type of work: 25% software 25% hardware 25% integration 25% testing
Subject(s): Software (JS), Hardware (Rpi-Z)
Responsible(s): Anand Bhaskaran, Przemyslaw Kornatowski