Student Projects

DISAL offers a range of student projects its three main areas of expertise: Distributed Robotic Systems, Sensor and Actuator Networks, and Intelligent Vehicles. For more information about supervision guidelines and different types of student projects please refer to this page.

Simulation-Based Policy Improvement
for Automatic Design of Behavioral Arbitrators for Khepera IV Robots

Assigned to: Jiaxuan Zhang

With the increased deployment of mobile robotics in a multitude of real-life scenarios, robots must face more and more complex situations. However, every robotic platform needs to be carefully programmed to fulfill its task, which, today, is often still done manually. Such programming operation usually represents an intense time investment and requests high expertise to eventually achieve the targeted performances in the real deployment, even more so if the application involves a multi-robot system. There are multiple approaches for the automatic generation of controllers (or specific components), whereas the nature of the control components ranges from Probabilistic Finite State Machines (PFSM) [1] to Artificial Neural Networks (ANNs) [2].  However, to automatically generate any of these controllers, powerful machine-learning techniques are necessary.

This project aims at leveraging Simulation-Based Policy Improvement (SBPI) algorithms [3] to automatically generate a PFSM arbitrator for a behavior-based controller for Khepera IV robots. Starting with a thorough literature research on the state-of-the-art of SBPI algorithms, the most promising candidate(s) shall then be selected, implemented and applied to the problem at hand. Using Webots, a high-fidelity robotics simulator, PFSM arbitrators for increasingly challenging scenarios shall then be generated and compared to both manually implemented and machine learned PFSM using different approaches. The expected effort and outcome of this work will be adjusted based on the type of the project.

Recommended type of project: Semester project / Master project

Work breakdown: 20% theory, 50% software, 30% experimentation

Prerequisites: broad interest in machine learning, control algorithms and robotics, good command of C/C++, course on Distributed Intelligent Systems recommended

Keywords: machine learning, robot control, automatic design and optimization

Contact: Cyrill Baumann

Reference:

[1] Francesca G., Brambilla M., Brutschy A., Trianni V., and Birattari M., “AutoMoDe: A novel approach to the automatic design of control software for robot swarms” Swarm Intelligence 8 (2014), 89-112.
[2] Duarte M., Sancho Moura O., and Anders Lyhne C., “Evolution of Hybrid Robotic Controllers for Complex Tasks” Journal of Intelligent & Robotic Systems (2015) 78: 463
[3] D. Wu, Q. S. Jia, and C. H. Chen, “Sample path sharing in simulation-based policy improvement,” Proc. IEEE Int. Conf. Robot. Autom. (ICRA), pp. 3291-3296, 2014.

Automatic Design of Parallel Behavioral Arbitrators for Khepera IV Robots

Assigned to: Maxime Zufferey


With the increased deployment of mobile robotics in a multitude of real-life scenarios, robots must face more and more complex situations. However, every robotic platform needs to be carefully programmed to fulfill its task, which, today, is often still done manually. Such programming operation usually represents an intense time investment and requests high expertise to eventually achieve the targeted performances in the real deployment, even more so if the application involves a multi-robot system. There are multiple approaches for the automatic generation of controllers (or specific components), whereas the nature of the control components ranges from Probabilistic Finite State Machines (PFSM) [1] to Artificial Neural Networks (ANNs) [2].  However, to automatically generate any of these controllers, powerful machine-learning techniques are necessary.

This project aims at exploring the area in between PFSMs and ANNs by leveraging the Mixed-Discrete Particle Swarm Optimization (MDPSO) algorithm  to automatically generate parallel behavior arbitrators for Khepera IV robots. After an in-depth literature research on the state-of-the-art, a grammar shall be defined which describes both involved behaviors as well as their (potentially condition-depending, non-constant) weights in format compatible with the MDPSO algorithm. Using Webots, a high-fidelity robotics simulator, an arbitrator for a flocking scenario without obstacles shall then be designed using MDPSO and compared against a manually designed solution. If successful, the approach shall then be challenged with a flocking scenario including obstacles.

 

Recommended type of project:  Semester project

Work breakdown: 20% theory, 50% software, 30% experimentation

Prerequisites: broad interest in machine learning, control algorithms and robotics, good command of C/C++, course on Distributed Intelligent Systems recommended

Keywords: machine learning, robot control, automatic design and optimization

Contact: Cyrill Baumann

Reference:

[1] Francesca G., Brambilla M., Brutschy A., Trianni V., and Birattari M., “AutoMoDe: A novel approach to the automatic design of control software for robot swarms” Swarm Intelligence 8 (2014), 89-112.
[2] Duarte M., Sancho Moura O., and Anders Lyhne C., “Evolution of Hybrid Robotic Controllers for Complex Tasks” Journal of Intelligent & Robotic Systems (2015) 78: 463

Multi-Robot Gas Distribution Mapping in Simulation

Assigned to: Thomas Peeters

Gas Distribution Mapping (GDM) is a useful technique for real world applications such as environmental monitoring. Implementations so far have focused on 2D mapping [1][2], but a 3D approach would be very valuable, given the tridimensional nature of the plume [3]. Researchers at DISAL are focusing on exploiting quadrotors in gas mapping algorithms to obtain a 3D map of the environment [4]. Moreover, using an array of robotic assets (e.g., sensor nodes, UAVs, wheeled robot), could be beneficial for improving the final gas map.

This project will explore the benefits of a multi-robot architecture in a gas distribution mapping task. Two UAVs will be employed and will move simultaneously in the environment and will both contribute to the creation of a gas map. The student will choose an appropriate strategy for navigation and path planning for this multi-robot system and will adapt the current mapping strategy to fit the new system.

Recommended type of project: semester/master project

Work breakdown: 40% theory, 60% simulation

Prerequisites: Broad interest in robotics, very good programming skill (Python/C/C++), knowledge of ROS, interest in research

Keywords: gas source mapping, quadrotors, multi-robot system

Contact: Chiara Ercolani

References:

[1] Lilienthal, Achim, and Tom Duckett. “Creating gas concentration gridmaps with a mobile robot.” Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems Vol. 1. 2003, p. 118-123.

[2] Lilienthal, Achim J., et al. “A statistical approach to gas distribution modelling with mobile robots-the kernel dm+ v algorithm.” Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems, 2009, p. 570-576.

[3] Bing, Luo, et al. “Three-dimensional gas distribution mapping with a micro-drone.” 34th Chinese Control Conference. IEEE, 2015, p. 6011-6015.

[4] Ercolani, Chiara and Martinoli, Alcherio “3D Odor Source Localization using a Micro Aerial Vehicle: System Design and Performance Evaluation” Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems, 2020, p. 6194-6200

Wind Estimation on a Quadrotor

Assigned to: Nicolaj Schmid

Gas Distribution Mapping (GDM) is a useful technique for real world applications such as environmental monitoring. Implementations so far have focused on 2D mapping [1][2], but a 3D approach would be very valuable, given the tridimensional nature of the plume [3]. Researchers at DISAL are focusing on exploiting quadrotors in gas mapping algorithms to obtain a 3D map of the environment [4]. Coupling wind direction and intensity information with gas concentration readings can enhance the performance of GDM algorithms by drawing more accurate conclusions about the plume presence and dispersion model.

The objective of this student project is to implement a wind estimation technique on a Crazyflie robot using only on-board sensors and without the help of an anemometer. The student will carry out a careful literature review, choose a suitable wind estimation technique, implement it and test it on the drone.

Recommended type of project: semester/master project

Work breakdown: 40% theory, 60% implementation

Prerequisites: broad interest in robotics, very good programming skill (Python/C/C++), knowledge of ROS, interest in research

Keywords: gas source mapping, quadrotors, wind estimation

Contact: Chiara Ercolani

References:

[1] Lilienthal, Achim, and Tom Duckett. “Creating gas concentration gridmaps with a mobile robot.” Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems Vol. 1. 2003, p. 118-123.

[2] Lilienthal, Achim J., et al. “A statistical approach to gas distribution modelling with mobile robots-the kernel dm+ v algorithm.” Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems, 2009, p. 570-576.

[3] Bing, Luo, et al. “Three-dimensional gas distribution mapping with a micro-drone.” 34th Chinese Control Conference. IEEE, 2015, p. 6011-6015.

[4] Ercolani, Chiara and Martinoli, Alcherio “3D Odor Source Localization using a Micro Aerial Vehicle: System Design and Performance Evaluation” Proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems, 2020, p. 6194-6200

[5] Neumann, P. P., et al. “Autonomous gas-sensitive microdrone: Wind vector estimation and gas distribution mapping.” IEEE robotics & automation magazine, 19(1),2012, 50-61.

Flocking of Multi-rotor Micro Aerial Vehicles via Model Predictive Control with Collision Avoidance

Assigned to: Tiffany Portela

Flocking of multi-robot systems becoming popular among the researchers from diverse fields due to its extensive applications including mobile sensor networks, coordinated payload transportation, distributed surveillance etc. [1] One of the promising methods to apply this type of flexible collective behavior is the distributed model predictive control (D-MPC) due to its optimized performance, direct handling of constraints and architectural flexibility.   Specially with the aid of the second characteristics, obstacle and collision avoidance capabilities are generally integrated into the flocking behavior [2, 3, 4].

In this project, the final aim is to design an efficient flocking algorithm for a multi-rotor Micro Aerial Vehicle (MAV) system by elaborating D-MPC and showing its performance, stability and feasibility. The sub-tasks of this project includes, first, in-depth literature survey about D-MPC algorithms suited for a flocking mission. Next, the selected algorithm should be prototyped in a quadrotor simulator that will be provided in MATLAB. After assessing the results in terms of designed metrics, a similar algorithm should be implemented in a high-fidelity, open-source framework consisting of the Webots simulator and the Robotic Operating Systems (ROS). Finally, the algorithm should be compared with one of the benchmark algorithms such as Reynold’s flocking algorithm.

Recommended type of project: Semester/master project

Work breakdown: 40% theory, 60% software

Prerequisites: Good command of MATLAB, C/C++, prior experience with ROS and Webots, knowledge about MPC will be an asset.

Keywords: Multi-rotor MAVs, flocking, distributed MPC architectures, collision avoidance

Contact: Izzet Kagan Erunsal

References: [1] J. Zhan and X. Li, “Flocking of Multi-Agent Systems Via Model Predictive Control Based on Position-Only Measurements,” in IEEE Transactions on Industrial Informatics, vol. 9, no. 1, pp. 377-385, Feb. 2013, doi: 10.1109/TII.2012.2216536.

[2] Y. Lyu, J. Hu, B. M. Chen, C. Zhao and Q. Pan, “Multivehicle Flocking With Collision Avoidance via Distributed Model Predictive Control,” in IEEE Transactions on Cybernetics, vol. 51, no. 5, pp. 2651-2662, May 2021, doi: 10.1109/TCYB.2019.2944892.

[3] D. Huang, Q. Yuan and X. Li, “Decentralized flocking of multi-agent system based on MPC with obstacle/collision avoidance,” 2019 Chinese Control Conference (CCC), 2019, pp. 5587-5592, doi: 10.23919/ChiCC.2019.8865184.

[4] Model predictive control in aerospace systems: Current state and opportunities, Eren, Utku, et al., Journal of Guidance, Control, and Dynamics 40.7 (2017): 1541-1566.

Solver Comparison for Complex and Real-time Nonlinear Model Predictive Control Problems

Assigned to: Frank Centamori

In order to optimally control the systems that have fast and nonlinear dynamics such as multi-rotor Micro Aerial Vehicles (MAVs), Nonlinear Model Predictive Control (NMPC) algorithms have to be optimized for real-time operation. For this purpose, implicit, explicit or combined solvers can be leveraged [1].  In the literature, there exist several solver packages tailored for NMPC applications such as ForcesPRO [2], ACADO [3], YANE [4] and MATLAB MPC Toolbox [5]; they adopt various numerical optimization schemes such as Sequential Quadratic Programming (SQP), Interior Points (IP) and Active Set Methods (ASM).

In this project, the main objective is to classify and compare the most efficient solvers in the literature for MAVs applications such as trajectory tracking, formation control and flocking by taking constraints into account. For this benchmarking task, a MATLAB quadrotor simulator will be provided. After performing an extensive literature survey on available solvers, at least five solvers should be tested on the MATLAB simulator. Finally, a comparative study should be carried out in terms of various performance metrics such as computational complexity, optimality, implementation simplicity, etc. [6]

Recommended type of project: Semester/master project

Work breakdown: 30% theory, 70% software

Prerequisites: Good command of MATLAB, C/C++, knowledge about MPC will be an asset.

Keywords: Multi-rotor MAVs, solver types, numerical optimization, NLP, LP, QP

Contact: Izzet Kagan Erunsal

References: [1] Model predictive control in aerospace systems: Current state and opportunities, Eren, Utku, et al., Journal of Guidance, Control, and Dynamics 40.7 (2017): 1541-1566.

[2] https://www.embotech.com/products/forcespro/overview/, accessed on 05. 05.2021

[3] https://acado.github.io/, accessed on 05.05.2021

[4] http://www.nonlinearmpc.com/index.php/documentation, accessed on 05.05.2020

[5] https://ch.mathworks.com/products/model-predictive-control.html, accessed on 05.05.2020

[6] F. Aubeck, V. Kumar, N. Murgovski and S. Pischinger, “Performance Comparison of Real-Time Solver Implementations for Powertrain Nonlinear Energy Management Optimization with MPC,” 2020 European Control Conference (ECC), 2020, pp. 483-490, doi: 10.23919/ECC51009.2020.9143843.

Development of a Realistic Quadrotor Simulation

Assigned to: Nicolas Cavedon

Webots [1] is a powerful, realistic and easy to use simulator for mobile robots, supporting many programming languages. It can simulate numerous types of robots, including industrial and mobile (wheeled, legged, flying) platforms as well as numerous sensors. However, the simulator does not currently offer many functionalities when it comes to drones. Research at DISAL relies on drones for several topics, including: 3D odor mapping, model predictive control (MPC) and autonomous indoor inspection, and would therefore greatly benefit from a faithful and complete drone simulation in Webots.

Two packages, that were developed with the objective of a complete and realistic drone simulation in mind, already exist. The first one takes care of implementing an auto-pilot [2] inspired by the PX4 [3] control stack, while the second focuses on integrating a full state estimate pipeline (e.g., ROVIO [4]) into the simulator over ROS. The role of this project is therefore to bring together auto-pilot and state estimation developments to further complete the simulation of drones in Webots. This will include adjusting the auto-pilot implementation, interfacing the state estimation pipeline with the custom-built auto-pilot over ROS (following MAVROS [5] message conventions) and carrying out tests to evaluate how faithful the simulation is to reality. Various aspects can be explored, such as varying sensor configurations, sensor and actuator noise models or external perturbations (e.g., modeling ground effect or other aerodynamic effects).

Recommended type of project: semester project

Work breakdown: 30% theory, 70% simulation

Prerequisites: Broad interest in robotics, very good programming skills (C/C++, Python). Knowledge in: control theory, ROS, Webots, git. Having attended the course “Aerial robotics” is recommended.

Keywords: Webots, simulation, quadrotors, auto-pilot, flight-controller

Contact: Lucas Wälti

References:

[1] Webots: https://cyberbotics.com/doc/reference/index

[2] Mahony, Robert, Vijay Kumar, and Peter Corke. “Multirotor aerial vehicles: Modeling, estimation, and control of quadrotor.” IEEE Robotics and Automation magazine 19.3 (2012): 20-32.

[3] PX4 auto-pilot: https://px4.io/

[4] Bloesch, Michael, et al. “Robust visual inertial odometry using a direct EKF-based approach.” 2015 IEEE/RSJ international conference on intelligent robots and systems (IROS). IEEE, 2015. website: https://github.com/ethz-asl/rovio

[5] MAVROS: http://wiki.ros.org/mavros

Controlled Physical Interactions of Aerial Vehicles with Their Environment

Assigned to: Jean Decroux

Drone-based inspection is gaining attractivity in the industry, as such an approach allows to drastically reduce inspection costs, inspection times, as well as the exposure of technicians to potentially dangerous environments. Research at DISAL is carried out on this topic with the use case of elevator shaft inspection in mind. An elevator shaft presents numerous critical components that need not only regular visual, but also physical inspections. While this type of work is entirely carried out by technicians nowadays, simpler tasks could be left to robots in order to limit the intervention of a technician when an actual defect is encountered.

This project aims at evaluating the feasibility of letting a drone physically interact with its environment. Possible interactions can for instance include spraying a surface or operating a lever. An in-depth literature review on physical interactions between aerial vehicles and their environment shall be carried out (see for instance [1,2,3,4,5]), and selected approach(es) shall be implemented in Webots [6], a powerful, realistic and easy to use simulator for mobile robots. This project should also investigate how an object that requires physical interaction shall be localized, assuming only onboard sensing.

Recommended type of project: semester project

Work breakdown: 60% theory, 40% simulation

Prerequisites: Broad interest in robotics, very good programming skills (C/C++, Python), knowledge in: ROS, Webots, git

Keywords: Webots, simulation, quadrotors, auto-pilot, flight-controller, aerial physical interaction

Contact: Lucas Wälti

References:

[1] Kamel, M., Verling, S., Elkhatib, O., Sprecher, C., Wulkop, P., Taylor, Z., Siegwart, R. and Gilitschenski, I., The voliro omniorientational hexacopter: An agile and maneuverable tiltable-rotor aerial vehicle. IEEE Robotics & Automation Magazine, 25(4), pp.34-44, 2018.

[2] Albers, A., Trautmann, S., Howard, T., Nguyen, T.A., Frietsch, M. and Sauter, C., 2010, June. Semi-autonomous flying robot for physical interaction with environment. In 2010 IEEE Conference on Robotics, Automation and Mechatronics (pp. 441-446).

[3] Fumagalli, M. and Carloni, R., 2013, November. A modified impedance control for physical interaction of UAVs. In 2013 IEEE/RSJ International Conference on Intelligent Robots and Systems (pp. 1979-1984).

[4] Fumagalli, M., Naldi, R., Macchelli, A., Forte, F., Keemink, A.Q., Stramigioli, S., Carloni, R. and Marconi, L. Developing an aerial manipulator prototype: Physical interaction with the environment. IEEE Robotics & Automation magazine, 21(3), pp.41-50, 2014.

[5] Suarez, A., Vega, V.M., Fernandez, M., Heredia, G. and Ollero, A., 2020. Benchmarks for aerial manipulation. IEEE Robotics and Automation Letters, 5(2), pp.2650-2657.
See also: https://griffin-erc-advanced-grant.eu/publications/ for more publications from this group.

[6] Webots: https://cyberbotics.com/doc/reference/index