Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Distributed Control For Platoons of Autonomous Vehicles (Semester project/Master thesis - ETH)
In this project, you will first investigate the performance of several distributed control architectures for the deployment of formations of autonomous vehicles (e.g. platoons). You will then develop solutions to deal with a maliciously controlled vehicle hidden in the platoon.
Keywords: Control theory, Distributed control, Optimization, Vehicle platoons, Robust control, Control in adversarial environments
The last few decades have witnessed an increasing traffic demand, which often results in persistent traffic jam and serious casualties. A very promising solution is offered by the deployment of autonomous and cooperating formations of vehicles, which have the potential to significantly improve traffic capacity and reduce fuel consumption.
One fundamental task in operating large networks of autonomous vehicles is to ensure that they act as a single unit through efficient regulation of their relative distance. A significant challenge is that each vehicle can only observe a small subset of the vehicles in the network, for instance, it can only measure the relative distance to the vehicle in front of it. This framework, known as platooning, requires the development of **distributed control algorithms** and it continues to be of interest to academics, governments, and industry. Furthermore, it has been shown that a single, **maliciously controlled vehicle** can catastrophically destabilize an autonomous vehicular platoon, for example by inducing resonating oscillations that can result in collisions.
The purpose of the project is twofold. First, you will investigate and compare the different distributed and optimal control solutions for the platooning problem that have been developed in the literature. This part can be substantial and will include developing deep theoretical understanding of the challenges of optimal distributed control and thoroughly testing the algorithms in simulation. Second, you will consider the scenario of an adversarial, malicious vehicle hidden in the platoon. You will develop control strategies aimed at quickly identifying the attacker and/or restoring platoon security. Other directions can be discussed with the student on-the-go and may include: game-theoretic setting, learning-based control with unknown vehicle models, complex control tasks beyond platooning.
The last few decades have witnessed an increasing traffic demand, which often results in persistent traffic jam and serious casualties. A very promising solution is offered by the deployment of autonomous and cooperating formations of vehicles, which have the potential to significantly improve traffic capacity and reduce fuel consumption.
One fundamental task in operating large networks of autonomous vehicles is to ensure that they act as a single unit through efficient regulation of their relative distance. A significant challenge is that each vehicle can only observe a small subset of the vehicles in the network, for instance, it can only measure the relative distance to the vehicle in front of it. This framework, known as platooning, requires the development of **distributed control algorithms** and it continues to be of interest to academics, governments, and industry. Furthermore, it has been shown that a single, **maliciously controlled vehicle** can catastrophically destabilize an autonomous vehicular platoon, for example by inducing resonating oscillations that can result in collisions.
The purpose of the project is twofold. First, you will investigate and compare the different distributed and optimal control solutions for the platooning problem that have been developed in the literature. This part can be substantial and will include developing deep theoretical understanding of the challenges of optimal distributed control and thoroughly testing the algorithms in simulation. Second, you will consider the scenario of an adversarial, malicious vehicle hidden in the platoon. You will develop control strategies aimed at quickly identifying the attacker and/or restoring platoon security. Other directions can be discussed with the student on-the-go and may include: game-theoretic setting, learning-based control with unknown vehicle models, complex control tasks beyond platooning.
The main steps of the project are as follows:
1) _Skill-development_: basics of distributed control, optimal control, convex programming for constrained control, platoon modeling
2) _Literature review_: thorough theoretical understanding of the most prominent platoon control algorithms, their issues, objectives and performance.
3) _Detailed comparison_ of the chosen control solutions. Your task is to develop a theoretical framework that shows advantages and disadvantages of each solution in a unifying manner.
4) _Thorough testing_ of the algorithms in simulation (MATLAB/Simulink).
5) Explore the recent literature on _adversarial environments_. Devise a control solution that deals with the scenario of a maliciously controlled vehicle aimed at destabilizing the platoon or inducing collisions.
**Prerequisites**: Sufficient background in optimization and control theory, for instance by taking courses such as Control Systems 1 and 2, Linear System Theory, Convex Optimization, Game Theory, and performing well. Good working knowledge of MATLAB/Simulink. High motivation and interest in addressing relevant theoretical challenges at the edge of control-theory, optimization and game-theory.
The main steps of the project are as follows:
1) _Skill-development_: basics of distributed control, optimal control, convex programming for constrained control, platoon modeling 2) _Literature review_: thorough theoretical understanding of the most prominent platoon control algorithms, their issues, objectives and performance. 3) _Detailed comparison_ of the chosen control solutions. Your task is to develop a theoretical framework that shows advantages and disadvantages of each solution in a unifying manner. 4) _Thorough testing_ of the algorithms in simulation (MATLAB/Simulink). 5) Explore the recent literature on _adversarial environments_. Devise a control solution that deals with the scenario of a maliciously controlled vehicle aimed at destabilizing the platoon or inducing collisions.
**Prerequisites**: Sufficient background in optimization and control theory, for instance by taking courses such as Control Systems 1 and 2, Linear System Theory, Convex Optimization, Game Theory, and performing well. Good working knowledge of MATLAB/Simulink. High motivation and interest in addressing relevant theoretical challenges at the edge of control-theory, optimization and game-theory.
Please apply by writing to Luca Furieri at _furieril@control.ee.ethz.ch_ and including your CV and transcript of grades at Bachelor and Master's level. During the project, you will work under the supervision of Luca Furieri (PhD student at IfA) and Prof. Maryam Kamgarpour.
Please apply by writing to Luca Furieri at _furieril@control.ee.ethz.ch_ and including your CV and transcript of grades at Bachelor and Master's level. During the project, you will work under the supervision of Luca Furieri (PhD student at IfA) and Prof. Maryam Kamgarpour.