Logic List Mailing Archive

PhD student position in situated autonomous systems, Gothenburg (Sweden)

A fully funded 5-year Ph.D. position in Computer Science and Engineering
with the subject Space-Aware Synthesis of Situated Autonomous Systems
open at the Department of Computer Science and Engineering, Chalmers |
University of Gothenburg, Sweden. We are looking for excellent and
motivated candidates. This is a joint call at our department with several
projects and is limited to a maximum of 4 projects being funded in this
round. Thus, only highly qualified applicants will be funded.

The Ph.D. position that I am advertising will be hosted by the Formal
Methods (FM) research unit and supervised by Dr. Yehia Abd Alrahman.  Our
unit (and department as a whole) has excellent research groups that work on
Modelling and Verification problems, that may intersect with research in
control and AI.

The objective of the Ph.D. project is to develop novel methods and 

algorithms for automatic production of correct-by-design situated 

autonomous systems. That is, systems that live on autonomous machines 

(such as robots) and populate a shared physical space. Such machines can 

influence each other by merely moving/modifying the space. Situated 

Autonomous systems are expected to operate while co-existing with other 

machines and/or humans. Thus, there is a pressing demand to make them 

safe, reliable, and robust to sudden changes.

Consider for example autonomous cars: these cars share the road as a 

driving space and need to handle all traffic situations such as changing 

lanes, negotiating junctions, etc., and additionally to respond robustly 

to the driver and/or a pedestrians sudden intervention, while ensuring 

the safety on the road. According to the automotive industry standard 

ISO26262, a self-driving car may need to be tested for 255 million miles 

without having accident to be verified fully safe on the road. Clearl

testing is very expensive and is shown to have low coverage in distributed 

settings. Indeed, it is very hard to reproduce the error by simply testing 

for a longer time.

We will use Formal Methods instead, and more specifically Controller 

Synthesis as a design technique. Our goal is to extend existing synthesis 

techniques to permit native reasoning about time-space related design 

properties. The work in this project is on the intersection of computer 

science, control theory, and AI.

   The thesis will be focused on several aspects of modelling, reasoning, 

and verification of Situated Autonomous Systems. More specifically, we 

will target the following:

     - Define mathematical formalisms that describe the dynamics (or the
     semantics) of the system both in Space and Time. The main challenge he
re is
     to exploit the finite variability of the physical space to provide
     representative and compact discrete abstractions that account to the
     movement in space.
     - We will study Spatio-Temporal logics and identify efficient fragments
     that permit reasoning about space mobility as time passes.
     - Provide online synthesis algorithms that natively analyse a snapshot
     of the currently observed space and provide short-time horizon control

Ideal candidates are expected to have knowledge in one or more of the
following topics:

     - Spatio-temporal reasoning.
     - Controller synthesis or Supervisory control (Knowledge in model-based
     reinforcement learning techniques is a plus)
     - Concurrency or applications of Automata in Synthesis and Control.

Please use the instructions in the Joint call (Link below) to apply. More
information about the project are available in the project link below.


Project description: can be found here Inst.doktorander

Joint call:

Note: If you are interested, please forward any informal inquiries directly
to me, before submitting your application.

Many Thanks,

Plain text email: yehiaa [at] chalmers [dot] se
[LOGIC] mailing list
Archive: http://www.illc.uva.nl/LogicList/

provided by the Division for Logic, Methodology and Philosophy of Science a
nd Technology (DLMPST) in cooperation with the ILLC at the Universiteit van