Integrated Sensing and communications for future vehicuLAr systems..

ISLANDS – Doctoral Candidate (DC)

Chalmers University of Technology is seeking to appoint one Doctoral Candidate (DC) to join the Marie Skłodowska-Curie Doctoral Network on “Integrated Sensing and communications for future vehicuLAr systems a Network of Doctoral Students” (ISLANDS).


Position: Doctoral Candidate (DC)

Location: Gothenburg-Sweden

Working time: Full-time

Duration: Fixed term (3 years)

Living, mobility, family, and research allowances:
In agreement with the MSCA Doctoral Network financial regulations
https://ec.europa.eu/info/funding-tenders/opportunities/docs/2021-2027/horizon/wp-call/2023-2024/wp-2-msca-actions_horizon-2023-2024_en.pdf (Section 1. MSCA DOCTORAL NETWORKS, page 81)

About ISLANDS
For decades communication systems have been developed independently to radar systems, leading to a duplication of systems and devices that exploit the electromagnetic spectrum in common ways. Yet, the future wireless infrastructure will need to do more than just communications to support smart cities, intelligent mobility, infrastructure monitoring, security. It will need to perform multiple functions and will rely on high-reliability communication and sensing. The independent growth of radar and communication systems is no longer sustainable and will lead to a congestion of devices, emitters and sensors. There is a skills gap in the community to address this as communication engineers work isolated to radar engineers, and a new set of skills need to be developed. ISLANDS is a doctoral network that focuses on the theoretical and algorithmic foundations of integrated sensing and communication for the automotive sector, with the objective of developing new physical-layer and network-level solutions, to explore the fundamental limits of such technology, and to provide experimental validation and testing for the developed techniques. Specifically, ISLANDS will: 1) develop new transceiver algorithms, capable of integrating and leveraging the communication and sensing functionalities, with the purpose of achieving superior performance and energy and hardware efficiency; 2) investigate the ultimate network performance limits that the integration of communication and sensing can achieve in environments with extreme mobility; and 3) provide experimental validations of the developed techniques with proof-of-concept testbeds and realistic system-level simulators. ISLANDS will train the next generation of EU experts and leaders with specific interdisciplinary expertise, combining sensing and communications, with the aim of reinforcing European leadership in the automotive sector of the next decades. 

The Role
Supervised end-to-end learning approaches have recently been proposed to simultaneously design transmitters and receivers for ISAC and mmWave positioning systems. However, they often assume the existence of differentiable channel models for sensing and communications to be able to use them for joint transmitter-receiver training via neural networks. For the receiver side, the gradients can be computed from the output of the radar and communication channels, while access to the likelihood function for transmitter learning is not possible. Suitable alternative methods, such as reinforcement learning (RL) techniques (e.g., using policy gradient), or other methodologies, are thus needed to realize AI-based design of ISAC transmit waveforms.

The DC will reconsider the entire vehicular ISAC problem from a data-driven AI approach. This activity will enable the discovery of new transmitter and receiver structures that are optimized for scenarios for which no suitable model-based approaches exist (e.g., under severe compound hardware impairments common in resource constrained vehicular settings) or where model-based solutions are sub-optimal. The DC will also investigate supervised end-to-end learning solutions for cooperative and sidelink-based vehicular ISAC systems and the design of data-driven approaches to ISAC in cell-free massive MIMO networks.

Further information about the Ph.D. projects can be found in the following tables.





Position: CHAL-1





Title: AI-based system design for integrating sensing and communications





Scientific context: ISAC in vehicular environments





Objectives: Design data-driven ISAC transmitters and receivers for connected and automated mobility (CAM) services using reinforcement learning and supervised learning





Expected results: Novel transceivers schemes based on AI over differentiable channel models in automotive scenarios. OTA experiments





Acquire knowledge: ISAC, AI, statistical signal processing





Planned secondment(s): CNIT-3 months: design of data-driven approaches to ISAC in cell-free networks; BOSCH-3 months: investigate supervised end-to-end learning solutions for cooperative and sidelink-based vehicular ISAC systems





Ph.D. enrolment: Chalmers University of Technology





For more information about how to apply and more, please visit Chalmers website.