PhD Research Project: Optimization of Real-Time Communication in Data-Intensive Cyber-Physical Systems

Project leader: Soheil Samii, Senior Associate Professor

Context and Research Area

Data communication and computing needs in cyber-physical systems (CPSs) are increasing rapidly. This is caused by the enormous evolution of embedded software and electronics based features in many industries, including the automotive, automation, and aerospace sectors. This drives unprecedented increases in data bandwidth requirements from sensors and actuators, but also between the embedded computing system and edge/cloud nodes. For example, in automotive systems with driver assistance or autonomous driving functions, the in-vehicle data communication network must support a massive amount of bandwidth to support computer vision, sensor fusion, and perception and planning tasks. This includes the integration of sensors like cameras, radars, and lidars. In addition to supporting high data bandwidths, there are strict real-time and reliability requirements on the network. As vehicles are more and more connected to the outside world, several new application areas drive the need for low-latency, high-reliability cellular communication. The implementation of many future CPS is enabled by Time Sensitive Networking (TSN) for Ethernet and 5G/6G cellular. These technologies are becoming mature to be used in real-time and safety-critical applications. In addition, several other communication technologies and architectures are being considered in CPS. Examples include Serializer/Deserializer technologies, PCI Express, Remote DMA, and distributed shared memory architectures. The use of the aforementioned technologies and protocols introduced a huge and complex design space that needs to be explored. Towards this, the project will address analysis, design, and optimization of real-time communication in future CPS. TSN Ethernet networks have several protocols and configuration parameters that need to be carefully optimized to meet application requirements on real-time performance, reliability, and availability. The project will address optimization problems related to scheduling and routing, queue allocation, static or dynamic preemption decisions, redundancy allocation, and synchronization. To expand beyond the embedded communication network of a CPS, the project will consider wireless and cellular networking technologies, which introduce allocation decisions that need to be optimized by considering trade-offs related to task allocation to edge/cloud devices, feature availability, and system and data costs.

Research Methodology and Expected Results

The research project is expected to develop new analysis, synthesis, and optimization methods as part of the overall system design process. New formal models will be developed to capture the key characteristics of applications, system, and communication technologies. The models will be validated through simulation as well as through realistic use cases and case studies inspired by practical applications in industry. The researcher will analyze key trade-offs through comprehensible examples and use cases, followed by the application of state-of-the-art optimization methods and tools to optimize the large and complex space of configurations parameters related to communication synthesis in future CPS. Simulation frameworks and proof-of-concept demonstrators will be developed to validate and illustrate applicability to real-world problems and use cases. The project results will be presented at international conferences and, as applicable, at industry conferences and standardization forums. PhD students in this project can expect to become well equipped to pursue a research career in both academia and industry.

Qualifications of Prospective PhD Student

For general qualifications, see the CUGS announcement of the PhD student positions. It is helpful if candidates have some background in computer networks discrete mathematics and discrete optimization techniques. The project will rely on the candidate having the ability to develop and debug software independently. When selecting the candidate, we will also consider verbal and written English language proficiency.

Page responsible: Soheil Samii
Last updated: 2023-05-02