The Institute for Communication Technologies and Embedded Systems (ICE) is jointly directed by Prof. Rainer Leupers (Chair for Software for Systems on Silicon, SSS) and Prof. Haris Gačanin (Chair for Distributed Signal Processing, DSP), who collaborate closely in research and complement each other with their expertise.
The mission of ICE is to contribute to the advancement of application-specific computing and communication systems. Emphasis is on compute-intensive embedded application domains like wireless communication infrastructure and terminals, multimedia processing, advanced driver assistance, machine learning, and industrial internet (Industry 4.0). The institute is structured into three tightly connected research groups (applications, HW/SW implementation, and design methodology, algorithms and tools), thereby having the critical mass to cover all major aspects of system design.
Key research areas at ICE:
- Wireless signal processing and networking for wireless communication systems: from algorithm to implementation
- Intelligent systems and short-range communication systems (e.g. AI-based RRM, D2D, Cyber Physical Systems, M2M, Wireless 2.0)
- Design of sensors with convergence of signal processing and machine learning methodologies
- Efficient hardware accelerators and processor architectures for embedded applications, in particular multiprocessor systems-on-chip (MPSoC)
- Electronic system-level (ESL) design tools for application-specific HW/SW systems
- Software development tools (e.g. compilers and fast simulators) for heterogeneous multicore architectures
- Hardware security
There is a tight relation between these areas, since already in the design phase of new applications and algorithms, implementation and design methodology aspects have to be considered. Furthermore, various demands have to be balanced:
- High amount of flexibility, e.g. by employing processors with customized instruction sets (ASIPs, Application Specific Instruction Set Processors)
- Extremely complex signal processing and resource-demanding machine learning methodologies, to achieve best possible transmission quality and reliability with least possible resource utilization (bandwidth, transmit power, processing)
- Transmitter/receiver algorithmic and system designs (with proof-of-concepts) for low-latency and high-reliability applications.
- Low power consumption for maximum battery lifetime and heat minimization
- Low NRE costs
- Safe and secure system operation
For industry clients, ICE offers individual consultation and training in the above domains, as well as technology transfer and case studies, such as application-specific hardware design and advanced design tools for multicore HW/SW systems.