Search

Health care platforms rapidly shift toward the ICT-based solutions. In this context, a wide range of consumer electronics technologies come into play ranging from robotics, embedded systems, sensors, and communication infrastructures. Driven by such observations, the H2020 RADIO project set forward a service-oriented, easily expandable design paradigm that emphasizes on how heterogenous commercial off-the-shelf (COTS) ICT technologies can be used as enablers of existing and new health-care services. Such services focus on activities of daily life (ADL) monitoring algorithms, on facilitating the indoor everyday life activities of the end user emphasizing on energy multifaceted conservation and security provision. All software components produced during the RADIO project with relative guidelines are freely available through dedicated GitHub repository (https://github.com/RADIO-PROJECT-EU).

Motivated by an increasing interest from startups in embedded Artificial Intelligence (AI) and by their limited expertise, the AITIA Project targets the development of embedded AI techniques for industrial applications. This extended abstract presents the motivation and the solutions being developed towards four use cases: smart sensors, network intrusion detection, driver-assistance systems, and Industry 4.0.

The near-data processing (NDP) paradigm has emerged as a promising solution for the memory wall challenges of future computing architectures. Modern 3D-stacked DRAM systems can be exploited to prevent unnecessary data movement between the main memory and the CPU. To date, no standardized simulation frameworks or benchmarks are available for the systematic evaluation of NDP systems. Identifying which type of high-performance 3D memory is suitable to use in an NDP system remains a challenge. This is mainly due to the fact that understanding the interactions between modern workloads and the memory subsystem is not a trivial task. Each memory type has its advantages and drawbacks. Additionally, memory access patterns vary greatly across applications. As a result, the performance of a given application on a given memory type is difficult to intuitively predict. There is no specific memory type that can effectively provide high performance for all applications.In this work, we propose a machine learning framework that can efficiently decide which NDP system is suitable for an application. The framework relies on performance prediction based on an input set of application characteristics. For each NDP system we are examining, we build a machine learning model that can accurately predict performance of previously unseen applications on this system. Our models are on average 200x faster than architectural simulation. They can accurately predict performance with coefficients of determination ranging between 0.88 and 0.92, and root mean square errors ranging between 0.08 and 0.19.

From caretaking activities for elderly people to being assistive in healthcare setup, mobile and non-mobile robots have the potential to be highly applicable and serviceable. The ongoing pandemic has shown that human-to-human contact in healthcare institutions and senior homes must be limited. In this scenario, elderlies and immunocompromised individuals must be exclusively protected. Robots are a promising way to overcome this problem in assisted living environments. In addition, the advent of AI and machine learning will pave a way for intelligent robots with cognitive abilities, while enabling them to be more aware of their surroundings. In this paper, we discuss the general perspectives, potential research opportunities, and challenges arising in the area of robots in assisted living environments and present our research work pertaining to certain application scenarios, i.e., robots in rehabilitation and robots in hospital environments and pandemics, which, in turn, exhibits the growing prospects and interdisciplinary nature of the field of robots in assisted living environment.

A network-on-chip (NoC) offers high performance, flexibility and scalability in communication infrastructure within multi-core platforms. However, NoCs contribute significantly to the overall system’s power consumption. The double-layer energy efficient synchronous-asynchronous circuit-switched NoC (CS-NoC) is proposed to enhance the power utilization. To reduce the dynamic power consumption, single-rail asynchronous protocols are utilized. The two-phase and four-phase encoding algorithms are analyzed to determine the most efficient technique. For the data layer, the two asynchronous protocols reduced the power consumption by 80%, with an increase in latency when compared with the fully synchronous protocol. However, the two-phase single-rail protocol had better performance compared with the four-phase protocol by 38%, with the same power consumption and a slight increase in area of 5%. Based on this conducted analysis, the asynchronous two-phase layer had significant power reduction yet operated at a moderate frequency. Therefore, the proposed NoC is divided into two data transfer layers with a single control layer. The data transfer layers are designed using synchronous and asynchronous protocols. The synchronous layer is designated to high-frequency loads, and the asynchronous layer is confined to low-frequency loads. The switching between the layers creates a trade-off between the maximum allowed frequency and the power consumption. The proposed NoC reduces the overall power consumption by 23% when compared with recent previous work. The NoC maintains the same system performance with an 8% area increase over the fully synchronous double-layer in the literature.

Terahertz spectrum is an emerging and promising spectrum as it provides sub-mm spatial resolution and many materials have unique fingerprints at this frequency range. However, the THz spectra suffer from high atmospheric absorption and free space path loss, limiting the propagation range. State-of-the-art THz testbed demonstrations with a bandwidth of >100 GHz are focused on a line of sight (LoS) sensing and limited to a propagation range of sub-2 m. This paper demonstrates non-LoS sensing up to the fifth order of reflection for the frequency band of 220-330 GHz with a range of >6 m.

In radar systems, the frequency range is being extended to high frequencies such as THz for sub-mm resolution. The spectrum offers high resolution but on the contrary, propagation distance and penetration depth are limited because of smaller wavelength. It suffers from higher atmospheric absorption in comparison to sub-GHz systems. In comparison to optical technology, the radar technique majorly benefits with respect to the penetration property such as cloud/smoke cover penetration and detection of concealed objects. However, the THz range and synthetic aperture radar (SAR) imaging of concealed objects are not very well established. Therefore, this paper examines this property at THz. A testbed has been set up with a bandwidth of 110 GHz at a carrier frequency of 275 GHz. The imaging is performed of a very small metal object. Firstly, the sub-mm resolution is validated with the experiment after that the range and SAR imaging are performed in which this object is covered with different types of materials. The backscattered data is processed with the image reconstruction algorithms and the results are presented in this paper with respect to sub-mm resolution and detection. © 2020 IEEE.

Indoor THz Synthetic Aperture Radar (SAR) is an emerging research domain. It expands the conventional SAR applications to sub-mm resolution imaging and localization due to high available bandwidth and compact antennas at this spectrum. In general, high resolution is always preferred but it comes with the cost of complexity of the radio front-end chain. Therefore, in this paper three SAR systems with different carrier frequency $f_{c}$ and bandwidth $B_{w}$ are studied. These systems are set-up with $f_{c}= 230\ \text{GHz}$ and $B_{w}= 20 \text{GHz}, f_{c}= 245\ \text{GHz}$ and $B_{w}= 50\ \text{GHz}$ and lastly, $f_{c}=275\ \text{GHz}$ and $B_{w}= 110\ \text{GHz}$ . The paper presents the comparison between these systems w.r.t. the focused SAR image quality associated with the spatial resolution.

The advancement in technology has led to the integration of multiple cores onto a single chip, calling for new means of on-chip communication substituting the simple physical wires. Networks-on-Chip (NoCs) were introduced as the emerging solution for a high performance and scalable communication infrastructure in the multi-core platforms. However, with the increasing chip complexity heading to thousand cores, NoCs are significantly contributing in the chip power consumption. This paper presents the energy efficient sync-async Circuit-switched NoC (CS-NoC) with a synchronous control sub-router and an asynchronous data transfer subrouter. CS was chosen to benefit from the fact that CS NoC presets the data-path. The proposed NoC is synthesized using Synopsys Design Compiler for 65nm technology. The obtained results for 65nm show a reduction of 80% in the dynamic power and 7% for the leakage power as compared to pure synchronous CS-NoCs.

This paper presents a novel signal processing implementation for a dielectric materials characterization method using the hardware/software co-design methodology for application acceleration and rapid deployment on heterogeneous embedded platforms. The dielectric materials characterization method uses Frequency Modulated Continuous Wave radar systems for measurement purposes. Since the dielectric materials characterization method is intended to be used in a mobile materials characterization setup. The required signal processing algorithm of the materials characterization method is implemented on a Xilinx Zynq Ultrascale+ Multiprocessor System-on-Chip. Whereas, the intensive computing part of the algorithm is implemented on the hardware partition using high level synthesis design methodology for purposes of acceleration. The remaining parts of the algorithm are executed in software on the processing system part of the multiprocessor system on chip platform. The execution time, hardware resources utilization and energy consumption of the proposed co-design are determined experimentally. Furthermore, several optimization techniques are explored in order to reduce the total energy consumption and increase the data transfer bandwidth between the hardware and software partitions. The experimental results show a maximum speed up of 7.25 for the proposed co-design over the software reference model with an energy consumption less than 10 Joule.

Indoor Terahertz Synthetic Aperture Radar (SAR) is an emerging technology for material characterization, high resolution imaging and localization. In comparison to optical technology, it provides benefits in hazardous scenarios such as fire in a building as objects inside the building can be characterized and localized. The principles of SAR are well established but the main challenge lies with extending this technology to high frequencies and indoor environment. To investigate this technology, imaging geometry and system parameters have to be evaluated respectively. This paper explains the signal processing of SAR and presents the imaging geometry for an indoor scenario. Further, it evaluates the parameters for high resolution imaging and localization. Based on these parameters, system design has been simulated and results of 2D high resolution indoor SAR imaging at 300 GHz are presented. Additionally, the proof of theoretical resolution across the range and azimuth is shown with the simulation results.