ERC FUNDED PROJECTS

Recent technological advances in optical microscopy have vastly broadened the possibilities for applications in the biomedical sciences. Fluorescence microscopy is the central tool for investigation of molecular structures and dynamics that take place in the cellular and tissue environment. Coupled with progress in labeling methods, these microscopes permit observation of biological structures and processes with unprecedented sensitivity and resolution. This work has been enabled by the engineering development of diverse optical systems that provide different capabilities for the imaging toolkit. All such methods rely upon high fidelity optics to provide optimal resolution and efficiency, but they all suffer from aberrations caused by refractive index variations within the specimen. It is widely accepted that in many applications this fundamental problem prevents optimum operation and limits capability. Adaptive optics (AO) has been introduced to overcome these limitations by correcting aberrations and a range of demonstrations has shown clearly its potential. Indeed, it shows great promise to improve virtually all types of research or commercial microscopes, but significant challenges must still be met before AO can be widely implemented in routine imaging. Current advances are being made through development of bespoke AO solutions to individual imaging tasks. However, the diversity of microscopy methods means that individual solutions are often not translatable to other systems. This proposal is directed towards the creation of theoretical and practical frameworks that tie together AO concepts and provide a suite of scientific tools with broad application. This will be achieved through a systems approach that encompasses theoretical modelling, optical engineering and the requirements of biological applications. Additional outputs will include practical designs, operating protocols and software algorithms that will support next generation AO microscope systems.

3 234 789 €

Start date: 2016-09-01, End date: 2021-08-31

The objective of the BEACON project is to (re-)introduce analog communications into the design of modern wireless networks. We argue that the extreme energy and latency constraints imposed by the emerging Internet of Everything (IoE) paradigm can only be met within a hybrid digital-analog communications framework. Current network architectures separate source and channel coding, orthogonalize users, and employ long block-length digital source and channel codes, which are either suboptimal or not applicable under the aforementioned constraints. BEACON questions these well-established design principles, and proposes to replace them with a hybrid digital-analog communications framework, which will meet the required energy and latency constraints while simplifying the encoding and decoding processes. BEACON pushes the performance of the IoE to its theoretical limits by i) exploiting signal correlations that are abundant in IoE applications, given the foreseen density of deployed sensing devices, ii) taking into account the limited and stochastic nature of energy availability due to, for example, energy harvesting capabilities, iii) using feedback resources to improve the end-to-end signal distortion, and iv) deriving novel converse results to identify fundamental performance benchmarks. The results of BEACON will not only shed light on the fundamental limits on the performance any coding scheme can achieve, but will also lead to the development of unconventional codes and communication protocols that can approach these limits, combining digital and analog communication techniques. The ultimate challenge for this project is to exploit the developed hybrid digital-analog networking theory for a complete overhaul of the physical layer design for emerging IoE applications, such as smart grids, tele-robotics and smart homes. For this purpose, a proof-of-concept implementation test-bed will also be built using software defined radios and sensor nodes.

1 496 350 €

Start date: 2016-10-01, End date: 2021-09-30

The aim of this project is to develop the next generation of compressive and computational sensing and processing techniques. The ability to identify and exploit good signal representations is pivotal in many signal and data processing tasks. During the last decade sparse representations have provided stunning performance gains for applications such as: imaging coding, computer vision, super-resolution microscopy and most recently in MRI, achieving many-fold acceleration through compressed sensing (CS). However in most real world sensing it is generally not possible to fully adopt the random sampling strategies advocated by CS. Systems are often nonlinear, measurements have limited dynamic range, noise is rarely Gaussian and reconstruction is not always the final goal. Furthermore, iterative reconstruction techniques are often not adopted in commercial imaging systems as they typically incur at least an order of magnitude more computation than traditional techniques. Thus there is a real need for a new framework for generalized computationally accelerated sensing and processing techniques. The research proposed here will build on the PIs recent work in this area and will develop and analyse a much richer class of hierarchical low dimensional signal models, accommodating everything from physical laws to data-driven models such as deep neural networks. It will provide quantitative guidance for system design and address sensing tasks beyond reconstruction including detection, classification and statistical estimation. It will also exploit low dimensional structure to reduce computational cost as well as estimation accuracy, challenging the notion that exploiting prior information must come at a computational cost. This research will result in a new generation of data-driven, physics-aware and task-orientated sensing systems in application domains such as advanced radar, CT and MR imaging and emerging sensing modalities such as multispectral time-of-flight cameras.

2 212 048 €

Start date: 2016-09-01, End date: 2021-08-31

The widespread adoption of the Internet and its influence on our daily life is unquestioned. Global Zettabyte traffic has rendered photonics as indispensable for the communication infrastructure. While direct signal detection has been dismissed in radio communications decades ago, it prevails in short- and medium-reach optics in virtue of its simplicity. In such an environment photonics can only rely on incremental improvements, whereas it desperately seeks for disruptive concepts. COYOTE envisions a novel coherent homodyne transceiver concept for analogue signals and access to higher-order formats with efficiencies of 10 bits/symbol. On top of this, high-fidelity transport of multi-band 5G radio signals in the millimetre-wave range up to 100 GHz will be enabled by analogue coherent photonics while mitigating energy-hungry digital signal processing. COYOTE takes one more leap and dares the contradictory full-duplex data transmission in virtue of its novel reception engine to ultimately guarantee a lean solution with greatly simplified yet flexible “hardware”. The key asset of COYOTE’s coherent engine will be a locked laser with improved coherence characteristics together with a flexible modulator-detector element, which is capable to emulate direct-detection systems in a transparent way while giving birth to novel networking concepts. Exploration of the 3D Stokes and 2D quadrature spaces through a segmented receiver architecture will boost the spectral efficiency to >10 bits/s/Hz. It is the lean and yet efficient coherent transceiver methodology of COYOTE that will remove the currently existing boundary between direct-detection and coherent systems in the midst of network reaches. By coherently “reviving” these telecom segments of integrated wireline-wireless access networks, optical interconnects for intra-datacentre connectivity and even quantum communication, an order-of-magnitude improvement in terms of spectral efficiency x reach product will be gained.

1 500 000 €

Start date: 2019-01-01, End date: 2023-12-31