Aarhus University Seal / Aarhus Universitets segl

Electrical and Computer Engineering

Machine learning for pose estimation of soft biological items

Jens Overby
Jens Overby

This PhD project will research methods in Machine Learning for Pose Estimation of multiple, partially occluded, soft biological items in camera images. The application of such poses is in new robotic solutions for the fish and poultry processing industries, where currently, items are transported between processing equipment by static conveyor systems.

In the future, these logistic systems will be dynamic by use of mobile robots at the same high level of automation as in the manufacturing industry today. The handling, alignment and process-infeed of soft items picked from bins and conveyors will be done by robots.

Identifying spatial properties of soft items, however, is challenging from a research point of view. Also, an algorithm for this purpose needs to be flexible and easily configurable for new tasks. Therefore, extended research is needed in the field of unsupervised and semi-supervised machine learning.

 

ABOUT THE PROJECT


Project title:
Machine learning for pose estimation of soft biological items

PhD student: Jens Overby

Contact: jens.overby@eng.au.dk

Project period: May 2019 to April 2022

Main supervisor: Prof. (Docent) Henrik Karstoft

Co-supervisor: Rasmus Skovgaard Andersen, Marel A/S

Section: Electrical and Computer Engineering


Secure data compression and analytics for Internet of Things

Gajraj Kuldeep
Gajraj Kuldeep

The IoT ecosystem enables massive sensor data acquisition, transmission and storage, as well as computation to perform data analytics for diverse smart monitoring, process automation and control services. The ever-increasing data generation brings many critical challenges to the communication, storage and computing infrastructure. The current infrastructure is not adequate to tackle these challenges in the long run. Firstly, the vast amount of generated data has to be represented more efficiently. Secondly, light-weight communication and storage secrecy is needed.

The goal of the project is to design novel schemes which provide both compression and security by using the advanced signal processing techniques. The proposed schemes will be implemented in IoT devices and a prototype will be delivered. The performance of the proposed scheme will be tested and evaluated through standard randomness tests, and energy constraints and computational requirements will be considered.

 

ABOUT THE PROJECT


Project title:
Secure data compression and analytics for Internet of Things

PhD student: Gajraj Kuldeep

Contact: gkuldeep@eng.au.dk

Project period: April 2019 to March 2022

Main supervisor: Assoc. Prof. Qi Zhang

Section: Electrical and Computer Engineering


Low-power hardware implementation of Spiking Neural Networks (SNNs) for implantable neuromorphic chips

Margherita Ronchini
Margherita Ronchini

Epilepsy is a particularly challenging neurological disorder caused by relentless brain damage and aberrant rearrangements of brain wiring. In the light of the dynamic nature of the disease, the medications may stop working and neuromodulation may not completely suppress seizures. Above all, current treatments are symptomatic, while healing the epileptic brain remains unsolved so far.

This project is born from the desire to establish a new paradigm. The goal is to drive self-repair of dysfunctional brain circuits via by the symbiotic integration of bioengineered brain tissue, neuromorphic microeletronics and artificial intelligence.

Our team is taking care of the design, implementation and testing of a low power chip for brain signal recording and processing. The chip’s task is to control the way that the graft tissue develops and interacts with the host. The chip will be neuromorphic, meaning that it will mimic the behaviour of biological neural networks and resemble their structure. The neurons will be implemented through CMOS connected by memristors-based synapses.

 

ABOUT THE PROJECT


Project title:
Low-power hardware implementation of Spiking Neural Networks (SNNs) for implantable neuromorphic chips

PhD student: Margherita Ronchini

Contact:  m.ronchini@eng.au.dk

Project period: April 2019 to March 2022

Main supervisor: Assoc. Prof. Farshad Moradi

Co-supervisor: Prof. (Docent) Preben Kidmose

Section: Electrical and Computer Engineering


Optimal strategies for efficient autonomous drone collaboration

The operation, surveillance and maintenance of high voltage transmission lines in power systems have been a manual and costly process for decades. The process incurs not only operational expenditures but also human capital risks. With the adventure of the Internet of Things (IoT) and drone technology, this surveillance operation can be automated through the collaboration of autonomous systems. This PhD project aims to research the network structure and its protocol architecture of autonomous drones (swarm). Collaboration strategies for drones will be implemented that permit safe and efficient surveillance operation of transmission lines in the power grid.

Analytical and numerical simulations to validate proposed solutions will be considered. Promising solutions obtained from the analytical framework are expected to be prototyped and tested in a laboratory environment. There will be access (drones4energy.dk) to drone equipment and a drone test facility for piloting and field trials.

 

ABOUT THE PROJECT


Project title:
Optimal strategies for efficient autonomous drone collaboration

PhD student: Liping Shi

Contact: liping@eng.au.dk

Project period: March 2019 to Feb 2022

Main supervisor: Assoc. Prof. Rune Hylsberg Jacobsen

Section: Electrical and Computer Engineering


Design of self-powered embedded wireless corrosion instruments

Jaamac Hassan Hire
Jaamac Hassan Hire

This project aims at exploring, researching and developing a novel embeddable, long life-time, ultra-low power, high resolution sensing principle for structural health monitoring (SHM), mainly in reinforced concrete (RC). The objective of the sensor is to output the corrosion rate wirelessly to a remote device, e.g. a computer, where the data is available for the owners of the structure. Corrosion damages reduce the service life of structures and can create serious safety hazards leading to fatal consequences. Today, in corrosion measurements, only very indicative and error-prone sensors exist which are based on technology developed half a century ago with slow processing, hig power usage and large formfactor. They are therefore not an attractive commodity for owners to implement. Thus, this project will challenge this status quo.

In accordance with recent years’ increased development of Wireless Sensor Networks (WSN) and Internet of Things (Iot), utilising such platforms will have tremendous benefits, such as significant lower installation cost and direct real-time access to data. This makes it possible to centralise, process and analyse a large amount of data collected from remotely placed structures, for reliable data interpretation that complies with international standards. This can then be used to launch in-time repair and maintenance operations. By further utilising energy harvesting (EH) from omnipresent sources, as the corrosion process itself, power requirements for the sensor node will be met.

 

ABOUT THE PROJECT


Project title:
Design of self-powered embedded wireless corrosion instruments

PhD student: Jaamac Hassan Hire

Contact: jhh@eng.au.dk 

Project period: Feb 2019 to Jan 2022

Main supervisor: Assoc. Prof. Farshad Moradi

Co-supervisor: Morten Wagner, FORCE Technology

Section: Electrical and Computer Engineering


Meteorological forecasting using deep learning on satellite images

Andreas Holm Nielsen
Andreas Holm Nielsen.

Weather forecasting is an extremely important part of modern society and plays a vital role in many real-world applications. Modern weather forecasting relies on a combination of complex numerical computer models, observations of the atmosphere and pattern-recognition by meteorologists with an exceptional knowledge of the physics underlying atmospheric processes.  The numerical weather prediction (NWP) models use supercomputers to create extremely large simulations of the atmosphere and its evolution in the future based on our best understanding of physics and fluid dynamics. Due to an incomplete theoretical knowledge of these processes and the inherent chaotic nature of the atmosphere, learning latent representations of the atmospheric processes using unsupervised learning can potentially extend our understanding beyond physics and climatology to improve upon the forecast accuracy of the NWP models.

This project is written in collaboration with Danske Commodities A/S and will focus on weather forecasting using deep learning approaches on time-series satellite images, applying state-of-the-art architectures within Generative Adversarial Models and/or convolutional- and recurrent neural networks to generate future sequences of satellite images. From these images, several important meteorological variables can be extracted as features which can be combined with other types of atmospheric measurements to perform supervised learning tasks in relation to weather forecasting.

 

ABOUT THE PROJECT


Project title: 
Meteorological forecasting using deep learning on satellite images

PhD student: Andreas Holm Nielsen

Contact: ahn@eng.au.dk

Project period: Feb 2019 to Feb 2022

Main supervisor: Prof. (Docent) Henrik Karstoft

Co-supervisor: Assoc. Prof. Alexandros Iosifidis

Section: Electrical and Computer Engineering


Efficient Deep Learning approaches for Unmanned Aerial Vehicles

Negar Heidari
Negar Heidari

Recent advances in Machine Learning have enabled us to target and successfully solve many challenging problems, most notably problems related to computer vision applications including image/scene recognition, object detection and recognition and human action localisation and recognition.

However, the current state of the art solutions based on deep neural networks require heavy computations, high memory footprint and long training processes. These requirements are restrictive in many real-life application scenarios like when they are applied in Unmanned Aerial Vehicles – UAVs (e.g. drones).

This project will research new techniques and methodologies for reducing the computational cost of deep neural network architectures (Convolutional Neural Networks and Recurrent Neural Networks). We will focus on the proposal of novel techniques for creating compact network topologies that can achieve the same (or better) performance with the state of the art. Development of the proposed approaches in UAVs and testing in real application scenarios will also lead to exciting research directions for improving existing technology.

 

ABOUT THE PROJECT


Project title:
Efficient Deep Learning approaches for Unmanned Aerial Vehicles

PhD student: Negar Heidari

Contact: negar.heidari@eng.au.dk

Project period: Dec 2018 to Nov 2021

Main supervisor: Assoc. Prof. Alexandros Iosifidis

Co-supervisor: Prof. Peter Gorm Larsen

Section: Electrical and Computer Engineering


An AI-based system for health and safety constraint checking in large public buildings

Beidi Li

The wellbeing of occupants in the built environment has become a rising issue in today’s society. Primary user concerns in large-scale public buildings can be described as:

  • health and safety control
  • accessibility
  • easy wayfinding
  • sense of privacy
  • appropriate contingency plan
  • continuity in space
  • daylighting

The widely deployed digital process, BIM (Building Information Modelling), contains comprehensive information about buildings' design features. BIM adopts an object-oriented approach and represents building objects (storey, door, air duct, furniture, space, etc.) at a high abstraction level, instantiated by location, type, geometry and properties.

Although BIM investigates buildings’ pre-occupancy performances thoroughly, it fails to examine the “spatial artefacts” induced by people living in them. Namely, people’s lines-of-sight, physical and psychological reaction to the surroundings, and frequent activities which can vary greatly from a child to an adult, and to a disabled person with functional impairing.

This project aims to use declarative programming to translate semantic constraints expressed in natural language (e.g. students sitting in an amphitheater should be able to see the lecturer) as machine-readable geometric and topological relationships. Existing BIM specifications will be extended by entities and rules to capture real life scenarios where people perceive and interact with the spaces. Artificial intelligence and computer vision will be used to interpret low-level field data such as high-level human cognitive responses and to compare design alternatives in terms of user-friendliness.

The project outcome is expected to promote a holistic, integrated and human-centered design by constantly checking descriptive requirements throughout the building life-cycle. The AEC (Architecture, Engineering, Construction) industry will then have the opportunity to benefit from an intuitive, integrated and intelligible tool for validation of qualitative and numerical data.

 

ABOUT THE PROJECT


Project title:
An AI-based system for health and safety constraint checking in large public buildings

PhD student: Beidi Li

Contact: beidi.li@eng.au.dk

Project period: Dec 2018 to Nov 2021

Main supervisor: Prof. Peter Gorm Larsen

Co-supervisor: Assistant Prof. Carl Peter Leslie Schultz

Section: Electrical and Computer Engineering


Network coding in distributed storage of IoT

Xiaobo Zhao
Xiaobo Zhao

The Internet of Things (IoT) has gained increasing attention in various fields due to its ability to enable new devices to connect to the Internet. Many IoT-based applications, e.g. environmental monitoring, data collection and industrial automation, have a common challenge, namely, how to maintain data reliably when using vulnerable, resource-limited devices. Device failures are caused by a variety of different factors, including limited energy, hardware failures and software errors. IoT systems that are deployed in harsh or inaccessible environments face unique challenges to replace damaged devices or they may be unable to do so. Thus, it is critical to develop a theoretical framework and policies to maintain data reliably in such systems without expecting replacement devices. For this purpose, we plan to do the research from several aspects, e.g. tradeoff between storage redundancy and repair cost, and optimal data protection strategy with/without a leader node.

The goal of this project is to study network coding in distributed storage of IoT comprehensively and to provide thorough research results in this field, which can be significant contributions on both the theoretical and practical sides.

 

ABOUT THE PROJECT


Project title:
Network coding in distributed storage of IoT

PhD student: Xiaobo Zhao

Contact: xiaobo.zhao@eng.au.dk

Project period: Nov 2018 to Oct 2019

Main supervisor: Assoc. Prof. Daniel Rötter

Section: Electrical and Computer Engineering


Vision-based anomaly detection in unmanned aerial vehicle and ground robots

Ilker Bozcan
Ilker Bozcan

Anomaly detection is the classification of objects and events that are labeled as suspicious. Autonomous surveillance systems should be aware of what entities are anomaly in the environment. During this project, we will develop an autonomous surveillance system using cost-effective visual sensors and deep neural networks that are state-of-the-art object detection algorithms.

The system is suited for anomaly detection for both types of data: aerial and ground. Unmanned Aerial/Ground Vehicles (UAV/UGV) are possible robotic platforms to operate anomaly detection systems. As a use case, UGVs will be used for plant classification where weeds are considered as anomaly. In another use case, UAVs will be used for flying object detection where other drones are marked as anomaly.

 

ABOUT THE PROJECT


Project title:
Vision-based anomaly detection in unmanned aerial vehicle and ground robots

PhD student: Ilker Bozcan

Contact: ilker@eng.au.dk

Project period: Oct 2018 to Sep 2021

Main supervisor: Assoc. Prof. Erdal Kayacan

Section: Electrical and Computer Engineering


Massive-scale IoT dynamic data updating and compression for a scalable IoT infrastructure

Niloofar Yazdani
Niloofar Yazdani

Future communication networks and storage systems will face tremendous challenges to answer the increasing data traffic. Additionally, upcoming services and applications may also impose very low delay constraints on the transferred data.

Thus, developing technologies that increase the throughput, reduce the delay and operate in a distributed fashion are economically viable, and store and process data close to the end devices is crucial to the design and deployment of future communication networks and cloud computing.

This PhD project will focus on the development of novel coding theory designs for a more efficient management, updating, consistency assurance and storage of Internet of Things data at a massive scale.

In particular, this project will focus on the integration of (network) coding techniques and data deduplication, two approaches for reducing storage costs that have typically been attacked separately. This work is expected to open a new field at the intersection of traditional coding theory and distributed Cloud technologies and systems. The underlying goal of the results and designs of this project is to develop new technologies for Cloud, Edge and Local content management, transmission and consistency assurance.

 

ABOUT THE PROJECT


Project title: 
Massive-scale IoT dynamic data updating and compression for a scalable IoT infrastructure

PhD student: Niloofar Yazdani

Contact: n.yazdani@eng.au.dk

Project period: Sep 2018 to Aug 2021

Main supervisor: Assoc. Prof. Daniel Rötter

Section: Electrical and Computer Engineering


Massive-scale storage compression for a scalable IoT infrastructure

Lars Nielsen
Lars Nielsen

The strict requirements for 5G and the increasing amount of data generated by both end-user and IoT devices present a set of challenges for communication networks and storage systems.

The currently employed infrastructure is unable to handle this load increase, especially with the massive increase of data. Furthermore, with the expected increase of active devices, the cost of transferring data through the current networks to the destination will be problematic. The economical cost of both maintaining and providing networks and storage systems with the ability to handle this will grow significantly. Therefore, it is critical to develop new technologies to handle the increased data load.

The goal of the project is to design technologies and architectures for future storage and communication systems which can handle the increased data loads, using compression to decrease the amount of data actually stored without loss of information.

 

ABOUT THE PROJECT


Project title:
Massive-scale storage compression for a scalable IoT infrastructure (MSCSII)

PhD student: Lars Nielsen

Contact: lani@eng.au.dk

Project period: Aug 2018 to July 2021

Main supervisor: Assoc. Prof. Daniel Rötter

Section: Electrical and Computer Engineering


Adaptive network coding to develop future networks and cloud technologies

Rasmus Vestergaard
Rasmus Vestergaard

Future communication networks and storage systems will need to be able to handle the increasing traffic generated by end users, the strict requirements for 5G communications, and billions of sensing and actuating devices connected to the Internet of Things. 

Current infrastructure is not adequate for these demands, especially since the amount of data will increase massively. With the expected number of devices, the cost of transferring the data all the way through current networks to reach clouds is prohibitive, and the economic cost of deploying a network capable of handling such amounts would be significant with the current technologies. On other factors, such as delay, the current technologies are entirely unable to satisfy the requirements. It is critical to develop new technologies that can increase the throughput in networks, reduce delays and operate in a distributed fashion while being economically viable. 

The goal of the project is to design communication and storage technologies for these future networks and clouds using coding theory constructions. This will be done by applying error-control codes in the network, and developing novel, adaptive codes that are suitable for flexible management of complexity and performance.

 

ABOUT THE PROJECT


Project title:
Adaptive network coding to develop future networks and cloud technologies

PhD student: Rasmus Vestergaard

Contact: rv@eng.au.dk

Project period: Aug 2018 to July 2021

Main supervisor: Assoc. Prof. Daniel Rötter

Co-supervisor: Assoc. Prof. Qi Zhang

Section: Electrical and Computer Engineering


Biomechanical characterisation of aortic root properties for optimising repair techniques in the presence of aneurysmal disease

The pathophysiologic mechanisms governing aortic aneurysm progression in humans are not fully understood. Currently, aneurysm size remains the best criteria for recommending surgery in large aortic aneurysms. This has clear shortcomings as aneurysm size is not an absolute predictor of aneurysm expansion and risk of rupture.

This study comprises of a biomechanical characterisation and histological mapping of collagen and elastin architecture of exercised human aortic root tissue around its circumference. Knowledge from this study will be applied to develop a conceptual design of a dynamic annuloplasty ring. The expectation is that this ring could mimic and support the native functional and biomechanical properties addressing normal physiological aortic root dynamics. The hope is that these findings will lead to a significantly improved treatment for patients undergoing repair of the aortic root.

ABOUT THE PROJECT


Project title:
Biomechanical characterisation of aortic root properties for optimising repair techniques in the presence of aneurysmal disease - a clinical experimental study

PhD student: Mariam Abdi Noor

Contact: noor@eng.au.dk

Project period: May 2018 to April 2021

Main supervisor: Assoc. Prof. Peter Johansen

Co-supervisor: J. Michael Hasenkam

Section: Electrical and Computer Engineering


Low-power radio-frequency low-noise amplifier for future wireless transceivers

Michele Spasaro

Next-generation (5G and beyond) wireless networks and communications, such as Internet of Things (IoT) and enhanced Mobile Broadband (eMBB), will lead to a trillion devices dealing with the ever-growing demand of data communications between objects and people.

Their realisation requires overcoming severe scientific and technical challenges. Battery-operated devices require extremely low power consumption; next-generation mobile phones require developing wireless transceivers operating at much higher frequencies with respect to current generation; to name a few. The roadmap towards 5G considers the official launch of 5G new-radio phase two in 2020, but such an ambitious goal is far from reached yet.

The aim of the PhD project is to anticipate and develop innovative design solutions for radio-frequency low-noise amplifiers as key enabling step towards the implementation of next-generation low-power wireless transceivers and beyond.

ABOUT THE PROJECT


Project title:
Low-power radio-frequency low-noise amplifier for future wireless transceivers

PhD student: Michele Spasaro

Contact: michele.spasaro@eng.au.dk

Project period: May 2018 to April 2021

Main supervisor: Prof. Domenico Zito

Section: Electrical and Computer Engineering


Machine Learning for optimisation of baggage handling and sorter systems for logistics

René Ahrendt Sørensen

Big data and Machine Learning is currently a very popular field of research. In collaboration with the company BEUMER Group, I will utilise their large amounts of data and their emulator environments to optimise their Baggage Handling System (BHS) in airports.

I will compare their routing algorithms with a self-taught Reinforcement Learning (RL) system not unlike the system that in resent years have been able to beat professional players in games such as chess, backgammon and go, and play at superhuman level in Atari games.

Besides finding the shortest path through the BHS, the RL system might find patterns which could prevent deadlocks and other unwanted events. Currently such events are manually avoided by software developers.

One of the challenges is to describe exactly what such a system should optimise towards. Is it shortest path, lowest delay, highest throughput, etc.

Another very important part of such a system is the transparency, i.e. how well can we explain why the system does what it does. To address this part, I intend to use methods from the field of Explainable Artificial Intelligence.

ABOUT THE PROJECT


Project title:
Machine learning for optimisation of baggage handling and sorter systems for logistics

PhD student: René Ahrendt Sørensen

Contact: ras@eng.au.dk

Project period: May 2018 to April 2021

Main supervisor: Prof. (Docent) Henrik Karstoft

Co-supervisors: Peter Gorm Larsen, Michael Nielsen, Morten Granum

Section: Electrical and Computer Engineering


Ultra-Low power IC design for an implantable ultrasonically-powered device for neural stimulation

Amin Rashidi

Today you can't underestimate the important role of Biomedical Implantable Microsystems (BIM) in modern therapies. For instance, BIMs like cochlear implants and pacemakers have already become successful mass produced products. Nevertheless, other BMIs like deep brain microstimulators for treating malfunctions such as Parkinson Disease (PD) are still facing serious challenges. In 2005, about 4.5 million people in the EU suffered from PD, and it is anticipated that this number will be doubled by 2030.

In the STARDUST project, we will work on design and implementation of a BIM for PD treatment by stimulating neurons in Globous Pallidus through optogentics. To this end, it is proposed to design and fabricate a sub-millimeter biocompatible device to make it as little invasive as possible to the brain. Wireless powering and communications of the device will be done through ultrasonic waves, which is one of the most challenging parts of the project.

Design, fabrication and testing of low-power low-area circuits for optical stimulation of neurologically manipulated neurons, and recording of neural activity associated with PD are also parts of this work.

ABOUT THE PROJECT


Project title:
Ultra-low power IC design for an implantable ultrasonically-powered device for neural stimulation

PhD student: Amin Rashidi

Contact: a.rashidi@eng.au.dk 

Project period: Dec 2017 to Nov 2020

Main supervisor: Assoc. Prof. Farshad Moradi

Research section: Electrical and Computer Engineering


Energy-efficient ultrasonic energy harvesting for biomedical implants for neural stimulation

Seyedsina Hosseini

As time goes by, Wearable, Portable, Implantable (WPI) technologies find their way into the market. Continuous health monitoring with WPI technologies has opened a very promising, although challenging, door to the engineers for developing smarter and smaller technologies, where low-power and low area are the main concerns, especially for the implantable ones.

In this project, “STARDUST”, the main goal is to provide a novel wireless bio-implantable chip enabling in-vivo monitoring, stimulation and local drug delivery specifically for Parkinson's Disease (PD). In Europe, for instance, 4.5 million people suffer from PD.

The project is multidisciplinary and combines the advantages of MicroElectroMechanical Systems (MEMS), local drug delivery, integrated electrodes and micro-scale light emitting diodes (µLEDs). The device consists of a piezoelectric energy harvester, which can be powered with ultrasonic waves.

ABOUT THE PROJECT


Project title: 
Energy-efficient ultrasonic energy harvesting for biomedical implants for neural stimulation

PhD student: Seyedsina Hosseini

Contact: hosseini@eng.au.dk 

Project period: Dec 2017 to Nov 2020

Main supervisor: Assoc. Prof. Farshad Moradi

Research section: Electrical and Computer Engineering


Ultra-low power IC design for ultrasonically-powered implants

Kjeld Laursen

This project is part of a larger project called STARDUST. The broad goal of STARDUST is to create a biomedical implant that can help to alleviate some of the symptoms of Parkinson’s disease. This is to be accomplished by the use of optogenetics, where neurons in a specific area deep in the brain is modified to be sensitive to light of a certain wavelength.

The light source will be an LED in the biomedical implant that will be implanted amongst the modified neurons in the brain. The implant will be powered wirelessly by harvesting the energy from ultrasonic waves transmitted from a transducer placed outside the body.

My task will be on designing an energy harvester integrated circuit (IC) for the implant, and managing the power available to the LED for stimulating neurons. The energy harvesting IC will be used to maximize the energy efficiency of the harvested power by the piezoelectric crystal. The piezoelectric crystal acts as an acoustic power-receiver driven by the ultrasound from an external transducer.

In later stages, I will work on designing chips to monitor the activities of the neurons along with a wireless datalink to transmit the data to an external device for further analysis.

ABOUT THE PROJECT


Project title:
Ultra-low power IC design for ultrasonically-powered implants

PhD student: Kjeld Laursen

Contact: laursen@eng.au.dk 

Project period: Oct 2017 to April 2021

Main supervisor: Assoc. Prof. Farshad Moradi

Research section: Electrical and Computer Engineering


AI-assisted inspection of clover grass fields based on deep learning for targeted fertilisation

This PhD project is supported by the Innovation Fund project SmartGrass and the GUDP project CloverSense. The common goal of the projects is to develop the use of precision agriculture for clover grass fields. Adapting the fertilisation strategies to the local conditions of each field leads to higher quality yields while reducing the impact on the environment.

The aim of the PhD project is to develop the method for analysing the local conditions of the clover grass fields. Through the use of machine learning, gathered images of the field are to be transformed into accurate measures, such as yield estimates and clover to grass ratios, without human interaction.

The main focus of the research lies within the field of vision-based Deep Learning - specifically within the topics of semi-supervised learning, semantic segmentation and instance segmentation. To exemplify, this can allow the computer to translate a simple image into a precise map of objects and plant species present in the image.

ABOUT THE PROJECT


Project title: 
AI-assisted inspection of clover grass fields based on deep learning for targeted fertilisation

PhD student: Søren Skovsen

Contact: ssk@eng.au.dk 

Project period: August 2017 to July 2020

Main supervisor: Professor (Docent) Henrik Karstoft

Co-supervisor: Senior Researcher Rasmus Nyholm Jørgensen

Research section: Electrical and Computer Engineering


Smart industrial products

Tomas Kulik

A rising trend in the industry is to build sensors and small computers into the products that otherwise would not have any computing capability. These products can then share data from the sensors among themselves or with a service center. This is what is called an industrial Internet of Things (IoT).

Connecting IoT with the cloud computing provides new opportunities for design and operation of these industrial products. Sharing data within this kind of system, however, poses challenges that are not present in a traditional industrial IT. One of the main challenges is how to ensure that this environment is secure from data theft and system misuse.

The project focuses primarily on addressing this challenge by creating a model of an existing industrial communication system security standard. This model can then be used to verify that the system meets the security requirements needed to withstand specific cyber attacks. This verification in turn acts as a proof to the industry that deployment of such a system does not compromise the security of the business.

ABOUT THE PROJECT


Project title:
Smart industrial products

PhD student: Tomas Kulik

Contact: tomaskulik@eng.au.dk 

Project period: August 2017 to July 2020

Main supervisor: Professor Peter Gorm Larsen

Research section: Electrical and Computer Engineering


In-field traffic intensity reduction as a means to increase yield quantity and quality

Andrés Villa-Henriksen

The overall objective of this project is to study the site-specific relations between yield quantity and quality, and in-field traffic intensity. This information will be used to reduce the negative effects on soil of traffic intensity by in-field optimised route planning.

The forthcoming implementation of IoT applied to farming will allow access to large amounts of data. This site-specific data can be used to relate yield data to the increasing problem of soil degradation induced by in-field trafficking. Therefore, this can now be achieved without the need of the traditional time-consuming data sampling throughout a field.

The main objectives of the project are:

  • Investigate the state of the art of farm machine inter-operability and its potential use in vehicle logistics, documentation and traffic intensity reduction with the goal of increasing crop yields.
  • Research the relation between traffic intensity and yield quantity and quality by the use of site-specific data.
  • Evaluate how route planning and vehicle logistics can be used to reduce in-field traffic intensity, and consequently avoid yield decrease due to soil compaction problems.

ABOUT THE PROJECT


Project title: 
In-field traffic intensity reduction as a means to increase yield quantity and quality

PhD student: Andrés Villa-Henriksen

Contact: avh@eng.au.dk 

Project period: August 2017 to July 2020

Main supervisor: Senior Researcher Claus Grøn Sørensen

Co-supervisor: Gareth Edwards and Ole Green, AgroIntelli

Research section: Electrical and Computer Engineering


Towards autonomy in farming operations: Logistics optimisation

Rene Søndergaard Nilsson

This project aims to improve the efficiency of farming operations through logistics optimisation as a next step towards autonomous systems.

Logistics optimisation in an agricultural setting involves generation and optimisation of:

  • Path plans for all involved vehicles both in-field and out of fields.
  • Inter-vehicle interaction plans, such as loading/unloading of goods (seed, fertilizer, yield, bales, etc.)
  • Resource utilisation plans. For example minimising waiting time at conditioning/drying/storage facilities.

This project uses a model-based approach such that all farming operations can be simulated and the developed algorithms can be verified in an ‘off-line’ environment.

The developed algorithms will be deployed in a real-time system, providing input to auto-steering systems or guidance to drivers similar to a GPS in cars. Additionally, it will be able to cope with deviations and system failures, such as a vehicle breakdown, by re-planning and re-optimising for the remaining vehicles and resources involved in the operation.

ABOUT THE PROJECT


Project title:
Towards autonomy in farming operations: Logistics optimisation

PhD student: Rene Søndergaard Nilsson

Contact: rn@eng.au.dk

Project period: July 2017 to June 2020

Main supervisor: Prof. Peter Gorm Larsen

Co-supervisor: Senior Researcher Claus Grøn Sørensen

Research section: Electrical and Computer Engineering


Reinforced Weed Classification utilising utilising Context Data and Deep Generative Models

Simon Leminen Madsen

This project is part of the Innovation Fund Denmark project RoboWeedMaPS. The overall goal of RoboWeedMaPS is to substantially reduce the amount of herbicides in modern crop farming, which will benefit society, the environment and the farmer. To achieve this, a more efficient and precise deployment of herbicides is needed. The project will incorporate automated vision systems to assess the optimum weed treatment and thereby eliminate the need for intermediate manual decision-making and data processing.

This project seeks to apply machine learning, specifically deep learning, to automatically classify weed species and their current development stage from images. To improve the certainty of the classification, it should incorporate context relevant data such as site-specific cropping history, past weed registrations, etc. The project will also explore the potential of generating photo realistic image samples of weeds using deep generative models. These artificial samples are expected to be used for creating a more robust weed classification model. Generative models can potentially also be used to improve the quality of unfocused and blurred images.

ABOUT THE PROJECT


Project title:
Reinforced Weed Classification utilising utilising Context Data and Deep Generative Models

PhD student: Simon Leminen Madsen

Contact: slm@eng.au.dk

Project period: Feb 2017 to Jan 2020

Main supervisor: Prof. (Docent) Henrik Karstoft

Co-supervisor: Senior Researcher Rasmus Nyholm Jørgensen 

Research section: Electrical and Computer Engineering


Compact integrated terahertz sources

Pengli An

Science and technologies based on terahertz (THz) frequency electromagnetic radiation (100 GHz-30 THz) have developed rapidly over the last 30 years, to the extent that these now touch many areas from fundamental science to ‘real world’ applications. THz spectroscopy and imaging are currently active and rapidly expanding fields of research with applications in biological and medical fields, for example cancer tumor imaging and defence and security fields like the airport security scanners to detect explosives and illegal drugs.

However, many applications require hand-held or mobile THz devices but these are not commercially available yet. The most important challenge for THz technology is the development of compact and efficient THz sources providing continuous wave output power.

The aim of this project is to develop compact THz emitters. THz radiation can be generated by mixing photonic signals, but these optical components are bulky. Using photonic integration, we can put all these components on a single chip. This provides an efficient and reliable solution for THz sources in terms of cost and size. Therefore, we will develop a photonic integrated circuit (PIC) for THz signal generation in the optical domain. Our PIC will then be coupled to another chip containing THz antennas to convert the optical signal to a narrow beam of THz radiation.

ABOUT THE PROJECT


Project title:
Compact integrated terahertz sources

PhD student: Pengli An

Contact: pengli_an@eng.au.dk

Project period: Jan 2017 to Dec 2019

Main supervisor: Assoc. Prof. Martijn Heck

Research section: Electrical and Computer Engineering


Enabling ultra-reliable low latency communications in 5G

Jianhui Liu

The fifth generation (5G) systems are expected to underpin the Tactile Internet connectivity. The typical use cases of Tactile Internet include industrial automation, remote surgery and autonomous vehicle and so on. To enable these use cases, ultra-reliable and low latency communication (URLLC) is essential to promote the experiences of latency, reliability and availability into an unprecedented new stage.

The URLLC requires reliability of 99.999 percent with latency in the range of 1-10 ms depending on the use cases. It presents enormous challenges at multiple OSI layers and requires disruptive design approaches. One of the prospective solutions is the combination between Fog computing and Radio Access Networks (Fog-RAN). Fog computing extends the computing and storage capacity from distant cloud to the edge of networks which therefore can shorten the end-to-end latency significantly and is beneficial for latency-critical applications. Furthermore, Fog-RAN will provide a flexible network architecture for service-optimised network design and deployment. The aim of this project is to design novel protocols and algorithms to enable URLLC from the aspect of radio access and networking layer.

ABOUT THE PROJECT


Project title:
Enabling ultra-reliable low latency communications in 5G

PhD student: Jianhui Liu

Contact: jianhui.liu@eng.au.dk

Project period: Jan 2017 to Dec 2019

Main supervisor: Assoc. Prof. Qi Zhang

Research section: Electrical and Computer Engineering


Microwave Photonic Oscillators Integrated on an Optical Chip

Peter Tønning

”Timing is everything” is a saying usually reserved for delivering punch-lines. In the field of communication technology, the sentence is literally true and ultra-precise timing is key in advancing the present network society. The timekeepers of technology, oscillators, provide the clock frequency for computing systems, enable carriers for information in wireless communication and provide timing in GPS and radar systems. Oscillators can be realised in a number of different ways, electronic based, crystal based or, more recently, photonic based. 

The optoelectronic oscillator, realised for the first time 20 years ago, offers performance in the GHz-regime way beyond what can be realised by other oscillator types. An optoelectronic oscillator utilises a mix of electronic and photonic components as the name suggests to combine the existing electron-powered technological platform with the superior bandwidth and low-loss capabilities of photons in waveguides/fibers. So far, these systems have been reserved mainly for research purposes due to considerable size and cost as well as lack of robustness. Recent advance in photonic integrated circuits means that the realisation of an optoelectronic oscillator on a chip is within reach.

The goal of this project is to create an optoelectronic oscillator on a chip with a frequency in the GHz regime and phase noise performance better than the crystal based alternative. An oscillator at this frequency would allow for advances in e.g. Doppler radar technology and future high-speed wireless communication.

This research is done in close collaboration with PhD student Lars Nielsen so for another take on this project, see Lars Nielsen's project description.

ABOUT THE PROJECT


Project title:
Microwave Photonic Oscillators Integrated on an Optical Chip

PhD student: Peter Tønning

Contact: toenning@eng.au.dk

Project period: Nov 2016 to Oct 2019

Main supervisor: Assoc. Prof. Martijn Heck

Research section: Electrical and Computer Engineering


Silicon photonics for future spintronic-photonic memory

Hanna Becker

The use of information technology is exponentially growing and has become so frequent that information processing, transmitting and storing accounted for estimated 5 percent of world electricity production in 2012. As this growth is expected to continue, we need to find a way to decrease energy consumption.

The goal of this project is the development of an innovative optical on-chip network. This effort is part of a collaborative project which addresses the need for an energy-efficient data storage device. This network will switch and direct light pulses to spintronic memory elements as well as illuminate them to ‘write’ data. The energy consumption of this device will be reduced by two orders of magnitude compared to present-day memory technology.

Integrated photonics offers a promising platform for new energy-efficient on-chip technologies with the number of photonic elements on optical chips increasing at a rate comparable to Moore’s law. Together with the recent discovery of magnetization reversal by short optical pulses, an optically switchable spintronic memory element becomes feasible. This enables the unique integration of photonics and spintronic memory elements. The project is conducted in collaboration with IMEC/Ghent University (BE), Radboud University (NL), SpinTEC (FR) and QuantumWise (DK).

ABOUT THE PROJECT


Project title:
Silicon photonics for future spintronic-photonic memory

PhD student: Hanna Becker

Contact: hanna.becker@eng.au.dk

Project period: Oct 2016 to Sept 2019

Main supervisor: Assoc. Prof. Martijn Heck

Research section: Electrical and Computer Engineering


Microwave Photonic Oscillators Integrated on an Optical Chip

Lars Nielsen

Oscillations exist naturally everywhere in our everyday life, from sound waves to the periodic rotations of the earth, creating the dynamics of this world. In order to bring life into our electronic systems, we use electronic oscillators which generate alternating voltages and currents at a desired frequency. They are a fundamental part of almost every piece of electronic circuitry, and they enable, for example, the dynamic behaviour of computers, carriers for the transmission of data and time in clocks.

Electronic oscillators relying on the well-defined resonance of crystals, mainly quartz, have existed for decades. However, the increasing demands for low noise performance and high bandwidth in applications such as high speed analog to digital converters, radars and positioning systems, go beyond the limit of these oscillators.

Another type of oscillator is the optoelectronic oscillator which has gained more and more interest due to its ultra-low noise performance. An optoelectronic oscillator is based on the modulation of an optical carrier wave by a microwave signal which is generated through an optical feedback loop. The current optoelectronic oscillators outperform the crystal oscillators in the giga-hertz regime. However, currently they are based mainly on large discrete components such as optical fibers and are thus applicable to laboratory use only.

The aim of this project is to reduce the size of the optoelectronic oscillator by integrating it onto a photonic chip whilst retaining the ultra-low noise performance.

ABOUT THE PROJECT


Project title:
Microwave Photonic Oscillators Integrated on an Optical Chip

PhD student: Lars Nielsen

Contact: ln@eng.au.dk

Project period: May 2016 to April 2019

Main supervisor: Assoc. Prof. Martijn Heck

Research section: Electrical and Computer Engineering


Supporting Multidisciplinary Development of Cyber-Physical Systems

Casper Thule Hansen

Systems consisting of both software and hardware are becoming a vital part of society, where they constitute cars, trains, medical devices and so forth. Such systems can be called Cyber-Physical Systems as they often involve cyber elements controlling physical processes.

When developing Cyber-Physical Systems it can be useful to create models of components, a model being an abstract description of a component. These models are then used in a Co-Simulation which is a simulation of coupled technical systems. Simulating the constituents that make up a given system can help identify undesired behaviour. This study will involve the development of the Co-Simulation Orchestration Engine which is the software responsible for orchestrating a simulation using models of components.

The Co-Simulation Orchestration Engine is part of the INTO-CPS project which is short for Integrated Tool Chain for Model-based Design of Cyber-Physical Systems. The purpose of the INTO-CPS project is to create a family of interlinked tools that support development of Cyber-Physical Systems from requirements to realisation in hardware and software.

ABOUT THE PROJECT

Project title: Supporting Multidisciplinary Development of Cyber-Physical Systems

PhD student: Casper Thule Hansen

Contact: casper.thule@eng.au.dk

Project period: Feb. 2016 to Jan. 2019

Main supervisor: Prof. Peter Gorm Larsen

Research section: Electrical and Computer Engineering