Three PhD Scholarships in Autonomous Technology for Marine Craft
The Automation and Control Group at DTU Department of Electrical Engineering seeks highly qualified motivated candidates for a number of PhD positions. Applicants are invited to apply for a specific project from the list below. Additional information concerning individual project proposals can be obtained from the contact persons listed below. The procedure for applying is described below.
A total of five PhD projects are part of the ShippingLab Autonomy Work Package (WP). Three of these vacancies are advertised below. The ShippingLab Autonomy research will be a joint effort between senior faculty with Professor Mogens Blanke as Principal Investigator, experienced developers and navigators from partners, four PostDocs and five PhD candidates. The Autonomy WP is a joint effort between DTU (Kgs. Lyngby), SIMAC, Wärtsila-Lyngsø Marine, Logimatic Engineering, Danelec Marine, TUCO Marine, DFDS, Danske Færger and a Harbor Bus Operator. The Autonomy WP will provide the research and development necessary to obtain maritime autonomy over a range of levels, from decision support to enhance safety and voyage economy, to a ferry on demand solution that would be able to operate without a navigator on board.
PhD candidates will participate in measurements, test and validation at sea. Candidates are expected to be dedicated, productive and ambitious team players, who wish to obtain an international level in their area. Scientific publication and experimental validation of results will be part of all the PhD projects and candidates a will include two or more publications in international scientific journals plus a number of conference papers. Presentation at international conferences will be part of the research training and elements of didactics will be obtained through experience as teaching assistants.
The research of this project has civilian objectives. However, equipment restricted by export licenses and ITAR (International Traffic in Arms Regulations) is being used in this research. Applicants that are citizens of Denmark, Norway and other NATO countries, Sweden, Australia, New Zealand, or Japan are eligible. Other applicants should provide evidence of eligibility to use such equipment for their application to be considered.
PhD project 1: Autonomous Supervision for Safe Ship Operation with Temporarily Unmanned Bridge
This research will focus on creating overall awareness of own ship’s condition related to health and quality of sensors, machinery needed for manoeuvering and of the actuators used for manoeuvering. It will assess the risks related to the observed navigation of own ship and that of other vehicles. It will assess risks related to berthing capabilities given relevant weather information. Supervision will include diagnosis of not normal events including sensor artefacts and temporal or permanent faults in relevant devices on board. Implementation could be obtained using mixed deterministic and stochastic automata.
- MSc in automation and control or robotics with a profile that includes fault diagnosis and fault-tolerant control.
- You have interest in, and perhaps experience with supervision, event diagnosis and control
- Experience with marine control systems and with real time implementation would be an advantage
PhD project 2: Multi-modal sensor fusion and calibration for autonomous surface ships
The goal is to develop novel and reliable calibration and fusion of multiple sensors – employing robust state estimation techniques – in the context of situational awareness for autonomous vessels. A multitude of sensors and modalities are expected to be employed.
- MSc in Electrical/Electronics, Computer engineering, or equivalent with specialization in state estimation and sensor fusion.
- Experience with open source robotic tools, such as ROS, would be an advantage.
- Hands on experience with hardware for sensing would be an advantage.
The goal of this project is to develop novel and robust algorithms for object detection and categorization, to be used on ships during autonomous operation. The project will research methods and architectures of deep learning in combination with other techniques for analysis of images and radar plots such that detection and categorization errors are minimized. The research will include incremental updating of recognition models as experience is accumulated. The developed techniques need to enable operability over long periods of time, supporting safe operation at all times.
- MSc in Robotics, Computer Science or equivalent, with specialization in machine learning and robot vision.
- Experience with open-source tools for machine learning and robot vision would be an advantage.
- Familiarity with design of convolutional neural networks architectures would be an advantage.
- Knowledge about radar technology would be an advantage.
Applicants must have a two-year master's degree (120 ECTS points) or a similar degree with an academic level equivalent to a two-year master's degree. Applicants who will obtain such degree by 1 August 2019 can also apply, but they cannot begin before having received the degree.