TNO Science and Technology, Den Haag

Company profile:

TNO is a leading, independent knowledge company whose expertise and research make an important contribution to the competitiveness of companies and organisations, to the economy and to the quality of society as a whole. TNO's unique position comes through its versatility and capacity to integrate knowledge. Some 4300 professionals work at TNO.

Master Assignments:
Distributed optimization
Optimization techniques for dynamical systems is a well developed area of expertise with great practical significance (e.g. joint optimization of traffic lights, power network maintenance, control of renewable energy sources, etc.). In systems that are spatially distributed, i.e., a network of multiple subsystems, the observations from the process cannot be communicated centrally and thus optimization can only be done locally by each subsystems. The question is how the local optimizers should interact in order to achieve global optimum. The problem is especially interesting in wireless sensor networks, where the number of subsystems is high and the communication/power constraints are serious. The assignment should overview the different types of distributed optimization problems and develop solutions for particular large-scale mobility related optimization cases (typical cases: traffic light management, cooperative highway on-ramp control).

Behavioural analysis and anomaly detection in maritime situation
The Netherlands Coastguard is responsible for many operational tasks on the North Sea. One of these responsibilities is to intercept illegal activities such as smuggling. Behavioural analysis of vessel trajectories is performed to reduce the number of vessels to investigate. At TNO we run several projects that study automated behavioural analysis and anomaly detection, in which an anomaly is defined as a deviation from a common rule, pattern, or behaviour. In these studies we make a distinction between logical methods (e.g. rule based, case based reasoning, complex event processing) and probabilistic or statistical methods (e.g. Mahalanobis distance, Gaussian mixture models, nearest neighbour, local outlier factor, kernel density estimation). The goal of this master assignment is to evaluate probabilistic or statistical methods using real vessel trajectory data.

Vessel assessment for maritime situations
The Netherlands Coastguard is responsible for many operational tasks on the North Sea. One of these responsibilities is to intercept illegal activities, such as smuggling, by investigating the vessels traveling on the North Sea. Behavioural analysis of vessel trajectories is performed to reduce the number of vessels to investigate. At TNO we run several projects that study automated behavioural analysis and anomaly detection, in which an anomaly is defined as a deviation from a common rule, pattern, or behaviour. In these studies we make a distinction between logical methods (e.g. rule based, case based reasoning, complex event processing) and probabilistic or statistical methods (e.g. Mahalanobis distance, Gaussian mixture models, nearest neighbour, local outlier factor, kernel density estimation). The goal of this master assignment is to evaluate probabilistic or statistical methods using real vessel trajectory data.

Localization of animals using video
In this assignment you will work on a research project in the field of image processing, object detection and classification, and tracking. A research environment has been set up at the Apenheul in Apeldoorn (NL). Eleven cameras are positioned around the island with Gorillas . Your task will be to detect the gorillas at the island and track their movements. More specifically you will work on one or more of the following subjects: 1. Detect gorillas in video data; 2. Track the gorillas within the range of a camera; 3. Track the when moving between cameras; 4. Integrate (third party) functionality to classify and/or identify tracks; 5. Extract “useful” data from the gorilla tracks and integrate this into the map.

Identification of indoor climate
Controlling the indoor-climate in greenhouses increases the production of vegetables, as temperature , humidity and CO2 densities can be tuned with higher accuracies. However, this requires a mathematical model that describes the dynamics of the indoor climate and which can further be employed in a real-time controller. The currently available models are mainly based on CFD (computer-fluid-dynamics). They are far too computationally demanding and are thus infeasible with respect to the real-time requirements. TNO has created an alternative climate model that enjoys this real-time property. First simulations prove the potential of this novel climate model, although real experiments are yet to be done. To that extent, TNO has bought a small climate-room with control capabilities. Your assignment is to run these experiments and from that, identify the corresponding model parameters. Additionally, improvements on the model should be made if necessary.

Localization in ad-hoc networks
An ad-hoc sensor network is a network of nodes in which each node is equipped with a sensor, some processing power and a wireless communication device. A well known example of such a node is the Tmote or TelosB. One of the characteristics of such a network is that the nodes are randomly spatially distributed in a certain area and establish an ad-hoc communication topology. For a large amount of applications the position of each node is required when the sensory data is processed. Currently there are a number of methods to estimate the relative position from one node to another based on their communication, i.e. signal strength of the radio, the number of bits that were flipped or radio interferometer. Your task is to analyze these methods by implementing them in a network of Tmotes/TelosB. As a master thesis this assignment can be extended in improvement of the current localization methods.

Studierichting: Artificial Intelligence, Computer Science, Information Sciences, Mathematics, Business Analytics, Parallel and Distributed Computer Systems.

Information:
Above we describe several possibilities for internships at TNO Technical Sciences. Students who have other projects are welcome to contact us. For more information contact Joris Sijs (joris.sijs@tno.nl).