Semantic and fuzzy modelling for human behaviour recognition in smart spaces : a case study on ambient assisted living.

Human activity recognition in everyday environments is a critical, but challenging task in Ambient Intelligence applications to achieve proper Ambient Assisted Living, and key challenges still remain to be dealt with to realize robust methods. One of the major limitations of the Ambient Intelligence systems today is the lack of semantic models of those activities on the environment, so that the system can recognize the speci c activity being performed by the user(s) and act accordingly. In this context, this thesis addresses the general problem of knowledge representation in Smart Spaces. The main objective is to develop knowledge-based models, equipped with semantics to learn, infer and monitor human behaviours in Smart Spaces. Moreover, it is easy to recognize that some aspects of this problem have a high degree of uncertainty, and therefore, the developed models must be equipped with mechanisms to manage this type of information. A fuzzy ontology and a semantic hybrid system are presented to allow modelling and recognition of a set of complex real-life scenarios where vagueness and uncertainty are inherent to the human nature of the users that perform it. The handling of uncertain, incomplete and vague data (i.e., missing sensor readings and activity execution variations, since human behaviour is non-deterministic) is approached for the rst time through a fuzzy ontology validated on real-time settings within a hybrid data-driven and knowledgebased architecture. The semantics of activities, sub-activities and real-time object interaction are taken into consideration. The proposed framework consists of two main modules: the low-level sub-activity recognizer and the high-level activity recognizer. The rst module detects sub-activities (i.e., actions or basic activities) that take input data directly from a depth sensor (Kinect). The main contribution of this thesis tackles the second component of the hybrid system, which lays on top of the previous one, in a superior level of abstraction, and acquires the input data from the rst module's output, and executes ontological inference to provide users, activities and their in uence in the environment, with semantics. This component is thus knowledge-based, and a fuzzy ontology was designed to model the high-level activities. Since activity recognition requires context-awareness and the ability to discriminate among activities in di erent environments, the semantic framework allows for modelling common-sense knowledge in the form of a rule-based system that supports expressions close to natural language in the form of fuzzy linguistic labels. The framework advantages have been evaluated with a challenging and new public dataset, CAD-120, achieving an accuracy of 90.1% and 91.1% respectively for low and high-level activities. This entails an improvement over both, entirely data-driven approaches, and merely ontology-based approaches. As an added value, for the system to be su ciently simple and exible to be managed by non-expert users, and thus, facilitate the transfer of research to industry, a development framework composed by a programming toolbox, a hybrid crisp and fuzzy architecture, and graphical models to represent and con gure human behaviour in Smart Spaces, were developed in order to provide the framework with more usability in the nal application. As a result, human behaviour recognition can help assisting people with special needs such as in healthcare, independent elderly living, in remote rehabilitation monitoring, industrial process guideline control, and many other cases. This thesis shows use cases in these areas.

Probabilistic position and orientation estimation of rigid objects using an RGB-D sensor

The recent emergence of low-cost RGB-D sensors has brought new opportunities for robotics by providing affordable devices that can provide synchronized images with both color and depth information. In this thesis, recent work on pose estimation utilizing RGBD sensors is reviewed. Also, a pose recognition system for rigid objects using RGB-D data is implemented. The implementation uses half-edge primitives extracted from the RGB-D images for pose estimation. The system is based on the probabilistic object representation framework by Detry et al., which utilizes Nonparametric Belief Propagation for pose inference. Experiments are performed on household objects to evaluate the performance and robustness of the system.

Depth sensors in augmented reality solutions. Literature review

The emergence of depth sensors has made it possible to track – not only monocular cues – but also the actual depth values of the environment. This is especially useful in augmented reality solutions, where the position and orientation (pose) of the observer need to be accurately determined. This allows virtual objects to be installed to the view of the user through, for example, a screen of a tablet or augmented reality glasses (e.g. Google glass, etc.). Although the early 3D sensors have been physically quite large, the size of these sensors is decreasing, and possibly – eventually – a 3D sensor could be embedded – for example – to augmented reality glasses. The wider subject area considered in this review is 3D SLAM methods, which take advantage of the 3D information available by modern RGB-D sensors, such as Microsoft Kinect. Thus the review for SLAM (Simultaneous Localization and Mapping) and 3D tracking in augmented reality is a timely subject. We also try to find out the limitations and possibilities of different tracking methods, and how they should be improved, in order to allow efficient integration of the methods to the augmented reality solutions of the future.

Nutrient enrichment and water depth modify consumer control in rocky shore macroalgal communities

The objective of my thesis is to assess mechanisms of ecological community control in macroalgal communities in the Baltic Sea. In the top-down model, predatory fish feed on invertebrate mesograzers, releasing algae partly from grazing pressure. Such a reciprocal relationship is called trophic cascade. In the bottom-up model, nutrients increase biomass in the food chain. The nutrients are first assimilated by algae and, via food chain, increase also abundance of grazers and predators. Previous studies on oceanic shores have described these two regulative mechanisms in the grazer - alga link, but how they interact in the trophic cascades from fish to algae is still inadequately known. Because the top-down and bottom-up mechanisms are predicted to depend on environmental disturbances, such as wave stress and light, I have studied these models at two distinct water depths. There are five factorial field experiments behind the thesis, which were all conducted in the Finnish Archipelago Sea. In all the experiments, I studied macroalgal colonization - either density, filament length or biomass - on submerged colonization substrates. By excluding predatory fish and mesograzers from the algal communities, the studies compared the strength of the top-down control to natural algal communities. A part of the experimental units were, in addition, exposed to enriched nitrogen and phosphorus concentrations, which enabled testing of bottom-up control. These two models of community control were further investigated in shallow (<1 m) and deep (ca. 3 m) water. Moreover, the control mechanisms were also expected to depend on grazer species. Therefore different grazer species were enclosed into experimental units and their impacts on macroalgal communities were followed specifically. The community control in the Baltic rocky shores was found to follow theoretical predictions, which have not been confirmed by field studies before. Predatory fish limited grazing impact, which was seen as denser algal communities and longer algal filaments. Nutrient enrichment increased density and filament length of annual algae and, thus, changed the species composition of the algal community. The perennial alga Fucus vesiculosusA and the red alga Ceramium tenuicorne suffered from the increased nutrient availabilities. The enriched nutrient conditions led to denser grazer fauna, thereby causing strong top-down control over both the annual and perennial macroalgae. The strength of the top-down control seemed to depend on the density and diversity of grazers and predators as well as on the species composition of macroalgal assemblages. The nutrient enrichment led to, however, weaker limiting impact of predatory fish on grazer fauna, because fish stocks did not respond as quickly to enhanced resources in the environment as the invertebrate fauna. According to environmental stress model, environmental disturbances weaken the top-down control. For example, on a wave-exposed shore, wave stress causes more stress to animals close to the surface than deeper on the shore. Mesograzers were efficient consumers at both the depths, while predation by fish was weaker in shallow water. Thus, the results supported the environmental stress model, which predicts that environmental disturbance affects stronger the higher a species is in the food chain. This thesis assessed the mechanisms of community control in three-level food chains and did not take into account higher predators. Such predators in the Baltic Sea are, for example, cormorant, seals, white-tailed sea eagle, cod and salmon. All these predatory species were recently or are currently under intensive fishing, hunting and persecution, and their stocks have only recently increased in the region. Therefore, it is possible that future densities of top predators may yet alter the strengths of the controlling mechanisms in the Baltic littoral zone.

Multi-sensor Simultaneous Localization and Mapping

Simultaneous localization and mapping(SLAM) is a very important problem in mobile robotics. Many solutions have been proposed by different scientists during the last two decades, nevertheless few studies have considered the use of multiple sensors simultane¬ously. The solution is on combining several data sources with the aid of an Extended Kalman Filter (EKF). Two approaches are proposed. The first one is to use the ordinary EKF SLAM algorithm for each data source separately in parallel and then at the end of each step, fuse the results into one solution. Another proposed approach is the use of multiple data sources simultaneously in a single filter. The comparison of the computational com¬plexity of the two methods is also presented. The first method is almost four times faster than the second one.

Electrical characteristics of eddy current position sensor

In the thesis the principle of work of eddy current position sensors and the main cautions that must be taken into account while sensor design process are explained. A way of automated eddy current position sensor electrical characteristics measurement is suggested. A prototype of the eddy current position sensor and its electrical characteristics are investigated. The results obtained by means of the automated measuring system are explained.

Sensor Fusion of Proprioception, Force and Vision in Estimation and Robot

Sensor-based robot control allows manipulation in dynamic environments with uncertainties. Vision is a versatile low-cost sensory modality, but low sample rate, high sensor delay and uncertain measurements limit its usability, especially in strongly dynamic environments. Force is a complementary sensory modality allowing accurate measurements of local object shape when a tooltip is in contact with the object. In multimodal sensor fusion, several sensors measuring different modalities are combined to give a more accurate estimate of the environment. As force and vision are fundamentally different sensory modalities not sharing a common representation, combining the information from these sensors is not straightforward. In this thesis, methods for fusing proprioception, force and vision together are proposed. Making assumptions of object shape and modeling the uncertainties of the sensors, the measurements can be fused together in an extended Kalman filter. The fusion of force and visual measurements makes it possible to estimate the pose of a moving target with an end-effector mounted moving camera at high rate and accuracy. The proposed approach takes the latency of the vision system into account explicitly, to provide high sample rate estimates. The estimates also allow a smooth transition from vision-based motion control to force control. The velocity of the end-effector can be controlled by estimating the distance to the target by vision and determining the velocity profile giving rapid approach and minimal force overshoot. Experiments with a 5-degree-of-freedom parallel hydraulic manipulator and a 6-degree-of-freedom serial manipulator show that integration of several sensor modalities can increase the accuracy of the measurements significantly.

Energy harvesting methods for wireless sensor nodes in heavy-duty vehicles

The number of autonomous wireless sensor and control nodes has been increasing rapidly during the last decade. Until recently, these wireless nodes have been powered with batteries, which have lead to a short life cycle and high maintenance need. Due to these battery-related problems, new energy sources have been studied to power wireless nodes. One solution is energy harvesting, i.e. extracting energy from the ambient environment. Energy harvesting can provide a long-lasting power source for sensor nodes, with no need for maintenance. In this thesis, various energy harvesting technologies are studied whilst focusing on the theory of each technology and the state-of-the-art solutions of published studies and commercial solutions. In addition to energy harvesting, energy storage and energy management solutions are also studied as a subsystem of a whole energy source solution. Wireless nodes are also used in heavy-duty vehicles. Therefore a reliable, long-lasting and maintenance-free power source is also needed in this kind of environment. A forestry harvester has been used as a case study to study the feasibility of energy harvesting in a forestry harvester’s sliding boom. The energy harvester should be able to produce few milliwatts to power the target system, an independent limit switch.

The Cellular Oxygen Sensor PHD2 in Cancer Growth

Adequate supply of oxygen is essential for the survival of multicellular organisms. However, in several conditions the supply of oxygen can be disturbed and the tissue oxygenation is compromised. This condition is termed hypoxia. Oxygen homeostasis is maintained by the regulation of both the use and delivery of oxygen through complex, sensitive and cell-type specific transcriptional responses to hypoxia. This is mainly achieved by one master regulator, a transcription factor called hypoxiainducible factor 1 (HIF-1). The amount of HIF-1 is under tight oxygen-dependent control by a family of oxygen-dependent prolyl hydroxylase domain proteins (PHDs) that function as the cellular oxygen sensors. Three family members (PHD1-3) are known to regulate HIF of which the PHD2 isoform is thought to be the main regulator of HIF-1. The supply of oxygen can be disturbed in pathophysiological conditions, such as ischemic disorders and cancer. Cancer cells in the hypoxic parts of the tumors exploit the ability of HIF-1 to turn on the mechanisms for their survival, resistance to treatment, and escape from the oxygen- and nutrient-deprived environment. In this study, the expression and regulation of PHD2 were studied in normal and cancerous tissues, and its significance in tumor growth. The results show that the expression of PHD2 is induced in hypoxic cells. It is overexpressed in head and neck squamous cell carcinomas and colon adenocarcinomas. Although PHD2 normally resides in the cytoplasm, nuclear translocation of PHD2 was also seen in a subset of tumor cells. Together with the overexpression, the nuclear localization correlated with the aggressiveness of the tumors. The nuclear localization of PHD2 caused an increase in the anchorage-independent growth of cancer cells. This study provides information on the role of PHD2, the main regulator of HIF expression, in cancer progression. This knowledge may prove to be valuable in targeting the HIF pathway in cancer treatment.

Implementing gateway monitoring service for infrastructure sensor networks

Wireless sensor networks and its applications have been widely researched and implemented in both commercial and non commercial areas. The usage of wireless sensor network has developed its market from military usage to daily use of human livings. Wireless sensor network applications from monitoring prospect are used in home monitoring, farm fields and habitant monitoring to buildings structural monitoring. As the usage boundaries of wireless sensor networks and its applications are emerging there are definite ongoing research, such as lifetime for wireless sensor network, security of sensor nodes and expanding the applications with modern day scenarios of applications as web services. The main focus in this thesis work is to study and implement monitoring application for infrastructure based sensor network and expand its usability as web service to facilitate mobile clients. The developed application is implemented for wireless sensor nodes information collection and monitoring purpose enabling home or office environment remote monitoring for a user.

Electrical Equivalent Circuit for Dry Bacteriorhodopsin Sensor

Bacteriorhodopsin (BR) is a photosensitive protein which functions as a light-driven proton pump. Due to its photoactivity, BR could be used in photosensing and information processing which has inspired researchers to study the photoelectric response and the appropriate measurement instrumentation for BR. In this thesis, the measurement instrumentation connected to a dry BR sensor was confirmed to affect the photovoltage response measured by using voltage amplifiers. Changing of the input impedance of the measurement instrumentation was shown to alter a part of the measured photovoltage response. The photocurrent measurements using transimpedance amplifier and the presented electrical equivalent circuit were used to show that the photocurrent measurements have no significant effect on the photoelectric response. The photocurrent was shown to be a derivate of the photovoltage response measured from the dry BR sensor when it was compared to the response measured with a voltage amplifier. This confirmed that another part of the photovoltage response was not affected by the measurement instrumentation. The time-variant behavior of the dry BR sensor was confirmed in both the photocurrent and the photovoltage measurements. This was caused by the fact that the capacitance of the dry BR sensor changes with the excitation light intensity.

Heat flux measurement inside internal combustion engine with gradient heat flux sensor.

This master’s thesis is devoted to study different heat flux measurement techniques such as differential temperature sensors, semi-infinite surface temperature methods, calorimetric sensors and gradient heat flux sensors. The possibility to use Gradient Heat Flux Sensors (GHFS) to measure heat flux in the combustion chamber of compression ignited reciprocating internal combustion engines was considered in more detail. A. Mityakov conducted an experiment, where Gradient Heat Flux Sensor was placed in four stroke diesel engine Indenor XL4D to measure heat flux in the combustion chamber. The results which were obtained from the experiment were compared with model’s numerical output. This model (a one – dimensional single zone model) was implemented with help of MathCAD and the result of this implementation is graph of heat flux in combustion chamber in relation to the crank angle. The values of heat flux throughout the cycle obtained with aid of heat flux sensor and theoretically were sufficiently similar, but not identical. Such deviation is rather common for this type of experiment.

Synthesis, characterization and chemical sensor application of conducting polymers

Polymeric materials that conduct electricity are highly interesting for fundamental studies and beneficial for modern applications in e.g. solar cells, organic field effect transistors (OFETs) as well as in chemical and bio‐sensing. Therefore, it is important to characterize this class of materials with a wide variety of methods. This work summarizes the use of electrochemistry also in combination with spectroscopic methods in synthesis and characterization of electrically conducting polymers and other π‐conjugated systems. The materials studied in this work are intended for organic electronic devices and chemical sensors. Additionally, an important part of the presented work, concerns rational approaches to the development of water‐based inks containing conducting particles. Electrochemical synthesis and electroactivity of conducting polymers can be greatly enhanced in room temperature ionic liquids (RTILs) in comparison to conventional electrolytes. Therefore, poly(para‐phyenylene) (PPP) was electrochemically synthesized in the two representative RTILs: bmimPF6 and bmiTf2N (imidazolium and pyrrolidinium‐based salts, respectively). It was found that the electrochemical synthesis of PPP was significantly enhanced in bmimPF6. Additionally, the results from doping studies of PPP films indicate improved electroactivity in bmimPF6 during oxidation (p‐doping) and in bmiTf2N in the case of reduction (n‐doping). These findings were supported by in situ infrared spectroscopy studies. Conducting poly(benzimidazobenzophenanthroline) (BBL) is a material which can provide relatively high field‐effect mobility of charge carriers in OFET devices. The main disadvantage of this n‐type semiconductor is its limited processability. Therefore in this work BBL was functionalized with poly(ethylene oxide) PEO, varying the length of side chains enabling water dispersions of the studied polymer. It was found that functionalization did not distract the electrochemical activity of the BBL backbone while the processability was improved significantly in comparison to conventional BBL. Another objective was to study highly processable poly(3,4‐ethylenedioxythiophene) poly(styrenesulfonate) (PEDOT:PSS) water‐based inks for controlled patterning scaled‐down to nearly a nanodomain with the intention to fabricate various chemical sensors. Developed PEDOT:PSS inks greatly improved printing of nanoarrays and with further modification with quaternary ammonium cations enabled fabrication of PEDOT:PSS‐based chemical sensors for lead (II) ions with enhanced adhesion and stability in aqueous environments. This opens new possibilities for development of PEDOT:PSS films that can be used in bio‐related applications. Polycyclic aromatic hydrocarbons (PAHs) are a broad group of π‐conjugated materials consisting of aromatic rings in the range from naphthalene to even hundred rings in one molecule. The research on this type of materials is intriguing, due to their interesting optical properties and resemblance of graphene. The objective was to use electrochemical synthesis to yield relatively large PAHs and fabricate electroactive films that could be used as template material in chemical sensors. Spectroscopic, electrochemical and electrical investigations evidence formation of highly stable films with fast redox response, consisting of molecules with 40 to 60 carbon atoms. Additionally, this approach in synthesis, starting from relatively small PAH molecules was successfully used in chemical sensor for lead (II).

Track of HSMS Sophia tracing the extent of navigable water along the main northern ice between Spitzbergen and Greenland September 1868 : also showing the nature of the ice and temperature and depth of the sea

Kartta kuuluu A. E. Nordenskiöldin kokoelmaan