Ambient activity monitoring systems produce large amounts of data, which can be used for health monitoring. The problem is that patterns in this data reflecting health status are not identified yet. In this paper the possibility is explored of predicting the functional health status (the motor score of AMPS = Assessment of Motor and Process Skills) of a person from data of binary ambient sensors. Data is collected of five independently living elderly people. Based on expert knowledge, features are extracted from the sensor data and several subsets are selected. We use standard linear regression and Gaussian processes for mapping the features to the functional status and predict the status of a test person using a leave-oneperson-out cross validation. The results show that Gaussian processes perform better than the linear regression model, and that both models perform better with the basic feature set than with location or transition based features. Some suggestions are provided for better feature extraction and selection for the purpose of health monitoring. These results indicate that automated functional health assessment is possible, but some challenges lie ahead. The most important challenge is eliciting expert knowledge and translating that into quantifiable features.
Existing research on the recognition of Activities of Daily Living (ADL) from simple sensor networks assumes that only a single person is present in the home. In real life there will be situations where the inhabitant receives visits from family members or professional health care givers. In such cases activity recognition is unreliable. In this paper, we investigate the problem of detecting multiple persons in an environment equipped with a sensor network consisting of binary sensors. We conduct a real-life experiment for detection of visits in the oce of the supervisor where the oce is equipped with a video camera to record the ground truth. We collected data during two months and used two models, a Naive Bayes Classier and a Hidden Markov Model for a visitor detection. An evaluation of these two models shows that we achieve an accuracy of 83% with the NBC and an accuracy of 92% with a HMM, respectively.
MULTIFILE
In this work, a feasible and low-cost approach is proposed for level measurement in multiphase systems inside tanks used for petroleum-derived oil production. The developed level sensor system consisted of light-emitting diodes (LEDs), light-dependent resistor (LDR), and a low-cost microprocessor. Two different types of oil were tested: AW460 and AW68. Linear regression (LR) was applied for 11 scenarios and showed a direct correlation between the level of oil and the sensor’s output. The measurement with AW460 oil presented a perfect linear behavior, while for AW68, a higher standard deviation was obtained justifying the occurrence of the nonlinearity in several scenarios. In order to overcome the nonlinear effect, two machine learning (ML) techniques were tested: K-nearest neighbors regression (KNNR) and multilayer perceptron (MLP) neural network regression. The highest correlation coefficient ( R2 ) and the lowest root mean squared error (RMSE) were obtained for AW68 with MLP. Therefore, MLP was used for regression (level prediction for water, oil, and emulsion) as well as classification (identify the type of oil in the reservoir) simultaneously. The suggested network exhibited a high accuracy for oil identification (99.801%) and improved linear performance in regression ( R2 = 0.9989 and RMSE = 0.065).
Size measurement plays an essential role for micro-/nanoparticle characterization and property evaluation. Due to high costs, complex operation or resolution limit, conventional characterization techniques cannot satisfy the growing demand of routine size measurements in various industry sectors and research departments, e.g., pharmaceuticals, nanomaterials and food industry etc. Together with start-up SeeNano and other partners, we will develop a portable compact device to measure particle size based on particle-impact electrochemical sensing technology. The main task in this project is to extend the measurement range for particles with diameters ranging from 20 nm to 20 um and to validate this technology with realistic samples from various application areas. In this project a new electrode chip will be designed and fabricated. It will result in a workable prototype including new UMEs (ultra-micro electrode), showing that particle sizing can be achieved on a compact portable device with full measuring range. Following experimental testing with calibrated particles, a reliable calibration model will be built up for full range measurement. In a further step, samples from partners or potential customers will be tested on the device to evaluate the application feasibility. The results will be validated by high-resolution and mainstream sizing techniques such as scanning electron microscopy (SEM), dynamic light scattering (DLS) and Coulter counter.
Cell-based production processes in bioreactors and fermenters need to be carefully monitored due to the complexity of the biological systems and the growth processes of the cells. Critical parameters are identified and monitored over time to guarantee product quality and consistency and to minimize over-processing and batch rejections. Sensors are already available for monitoring parameters such as temperature, glucose, pH, and CO2, but not yet for low-concentration substances like proteins and nucleic acids (DNA). An interesting critical parameter to monitor is host cell DNA (HCD), as it is considered an impurity in the final product (downstream process) and its concentration indicates the cell status (upstream process). The Molecular Biosensing group at the Eindhoven University of Technology and Helia Biomonitoring are developing a sensor for continuous biomarker monitoring, based on Biosensing by Particle Motion. With this consortium, we want to explore whether the sensor is suitable for the continuous measurement of HCD. Therefore, we need to set-up a joint laboratory infrastructure to develop HCD assays. Knowledge of how cells respond to environmental changes and how this is reflected in the DNA concentration profile in the cell medium needs to be explored. This KIEM study will enable us to set the first steps towards continuous HCD sensing from cell culture conditions controlling cell production processes. It eventually generates input for machine learning to be able to automate processes in bioreactors and fermenters e.g. for the production of biopharmaceuticals. The project entails collaboration with new partners and will set a strong basis for subsequent research projects leading to scientific and economic growth, and will also contribute to the human capital agenda.
Low back pain is the leading cause of disability worldwide and a significant contributor to work incapacity. Although effective therapeutic options are scarce, exercises supervised by a physiotherapist have shown to be effective. However, the effects found in research studies tend to be small, likely due to the heterogeneous nature of patients' complaints and movement limitations. Personalized treatment is necessary as a 'one-size-fits-all' approach is not sufficient. High-tech solutions consisting of motions sensors supported by artificial intelligence will facilitate physiotherapists to achieve this goal. To date, physiotherapists use questionnaires and physical examinations, which provide subjective results and therefore limited support for treatment decisions. Objective measurement data obtained by motion sensors can help to determine abnormal movement patterns. This information may be crucial in evaluating the prognosis and designing the physiotherapy treatment plan. The proposed study is a small cohort study (n=30) that involves low back pain patients visiting a physiotherapist and performing simple movement tasks such as walking and repeated forward bending. The movements will be recorded using sensors that estimate orientation from accelerations, angular velocities and magnetometer data. Participants complete questionnaires about their pain and functioning before and after treatment. Artificial analysis techniques will be used to link the sensor and questionnaire data to identify clinically relevant subgroups based on movement patterns, and to determine if there are differences in prognosis between these subgroups that serve as a starting point of personalized treatments. This pilot study aims to investigate the potential benefits of using motion sensors to personalize the treatment of low back pain. It serves as a foundation for future research into the use of motion sensors in the treatment of low back pain and other musculoskeletal or neurological movement disorders.
