Douglas G. McIlwraith

Ambient and Wearable Sensor Fusion for Activity Recognition in Healthcare Monitoring Systems

The use of wearable sensors for home monitoring provides an effective means of inferring a patient’s level of activity. However, wearable sensors have intrinsic ambiguities that prevent certain activities to be recognized accurately. The purpose of this paper is to introduce a robust framework for enhanced activity recognition by integrating an ear-worn activity recognition (e-AR) sensor with ambient blob-based vision sensors. Accelerometer information from the e-AR is fused with features extracted from the vision sensor by using a Gaussian Mixture Model Bayes classifier. The experimental results showed a significant improvement of the classification accuracy compared to the use of the e-AR sensor alone.

A Flexible, Low Noise Reflective PPG Sensor Platform for Ear-Worn Heart Rate Monitoring

This paper presents a novel ear-worn reflective photoplethysmography (PPG) sensor that addresses the mechanical complexities of coupling the sensor to the surface of the skin and a detection circuit that minimises ambient noise artefacts. The flexible optoelectronic transducer structure can adapt to a variety of skin surface contours. Light emitting diode (LED) modulation and a unique integrating photocurrent demodulator reduce susceptibility to wideband noise and allow subtraction of ambient light from the desired PPG signal. Experimental results demonstrate that the sensitivity is robust to sensor location and application pressure variations. Simulations also show that the photodetection method is resilient against high levels of wideband noise.

An integrated multi-sensing framework for pervasive healthcare monitoring

Pervasive healthcare provides an effective solution for monitoring the wellbeing of elderly, quantifying post-operative patient recovery and monitoring the progression of neurodegenerative diseases such as Parkinsons. However, developing functional pervasive systems is a complex task that entails the creation of appropriate sensing platforms, integration of versatile technologies for data stream management and development of elaborate data analysis techniques. This paper describes a complete and an integrated multi-sensing framework, with which the sensing platforms, data fusion and analysis algorithms, and software architecture suitable for pervasive healthcare applications are presented. The potential value of the proposed framework for pervasive patient monitoring is demonstrated and initial results obtained from our current research experiences are described.

Body Sensor Networks for Monitoring Rowing Technique

This paper presents a prototype for monitoring the kinematics of the femur and lower back (sacrum and thoraco-lumbar junction) during rowing. Data is collected from inertial sensors attached to the rower and a simple yet relatively accurate method for determining the rotation of the lower back and femur in the sagittal plane is presented. We also present results from an initial validation experiment using an optical tracking system which demonstrate that it is possible to monitor rowers using the proposed sensors and identify some common poor rowing techniques. Due to their small size, wireless capability and lightweight characteristics, the proposed Body Sensor Network (BSN) system has the potential to be used during ergometer sessions and whilst training on the water.

ClimBSN: Climber performance monitoring with BSN

The recent growth in popularity in sport climbing is partly due to the safe environment provided by indoor climbing walls, particularly for novice climbers. Sport climbing involves a wide range of skills and abilities. The purpose of this paper is to present a wearable sensing platform and an analysis framework for assessing general climbing performance during training. To provide the required freedom of movement, a single miniaturized ear-worn 3D accelerometer-based sensor is used. Independent features derived from the accelerometer data are then translated into climbing-specific measures, such as motion fluidity, strength, as well as endurance. Based on these indices, the overall level of the climber and the associated climbing styles can be quantified.

Wearable and ambient sensor fusion for the characterisation of human motion

Home monitoring plays an important role within pervasive healthcare, particularly for monitoring the elderly and patients with chronic disease. For assessing activities of daily living, one of the most challenging problems for research remains that of accurate transition detection and characterisation. Early detection of a change in these transitions, such as difficulty getting up from a seated position, can be an indicator of further complications which often precede a fall. Such changes can also accompany early stage neurological disorders which can be treated effectively to improve quality of life. In this paper, we present a system for the accurate characterisation of motion based upon the fusion of ambient and wearable sensors. A probabilistic, privacy respectful method for the extraction of detailed 3D posture information is proposed and fusion with an ear-worn accelerometer and gyroscope is discussed. We present results detailing high accuracy in the recognition of complex motions over four subjects.

Body Sensor Networks for Sport, Wellbeing and Health

The last decade has witnessed rapid growth of high power, low cost mobile sensing platforms, finding successful application in a range of environments – from industrial process monitoring to structure management. To date, the most challenging deployment arena is the human body and extensive research is focusing on biocompatibility, signal propagation and power management to permit pervasive sensing of detailed physiological signals from implantable, wearable and ambient sensors. By involving users in the management of their own wellbeing, Body Sensor Networks (BSNs) aid in the delivery of preventative care by detecting the onset and systematic deterioration of “lifestyle” diseases. When used for sports performance monitoring, they can also record long-term progress whilst providing real-time training information that can be used to maximise the effectiveness of individual sessions. This chapter provides a detailed review of recent developments in BSNs, highlighting both the technical challenges and deployment issues related to autonomic sensing, context awareness and distributed inferencing.

Probabilistic decision level fusion for real-time correlation of ambient and wearable sensors

Fusing data from ambient and wearable sensors when performing in-home healthcare monitoring allows for high accuracy activity inference due to the complementary nature of sensing modalities. Where residences may house multiple occupants, we must automatically identify related data streams before fusion may occur, a process known as sensor correlation. In this paper a multi-objective variant of the Bayesian Framework for Feature Selection (BFFS) is used to construct small inter-sensor redundant feature sets which train efficient per-sensor activity classifiers. Probabilistic decision level fusion is then used to deal with noisy and erroneous sensor data and perform real-time correlation. The potential value of the proposed algorithm for pervasive sensing is demonstrated with both simulated and experimental data.

Distributed inferencing with ambient and wearable sensors

Wireless sensor networks enable continuous and reliable data acquisition for real-time monitoring in a variety of application areas. Due to the large amount of data collected and the potential complexity of emergent patterns, scalable and distributed reasoning is preferable when compared to centralised inference as this allows network wide decisions to be reached robustly without specific reliance on particular network components. In this paper, we provide an overview of distributed inference for both wearable and ambient sensing with specific focus on graphical models—illustrating their ability to be mapped to the topology of a physical network. Examples of research conducted by the authors in the use of ambient and wearable sensors are provided, demonstrating the possibility for distributed, real-time activity monitoring within a home healthcare environment. Copyright

Deep Sequence Learning with Auxiliary Information for Traffic Prediction

Predicting traffic conditions from online route queries is a challenging task as there are many complicated interactions over the roads and crowds involved. In this paper, we intend to improve traffic prediction by appropriate integration of three kinds of implicit but essential factors encoded in auxiliary information. We do this within an encoder-decoder sequence learning framework that integrates the following data: 1) offline geographical and social attributes. For example, the geographical structure of roads or public social events such as national celebrations; 2) road intersection information. In general, traffic congestion occurs at major junctions; 3) online crowd queries. For example, when many online queries issued for the same destination due to a public performance, the traffic around the destination will potentially become heavier at this location after a while. Qualitative and quantitative experiments on a real-world dataset from Baidu have demonstrated the effectiveness of our framework.