Towards Behaviour Recognition with Unlabelled Sensor Data

The problem of behaviour recognition based on data from sensors is essentially an inverse problem: given a set of sensor observations, identify the sequence of behaviours that gave rise to them. In a smart home, the behaviours are likely to be the standard human behaviours of living, and the observations will depend upon the sensors that the house is equipped with. There are two main approaches to identifying behaviours from the sensor stream. One is to use a symbolic approach, which explicitly models the recognition process. Another is to use a sub-symbolic approach to behaviour recognition, which is the focus in this chapter, using data mining and machine learning methods. While there have been many machine learning methods of identifying behaviours from the sensor stream, they have generally relied upon a labelled dataset, where a person has manually identified their behaviour at each time. This is particularly tedious to do, resulting in relatively small datasets, and is also prone to significant errors as people do not pinpoint the end of one behaviour and commencement of the next correctly. In this chapter, the authors consider methods to deal with unlabelled sensor data for behaviour recognition, and investigate their use. They then consider whether they are best used in isolation, or should be used as preprocessing to provide a training set for a supervised method.

Novelty Detection in Human Behavior through Analysis of Energy Utilization

The value of smart environments in understanding and monitoring human behavior has become increasingly obvious in the past few years. Using data collected from sensors in these environments, scientists have been able to recognize activities that residents perform and use the information to provide context-aware services and information. However, less attention has been paid to monitoring and analyzing energy usage in smart homes, despite the fact that electricity consumption in homes has grown dramatically. In this chapter, the authors demonstrate how energy consumption relates to human activity through verifying that energy consumption can be predicted based on the activity that is being performed. The authors then automatically identify novelties in human behavior by recognizing outliers in energy consumption generated by the residents in a smart environment. To validate these approaches, they use real energy data collected in their CASAS smart apartment testbed and analyze the results for two different data sets collected in this smart home.

Motion and Location-Based Online Human Daily Activity Recognition

In this chapter, the authors propose an approach to indoor human daily activity recognition that combines motion data and location information. One inertial sensor is worn on the thigh of a human subject to provide motion data while a motion capture system is used to record the human location information. Such a combination has the advantage of significantly reducing the obtrusiveness to the human subject at a moderate cost of vision processing, while maintaining a high accuracy of recognition. The approach has two phases. First, a two-step algorithm is proposed to recognize the activity based on motion data only. In the coarse-grained classification, two neural networks are used to classify the basic activities. In the fine-grained classification, the sequence of activities is modeled by a Hidden Markov Model (HMM) to consider the sequential constraints. The modified short-time Viterbi algorithm is used for real-time daily activity recognition. Second, to fuse the motion data with the location information, Bayes’ theorem is used to refine the activities recognized from the motion data. The authors conduct experiments in a mock apartment, and the obtained results prove the effectiveness and accuracy of the algorithms.

From Streams of Observations to Knowledge-Level Productive Predictions

The benefit to be gained by Ambient Assisted Living (AAL) systems depends heavily on the successful recognition of human intentions. Important indicators for specific intentions are behavior and situational context. Once a sequence of actions can be associated with a specific intention, assistance may be provided by anticipating the next individual step and supporting the human in its execution. The authors present a combination of Sequence Abstraction Networks (SAN) and IGOR to guarantee early and impartial predictions with a powerful detection for symbolic regularities. They first generate a hierarchy of abstract action sequences, where individual contexts represent subgoals or minor intentions. Afterwards, they enrich this hierarchy by recursive induction. An example scenario is presented where a table needs to be set for several guests. It turns out that correct predictions can be made while still executing the observed sequence for the first time. Support can therefore be completely individual to the person being assisted but nonetheless be very dynamic and quick in anticipating the next steps.

Monitoring Social Life and Interactions

The chapter describes the sociological perspective of monitoring technologies and debates its method for analysing social implications of scientific and technical developments. It is articulated in five sections dedicated to social and privacy aspects involved in social analysis of technologies. Particular attention is devoted to social network analysis, an emergent area of sociological research that focuses on the relational implications of technologies in organizations, small groups, and other contexts of social participation. The text integrates examples of technology implementation from healthcare automated assistance to mobile communication devices, video-surveillance, RFID, and smart-meter technology. Case studies, illustrated in separate textboxes, describe the advancements in this field of enquiry and highlight the main elements of the structure of interactions in virtual and technology-mediated communications. Finally, ethical implications of behaviour monitoring technologies are discussed together with recent perspectives of sociological research.

A Logic-Based Approach to Activity Recognition

The authors have been developing a system for recognising human activities given a symbolic representation of video content. The input of the system is a stream of time-stamped short-term activities detected on video frames. The output of the system is a set of recognised long-term activities, which are pre-defined spatio-temporal combinations of short-term activities. The constraints on the short-term activities that, if satisfied, lead to the recognition of a long-term activity, are expressed using a dialect of the Event Calculus. The authors illustrate the expressiveness of the dialect by showing the representation of several typical complex activities. Furthermore, they present a detailed evaluation of the system through experimentation on a benchmark dataset of surveillance videos.

Clustering Algorithm for Human Behavior Recognition Based on Biosignal Analysis

Time series unsupervised clustering is accurate in various domains, and there is an increased interest in time series clustering algorithms for human behavior recognition. The authors have developed an algorithm for biosignals clustering, which captures the general morphology of a signal’s cycles in one mean wave. In this chapter, they further validate and consolidate it and make a quantitative comparison with a state-of-the-art algorithm that uses distances between data’s cepstral coefficients to cluster the same biosignals. They are able to successfully replicate the cepstral coefficients algorithm, and the comparison showed that the mean wave approach is more accurate for the type of signals analyzed, having a 19% higher accuracy value. They authors also test the mean wave algorithm with biosignals with three different activities in it, and achieve an accuracy of 96.9%. Finally, they perform a noise immunity test with a synthetic signal and notice that the algorithm remains stable for signal-to-noise ratios higher than 2, only decreasing its accuracy with noise of amplitude equal to the signal. The necessary validation tests performed in this study confirmed the high accuracy level of the developed clustering algorithm for biosignals that express human behavior.

Sensors for Smart Homes

Sensors are increasingly being employed to determine different activities of a person living at home. Numerous sensors can be used to obtain a variety of information. While many sensors may be used to make a system, it is important to look into the availability, cost, installation, mechanism, and performance of sensors. This chapter investigates different sensors and their usefulness in a smart home monitoring system. A smart home monitoring system provides a safe, sound, and secure living environment for elderly people. Statistics show that the population of elderly people is increasing around the world and this trend is not going to change in the near future. The authors have developed a smart home that consists of an optimum number of wireless sensors that includes current flow, water flow, and bed usage sensors. The sensors provide information that can be used for monitoring elderly people by detecting abnormal patterns in their daily activities. The system generates and sends an early warning message to the caregiver when an unforeseen abnormal condition occurs.

Intention-Based Decision Making via Intention Recognition and its Applications

In this chapter, the authors present an intention-based decision-making system. They exhibit a coherent combination of two Logic Programming-based implemented systems, Evolution Prospection and Intention Recognition. The Evolution Prospection system has proven to be a powerful system for decision-making, designing, and implementing several kinds of preferences and useful environment-triggering constructs. It is here enhanced with an ability to recognize intentions of other agents—an important aspect not well explored so far. The usage and usefulness of the combined system are illustrated with several extended examples in different application domains, including Moral Reasoning, Ambient Intelligence, Elder Care, and Game Theory.

Behaviour Monitoring and Interpretation

This chapter describes the field of Behaviour Monitoring and Interpretation, BMI for short, which defines a framework for the analysis and design of systems for the monitoring and interpretation of human behaviour. As an example scenario which is analysed by means of that framework, a pedestrian navigation and service tool is presented. This scenario is about a mobile user who is wearing a hearing-aid similar device that instructs him while walking through the city. The navigation assistant can be equipped with specific application constraints in order to enrich the navigation system with an application context. The navigation system guides the user through the environment while taking care of the application constraints. One application context is a child at pre-school age: within this context the idea is to guide the child along a safe path to kindergarten. There are many challenges involved in the development of such a pedestrian navigation system. This chapter focuses on the analysis of the behaviour of the user that determines how the navigation assistant can provide help in an appropriate way. By this means, principles underlying the field of behaviour monitoring and interpretation are explained. More specifically, how the BMI framework aids in analysing is shown along with how top-down and bottom-up processes are to be involved in behaviour recognition; additionally, how the framework supports the identification of information fusion at different abstraction layers is shown.