Elsevier

Computers in Biology and Medicine

Volume 95, 1 April 2018, Pages 236-247
Computers in Biology and Medicine

Adaptive and technology-independent architecture for fault-tolerant distributed AAL solutions

https://doi.org/10.1016/j.compbiomed.2017.11.002Get rights and content

Highlights

  • Establishment of a system model with fault-tolerance for safety-critical AAL systems.

  • Partitioning of communication resources by implementing a Traffic Shaping Layer.

  • Establishment of a flexible message-based run-time environment using Microservices.

  • Extension of IEEE11073 for Bluetooth LE (BLE) devices.

Abstract

Today's architectures for Ambient Assisted Living (AAL) must cope with a variety of challenges like flawless sensor integration and time synchronization (e.g. for sensor data fusion) while abstracting from the underlying technologies at the same time. Furthermore, an architecture for AAL must be capable to manage distributed application scenarios in order to support elderly people in all situations of their everyday life. This encompasses not just life at home but in particular the mobility of elderly people (e.g. when going for a walk or having sports) as well. Within this paper we will introduce a novel architecture for distributed AAL solutions whose design follows a modern Microservices approach by providing small core services instead of a monolithic application framework. The architecture comprises core services for sensor integration, and service discovery while supporting several communication models (periodic, sporadic, streaming). We extend the state-of-the-art by introducing a fault-tolerance model for our architecture on the basis of a fault-hypothesis describing the fault-containment regions (FCRs) with their respective failure modes and failure rates in order to support safety-critical AAL applications.

Introduction

Aging of the population is an ongoing trend visible in the increasing share of older people in the structure of the population. According to the Aging report of the European Commission from 2015 [1], by 2030 around 25% of the EU population will be over 65 and the amount of people aged from 65 to 80 will rise by nearly 40% between 2010 and 2030. This demographic change in combination with the wish of elderly people for a self-determined life and the extensive costs for the assistance and care for elderly people led to an upcoming interest in systems for Ambient Assisted Living (AAL).

In order to provide assistance in activities of elderly people in their everyday life, systems for AAL are typically realized as cyber-physical systems (CPS) which must cope with many technical challenges like integration of heterogeneous sensors and multiple communication protocols. The integration of different types of sensors needs to offer a high degree of autonomy in order to enable easy handling by the elderly person or the caregiver. In addition, dynamic setups must be supported with variable numbers of users while supporting multiple sensors per user at the same time. In order to support everyday life, systems for Ambient Assisted Living must not focus on the apartment or house of the elderly people. Rather, an AAL system must also support an elderly person in outdoor scenarios. For example, when an elderly is visiting a doctor's office or is going for a walk. Integrating professional services like information or medical services from caregivers or a doctor should be supported by a system for AAL as well. These distributed AAL solutions requires a highly adaptive and technology-independent architecture.

The proposed architecture will provide flexibility, reliability and technology-independence by introducing a fault-tolerant, distributed message-oriented system architecture using a modern Microservices approach. The fault-tolerance model introduced within this paper is one of the major contributions beyond the state-of-the-art and permits safety-critical scenarios with high reliability.

The following sections of this paper are structured as follows. First, in section 2 the main technical challenges that architectures for AAL are facing today are introduced. Section 3 continues with the analysis of the state-of-the-art. Section 4 introduces the most important background technologies applied in our fault-tolerant system model which is explained in section 5 comprising the system architecture, the fault-tolerance concept and applied fault-tolerance techniques. Further, section 6 shows the implementation of the runtime environment and the traffic shaping layer. Finally, section 7 presents the evaluation of our fault-tolerant system model.

Section snippets

Challenges

In this section, we discuss the fundamental requirements and challenges of an adaptive and distributed architecture for elderly care. The requirements can be derived directly from the use cases where AAL systems are deployed. Many of the uses cases require the integration of heterogeneous devices, incorporating simple weighting scales and even complex medical sensors. Often, this integration has to be conducted seamlessly at run-time demanding a dynamic reconfiguration of the AAL system where

Related work

Research projects like SOPRANO [4], PERSONA [5] and universAAL [6] focused on building platforms for AAL. The SOPRANO project aimed at the development of an open platform for AAL solutions by creating an ontology-based architecture on top of OSGi (Open Services Gateway Initiative). Strictly predefined interfaces based on contracts should allow the extension of the system with low effort. The middleware of the PERSONA research project was also built on top of OSGi. To achieve extensibility of

Background

The following section introduces the background technologies and design approaches used in the proposed architecture including technologies for message dissemination, service registration, service discovery and sensor integration.

Fault-tolerant system model

The following section introduces the system model including the main parts of the proposed architecture. A high-level overview is given in Fig. 5. As depicted, the system model comprises local services and sensor ecosystems interconnected by a local area network. The sensor and service ecosystem comprises at least a service gateway which runs the later introduced runtime environment with a common set of core services for interconnectivity, message dissemination, sensor integration and service

Traffic shaping layer

As introduced in the fault tolerance concept, COTS switches cannot be used as such for setting up a fault-tolerant communication infrastructure because they are missing fault-tolerance features such as like traffic shaping and redundancy. Therefore, our approach introduces a traffic shaping layer as part of the runtime environment.

The traffic shaping layer prevents on the one hand side the nodes from producing more network traffic as allowed, which would cause for example buffer congestion in

Evaluation

The evaluation of our architecture was performed in four subsequent steps. At first, we evaluated our implementations in an experimental setup. The main goals during this phase was to evaluate the correctness, reliability and performance of the core services within the runtime environment. Subsequent to this phase we evaluated the architecture in a real world scenario by joining a living lab experiment in the ongoing BMBF (Federal Ministry of Education and Research) funded project Cognitive

Conclusion and future work

As shown in the results of the experimental setup and the living lab experiment, our architecture complies with the challenges introduced in section 2. Though, the support for the interconnection at Internet level is not evaluated yet and we are currently extending our efforts to address this challenge as well. After finishing this step, we will extend the architecture by implementing services for hard real-time communication based on time-triggered concepts and add further fault-tolerance

Acknowledgment

This work has been supported in part by the European project H2020 Safe power under project No. 687902 and the German BMBF project Cognitive Village (grant 16SV7223K).

References (33)

  • European Commission

    The 2015 Ageing Report

    (2015)
  • S. Newman

    Building Microservices

    (2015)
  • F. Zhu et al.

    Service discovery in pervasive computing environments

    IEEE Pervasive Comput.

    (2005)
  • P. Wolf et al.

    Soprano-an extensible, open aal platform for elderly people based on semantical contracts

  • M.-R. Tazari et al.

    The Persona Service Platform for Aal Spaces, in: Handbook of Ambient Intelligence and Smart Environments

    (2010)
  • S. Hanke et al.

    Universal–an open and consolidated aal platform

  • M. Memon et al.

    Ambient assisted living healthcare frameworks, platforms, standards, and quality attributes

    Sensors

    (2014)
  • Health informatics - personal health device communication part 00103: Overview

    (2012)
  • Signove Tecnologia

    Antidote, an Open-source IEEE 11073-20601 Stack

    (2014)
  • M. Sousa et al.

    Human tracking and identification using a sensitive floor and wearable accelerometers

  • Empatica Inc

    E4 Wristband

    (2015)
  • H. Kopetz

    Real-time Systems: Design Principles for Distributed Embedded Applications (Real-time Systems Series)

    (2011)
  • R. Obermaisser

    Time-triggered Communication

    (2011)
  • R. Obermaisser et al.

    A fault hypothesis for integrated architectures

  • 664P7–1 Aircraft Data Network, Part 7, Avionics Full-duplex Switched Ethernet Network, International Standard

    (2009)
  • J. Vila-Carbo et al.

    Analysis of switched ethernet for real-time transmission

  • Cited by (11)

    View all citing articles on Scopus

    This document is a collaborative effort.

    View full text