Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

In order to cope with our aging society, a current vision in the area of ICT-supported independent living of the elderly involves populating the smart home with connected electronic devices (“things”, such as sensors and actuators) and linking them to the Internet. The mission of the E-care@home project is to create such an Internet-of-Things (IoT) infrastructure with the ambition to provide automated information gathering and processing on top of which e-services for the elderly residing in their homes can be built [2].

While the things need to communicate, wiring them is, however, unfeasible, inflexible and costly. Recently, low-power wireless communication has made tremendous progress. Today, we can network at least tens of stationary, battery-driven sensors and actuators wirelessly with a lifetime of several years using low-power IP-based communication stacks [6] on top of the IEEE 802.15.4 standard. In a smart home for elderly such communication means can be used to provide a reliable infrastructure consisting of stationary nodes. Stationary nodes are, however, not enough as monitoring the health condition of elderly people also requires on-body sensors such as physical activity and weight monitoring, blood pressure, blood glucose, heart rate, and oxygen saturation. Most of these sensors do not come with support for IEEE 802.15.4, but use Bluetooth—in particular Bluetooth Low Energy (BLE). This leads to a hybrid communication architecture where stationary nodes communicate with each other using communication protocols on top of the IEEE 802.15.4 standards while on-body sensors use BLE. For communication between on-body and stationary nodes, some stationary nodes are also equipped with BLE radios.

Our healthcare application requires that in some situations, messages from sensor nodes must be delivered timely and with very high reliability. In the following we explain more about our architecture, which caters to these application requirements, and provide some initial evaluation results.

2 Architecture

The E-Care@Home software architecture, illustrated in Fig. 1, comprises different components that are built to execute in Contiki, an open-source operating system for the IoT [1]. To be able to meet specific application goals regarding metrics such as packet delivery rate, energy consumption, latency, and node lifetime, we employ TSCH (Time Slotted Channel Hopping MAC), one of the MAC protocols of the IEEE802.15.4-2015 standard [3]. At the routing layer, we use RPL [8], the routing protocol for low-power IPv6 networks standardized by the IETF ROLL working group.

Table 1 shows some results from our previous work [5], which reveals that even without centralized scheduling, TSCH achieves end-to-end delivery ratios of over 99.99 %. Hence, we improve reliability by two orders of magnitude compared to asynchronous low-power MAC protocols, while achieving a similar latency-energy balance. Furthermore, we can provide bounds on energy consumption. One of the challenges in our project is to extend the performance bounds to end-to-end latency and also include the on-body sensors that use BLE.

Fig. 1.
figure 1

The software architecture of static infrastructure nodes comprises a low-power IPv6 stack, built on top of TSCH. The application software runs in a virtualized environment that provides safe execution.

Full size image
Table 1. Performance of Contiki’s 6TiSCH implementation on a 98-node testbed. RPL + TSCH data collection at a 1-minute packet interval [7].
Full size table

Another key component of the architecture is a virtual machine, which executes bytecode in a safe manner so that application software cannot exceed their assigned privileges. Inside the virtual machine, the sensing application executes and communicates sensor samples and system status to a base station using CoAP and UDP over IPv6. Depending on the type of sensor, the communication can occur at regular intervals or in response to events. Some health parameters, such as blood glucose and weight, are measured sparsely; whereas others, such as electrocardiography and respiratory rate, are measured continuously at specific time periods [4].

The base station, which collects all incoming messages from the infrastructure nodes and the on-body nodes, contains a sensor database, which stores all sensor samples in a structured manner, and which can be used to extract data for context-aware data processing and reasoning. Another application residing in the virtual machine is the runtime assurance application, which continuously monitors the main parts of the system, and ensures that the performance stays within the guaranteed bounds.