Human Context Detection From Kinetic Energy Harvesting Wearables

Human Context Detection From Kinetic Energy Harvesting Wearables

Sara Khalifa (Data61, CSIRO, Australia & University of New South Wales, Australia), Guohao Lan (University of New South Wales, Australia & Data61, CSIRO, Australia), Mahbub Hassan (University of New South Wales, Australia & Data61, CSIRO, Australia), Wen Hu (University of New South Wales, Australia & Data61, CSIRO, Australia) and Aruna Seneviratne (University of New South Wales, Australia & Data61, CSIRO, Australia)
DOI: 10.4018/978-1-5225-3290-3.ch005
OnDemand PDF Download:


Advances in energy harvesting hardware have created an opportunity for realizing self-powered wearables for continuous and pervasive Human Context Detection (HCD). Unfortunately, the power consumption of the continuous context sensing using accelerometer is relatively high compared to the amount of power that can be harvested practically, which limits the usefulness of energy harvesting. This chapter employs and infers HCD directly from the Kinetic Energy Harvesting (KEH) patterns generated from a wearable device that harvests kinetic energy to power itself. This proposal eliminates the need for accelerometer, making HCD practical for self-powered devices. The authors discuss in more details the use of KEH patterns as an energy efficient source of information for five main applications, human activity recognition, step detection, calorie expenditure estimation, hotword detection, and transport mode detection. This confirms the potential sensing capabilities of KEH for a wide range of wearable applications, moving us closer towards self-powered autonomous wearables.
Chapter Preview


Recent advancements in wearable devices enable a wide era of human context-aware services in various domains, including healthcare (Osmani et al., 2008; Chipara et al.,2010), indoor positioning (Altun & Barshan, 2012; Khalifa et al., 2013), and fitness management (Albinali et al., 2010). Particularly, wearable sensors-based Human Context Detection (HCD) has recently become the focus of intense research and development, thus producing a wealth of tools and algorithms to accurately detect human context from data collected by the wearables (He et al., 2012). For example, a wearable sensor attached to the patient body can enable health care authorities to continuously monitor the current status of a patient from a remote centre. HCD then is expected to play a key role in reducing hospital costs by reducing the need for hospital admissions. Similarly, HCD can help individuals in monitoring their fitness level and having a better well-being by recognising various ambulation activities, such as walking, running, sitting, jogging, and so on. It has been confirmed that wearable technology coupled with HCD algorithms have the potential to improve the user's experience and quality of life.

The market of wearable devices is large, it has been found to be $20 billion in 2015 and expected to grow and reach $70 billion by 2025 (Harrop et al., 2015). Healthcare is considered the dominant sector of the wearable market, which combines medical, fitness, and wellness. It has big names such as apple, Fitbit, Google, Samsung, Nike, and Adidas. According to the International Data Corporation (IDC) Worldwide Quarterly Wearable Device Tracker report in 2016 the top leaders of the wearable market are Fitbit, Apple, Xiaomi, Samsung, and Garmin. A total of 78.1 million wearable units have been shipped in 2015, with 171.6% increase over 2014.

Almost all existing wearable products are powered by batteries. While battery technology has improved over the years, battery-powered devices cannot provide sustained operation without frequent charging. To achieve sustained operation, we either need to instrument the wearables with large batteries or be prepared to manually replenish the batteries when they die. Neither of these options is desirable because large batteries make the wearables heavy and less convenient to wear, while manual replacement is inconvenient and not a practical option for many elderly users, who may have to critically depend on such systems.

Over the past few years, a research trend in Energy Harvesting (EH) has emerged and gained the attention of the research community (Hamilton, 2012; Elvin & Erturk 2013). EH is commonly defined as the conversion of ambient energy such as vibrations, heat, wind, light, etc into electrical energy. EH devices can eliminate the need for battery replacement and significantly enhance the versatility of consumer electronics. In fact, significant advancements have been recently made in the EH hardware technology leading to many off-the-shelf products available at low cost. These developments point to future mobile devices that will be equipped with EH hardware to ease the dependence on batteries (Lee et al., 2013).

This means that it is conceptually possible to replace the battery of a wearable sensor with an EH unit to achieve perpetual sensing in many applications including HCD. Of all the ambient energy options, kinetic energy harvesting (KEH) is the most relevant option used for HCD because it can generate power directly from human motion and context. Advances in KEH hardware have motivated us to consider the concept of self-powered wearables for continuous and pervasive HCD, where numerous wearable tiny devices continue to sense and monitor the human on a permanent basis.

However, there is a caveat. KEH generally suffers from low power output (Bickerstaffe, 2015), which may challenge the power requirement of the wearable sensor's components, such as the accelerometer used for sampling human motion. Given that the sensor will also have to turn on its radio for occasional communications with a nearby sink, the power generated from energy harvesting is clearly too small to simply port the existing battery-powered wearables into energy-harvesting wearables. In fact, using energy harvesting to provide self-powered wearables is a very challenging problem that requires innovative sensing and communication solutions.

Complete Chapter List

Search this Book: