Abstract
Human activity recognition (HAR) has recently received significant attention due to its wide range of applications in health and activity monitoring. The nature of these applications requires mobile or wearable devices with limited battery capacity. User surveys show that charging requirement is one of the leading reasons for abandoning these devices. Hence, practical solutions must offer ultra-low power capabilities that enable operation on harvested energy. To address this need, we present the first fully integrated custom hardware accelerator (HAR engine) that consumes 22.4 μJ per operation using a commercial 65 nm technology. We present a complete solution that integrates all steps of HAR, i.e., reading the raw sensor data, generating features, and activity classification using a deep neural network (DNN). It achieves 95% accuracy in recognizing 8 common human activities while providing three orders of magnitude higher energy efficiency compared to existing solutions.
- Martín Abadi et al. 2015. TensorFlow: Large-Scale Machine Learning on Heterogeneous Systems. [Online] http://tensorflow.org/, accessed 31 July 2019.Google Scholar
- Muhammad Arif, Mohsin Bilal, Ahmed Kattan, and S. Iqbal Ahamed. 2014. Better physical activity classification using smartphone acceleration sensor. J. of Med. Syst. 38, 9 (2014), 95.Google ScholarDigital Library
- Ferhat Attal et al. 2015. Physical human activity recognition using wearable sensors. Sensors 15, 12 (2015), 31314--31338.Google ScholarCross Ref
- Akin Avci, Stephan Bosch, Mihai Marin-Perianu, Raluca Marin-Perianu, and Paul Havinga. 2010. Activity recognition using inertial sensing for healthcare, wellbeing and sports applications: A survey. In Proc. Int. Conf. on Arch. of Comput. Syst. 1--10.Google Scholar
- Ganapati Bhat, Kunal Bagewadi, Hyung Gyu Lee, and Umit Y. Ogras. 2019. REAP: Runtime energy-accuracy optimization for energy harvesting IoT devices. In Proc. of Annual Design Autom. Conf. 171:1--171:6.Google Scholar
- Ganapati Bhat, Ranadeep Deb, Vatika Vardhan Chaurasia, Holly Shill, and Umit Y. Ogras. 2018. Online human activity recognition using low-power wearable devices. In Proc. Int. Conf. on Comput.-Aided Design. 72:1--72:8.Google Scholar
- Ganapati Bhat, Ranadeep Deb, and Umit Y. Ogras. 2019. OpenHealth: Open source platform for wearable health monitoring. IEEE Design 8 Test (2019). To be published, DOI:10.1109/MDAT.2019.2906110.Google Scholar
- Ganapati Bhat, Jaehyun Park, and Umit Y. Ogras. 2017. Near-optimal energy allocation for self-powered wearable systems. In Proc. Int. Conf. on Comput.-Aided Design. 368--375.Google Scholar
- Judit Bort-Roig, Nicholas D. Gilson, Anna Puig-Ribera, Ruth S. Contreras, and Stewart G. Trost. 2014. Measuring and influencing physical activity with smartphone technology: A systematic review. Sports Medicine 44, 5 (2014), 671--686.Google ScholarCross Ref
- Wouter Bracke, Patrick Merken, Robert Puers, and Chris Van Hoof. 2006. A 1 cm3 modular autonomous sensor node for physical activity monitoring. In 2006 Ph.D. Research in Microelectronics and Electronics. 429--432.Google Scholar
- Yufei Chen and Chao Shen. 2017. Performance analysis of smartphone-sensor behavior for human activity recognition. IEEE Access 5 (2017), 3095--3110.Google ScholarCross Ref
- François Chollet et al. 2015. Keras. [Online] https://keras.io, accessed 31 July 2019.Google Scholar
- Ana Lígia Silva de Lima et al. 2017. Feasibility of large-scale deployment of multiple wearable sensors in Parkinson’s disease. PLOS One 12, 12 (2017), e0189161.Google ScholarCross Ref
- DMI International Distribution Ltd. 2017. Curved Lithium Thin Cells. [Online] http://www.dmi-international.com/data%20sheets/Curved%20Li%20Polymer.pdf Accessed 31 July 2019.Google Scholar
- Matthias Geisler et al. 2017. Human-motion energy harvester for autonomous body area sensors. Smart Materials and Structures 557, 1 (2017), 012024.Google Scholar
- Norbert Győrbíró, Ákos Fábián, and Gergely Hományi. 2009. An activity recognition system for mobile phones. Mobile Networks and Appl. 14, 1 (2009), 82--91.Google ScholarDigital Library
- Mostafa Haghi, Kerstin Thurow, and Regina Stoll. 2017. Wearable devices in medical Internet of Things: Scientific research and commercially available devices. Healthcare Informatics Research 23, 1 (2017), 4--15.Google ScholarCross Ref
- Mark Hall, Eibe Frank, Geoffrey Holmes, Bernhard Pfahringer, Peter Reutemann, and Ian H. Witten. 2009. The WEKA data mining software: An update. ACM SIGKDD Explorations Newsletter 11, 1 (2009), 10--18.Google ScholarDigital Library
- Nils Y. Hammerla, Shane Halloran, and Thomas Plötz. 2016. Deep, convolutional, and recurrent models for human activity recognition using wearables. In Proc. Int. Joint Conf. on Artificial Intell. 1533--1540.Google Scholar
- Dustin A. Heldman et al. 2017. Telehealth management of Parkinson’s disease using wearable sensors: An exploratory study. Digital Biomarkers 1, 1 (2017), 43--51.Google ScholarCross Ref
- Diederik P. Kingma and Jimmy Ba. 2014. Adam: A method for stochastic optimization. In Proc. Int. Conf. on Learning Representations. 1--13.Google Scholar
- Morwenna Kirwan, Mitch J. Duncan, Corneel Vandelanotte, and W. Kerry Mummery. 2012. Using smartphone technology to monitor physical activity in the 10,000 steps program: A matched case--control trial. J. of Med. Internet Research 14, 2 (2012).Google ScholarCross Ref
- Alicia Klinefelter et al. 2015. 21.3 A 6.45 μW self-powered IoT SoC with integrated energy-harvesting power management and ULP asymmetric radios. In Proc. IEEE Int. Solid-State Circuits Conf. (ISSCC) Dig. of Techn. Papers. 384--385.Google Scholar
- Raghuraman Krishnamoorthi. 2018. Quantizing deep convolutional networks for efficient inference: A whitepaper. arXiv preprint arXiv:1806.08342 (2018).Google Scholar
- Jennifer R. Kwapisz, Gary M. Weiss, and Samuel A. Moore. 2011. Activity recognition using cell phone accelerometers. ACM SigKDD Explorations Newsletter 12, 2 (2011), 74--82.Google ScholarDigital Library
- Oscar D. Lara and Miguel A. Labrador. 2013. A survey on human activity recognition using wearable sensors. IEEE Commun. Surveys 8 Tut. 15, 3 (2013), 1192--1209.Google Scholar
- Baihua Li and Horst Holstein. 2004. Perception of human periodic motion in moving light displays - A motion-based frequency domain approach. Interdisciplinary J. of Artificial Intell. and the Simulation of Behaviour (AISBJ) 1, 5 (2004), 403--416.Google Scholar
- Xin Liu et al. 2012. An ultra-low power ECG acquisition and monitoring ASIC system for WBAN applications. IEEE J. on Emerg. and Sel. Topics in Circuits Syst. 2, 1 (2012), 60--70.Google ScholarCross Ref
- Reza Lotfian and Roozbeh Jafari. 2013. An ultra-low power hardware accelerator architecture for wearable computers using dynamic time warping. In Proc. Conf. on Design, Autom. and Test in Europe. 913--916.Google ScholarDigital Library
- Yuxuan Luo et al. 2017. A 93μW 11Mbps wireless vital signs monitoring Soc with 3-lead ECG, bio-impedance, and body temperature. In Proc. IEEE Asian Solid-State Circuits Conf. 29--32.Google Scholar
- Walter Maetzler, Jochen Klucken, and Malcolm Horne. 2016. A clinical view on the development of technology-based tools in managing Parkinson’s disease. Movement Disorders 31, 9 (2016), 1263--1271.Google ScholarCross Ref
- Luis Morillo, Luis Gonzalez-Abril, Juan Ramirez, and Miguel de La Concepcion. 2015. Low energy physical activity recognition system on smartphones. Sensors 15, 3 (2015), 5163--5196.Google ScholarCross Ref
- Arsalan Mosenia, Susmita Sur-Kolay, Anand Raghunathan, and Niraj K. Jha. 2017. Wearable medical sensor-based system design: A survey. IEEE Trans. Multi-Scale Comput. Syst. 3, 2 (2017), 124--138.Google ScholarCross Ref
- Ben O’Brien, Todd Gisby, and Iain A. Anderson. 2014. Stretch sensors for human body motion. In Proc. Electroactive Polymer Actuators and Devices, Vol. 9056. 905618.Google Scholar
- Francisco Ordóñez and Daniel Roggen. 2016. Deep convolutional and LSTM recurrent neural networks for multimodal wearable activity recognition. Sensors 16, 1 (2016), 115.Google ScholarCross Ref
- Anneli Ozanne et al. 2018. Wearables in epilepsy and Parkinson’s disease-A focus group study. Acta Neurologica Scandinavica 137, 2 (2018), 188--194.Google ScholarCross Ref
- Jaehyun Park, Hitesh Joshi, Hyung Gyu Lee, Sayfe Kiaei, and Umit Y. Ogras. 2017. Flexible PV-cell modeling for energy harvesting in wearable IoT applications. ACM Trans. Embedd. Comput. Syst. 16, 5s (2017), 156:1--156:20.Google ScholarDigital Library
- Susanna Pirttikangas, Kaori Fujinami, and Tatsuo Nakajima. 2006. Feature selection and activity recognition from wearable sensors. In Proc. Int. Symp. on Ubiquitious Comput. Systems. 516--527.Google ScholarDigital Library
- Stephen J. Preece et al. 2009. Activity identification using body-mounted sensors--A review of classification techniques. Physiological Measurement 30, 4 (2009), R1.Google ScholarCross Ref
- Daniel Roggen et al. 2010. Collecting complex activity datasets in highly rich networked sensor environments. In Proc. Int. Conf. on Networked Sensing Syst. (INSS). 233--240.Google Scholar
- Muhammad Shoaib, Stephan Bosch, Ozlem Durmaz Incel, Hans Scholten, and Paul J. M. Havinga. 2015. A survey of online activity recognition using mobile phones. Sensors 15, 1 (2015), 2059--2085.Google ScholarCross Ref
- Texas Instruments Inc. 2016. CC-2650 Microcontroller. [Online] http://www.ti.com/product/CC2650, accessed 31 July 2019.Google Scholar
- US Department of Labor. 2017. American Time Use Survey. [Online] https://www.bls.gov/tus/, accessed 13 July 2019.Google Scholar
- Adrian Valenzuela. 2008. Energy Harvesting for No-Power Embedded Systems. [Online] http://focus.ti.com/graphics/mcu/ulp/energy_harvesting_embedded_systems_using_msp430.pdf, accessed 31 July 2019.Google Scholar
- Nick Van Helleputte et al. 2014. 18.3 A multi-parameter signal-acquisition Soc for connected personal health applications. In Proc. IEEE Int. Solid-State Circuits Conf. (ISSCC) Dig. of Techn. Papers. 314--315.Google Scholar
- Alan CW Wong et al. 2008. A 1 V, micropower system-on-chip for vital-sign monitoring in wireless body sensor networks. In Proc. IEEE Int. Solid-State Circuits Conf. (ISSCC) Dig. of Techn. Papers. 138--602.Google Scholar
Index Terms
- An Ultra-Low Energy Human Activity Recognition Accelerator for Wearable Health Applications
Recommendations
Energy-efficient human activity recognition for self-powered wearable devices
ACSW '17: Proceedings of the Australasian Computer Science Week MulticonferenceAdvances in energy harvesting hardware have created an opportunity for realizing self-powered wearables for continuous and pervasive human activity recognition (HAR). Unfortunately, the power requirements of the continuous activity sensing using ...
SensorGAN: A Novel Data Recovery Approach for Wearable Human Activity Recognition
Human activity recognition (HAR) and, more broadly, activities of daily life recognition using wearable devices have the potential to transform a number of applications, including mobile healthcare, smart homes, and fitness monitoring. Recent approaches ...
A survey on wearable sensor modality centred human activity recognition in health care
Highlights- Discuss limitations and strengths of different sensor modalities.
- Focus on the ...
AbstractIncreased life expectancy coupled with declining birth rates is leading to an aging population structure. Aging-caused changes, such as physical or cognitive decline, could affect people's quality of life, result in injuries, mental ...
Comments