Feasibility Study on iPhone Accelerometer for Gait Detection Herman K.Y. Chan, Huiru Zheng, Haiying Wang, Rachel Gawley, Mingjing Yang, Roy Sterritt School of Computing and Mathematics University of Ulster Northern Ireland, UK chan-h@email.ulster.ac.uk Abstract—Falls amongst the elderly is becoming a major problem with over 50% of elderly hospitalizations due to injury from fall related accidents. Healthcare expenses are dramatically rising due to growing elderly population. Many current technologies for gait analysis are laboratory-based and can incur substantial costs for the healthcare sector for treatment of falls. However utilization of alternative commercially available technologies can potentially reduce costs. Accelerometers are one such option, being ambulatory motion sensors for the detection of orientation and movement. Smart mobile devices are considered as non- invasive and increasingly contain accelerometers for detecting device orientation. This study looks at the capabilities of the accelerometer within a smart mobile device, namely the iPhone, for identification of gait events from walking along a flat surface. The results prove that it is possible to extract features from the accelerometer of an iPhone such as step detection, stride time and cadence. I. INTRODUCTION In the UK, life expectancy in 2007-2008 has risen to 77.7 years and 81.9 years for male and female respectively compared to 73 years and 79 years in 1990 [1]. The percentage of injuries due to falls increase with ageing; 86% of injuries in the over 85s are due to falls [2]. Up to one in three over 65s suffers from a fall each year, costing up to £4.6 million a day, adding up to an estimated £1.7 billion per year [3]. These statistics and costs have resulted in a growing interest in the field of fall prevention. Gait analysis can assist by assessing a patient’s walking pattern to underline abnormalities and potentially determine causes of injuries by abnormal gait. Gait analysis in the clinical environment requires specialists to assess patients with optical systems such as Codamotion [5], Vicon [6] and Qualisys [7]. Using these systems require spatial gait labs, manual settings and assessments by specialists. Accelerometer and other inertial sensor technologies have been developed and applied to explore gait pattern changes and recognizing activities of a user. Accelerometers are relatively small and low-cost devices that provide quantitative measurements. There are many types of accelerometer that can measure single to multi axis to determine acceleration from different direction to sense orientation. In using the tri-axial output of the accelerometers, basic daily movements can be classified by measuring postural orientations. For example, activPAL™ [4] is an accelerometer based system for measuring the gross activity; however it does not provide in depth information such as features of gait. The IDEEA device [8] can provide a portable solution to gross activity and gait analysis. It can identify and differentiate between more than 40 types of activities providing 17 different parameters of gait such as single or double limb support, cycle, swing and step duration, ground impact, speed, cadence, step and stride length via LifeGait system. However, this system, similar to the activPAL™ device, still requires clinical professional’s assist and it is expensive for individual users. Yang et al. has identified a subset of gait features extract from a standalone accelerometer for gait analysis [9]. Compared with standalone accelerometers, smart mobile phone is low cost and provides a user interface for ease of use. Smart mobile phones have become widely available commercially over the past decade since their first appearance in the early 90s. Smart mobile devices provide a rich technology platform for users and developers to explore mobile computing possibilities. In recent years, smartphone manufacturers have adapted micro technologies such as micro electromechanical systems (MEMS) accelerometers and gyroscopes to determine device orientation. Potentially, the MEMS accelerometers in these devices can provide meaningful data of user’s movement and may be useful for gait analysis, assessment and monitoring in replacing the conventional piezo- electric crystal based accelerometers which are considered to be large and clumsy. The objective of this research is to investigate the feasibility of using a smart mobile device to gather the acceleration raw data during free walking and analyses the data using the absolute distance algorithm. A smartphone based gait analysis system will be developed for supporting gait monitoring and assessment. The remainder of the paper is organized as follows. In Section 2, we introduce methodology used in this research. Section 3 describes the results and the paper is concluded by discussion and summary in Section 4. II. METHODOLOGY The approach used in this study follows the workflow shown in Figure 1, which includes devices, design and development of device application, data collection and analysis of data gathered. PervasiveHealth 2011, May 23-26, Dublin, Republic of Ireland Copyright © 2012 ICST DOI 10.4108/icst.pervasivehealth.2011.245995