Abstract— Gait analysis methods to evaluate the spatiotemporal variability based on Inertial Measurements Units (IMUs) are present in literature. In this paper a preliminary study on the time gait variability as a tool for early diagnosis is presented. The system is non-invasive as it is composed by only a single axis gyroscope positioned on the patient shin. The unit uses a Bluetooth Low Energy interface for automatic data download and analysis. Twenty cognitively normal subjects have been used in order to exploit the system’s capability, with very good results. They were asked to use the device during the day in the normal activity. The algorithm is able to isolate the walk session and to calculate the global duration, the mean and the standard deviation of the gait duration. The proposed wearable gait analysis device is a promising tool for clinical study. Keywords—Alzheimer’s disease, gait analysis, inertial systems, Parkinson’s disease, wearable sensors. I. INTRODUCTION HE life span expectation strongly increased in the second half of 20th century. As a result, many diseases strongly bound to the aging of the society, such as senile dementia and Alzheimer’s disease (AD), become more commune and apparent. This latter disease, AD, implies severe limitation as patients’ conditions worsen and gait disorder, balance and cognitive problems, dementia and memory evanescence become extremely apparent [1-2]. The gait disorders and balance by themselves represent a threat to patient and strongly limit their independence. Recent literature reported that the gait and balance monitoring not only represent a clinical tool but also furnish a predictive tool for potential AD diagnosis [3]. A comparison between groups of early AD patients and healthy people showed that the first group had slower velocity, slower cadence, and shorter stride length than the second group [4]. Several techniques have been introduced to support the researches in this field, mainly based video monitoring or special pressure sensitive mats. The availability of ever integrated and compact electronic devices furnishes new fields of investigation to the research. In particular, the availability of three-axial gyroscopes and accelerometers, joint to micro-baric sensors gives some basic bricks to be integrated in a very compact and wearable diagnostic device. Some preliminary paper reports the potentialities of such approach [5]. The present paper introduces an experimental wearable system based on a single integrated unit which integrates a single-axis gyroscope with an RF link, creating an ad-hoc wireless sensor network (WSN). The data collected by the system are remotely processed using a proprietary algorithm that furnishes, as a result, indications regarding patient gait variability. In the forthcoming sections, the system architecture will be discussed, giving some insight on the hardware and algorithm organization. At the end, some preliminary data are reported and discussed. II. SYSTEM ARCHITECTURE The systems basic brick is a wearable, inexpensive and automated Inertial Measurement Unit (IMU) able independently measures the movement patterns during daily life. Starting from the measured inertial data an algorithm allows an accurate gait analysis and the assessment of spatiotemporal gait parameters (stance-swing phase, stride time and its variability, etc.). An expert neurological team has supported the IMU development and the final objective is the definition of an automated and impartial procedure based on IMU and software for the preliminary variability gait analysis of patients with Parkinson’s disease, validated by authorized health service providers. The main challenge was to obtain a system able to accurately measure gait parameters during daily life activities. The clinical picture assessed during an outpatient check up in the medical office poorly represents the real (actual) clinical status, especially in fluctuating patients. For this reason, recently researches on automatic systems based on wearable sensors to evaluate motor of PD patients have been developed. In general, the most common evaluated performances are motor tasks, such as “sit-to-stand”, gait cycle. There are alternative gait analysis systems available, but none meet all these criteria. Camera-based systems give high accuracy but are relatively costly, time-consuming to use, and gait analysis is typically restricted to laboratory settings. Pressure sensing mat systems are also relatively expensive, with data collection limited to 5–6 walking strides per pass, Preliminary study on gait variability analysis with a single axis gyroscope for Alzheimer and Parkinson’s diseases G. Avitabile, G. Coviello, N. Margiotta T INTERNATIONAL JOURNAL OF BIOLOGY AND BIOMEDICAL ENGINEERING Volume 9, 2015 ISSN: 1998-4510 155