**1. Introduction**

Early detection of symptoms enables better chances of treatment and sometimes complete cure of many diseases. Many health-related symptoms can be diagnosed in their primitive stages by monitoring body functions continuously. With the advancement in wearable sensor technology, many new pervasive monitoring systems are available in the market that can track physiological signals and wirelessly transmit them [1]. Many issues are related to gait disorders such as asynchronous body movement, asymmetry, loss of smoothness, and slowing of gait speed, etc. The understanding and identification of the human gait have generated much interest in many research fields beyond medical sciences such as biomechanics, etc. For the study of human gait, researchers have used various techniques in lab environments [2].

A gait cycle is described as the sequence and rhythm of foot/limb movements in the walking pattern of a person to achieve the motive of locomotion by propelling the body's center of gravity forward. A gait cycle comprises of a stance phase and a swing phase. The stance phase corresponds to the instance while walking when the body balances on both the feet, whereas the swing phase is the one in which the body balances on one foot while the other foot is off the ground to take the next step. Normal or healthy individuals have a normal gait, but certain neurological and musculoskeletal

disorders cause abnormality in gait pattern [3,4]. Notably, mobility and disability problems, including the gait disorder, tend to increase within the elderly population. Some gait cycle patterns generally change with age, such as gait velocity, double stance time, walking posture, and joint motion [5].

Gait assessment is one of the essential steps when diagnosing neurological conditions, orthopedic problems and medical conditions. [6]. Manifestations of an abnormal gait are the loss of symmetry of motion and timing between the left and right sides, which can cause difficulty in initiating or maintaining pace. The three-dimensional gait analysis enables the detection of angular deflections during gait; this is important for clinical purposes to be able to identify different disorders accurately [7]. A standard method for measuring gait cycle is to use optical tracking systems such as the Vicon motion analysis system. A set of cameras track the markers attached to the human body to measure and calculate the three-dimensional positions of the body segments. However, this tracking approach needs controlled environments and limits the area of operation in a laboratory room, where the camera and the equipment are located [8].

As an alternative solution, the application of the accelerometer provides a simple means for analyzing gait in real-time for monitoring physical activity levels and classifying movements [9–12]. Gyroscopes were applied in the measurement of angular rate and the angle of various joints by mounting multiple devices on the lower or upper extremities [12–15]. In this paper, a wireless IMU sensor network (i.e., two IMUs) is used to enable a three-dimensional angular assessment of the thigh and the knee for gait analysis [16]. The differential changes of the rotation angles are also used to reduce the influences of the drift and offset noises that are common in IMU raw data. In particular, this paper presents gait phase differences using the sagittal, frontal, and transverse directional angles of the knee joint as a novel parameter to classify normal and abnormal gait patterns. The preliminary experimental results demonstrate that the proposed parameter is promising and can carry out gait pattern identification.
