Samer Mohammed

An Unsupervised Approach for Automatic Activity Recognition Based on Hidden Markov Model Regression

Using supervised machine learning approaches to recognize human activities from on-body wearable accelerometers generally requires a large amount of labeled data. When ground truth information is not available, too expensive, time consuming or difficult to collect, one has to rely on unsupervised approaches. This paper presents a new unsupervised approach for human activity recognition from raw acceleration data measured using inertial wearable sensors. The proposed method is based upon joint segmentation of multidimensional time series using a Hidden Markov Model (HMM) in a multiple regression context. The model is learned in an unsupervised framework using the Expectation-Maximization (EM) algorithm where no activity labels are needed. The proposed method takes into account the sequential appearance of the data. It is therefore adapted for the temporal acceleration data to accurately detect the activities. It allows both segmentation and classification of the human activities. Experimental results are provided to demonstrate the efficiency of the proposed approach with respect to standard supervised and unsupervised classification approaches.

Ubiquitous robotics: Recent challenges and future trends

Ambient intelligence, ubiquitous and networked robots, and cloud robotics are new research hot topics that have started to gain popularity among the robotics community. They enable robots to acquire richer functionalities and open the way for the composition of a variety of robotic services with three functions: semantic perception, reasoning and actuation. Ubiquitous robots (ubirobots) overcome the limitations of stand-alone robots by integrating them with web services and ambient intelligence technologies. The overlap that exists now between ubirobots and ambient intelligence makes their integration worthwhile. It targets to create a hybrid physical-digital space rich with a myriad of proactive intelligent services that enhance the quality and the way of our living and working. Furthermore, the emergence of cloud computing initiates the massive use of a new generation of ubirobots that enrich their cognitive capabilities and share their knowledge by connecting themselves to cloud infrastructures. The future of ubirobots will certainly be open to an unlimited space of applications such as physical and virtual companions assisting people in their daily living, ubirobots that are able to co-work alongside people and cooperate with them in the same environment, and physical and virtual autonomic guards that are able to protect people, monitor their security and safety, and rescue them in indoor and outdoor spaces. This paper introduces the recent challenges and future trends on these topics.

Lower Limb Wearable Robots for Assistance and Rehabilitation: A State of the Art

Neurologic injuries, such as stroke, spinal cord injuries, and weaknesses of skeletal muscles with elderly people, may considerably limit the ability of this population to achieve the main daily living activities. Recently, there has been an increasing interest in the development of wearable devices, the so-called exoskeletons, to assist elderly as well as patients with limb pathologies, for movement assistance and rehabilitation. In this paper, we review and discuss the state of the art of the lower limb exoskeletons that are mainly used for physical movement assistance and rehabilitation. An overview of the commonly used actuation systems is presented. According to different case studies, a classification and comparison between different types of actuators is conducted, such as hydraulic actuators, electrical motors, series elastic actuators, and artificial pneumatic muscles. Additionally, the mainly used control strategies in lower limb exoskeletons are classified and reviewed, based on three types of human-robot interfaces: the signals collected from the human body, the interaction forces between the exoskeleton and the wearer, and the signals collected from exoskeletons. Furthermore, the performances of several typical lower limb exoskeletons are discussed, and some assessment methods and performance criteria are reviewed. Finally, a discussion of the major advances that have been made, some research directions, and future challenges are presented.

Joint segmentation of multivariate time series with hidden process regression for human activity recognition

The problem of human activity recognition is central for understanding and predicting the human behavior, in particular in a prospective of assistive services to humans, such as health monitoring, well being, security, etc. There is therefore a growing need to build accurate models which can take into account the variability of the human activities over time (dynamic models) rather than static ones which can have some limitations in such a dynamic context. In this paper, the problem of activity recognition is analyzed through the segmentation of the multidimensional time series of the acceleration data measured in the 3-d space using body-worn accelerometers. The proposed model for automatic temporal segmentation is a specific statistical latent process model which assumes that the observed acceleration sequence is governed by sequence of hidden (unobserved) activities. More specifically, the proposed approach is based on a specific multiple regression model incorporating a hidden discrete logistic process which governs the switching from one activity to another over time. The model is learned in an unsupervised context by maximizing the observed-data log-likelihood via a dedicated expectation–maximization (EM) algorithm. We applied it on a real-world automatic human activity recognition problem and its performance was assessed by performing comparisons with alternative approaches, including well-known supervised static classifiers and the standard hidden Markov model (HMM). The obtained results are very encouraging and show that the proposed approach is quite competitive even it works in an entirely unsupervised way and does not requires a feature extraction preprocessing step.

Recognition of gait cycle phases using wearable sensors

The analysis and monitoring of the human daily living activities plays an important role for rehabilitation goals, fall prevention rehabilitation and general health-care treatments. Among these activities, walking is the most important daily motion. Studying the evolution of the gait cycle through the analysis of the human center of force is beneficial to predict any abnormal walking pattern. The analysis is based on the use of pressure-based mapping system that collects pressure and force measurement during the gait cycle. This paper deals mainly with the detection of the main characteristics of the gait phases. To this end, a segmentation of the center of force of the human body measure through the in-shoe pressure mapping system is performed to identify the gait phases. The proposed segmentation approach consists in modeling each segment by a regression model and using logistic functions to model the transitions between segments. This flexible modeling through the logistic functions has the advantage of detecting abrupt and smooth transitions between segments. This paper deals with the detection of the main characteristics of the gait phases.A segmentation of the human body COF is done using the in-shoe pressure mapping system.Identification of the gait phases is done using a regression model.The proposed method has the advantage of detecting abrupt and smooth transitions.The proposed method is implemented and verified by experiment tests.

Posture estimation and human support using wearable sensors and walking-aid robot

In this paper, an omni-directional walking-aid robot is developed to assist the elderly in the daily living movements. A motion control strategy of walking-aid robot based on the observation of the human status through wearable sensors is proposed. During normal walking, the robot is controlled using a conventional admittance control scheme. When the tendency of a fall is detected, the robot will immediately react to prevent the user from falling down. The distance between the human Center of Pressure (COP) and the midpoint of the human feet is assumed to be a significant feature to detecting the fall events. When the user is in a quasi-static state or walking slowly, the COP can be approximated by the projection of Center of Gravity (COG) of the users body. A simple and low-cost wearable sensor system is proposed to measure online the COG of the user. A limitation of the proposed wearable sensor system is that the Head-Arms-Torso (HAT) of the user is assumed to be always in upright position, which may generate measurement error. From comparison experiments with a reference optical system it is found that the measurement error is acceptable especially at the early stage of fall event. Dubois possibility theory is applied to describe the membership function of normal walking state. A threshold based fall detection approach is obtained from online evaluation of the walking status. Finally, experiments demonstrate the validity of the proposed strategy. An omni-directional walking-aid robot is developed to assist and support the elderly.A wearable sensor system was designed to estimate online the human posture.A fall detection method by wearable sensor is obtained based on possibility theory.Normally the robot uses admittance control while it is braked when fall detected.Experiments were conducted to test the wearable sensor and walking-aid robot.

Active Impedance Control of a lower limb exoskeleton to assist sit-to-stand movement

As an important movement of the daily living activities, sit-to-stand (STS) movement is usually a difficult task facing elderly and dependent people. To provide appropriate power assistance for the sit-to-stand movement, a novel intention-based Active Impedance Control (AIC) strategy applied on a lower limb exoskeleton is proposed in this paper. The AIC is able to adapt the mechanical impedance of the human-exoskeleton system towards a desired one using the exoskeletons power assistance. In the AIC structure, a human joint torque observer is designed to estimate the human joint torques using joint angles information instead of electromyography (EMG) or force/torque sensors; a time-varying desired impedance model is proposed according the wearers lower limb motion ability. Simulations were implemented to illustrate the characteristics and performances of the proposed approach. Experiments with a healthy subject were carried out to evaluate the effectiveness of the proposed method. The experiments show satisfactory results in terms of appropriate power assist based on the wearers motion intention.

Observer-based active impedance control of a knee-joint assistive orthosis

In this study, a new active impedance control method of a knee joint orthosis is proposed by using a nonlinear observer to estimate the human joint torque. The proposed method is an alternative to the sensor-based impedance controller in which the human joint torque is estimated by using the electromyography (EMG) sensor, force/torque sensors, etc. The use of the nonlinear observer can efficiently overcome the shortcomings of the EMG and force/torque sensors, such as complexity of use, encumbrance, modeling, sensitivity to noise, high cost, etc. Additionally, a main goal of the proposed active impedance control is to decrease the impedance of the human-orthosis system to a desired level that the wearer can easily achieve. To achieve this goal, a virtual desired impedance model is designed to estimate human intention movements; a nonlinear disturbance observer based sliding mode controller is proposed to track the human estimated movements. Two experiments were carried out to verify the performance of the proposed nonlinear observer and sliding mode controller, as well as the effectiveness of the active impedance assistance strategy. The results show that the proposed method can effectively decrease the impedance of the human-orthosis system according to the desired impedance model. At the same time, the required human muscle torque for ensuring the knee joint flexion/extension movement can be also significantly decreased with the active impedance assistance.

Recognition of different daily living activities using hidden Markov model regression

The human activity recognition is widely used for human behavior prediction especially for dependent people. This is achieved to provide safety, health monitoring, and well being of this population at home. In this paper, the problem of human activity recognition is reformulated as joint segmentation of multidimensional time series. The hidden Markov model regression (HMMR) is used to perform unsupervised segmentation strategy between activities using the expectation-maximization algorithm. This is accomplished over six logical scenarios of twelve daily activities such as stair descent, standing, sitting down, sitting, From sitting to sitting on the ground and sitting on the ground. To evaluate the performance of HMMR model, other unsupervised methods are used including K-means, Gaussian mixtures model and the hidden Markov model. The results show that the HMMR model provides the best results for the different scenarios with up to 97% in terms of correct classification rate.

Human static postures analysis using empirical mode decomposition

The goal of this paper is to analyze the human stability during static postures using stabilometric signals. The effects of subjects visual input, feet position, age and gender are analyzed. Twenty eight healthy subjects have participated in this study. The center of pressure displacements were measured along the Medio-Lateral (ML) and Antero-Posterior (AP) directions. Empirical mode decomposition method is used to decompose the stabilometric signal into several elementary signals called Intrinsic Mode Functions (IMF). A stabilogram-diffusion method is used to generate the related diffusion curve of each IMF and a resulting index called Critical Point (CP) is calculated. The CP parameter showed significant differences between groups using repeated measure ANOVA, particularly in the ML direction in terms of visual modality, feet position, age and gender. The present findings may guide the rehabilitation process. Our proposed method compares favorably to conventional stabilometric analysis based on the center of pressure excursion calculation.The goal of this paper is to analyze the human stability during static postures using stabilometric signals. The effects of subjects visual input, feet position, age and gender are analyzed. Twenty eight healthy subjects have participated in this study. The center of pressure displacements were measured along the Medio-Lateral (ML) and Antero-Posterior (AP) directions. Empirical mode decomposition method is used to decompose the stabilometric signal into several elementary signals called Intrinsic Mode Functions (IMF). A stabilogram-diffusion method is used to generate the related diffusion curve of each IMF and a resulting index called Critical Point (CP) is calculated. The CP parameter showed significant differences between groups using repeated measure ANOVA, particularly in the ML direction in terms of visual modality, feet position, age and gender. The present findings may guide the rehabilitation process. Our proposed method compares favorably to conventional stabilometric analysis based on the center of pressure excursion calculation.