Nabil Hajj Chehade

Feature selection based on mutual information for human activity recognition

In this work, we consider a classification problem of 14 physical activities using a body sensor network (BSN) consisting of 14 tri-axial accelerometers. We use a tree-based classifier, and develop a feature selection algorithm based on mutual information to find the relevant features at every internal node of the tree. We evaluate our algorithm on 31 features per accelerometer (total of 434), and we present the results on 8 subjects with a 96% average accuracy.

Detecting stumbles with a single accelerometer

Falls are a common problem in the elderly population, and their prediction has been a major interest for the medical field. The relationship between stumbles and falls has not been very well understood yet. A critical requirement in advancing the study of this relationship is the realization of a realistic and effective stumble detection system. In this paper, we present a system for the detection of stumbles during walking. Our system consists of a single low cost triaxial accelerometer that may be worn by patients and is convenient for a wide range of subjects. We formulate the problem as an anomaly detection and we validate our system with a large data set collected from 9 subjects. The data set contains a total of 100 stumbles and 45 minutes of walking. We compare 7 different placements for the accelerometer, and show that our system achieves a 99% detection rate, with a 0.2% false alarm rate using an accelerometer worn on the chest.

Estimation of accelerometer orientation for activity recognition

Tri-axial accelerometers have been widely used for human activity recognition and classification. A main challenge in accelerometer-based activity recognition is the system dependence on the orientation of the accelerometer. This paper presents an approach for overcoming this challenge by calibrating the accelerometer orientation using pre-defined activities alongside automated correction algorithms. This method includes manipulation of data via rotation matrices estimated from the pre-defined activities. The system is subsequently tested with real data where sensors were placed in the wrong orientation. A control set of correctly oriented sensors were also placed for validation purposes. We show that our approach improves the accuracy from 38% to 92% for the wrongly oriented sensors, when the control sensors achieve 95%. A GUI was also created in order to make the tool easily available to other researchers.

Multi-class SVM for forestry classification

In this paper we propose a method for classifying the vegetation types in an aerial Color Infra-Red (CIR) image. Different vegetation types do not only differ in color, but also in texture. We study the use of four Haralick features (energy, contrast, entropy, homogeneity) for texture analysis, and then perform the classification using the One-Against-All (OAA) multi-class Support Vector Machine (SVM), which is a popular supervised learning technique for classification. The choice of features (along with their corresponding parameters), the choice of the training set, and the choice of the SVM kernel highly affect the performance of the classification. The study was done on several CIR aerial images provided by the French National Forest Inventory (IFN). In this paper, we will show one example on a national forest near Sedan (in France), and compare our result with the IFN map.

Distributed coordination for fast iterative optimization in wireless sensor/actuator networks

Large-scale coordination and control problems in sensor/actuator networks are often expressed within the networked optimization model. While significant advances have taken place in both first- and higher-order optimization techniques, their widespread adoption in practical implementations has been hindered by a lack of adequate programming and evaluation support. This motivates the two major contributions of this paper. First, we extend the distributed programming framework proposed in [1] with a synchronization primitive to implement different versions of the subgradient technique and perform extensive evaluation with varying deployment and algorithmic parameters. Second, the insights — obtained by observing the variability in practical metrics such as response time and incurred message cost — lead us to exploit the spatial locality inherent in these large-scale actuator control applications, and propose a novel consensus algorithm applied to the subgradient method. We show using simulations that there is at least 99% improvement in response time and the message cost is reduced by more than 90% over prior consensus based algorithms.