Betty Martin

Wireless implementation of mems accelerometer to detect red palm weevil on palms

Red palm weevil infests palm even in its very young stage and completes generation leaving the tree unhealthy or leading to its death [1-2]. Though there are various growing interests implemented on the nondestructive methods for detection of the pest, none of the technology has come into general use of detection. It suffered from logistic or implementation issues. Here the article explains a novel detection of the pest. A simple 3 axis accelerometer which is capacitive micro machined is used to measure the acceleration. The output is in terms of analog voltage which is directly proportional to the acceleration. Deflection of structure is measured using differential capacitor which consists of differential fixed plates with central plate attached to moving mass. Capacitor accelerometers are superior in low frequency ranges. The overview of the system is simple yet a complete process of three blocks which incorporate a sensing device (MEMS accelerometer), a computational core (Micro controller) and serial port for wireless communication (Bluetooth technology).

A Novel Approach to Identify Red Palm Weevil on Palms

Red Palm Weevil is a deadly pest on palms. It destroys thousands of palms every year by feeding on the soft tissues of palm trunk which are less than 15-20 years. The productivity decreases resulting in huge loss. There is no accurate or exact solution to destroy this infestation in the early stages. Hence there arises a technology to identify the presence of RPW along with its larva by sensing its sound. To achieve this task, we need to devise an acoustic sensor so that defected trees can be easily picked up and proper measures can be engaged to destroy the pest. This technology depends on the acoustic emissions from pests. Clustering is also done on input data recorded which gives similarity measures between two signals. The Euclidean distance for the least significant output with the database will match and help in identifying the insect species.

Extraction of feature from the acoustic activity of RPW using MFCC

RPW is a destructive pest which infests palms of age ranging from 5–20 yrs. Control measure if adopted timely on infestation can lead to saving of palm, else ends up in outright death of palm. RPW usually infests palm in its younger stage while adult causes damage through feeding. The burrowing of the larva into palm heart causes more mortality. Major symptoms such as crown loss or leaf wilt are usually visible only after palm has become infected. By the time these symptoms are observed the damage is usually high to kill the tree. Hence in order to detect the presence of weevil even at the early stage the sound of larvae burrowing and chewing are recorded using digital voice recorders. The recorded sound file is further analyzed. In sound processing, Mel frequency cepstrum is a representation of short term power spectrum of a sound based on a linear cosine transform of a log power spectrum on a nonlinear Mel scale of frequency. MFCCs have been used for feature extraction. In frequency domain, the most often used feature extraction is Mel frequency cepstral coding using Mel scale.

Identifying Sound of RPW In Situ from External Sources

Over the last decade, speech recognition has been used in the field of security system, gender identification for automatic speech recognition, pattern recognition, biometrics, voice finger, dragon naturally speaking, etc. In the recent past, lots of research works are being carried out in these fields. The proposed research work also deals with one such interesting system, wherein the characteristics of the sound generated by red palm weevil (RPW) for recognition of their presence in the palm in a nondestructive way is done. For this work, a text-independent identification system makes use of feature extraction and feature matching technique. The sound of RPW recorded is compared against external sources for easy detection. Out of the several techniques available for feature extraction and comparison, mel-frequency cepstral coding (MFCC) technique has been utilized for feature extraction and the comparison is being carried out using vector quantization (VQ).

Energy Distribution and Coherence-Based Changes in Normal and Epileptic Electroencephalogram

In endeavor toward better understanding of brain functions, the analysis of information transfer between the various brain lobes plays a crucial role. Electroencephalogram (EEG) is an electrical brain signal in microvolts, which provides unique and important information about the brain dynamics. Frequency of EEG signal lies between 0 and 100 Hz. In epileptic or seizure related studies, decomposition of EEG signal into various frequency sub-bands such as α, β, \( \delta \), \( \theta \), and γ is essential. EEG plays a key role in diagnosis of neurological disorders such as epilepsy. In this paper, we explore decomposition of EEG by db18 wavelet, power spectral density, coherence, energy distribution, and empirical cumulative distribution function of EEGs. This work was carried out to study the changes in the normal and epileptic EEGs with respect to PSD, coherence, energy, and ECDF to check the suitability of these parameters as an input to the classifiers. The methodology was applied mainly to three groups consisting of male and females between the age group of 01–107 years: (1) healthy subjects (normal), (2) subjects with focal seizures, (3) subjects with generalized seizures. The work was carried out on the signals obtained from real subjects to study the EEG-based brain connectivity analysis. It was observed that PSD and coherence study related to the sub-bands reveal more accurate information than the study of complete EEG with or without the seizures.

Welch’s Power Spectral Density of Cranial PPG Signal Using AVR ATmega 8535 Microcontroller

Functional near-infrared spectroscopy is an optical non-invasive technique for measurement of neural activity and hemodynamic response and has a potential for brain mapping. In this work, we aimed to develop a system to capture the cranial photoplethysmogram (CPPG) using IR source (860 nm) and detector (OPT 101). AC excitation for the IR source in the range of 1−2 MHz plays a major role in the CPPG sensor. Brain functional activity in prefrontal lobe is detected by placing the sensor on the forehead. The CPPG signal is captured with eyes open (EO) and eye blinking (EB) activity for various emotions (happy and sad) on prefrontal lobe. Data acquisition is done using AVR ATmega 8535 microcontroller at a sampling rate of 500 Hz. The data are acquired from subjects in the age group of 20−60 years. Using Daubechies 9, six-level wavelet decomposition of CPPG signal is performed, and spectral analysis of each level is done using Welch’s method. From the spectral analysis, it is found the centre frequency of A6 is at 1.953 Hz. The higher frequency part of the signal is found in D4 scale and the centre frequency is 17.58 Hz. All other scales D3−D1 show predominantly the noise part of the signal. Statistical features of A6 and D6 decomposition levels are important for analysis and classification of activity.

Robust Estimation of Brain Functional Connectivity from Functional Magnetic Resonance Imaging Using Power, Cross-Correlation and Cross-Coherence

Functional Magnetic Resonance Imaging (fMRI) is a non-invasive method for investigating the structure and function of the brain. Using fMRI, brain functions and areas responsible for the particular activities are investigated. The objective of the image processing methods using fMRI is to investigate the functional connectivity. To localize mental functions of specific brain regions and to identify the brain regions, those are activated simultaneously. Correlation and cross-coherence of the time series of the pixels are used for the detection of functional connectivity in fMRI images for the different motor movements (upper and lower limb movement and finger tapping action). The methodology was applied to three groups (six subjects) consisting aged between 10 and 75 years: (1) Normal and healthy subject performing finger tapping actions, (2) brain tumour patient performing lower limb movement (LL), and (3) brain tumour patient performing upper limb movements (UL). The threshold applied for the cross-correlation is 5000. Similarly, the threshold applied for cross-coherence and power parameters is in the range of (0.6–0.9). The algorithm implemented is found to be non-destructive, and there is no loss of temporal or spatial data. The result shows that for the normal subject, functionally connected pixels are more as compared to the brain tumour patients.

Efficient Set Routing for Continuous Patient Monitoring Wireless Sensor Network with Mobile Sensor Nodes

Static sensor nodes are effectively and efficiently replaced by mobile sensor nodes in WSN applications like tracking, periodic weather monitoring, etc. In this paper, challenges in WSN routing are focused with mobile nodes. A new routing protocol called as set routing is proposed for wireless sensor networks with mobile nodes. In set routing, the sets of nodes are constructed after deployment. Routing overhead is fully given to static sink node or base station. Direction based linear mobility pattern is chosen for its greater coverage area with low energy spent and is used in set routing. Set routing is simulated using Omnet++ and Matlab and the performance is studied. The results are compared with cluster based routing and mobile leach protocols and found that delivery ratio is higher in our set routing.

Route maintenance in set routing using FLT for mobile wireless sensor networks

The goal of wireless sensor networks is the maximum data gathering. Factors like energy, memory, bandwidth and processing capability affects the characteristics of the network which implies in reduced data collection. To avoid the reduction in data collection, mobile sensor nodes are used. They cover large area in short time and these networks are called as mobile wireless sensor networks (MWSN). In MWSN, set routing is used as one of the routing method for maximum data collection. In which the entire network is divided into many sets. This routing algorithm is applied for each set with the consideration that there is no node failure in the network. Here this work shows that how a new route is discovered if a node fails in a set using the Foldable Lookup Table (FLT) method. Also FLT method is compared with the routing table developed by fuzzy variables. Simulation is performed with OMNET++ simulator. Result shows the effectiveness of node failure in terms of delay and packet drops and how the route is maintained during node failure is shown with respect to delay.

Detection of pest infestation by preprocessing sound using vector quantization

In the recent past, economic damage due to infestation of pest on palms could be mitigated significantly by early detection and treatment. Tests were conducted with currently available acoustic instrumentation and software to assess the sound impulses produced by larvas locomotory and feeding activities. The incorporation of bursts into analysis significantly assisted in applications where consistent activity patterns of hidden pest could be identified. The purpose of this research is to realize the effectiveness of a text independent identification system making use of cepstral coefficients and vector quantization. The identification system will be making use of MFCC. The MFCC extracted was then matched to all available sound codebooks that have been stored. The codebook that returns the lowest quantization error will belong to sound contained in audio input file. This confirms the result of detection of the particular species of interest. The data resulting from the analysis will serve as key characteristic in identifying presence of the pest species to which sound belongs.