Distracted driving is any activity that deviates an individual’s attention from driving. Some of these activities include talking to people in the vehicle, using hand-held devices such as mobile phones or tablets, eating or drinking, and adjusting the stereo or navigation systems while driving. To counter the effects caused by distracted driving, many countries around the world have imposed rules and charged fines on distracted drivers in order to ensure safe driving. Owing to the technological advancement in recent times, modern-day technologies such as computer vision, image processing, and machine learning techniques can further support the efforts of governments to prevent accidents caused by distracted driving. In this paper, an efficient distracted driver detection scheme (DDDS) has been proposed using two robust deep learning architectures, mainly visual geometric groups (VGG-16) and residual networks (ResNet-50). The proposed DDDS scheme contains the pre-processing module, image augmentation techniques, and two classification modules based on deep learning architectures. Both the architectures are implemented, and the results have been compared in terms of performance indices, namely accuracy, and logarithmic loss. The two-dimensional (2D) dashboard images derived from the State-Farm dataset are pre-processed and are used for training, testing, and validation of the proposed architectures. Accuracy of 86.1% and 87.92% are achieved with VGG-16 and ResNet-50 models, respectively, and it is observed that the DDDS scheme is found highly efficient for c4, c5, and c7 categories of the State-Farm dataset. The results obtained with the proposed DDDS methodology are compared with existing literature and found to be satisfactory. The algorithms developed and discussed for the proposed DDDS can be instrumental in reducing the fatalities and injuries caused due to distracted driving.

Food is one of the basic essential elements for human beings. Its quality is directly related to our physical, emotional, and social well-being. An electronic nose is a non-destructive instrument and is an effective solution for quick and easy determination of food quality. In this paper, an integrated soft E-nose methodology focusing on performance enhancement and complexity reduction is proposed and implemented for the effective classification of beef quality and prediction of microbial population in beef. The soft E-nose methodology is tested on 18 available datasets. Ours is a four-step approach that includes validated data collection, signal processing with a moving average filter and principal component analysis for feature reduction, efficient pattern recognition using machine learning and deep learning techniques like support vector machine, k-nearest neighbors, artificial neural network, extreme learning machine, and deep neural network for classification and regression, and an output stage for classification and prediction of the microbial population of beef. This methodology could provide accuracy greater than 98% using k-nearest neighbors, artificial neural network methods, and a correlation coefficient of greater than 0.98 using support vector machine and artificial neural network methods. The proposed methodology is validated using 10-fold cross-validation.

Food adulteration is the most serious problem found in the food industry as it harms people’s healths and undermines their beliefs. The present study is focused on designing and developing a smart electronic nose (SE-Nose) for the qualitative and quantitative fast-track detection of food adulteration. The SE-Nose methodology is comprised of a dataset, sample slicing window protocol, normalization, pattern recognition, and output blocks. The dataset pork adulteration in beef is used to validate the SE-Nose methodology. The sample slicing window protocol extracts the early part of the signal. The sample slicing window protocol and pattern recognition models (classification and regression models) together achieved the high-performance and fast-track detection of pork adulteration in beef. With classification models, the qualitative analysis of adulteration is measured, and with regression models, the quantitative analysis of adulteration is measured. An accuracy of 99.996% and an RMSE of 0.02864 were achieved with the SVM classification and regression model. The recognition time in detecting pork adulteration in beef with SVM models is 40 s. With the proposed SE-Nose methodology, the recognition time is reduced by one-third. To validate the classification and regression models, a 10-fold cross-validation method was used.

Tactile pressure sensors have gained ever increasing attention in designing flexible electronics, such as wearable devices. Building a pressure sensor characterized by high sensitivity and low implementation costs is a very critical challenge to satisfy. Recently, a new low-cost pressure sensor, namely Velostat, has been proposed. Due to its properties of changing its resistance with either flexing or pressure, it has become popular for making inexpensive sensors for microcontroller experiments. Here, we design a smart sleeve made of Velostat to perform rudimentary controls on a smartphone. We first evaluate the optimum layout of the sensor by conducting three tests on the sensor property, such as area of contact, Velostat area, and layers of Velostat. Based on the test performance, a sensor patch is designed. Then, we develop a novel resistive signal conditioning circuit to convert the pressure on the sensor patch into an electrical signal to be transmitted to smartphones via Bluetooth. Finally, we have also realized an Android application that performs the required operations (e.g., multimedia player controller and an SOS alert system) on the smartphone based on the received sensor data. The functionality of our smart sleeve system is validated in real-time, showing promising results for the design of low-cost smart wearables.