To understand human behavior and intrinsically anticipate intentions, research into activity recognition HAR) using sensors in wearable handheld devices has intensified. The ability for a system to use as few resources possible recognize user's from raw data is what many researchers are striving for. In this paper, we propose holistic deep learning-based architecture, convolutional neural network-long short-term memory network (CNN-LSTM). This CNN-LSTM approach not only improves the...

Neural Networks (NN) are a family of models for broad range emerging machine learning and pattern recondition applications. NN techniques conventionally executed on general-purpose processors (such as CPU GPGPU), which usually not energy-efficient since they invest excessive hardware resources to flexibly support various workloads. Consequently, application-specific accelerators neural networks have been proposed recently improve the energy-efficiency. However, such were designed small set...

Wearable inertial measurement unit (IMU) sensors are powerful enablers for acquisition of motion data. Specifically, in human activity recognition (HAR), IMU sensor data collected from categorically combined to formulate datasets that can be used learning activities. However, successful activities involves the design and use proper feature representations suitable classifiers. Furthermore, scarcity labelled is an impeding factor process understanding performance capabilities data-driven...

Position-estimation systems for indoor localization play an important role in everyday life. The global positioning system (GPS) is a popular system, which mainly efficient outdoor environments. In scenarios, GPS signal reception weak. Therefore, achieving good position estimation accuracy challenge. To overcome this challenge, it necessary to utilize other position-estimation localization. However, existing systems, especially based on inertial measurement unit (IMU) sensor data, still face...

Advances in deep learning (DL) model design have pushed the boundaries of areas which it can be applied. The fields with an immense availability complex big data been beneficiaries these advances. One such field is human activity recognition (HAR). HAR a popular area research connected world because internet-of-things (IoT) devices and smartphones are becoming more prevalent. A major goal recent work has to improve predictive accuracy for limited computational resources. In this paper, we...

Localization using ultra-wide band (UWB) signals gives accurate position results for indoor localization. The penetrating characteristics of UWB pulses reduce the multipath effects and identify user with precise accuracy. In UWB-based localization, localization accuracy depends on distance estimation between anchor nodes (ANs) tag based time arrival (TOA) pulses. TOA errors in system, from ANs to adds error system. a system also line sight (LOS) conditions anchors tag, computational...

Neural Networks (NN) are a family of models for broad range emerging machine learning and pattern recondition applications. NN techniques conventionally executed on general-purpose processors (such as CPU GPGPU), which usually not energy-efficient since they invest excessive hardware resources to flexibly support various workloads. Consequently, application-specific accelerators neural networks have been proposed recently improve the energy-efficiency. However, such were designed small set...

Smartphone camera or inertial measurement unit (IMU) sensor-based systems can be independently used to provide accurate indoor positioning results. However, the accuracy of an IMU-based localization system depends on magnitude sensor errors that are caused by external electromagnetic noise drifts. based depend experimental floor map and poses. The challenge in smartphone camera-based is rapidness changes user's direction. In order minimize both systems, we propose hybrid combine IMU...

Sensor fusion frameworks for indoor localization are developed with the specific goal of reducing positioning errors. Although many conventional without have been improved to reduce error, sensor generally provide a further improvement in accuracy. In this paper, we propose framework using smartphone inertial measurement unit (IMU) data and Wi-Fi received signal strength indication (RSSI) measurements. The proposed uses location fingerprinting trilateration positioning. Additionally,...

Access to reliable estimates of the wireless channel, such as channel state information (CSI) and received signal strength would open opportunities for timely adaptation transmission parameters consequently increased throughput efficiency in vehicular communications. To design adaptive schemes, it is important understand realistic properties, especially environments where mobility communication devices causes rapid variation. However, getting CSI challenging due lack support obtaining from...

Facial emotion recognition (FER) systems play a significant role in identifying driver emotions. Accurate facial of drivers autonomous vehicles reduces road rage. However, training even the advanced FER model without proper datasets causes poor performance real-time testing. system is heavily affected by quality than algorithms. To improve for vehicles, we propose image threshing (FIT) machine that uses features pre-trained and from Xception algorithm. The FIT involved removing irrelevant...

Facial emotion recognition (FER) detects a user's facial expression with the camera sensors and behaves according to emotions. The FER can apply entertainment, security, traffic safety. system requires highly accurate efficient algorithm classify driver's The-state-of-art architectures for FER, such as visual geometry group (VGG), Inception-V1, ResNet, Xception, have some level of performance classification. Nevertheless, original VGG suffer from vanishing gradient, limited improvement...

The problem in the process of spectrum sensing that detection rate primary user (PU) signal is low environment signal-to-noise (SNR) present, a novel algorithm based on convolution neural network (CNN) proposed. CNN widely used image recognition and speech recognition, has good classification performance. Therefore, employed to solve which can be viewed as binary hypothesis-testing problem. Firstly, feature presence PU only noise are extracted, including cyclostationary energy feature. And...

Multi-sensor data fusion for advanced driver assistance systems (ADAS) in the automotive industry has received much attention recently due to emergence of self-driving vehicles and road traffic safety applications. Accurate surroundings recognition through sensors is critical achieving efficient (ADAS). In this paper, we use radar vision accurate object recognition. However, since sensor-specific have different coordinates, coordinate calibrate essential. introduce calibration algorithms...

Positioning using Wi-Fi received signal strength indication (RSSI) signals is an effective method for identifying the user positions in indoor scenario. RSSI autonomous system can be easily used vehicle tracking underground parking. In based positioning, positioning estimates of access points (APs) to receiver and identifies user’s positions. The existing systems use raw obtained from APs estimate These fluctuate interfered with by channel conditions. This interference condition reduces...

Localization is one of the current challenges in indoor navigation research. The conventional global positioning system (GPS) affected by weak signal strengths due to high levels interference and fading environments. Therefore, new solutions tailored for environments need be developed. In this paper, we propose a deep learning approach localization. However, performance depends on quality feature representation. This paper introduces two novel set extractions based continuous wavelet...

In recent days, research in human activity recognition (HAR) has played a significant role healthcare systems. The accurate classification results from the HAR enhance performance of system with broad applications. are useful monitoring person’s health, and predicts abnormal activities based on user movements. system’s predictions provide better reduce users’ health issues. conventional systems use wearable sensors, such as inertial measurement unit (IMU) stretch sensors for recognition....

The detection and classification of emotional states in speech involves the analysis audio signals text transcriptions. There are complex relationships between extracted features at different time intervals which ought to be analyzed infer emotions speech. These can represented as spatial, temporal semantic tendency features. In addition that exist each modality, modality consists grammatical tendencies uttered sentences. Spatial have been sequentially deep learning-based models using...

The success of deep learning in speech emotion recognition has led to its application resource-constrained devices. It been applied human-to-machine interaction applications like social living assistance, authentication, health monitoring and alertness systems. In order ensure a good user experience, robust, accurate computationally efficient models are necessary. Recurrent neural networks (RNN) long short-term memory (LSTM), gated recurrent units (GRU) their variants that operate...

Ultra wide band (UWB) systems use time information instead of the popular received signal strength indication (RSSI). UWB is known for its high position accuracy in localization. RSSI-based localization easily affected by attenuation and has a poor as compared to arrival (TOA) technique. In this paper, different algorithms system were analytically reviewed. The performance discussed terms root mean square cumulative distribution function errors. experiment results demonstrate effectiveness...

Efficient indoor positioning requires accurate heading and step length estimation algorithms. Therefore, in order to improve the position accuracy, it is necessary estimate both user with minimal error. These include errors from accelerometer, magnetometer gyroscope of smartphone sensors. Fusing different sensor data has a high impact on improving accuracy. In this paper, we present comparative analysis fusion techniques for using The performance discussed terms root mean square error...

UWB-based positioning systems have been proven to provide a significant high level of ac-curacy hence offering huge potential for variety indoor applications. However, the major challenges related UWB localization are multipath effects, excess delay, clock drift, signal interferences and system computational time estimate user position. To compensate these challenges, uses multiple anchors in experiment area this gives accurate position results with minimum errors. use means processing large...

Wearable sensor-based human activity recognition (HAR) is the study that deals with sensor data to understand movement and behavior. In a HAR model, feature extraction widely considered be most essential challenging part as signals contain important information in both spatial temporal contexts. addition, because people often carry out an for awhile before changing another activity, also long-term context dependencies. this paper, order enhance long, short-term features from data, we propose...

The facial emotion recognition (FER) system has a very significant role in the autonomous driving (ADS). In ADS, FER identifies driver's emotions and provides current mental status for safe driving. determines safety of vehicle prevents chances road accidents. FER, such as happy, sad, angry, surprise, disgust, fear, neutral. To identify these emotions, needs to train with large datasets system's performance completely depends on type dataset used model training. recent uses publicly...