Affective computing is becoming more and more popular, and the need to find a user-friendly and reliable method of estimating people’s emotions, in their everyday life, is growing. Traditional methods have reached their limits, and this thesis presents a new system of emotion recognition, though physiological signals. With a user-friendly, wearable device, the system can be deployed in a number of fields. A model for our emotion classification is presented and includes the following emotions: cheerfulness, sadness, erotic, horror, and neutral. An experiment of emotion elicitation is also described in this work. Three analysis models applied in our system in order to recognize emotions, including nearest neighbor, discriminant analysis, and multi-layer perception, are discussed in detail. The final test results show that the system has the average recognition rates of 40%, 55.7%, and 77.34% for nearest neighbor, discriminant analysis, and multi-layer perception respectively.

To address the feasibility of ambulatory emotion recognition, characteristics of biosignals were compared between sitting and controlled walking using different stimulus modalities. Emotional stimulus items were drawn from the International Affective Pictures System and International Affective Digitized Sounds libraries to elicit five basic emotions. To assess which emotion was elicited, participants (n=15) completed self-report scales using the Self-Assessment Manikin and discrete emotion ratings following the presentation of each stimulus item. Autonomic activity was monitored using electrocardiogram, electrodermal activity, and thoracic and abdominal respiration. Multivariate analysis of variance was employed to test for differences in biosignal features and supervised classifiers were trained to predict the elicited emotion using physiological data. The study revealed differences between sitting and walking states but no effect was found for stimulus modality. Self-reported emotions were poorly predicted using our methodology and a discussion of potential directions and recommendations for future research was presented.

To address the feasibility of ambulatory emotion recognition, characteristics of biosignals were compared between sitting and controlled walking using different stimulus modalities. Emotional stimulus items were drawn from the International Affective Pictures System and International Affective Digitized Sounds libraries to elicit five basic emotions. To assess which emotion was elicited, participants (n=15) completed self-report scales using the Self-Assessment Manikin and discrete emotion ratings following the presentation of each stimulus item. Autonomic activity was monitored using electrocardiogram, electrodermal activity, and thoracic and abdominal respiration. Multivariate analysis of variance was employed to test for differences in biosignal features and supervised classifiers were trained to predict the elicited emotion using physiological data. The study revealed differences between sitting and walking states but no effect was found for stimulus modality. Self-reported emotions were poorly predicted using our methodology and a discussion of potential directions and recommendations for future research was presented.

Home medical equipment becomes increasingly popular as VLSI fabrication technology advances. However, there are two important factors for realizing a miniaturized biochip: low noise [1] and low power. Firstly, physiological signals are very susceptible to interference while the amplitude of the signal is only a few millivolts or less. If the circuit cannot reject noise effectively, it is hard to amplify the signal and obtain the output voltage of the recording system accurately. Secondly, it is not convenient to replace the batteries frequently when using the portable measurement instrument for the patients. This thesis is focused on the measurement of physiological signals, such as electrocardiography (ECG) [2], electroneurogram (ENG) [3] and electromyography (EMG) [4] , and designing an all-in-one recording system to measure the different physiological signals in a chip. For this purpose, a programmable multi-parameter system for recording of the wide range of physiological signals is designed. The system provides two types of input transconductance stages, BiCMOS and CMOS. BiCMOS amplifiers provide high gain , low noise [5] and low offset voltage suitable for the small amplitude of the physiological signal. On the other hand, CMOS amplifiers provide practically infinite input impedance and ultra-low leakage current. The system also provides three selectable amplifier modes: (a) double-differential amplifier, (b) single-differential amplifier in channel 1, (c) single-differential amplifier in channel 2. The double-differential amplifier provides a high common-mode rejection and adjustable gain for each channel to further reduce common-mode interference. The single-differential amplifier (channel 1 or channel 2) in the recording system are also accessible as differential-input and single-ended output channels. Moreover, the system provides an offset compensation structure to prevent the amplifier from exceeding the input range. The offset compensation system can selectively be turned off to reduce the power consumption.

A WIoT is a wireless network of low-power sensing nodes placed on the human body. While operating, these networks routinely collect physiological signals to send to offsite medical professionals for review. In this manner, these networks support a concept known as pervasive healthcare in which patients can be continuously monitored and treated remotely. Given that these networks are used to guide medical treatment and depend on transmitting sensitive data, it is important to ensure that the communication channel remains secure. Symmetric pairwise cryptography is a traditional scheme that can be used to provide such security. The scheme functions by sharing a cryptographic key between a pair of sensors. Once shared, the key can then be used by both parties to encrypt and decrypt all future messages. To configure a WIoT to support the use of symmetric pairwise cryptography a key distribution protocol is required. Schemes for pre-deployment are often used to perform this distribution. These schemes usually require inserting key information into WIoT devices before they can be used in the network. Unfortunately, this need to manually configure WIoT devices can decrease their usability. In this thesis we propose and evaluate an alternative approach to key distribution that uses physiological signals derived from accelerometer and gyroscope sensors. The evaluation of our approach indicates that more study is required to determine techniques that will enable ballistocardiography-derived physiological signals to provide secure key distribution.

A safe electrical connection between the human body and the recording circuit is required for the acquisition of physiological signals such as the electrocardiogram (ECG), electroneurogram (ENG), or electromyogram (EMG). The recording chip is conventionally connected to the human body through a blocking capacitor. The capacitor avoids any DC current flowing from the recording system into the patient¡¦s body in the case of chip failure. However, the large capacitor area in an integrated chip and its effect on the signal transform function make the use of a coupling capacitor undesirable. In principle, a DC-coupled system can be used to overcome this limitation. The DC-coupled amplifier connects directly to the patient. However, a DC failure current caused, for example, by a gate-oxide short failure could harm the patient. To detect a dangerous condition, a safety monitoring system is proposed in this thesis. The safety monitoring system applies a test signal and physiological signals to the amplifier input. The disappearance of the test signal in the event of circuit failure is detected at the amplifier output. The recording system can then be switched into a safe state. The analysis of the monitoring system, its design procedure and simulation results are presented in this thesis. Moreover, the first measured results are reported for a system realized as an integrated circuit in TSMC 0.35 £gm 2P4M CMOS process technology.

abstract: Monitoring vital physiological signals, such as heart rate, blood pressure and breathing pattern, are basic requirements in the diagnosis and management of various diseases. Traditionally, these signals are measured only in hospital and clinical settings. An important recent trend is the development of portable devices for tracking these physiological signals non-invasively by using optical methods. These portable devices, when combined with cell phones, tablets or other mobile devices, provide a new opportunity for everyone to monitor one’s vital signs out of clinic. This thesis work develops camera-based systems and algorithms to monitor several physiological waveforms and parameters, without having to bring the sensors in contact with a subject. Based on skin color change, photoplethysmogram (PPG) waveform is recorded, from which heart rate and pulse transit time are obtained. Using a dual-wavelength illumination and triggered camera control system, blood oxygen saturation level is captured. By monitoring shoulder movement using differential imaging processing method, respiratory information is acquired, including breathing rate and breathing volume. Ballistocardiogram (BCG) is obtained based on facial feature detection and motion tracking. Blood pressure is further calculated from simultaneously recorded PPG and BCG, based on the time difference between these two waveforms. The developed methods have been validated by comparisons against reference devices and through pilot studies. All of the aforementioned measurements are conducted without any physical contact between sensors and subjects. The work presented herein provides alternative solutions to track one’s health and wellness under normal living condition. / Dissertation/Thesis / Doctoral Dissertation Electrical Engineering 2016

Background: Human-computer interaction (HCI) is one of the daily triggering emotional events in today’s world and researchers in this area have been exploring different techniques to enhance emotional ability in computers. Due to privacy concerns and the laboratory's limited capability for gathering data from a large number of users, common machine learning techniques that are extensively used in emotion recognition tasks lack adequate data collection. To address these issues, we propose a decentralized framework based on the Federated Learning architecture where raw data is collected and analyzed locally. The effects of these analyses in large numbers of updates are transferred to a server to aggregate for the creation of a global model for the emotion recognition task using only Galvanic Skin Response (GSR) signals and their extracted features.  Objectives: This thesis aims to explore how the CNN based federated learning approach can be used in emotion recognition considering data privacy protection and investigate if it reaches the same performance as basic centralized CNN.Methods: To investigate the effect of the proposed method in emotion recognition, two architectures including centralized and federated are designed with the CNN model. Then the results of these two architectures are compared to each other. The dataset used in our work is the CASE dataset. In federated architecture, we employ neurons and weights to train the models instead of raw data, which is used in the centralized architecture.  Results: The performance results indicate that the proposed model not only can work well but also performs better than some other related work methods regarding valance accuracy. Besides, it also has the ability to collect more data from various sources and also protecting sensitive users’ data better by supporting tighter privacy regulations. The physiological data is inherently anonymous but when it comes to using it with other modalities such as video or voice, maintaining the same anonymity is challenging.  Conclusions: This thesis concludes that the federated CNN based model can be used in emotion recognition systems and obtains the same accuracy performance as centralized architecture. Regarding classifying the valance, it outperforms some other state-of-the-art methods. Meanwhile, its federated nature can provide better privacy protection and data diversity for the emotion recognition system.

In this thesis we explore if a proposed random binary sequence generation algorithm can be combined with a separately proposed symmetric key agreement protocol to provide usable security for communications in Wireless Body Area Networks (WBAN). Other previous works in this area fall short by only considering key generation between two of the same signals or allowing for key generation between two different types of signals but with the cost of a significant signal collection time requirement. We hoped to advance this area of research by making secure key generation more efficient with less signal collection time and allowing keys to be generated between two sensors that measure two different physiological signals. However, while the binary sequence generation algorithm and key agreement protocol perform well separately, they do not perform well together. The combined approach yields keys that have good properties for use in a WBAN, but the generation rate is low.

Mémoire numérisé par la Division de la gestion de documents et des archives de l'Université de Montréal.