Abstract: Sleep is an important part of the human daily routine, as vital to survival as food and water. In fact, restorative sleep is closely linked to improved physical, cognitive, and psychological well-being. Conversely, poor or disordered sleep can result in cognitive and psychological impairments, as well as a decline in overall physical health, thereby increasing the risk of falls. To address these problems, sleep monitoring and detecting bed falls have thus become crucial areas of research, given their significant impact on human health and well-being. This thesis focuses on providing a foundational understanding of sleep and falls, in addition to proposing non-obtrusive sleep monitoring and fall detection devices. First, we provide a foundational understanding of sleep, including its stages, the significance of sleep monitoring, and the consequences of low sleep quality, including falls. Next, we present a basic understanding of falls in the elderly, their causes, and their consequences on the person as well as on the environment. We also propose a generic classification of sleep monitoring and fall-related systems based on their sensor deployment, aimed at offering scientists and engineers in this field a better understanding and providing a solid foundation for existing systems.Additionally, we propose a compact wearable device that can be attached to the person’s nightwear to measure body acceleration for sleep monitoring and fall detection. We also propose a novel real-time sensor node system that is non-wearable and mounted on the contour of the bed mattress to continuously monitor sleep activities and detect nocturnal falls. The performance of the proposed devices is validated through simulations, experiments, and comparison with existing devices.

Abstract: This thesis addresses the enhancement of digital sampling oscilloscopes (DSOs) through innovative techniques aimed at mitigating measurement uncertainties and improving overall accuracy. The first contribution focuses on minimizing sampling jitter in DSOs by introducing a phase-trigger sampling technique. The proposed approach utilizes a clock phase-shifting trigger to replace the Phase-Locked Loop (PLL) for sampling frequency generation. Implemented on an FPGA, the phase-trigger DSO demonstrates significant improvement in measurement accuracy, with improvements of up to 50% depending on data rate and pattern length. Building upon this, the second contribution investigates the implementation of a Precision Time Base (PTB) technique on an FPGA to further enhance the measurement jitter performance of DSOs. This technique uses two reference signals to correct timebase errors and reduce jitter introduced by the DSO itself. Experimental results show a substantial reduction in measurement jitter, from 1.3 picoseconds (ps) to 280 femtoseconds (fs), depending on the data rate and pattern length of the high-speed DSO.

Abstract: In 2050, France will have nearly 4 million people over 60 years old with a loss of autonomy, an increase of more than 60% compared to the last census in 2015. These few figures give an idea of the stakes in the medico-social sector and the challenge that the aging population represents for our society in various aspects. In an attempt to respond to the desire of the French people to be able to stay at home, expressed through the ''Concertation Grand Age'' in 2019, the objective of the thesis is a contribution in helping the elderly stay at home. The research approach consists first of all in identifying the main causes of departure from home to independent residences in order to focus the thinking and work on the very first causes representing the majority of cases of departure from home. The ambition of the project is to be able to detect the first warning signs of the loss of autonomy of elderly people: the physical and functional variables linked to the increase in their state of fragility, governing in particular their balance, are certainly predominant factors. However, a person's fragility is a complex and multifactorial phenomenon that is difficult to quantify. Health practitioners are able to evaluate an evolution of fragility through clinical observation of individuals, using questionnaires, but the parameters observed are not objectified in the daily clinical approach, and the frequency of observations, which is difficult to implement, is not individualized and does not allow for an adapted follow-up. We will therefore seek to link the relevant parameters on which professionals rely to detect the evolution of the state of fragility with the main causes of leaving home. The aim of the study is to equip the home, and not the individual, with devices in order to carry out continuous measurements in the home, a natural, non-stigmatizing and a priori non-intrusive environment, with the aim of obtaining relevant and objective information about the individual's behavior. This approach will be validated by the definition of a kit of connected objects associated with intelligent data processing. The work is thus organized in 4 phases: (1) determine the main causes of the departure from home; (2) analyze the professional practices used by practitioners to characterize the state of fragility of an elderly subject, and identify the relevant objectifiable parameters ; (3) to validate these parameters relevant to the diagnosis through preliminary measurements carried out in close collaboration with the practitioners, and to identify measures that can be carried out in the home; (4) to implement a proof of concept in the form of an experimental demonstrator based on connected objects to be placed in the home measuring balance and other parameters determined during the previous steps.

Abstract: This thesis explores the optimization of specialized education and home care services in France. It addresses the practical challenges encountered in these fields, focusing primarily on the allocation and planning of professionals to meet the diverse needs of people with, for example, visual or hearing impairments. The research is structured around three configurations: assignment and planning issues in specialized education services, the integration of specialized education and home care services with assignment, planning, and ergonomic challenges, and the optimization of multi-center scenarios. Each configuration is addressed using mathematical models and multi-objective approaches, with the aim of achieving equitable resource allocation and improving service efficiency. In the first configuration, a mixed integer linear programming model with two multi-objective approaches (a weighted sum method and an epsilon-constraint-based model) is employed to balance the workload among educators and ensure student satisfaction. The second configuration extends this approach to the integration of specialized education and home care services and the resolution of their multi-day assignment and planning problems while considering travel times and distances. We provided an exact solution using a mixed integer linear programming model to solve the problem studied. In addition, we implemented a greedy heuristic and two metaheuristic approaches (a genetic algorithm and a discrete invasive weed optimization algorithm) to solve large-size instances. We considered seven objectives: specialization of assignments, equitable distribution of unproductive hours and overtime hours among the employees, balancing of traveled distances among the employees and minimization of total distance traveled, highest distance traveled, and number of unproductive and overtime hours. In addition, assignment, planning, and ergonomic constraints were taken into account, such as skill qualification, lunch breaks, quota restrictions, tolerated overtime, and travel time. The final configuration focuses on the optimization of multi-center scenarios. A two-phase approach has been implemented. The first phase allocates missions to centers on the basis of a hierarchical multi-objective mathematical model, taking into account qualification and capacity constraints. The second phase assigns missions to employees in each center and optimizes the planning of schedules.