Sensor distribution optimisation in smart environments

The worldwide human population is set to inflate by over 2 billion within thecoming decades, with the ratios of old and older old forecast to enlarge rapidly in bothdeveloped and developing countries. Chronic and progressive disease and disability arepositively correlated with advanced ageing. Global financial instability, fiscal deficits,high unemployment and an enormous predicted energy shortage represent the globalbackdrop to the task of providing safe and secure accommodation and/or medicalsupervision to elderly cohorts, both now and in the future.Smart homes are living spaces facilitated with technology that allow individuals toremain in their own homes for longer, rather than be institutionalised. Sensors are thefundamental physical layer with any smart home as the data they generate is used toinform decision support systems, which is turn facilitate appropriate actuator actions.The appropriateness of actuator actions is thus crucial in providing safety and securityto a smart home inhabitant. The positioning, amount(s) and field of view(s) of sensors istherefore a fundamental characteristic of a smart home. Contemporary smart homesensor distribution is aligned to either a) a total coverage approach, or b) a humanassessment approach. These methods for sensor arrangement are not data drivenstrategies, are unempirical, and frequently irrational.The overall aim of this thesis was to develop and test a novel sensor distributionparadigm for the optimisation of resources. Four chronological studies were conducted.Studies 1 and 2 investigated the problem of how to acquire inhabitant spatial frequencydata from a smart home environment, which was robust to the range of ambientconditions prevalent in an indoor scene, unobtrusive, efficient and which providedprivacy to the inhabitant and their home. Studies 3 and 4 investigated what methodscould be applied to these spatial frequency data to determine the most optimal sensordistribution per inhabitant, unique to each inhabitant’s movement behaviour andenvironmental physicalities. These investigations propose novel application of existingmethods. In addition, the presented sensor distribution optimisation paradigm isproposed as a novel concept in the implementation and dissemination of smart hometechnology.