In this PhD thesis authored by Santiago Iregui, a novel approach to personalised assistive robotics is presented to improve the quality of life of people with motor impairments, particularly those resulting from spinal cord injuries. Recognising that fully autonomous systems may diminish users’ sense of control, the work advances a shared-control paradigm through a constraint-based reactive framework that enables adaptable and user-centred assistance for devices such as wheelchair- or table-mounted robotic arms. The framework structures assistive behaviours into reusable building blocks that continuously integrate input from human–machine interfaces and environmental sensors, allowing assistance levels to be modulated according to user needs and changing contexts in unstructured settings. A collection of combinable components for daily living activities is introduced and validated across eleven diverse use cases involving multiple robots, interfaces, sensors and autonomy configurations. A user study further demonstrates both qualitative and quantitative improvements in tasks such as opening articulated objects with unknown dimensions, highlighting the potential of flexible and personalised robotic systems to better support evolving user requirements.