As life expectancy is growing all around the world, a critical issue for many governments is how to deliver good health assistance to a great amount of elderly people, without affecting their daily life. Due to the COVID Pandemia, this problem becomes more urgent, and many new solutions are required for helping them. In this context, Home Care Systems (HCS) can proactively help people in preventing problems, e.g., supporting them in critical situations, such as loss of consciousness or physical disabilities. HCSs consist of services for health-cares and relatives, in a trustful and friendly way, combining both software and hardware technologies. They also should provide a valid approach that reduces the probability of false-positive or false-negative alarms. In this work, we propose drivers that should be taken for the designing of such a system. For that reason, we design an architecture that combines both computer vision algorithms and signal analysis, in order to detect home accidents and abnormal situations. One of the main issues is that integrated sensors should be easily handled and maintained, so we carry out a usability and technological study, detecting which features are necessary for these kinds of solutions. © 2021, The Author(s), under exclusive license to Springer Nature Switzerland AG.