Spain has more than 10 million inhabitants over the age of 60 (21.71% of the Spanish population), of which 50.9% of the elderly state they have suffered falls and a minimum of 8% of falls with immediate physical consequences ended in fractures or serious injuries (+200,000 inhabitants). From a gender point of view, we want to emphasize that women are the largest group among the elderly. Robots and social agents can be a tool to alleviate the loneliness of this group. But when designing them for older adults, developers need to keep in mind that older people have diminished functional abilities. The accessibility and customization of the interfaces are necessary: the robot or agent has to be adjusted to the user's skills, preferences, gender and personality to ensure acceptance by older adults. Robots or agents can play an important role as assistive technologies for older adults that provide relief to caregivers and social systems, but also to reduce loneliness and promote social engagement. Physical robots may be essential to solving the problem of isolation among the elderly; there has been a preference for physical robots over digital assistants. Devices like smartwatches can help robots determine health status and biorhythm. This data can be used by social robots to register, help and advise. Other data channels, such as cameras, depth cameras, and microphones, can be used to determine situations, recognize emotional states based on facial expression and speech intonation, or used to detect falls. In this project we propose to help, monitor and care for the elderly using a social agent or robot. Taking into account the above, the specific objective of this project, framed in a coordinated project, is to develop a framework to collect information from adult users, process and generate feedback: Multimodal Emotional Detection Collection: Data collection from biosensors worn by the user and from robotic cameras and microphones. Process data to generate high-level information. Recognize the identity of the user and detect emotions through it. facial expression recognition and voice/speech analysis. Perform activity recognition to detect dangerous situations. Fight loneliness. Interact with the user helping with daily tasks and proposing activities. In previous projects, and the last coordinated project JUGUEMOS (TIN2015-67149-C), we made progress in the areas of social robots, facial expression recognition, and activity recognition. We will apply the experience gained in this project and in others in which we have already worked with the elderly and disruptive technologies.