The integration of anticipatory robotic assistance within Ambient Assisted Living (AAL) environments represents a critical challenge in modern biomedical engineering. Traditional assistive interfaces, such as wearable sensors or vision-based systems, often suffer from physical intrusiveness, privacy issues, or high computational latency. This thesis proposes a non-invasive and real-time robotic control system based on Ground Reaction Force (GRF) analysis for the recognition of human postural intentions. The goal is to develop an intelligent framework that, beyond intention recognition, provides proactive support through a robot that adapts its movements based on both the surrounding environment and the specific kinetic signals received from the force platforms.