Personal emergency response (PER) system

What is this patent about?

’796 is related to the field of activity recognition and remote health monitoring. The invention addresses the growing need for continuous, unobtrusive supervision of elderly or disabled individuals, where traditional home care is often sporadic and expensive. By leveraging wearable sensors and automated analysis, the system aims to detect dangerous conditions, such as falls or medical emergencies, without requiring constant human intervention.

The underlying idea behind ’796 is the decomposition of complex physical behaviors into a structured hierarchy of elemental motions. Rather than attempting to identify a high-level activity from raw data alone, the system breaks down movement into discrete, identifiable segments. By classifying these segments into groups of similar postures, the invention creates a mathematical roadmap that allows a processor to reconstruct and recognize specific activities based on the order and timing of these basic building blocks.

The claims of ’796 focus on a system that utilizes one or more sensors and a wireless transceiver to monitor a mobile object. The core legal protection covers the specific mechanism of posture modeling, where sequences of movements are classified into groups. The independent claim specifically requires the processor to identify individual elemental motions within a sequence and then match that entire sequence against stored templates to identify a corresponding activity.

In practice, the invention works by capturing data from devices like accelerometers or EMG sensors and feeding it into a Hidden Markov Model (HMM) or similar pattern recognizer. The system identifies transitions between different states—such as moving from a sitting posture to a lying posture—to infer what the user is doing. If the sequence of motions deviates from a normal pattern or matches a known dangerous profile, such as a sudden acceleration followed by prolonged inactivity, the system can automatically trigger an emergency alert.

This approach differs from prior solutions by moving beyond simple threshold-based alerts, which are often prone to false alarms. By focusing on the sequence of elemental motions, the system gains a more nuanced understanding of human behavior, allowing it to distinguish between a user intentionally lying down and an accidental fall. Furthermore, the integration of a wireless mesh network ensures that data is relayed reliably even if a single communication node fails, providing a robust infrastructure for long-term independent living.