Dementia assistive system as a dense network

As elderly population increases, portion of dementia patients becomes larger. Thus social cost of caring dementia patients has been a major concern to many nations. This article introduces a dementia assistive system operated by various sensors and devices installed in body area and activity area of patients. Since this system is served based on a network which includes a number of nodes, it requires techniques to reduce the network performance degradation caused by densely composed sensors and devices. This article introduces existing protocols for communications of sensors and devices at both low rate and high rate transmission.

💡 Research Summary

The paper presents a comprehensive Dementia Assistive System (DAS) designed to support the daily lives of dementia patients and to provide continuous health monitoring for remote caregivers. DAS is architected as a three‑layer network: (1) a Body Area Network (BAN) composed of wearable sensors that measure vital signs such as heart rate, temperature, blood pressure, ECG, EEG, and EMG; (2) an Activity Area Network (AAN) that deploys environmental sensors (temperature, humidity, illumination) together with higher‑bandwidth devices such as cameras and microphones throughout the patient’s living space; and (3) a tele‑medical network that aggregates data from BAN and AAN, connects to medical databases, and enables real‑time alerts and emergency handling.

The authors emphasize that the dense deployment of nodes in both BAN and AAN creates severe channel contention and latency challenges. To address these, they propose a layered communication strategy that matches traffic characteristics with appropriate PHY/MAC standards. Low‑rate, low‑power BAN traffic (typically <10 kbps, latency 100–250 ms) is mapped to IEEE 802.15.4 (ZigBee) or IEEE 802.15.1 (Bluetooth). 802.15.4 offers a hybrid of TDMA and CSMA/CA with beacon and non‑beacon modes, providing excellent energy efficiency but limited contention‑free slots (maximum 7), which can become a bottleneck in very dense deployments. Bluetooth’s polling and frequency‑hopping mechanisms mitigate interference but restrict the number of slaves per master to seven and occupy the crowded 2.4 GHz band, potentially colliding with other systems.

High‑rate services required by AAN—such as video streaming, audio, and large‑scale sensor aggregation—necessitate IEEE 802.11 (Wi‑Fi) technologies (b/g/n/ac/ad). While Wi‑Fi delivers multi‑Mbps throughput, its CSMA/CA access suffers from exponential back‑off, causing average throughput to drop sharply as the node count rises (e.g., a 10‑node scenario on a 72 Mbps channel yields only ~3.5 Mbps per node). Wi‑Fi also consumes more power and can interfere with BAN frequencies. To alleviate cross‑technology interference, the paper recommends allocating 2.4 GHz exclusively for BAN/AAN low‑rate links and reserving the 5 GHz band for Wi‑Fi high‑rate links.

Beyond frequency separation, the authors discuss MAC‑level scheduling techniques. Predictable, periodic BAN traffic can be assigned deterministic TDMA slots, reducing contention and guaranteeing latency. Non‑urgent video and audio streams can exploit 802.11’s Enhanced Distributed Channel Access (EDCA) with priority queues, ensuring that critical health alerts pre‑empt bulk data. The paper also notes that more advanced OFDM‑based high‑speed schemes (e.g., SLM‑OFDM) are compatible with the DAS architecture when even higher data rates are required.

In conclusion, the study identifies the need for a hybrid networking approach that aligns low‑rate, energy‑constrained sensors with ultra‑low‑power standards, while routing bandwidth‑intensive media through Wi‑Fi on a separate spectrum. By carefully partitioning time and frequency resources and employing appropriate MAC mechanisms, DAS can maintain reliable performance even in densely populated sensor environments, thereby supporting continuous patient monitoring, decision‑making assistance, navigation, and rapid emergency response for dementia care.