VitalCloud: Weaving a Pervasive Life Network for Healthcare

Motivations

There are about one billion visits to see physicians in offices per year. Millions of these are likely unnecessary. On the other hand, chronic diseases, such as stroke, Parkinson disease, urgently need the continuous monitoring and diagnosis of patents’ daily activities and mobility in rehabilitation rather than one time office visit and the self-report questionnaire that is incomplete, inaccurate, and inconvenient. Current limited capacity and access time of clinical resources are insufficient for better services, and will be overwhelmed by the ever increasing elderly population and the high demand health care services. The technological interventions in patient and elderly monitoring have long been needed to improve the medical services. If patients or elders are equipped with advanced wireless self-diagnosis technologies, important changes of health status or vital signs can be monitored and processed without connected to a wired machine with wired cables on body. Moreover, with automatic health monitoring, some emergency situations could be detected and lots of life could be saved. For example, the risk of falls in the home account for 71 percent of all fatal accidents to those aged 65 and over, and most of the tragedy could be avoided if the accident is detected and served immediately.

The technology that enabling smart monitoring in home and community, and integrating with the advanced wireless pervasive networks, has the potential to revolutionize the health care services, rehabilitation and clinical practices. We propose a Medical Cyber-Physical System (mCPS) that aims to provide medical and healthcare service using wireless sensing and communication techniques along with cloud data and analytics services.

Overview

The realization of mCPS hinges on solutions to the following key challenges. 1) Noncontact continuous sensing techniques. Monitoring and collecting accurate physiological signals without forcing subjects to wear devices with complicated wires is attractive. Noncontact accurate sensing is the prerequisite to enable such a transformative paradigm shift for patient monitoring, diagnosis, research and feedback-based control. However, current noncontact sensing techniques, e.g., Radar, is not accurate and sensitive enough for capturing the extremely weak vital sign signal. 2) Accurate indoor localization techniques. To continuous tracking and monitoring a person’s activity data, locating the target in indoor environment is a turnkey solution. Otherwise, the activity and vital sign data measured is incorrect and unmatched to the real situations. However, current indoor localization techniques are immature and the resolution achieved is several meters. 3) Low power and low complexity realization of mCPS. The communication layer should support scalable data rate requirement of several medical applications, i.e., adapts the transmission speed to the application at the restraint of minimum energy consumption. The smart monitoring could only be accessible by ordinary consumers with the device cost as little as $5 to $10 each and can operate without changing battery frequently. Such demand requires a low-complexity and low-cost design for its hardware as well as maintaining its capability in sensing and monitoring.

Research

Leveraging the existing research and work, we plan to explore novel algorithms, hardware, and software system design mechanisms to realize the vision of the smart monitoring initiative. In particular, we will engineer a complete smart monitoring mCPS system, equipped with our hardware, MIMO radar sensing, localization, recognition and autonomic management algorithms.

Although some interesting work has been done to utilize radio transmission signatures for monitoring and localization, we need to conduct research on going beyond simple locations to infer activities and higher-level behaviors through multi-dimensional sensor nodes. Utilizing wideband MIMO radar to extract patient’s activity and perform localization simultaneously, higher level features, e.g., the walking speed, Cadence, activity level, or emergency situations of falling down, could be detected and analyzed. The validity of rehabilitation trails for patients after stroke could be monitored and accessed, along with the early detection of disease exacerbation.

Incorporating the advanced sensing to the BLE-based medical body area network, important physiological information could be transmitted to the App that running on user’s smartphone, e.g., iOS or Android version. The mobile App collects the data and performs recognition and reports back to the medical center. A small group of healthcare practitioners can take care of hundreds or thousands of elders with finer details about their vital signs without spatial or temporal barriers. If some abnormal situation happens to the object, immediate medical services could be called. With wireless no-contact monitoring, patients will no longer to be tethered to large machines, and the blind point of monitoring will be avoided. We plan to make the smart monitoring mCPS service affordable at homes with our health status monitored wirelessly, and in turn reduce the risk of danger and lower the expenditures in health care.

Significance and Innovation

Enabling the long-term smart monitoring could bridge the gap between the current physical barriers to health monitoring and the growing need for advanced and automatic monitoring capabilities. Combining the smart monitoring and the mCPS, a complete solution could be envisioned and make transformative contributions in health monitoring. The activities information inferred from sensors can be used for chronic diseases rehabilitation and early detection of diseases. Other applications include baby care, emergency alarm, trails evaluation, and interventions all need our mCPS services to improve the quality of monitoring and transmission. All these envisioned applications will usher in new research opportunities and competitions that enable such technology.

Preliminary Results