— A small network of low cost non-imaging sensors enhances critical patient monitoring without the invasiveness of cameras. —
(Troy, New York) Health and wellness care facilities often struggle to balance trade-offs between affordable, necessary patient monitoring and patient privacy. Camera surveillance, though useful in many situations, is not always the best option for monitoring occupants in a facility where preserving privacy is also an important function. Researchers at LESA (the Lighting Enabled Systems & Applications Center at Rensselaer Polytechnic Institute) are solving the camera conundrum in healthcare settings by demonstrating how a small network of low cost time-of-flight sensors use the speed of light to measure distances. The technology is similar to automotive LIDAR used for self-driving cars and offers an inexpensive ‘cameraless’ monitoring solution for health and wellness facilities.
Cameras are inherently invasive and require continuous human monitoring for real-time feedback. Recent research at LESA is showing how LIDAR can be applied indoors using a simple module with 9 infrared (IR) time-of-flight sensors to interpret patient and staff movement. The sensor modules are being integrated with machine learning algorithms to provide real-time feedback on patient pose and movement without a human in the loop. Rather, the sensor network constructs a dynamic, low resolution 3D map of the people and objects in the room. The system accurately counts and tracks occupants, detecting changes in position (e.g. falls) and monitoring patient/staff interactions. Unlike radio frequency monitoring systems that can have limited occupancy resolution, IR LIDAR signals do not penetrate walls and can accurately monitor the movements of multiple people in a small area. By training the sensor network to recognize normal movement patterns for a wide variety of typical patient activities and patient-staff interactions, it will be possible to automatically report activities or identify issues as they occur.
Using a combination of simulated and choreographed patient, patient/visitor and staff interactions to train the sensor network, machine learning algorithms “learn” to recognize occupants’ movements and interactions. Though LESA’s sensor network training process is just getting started, 5 distinct poses can already be accurately detected through simple privacy-preserving, LIDAR distance measurements. With further training, the researchers believe the system will automatically detect a wide variety of patient movements and will even be able to log staff/patient interactions to discern whether or not the interactions are normal (someone receiving treatment) or abnormal (someone being assaulted). Once an abnormal deviation is detected, the system autonomously and instantly notifies personnel if it requires caregiver inquiry.
The Patient Safety & Protection System or PSPS, could also be integrated with patient movement data via online medical records with non-image patient tracking to notify remote staff of the frequency and duration of visit to a patient by staff. It would increase the accuracy of patient/provider activity logs and help elevate the time spent filing out reports that may be better spent attending patients. With further development, the PSPS could interface directly with existing software platforms such as Vocera or other mobile communication devices typically used by nurses.
Perhaps most significantly, is that LESA’s proposed occupancy tracking system offers an extremely effective alternative where laws may be too vague or stringent on the use of electronic/camera surveillance in the healthcare setting. Continuous cameras surveillance in general is not HIPAA compliant and the Electronic Communications Privacy Act which regulates how an individual’s personal data can be retrieved, stored and distributed, can make facility administrators shy away from monitoring all together.
Citing the Feb. 8, 2019 Associated Press article that prompted a push for health facility video surveillance in Arizona, Claire Karlicek, a nurse (and wife of LESA Center Director Dr. Robert Karlicek), realized that the Center’s research on building occupancy sensing systems initially developed for improving building energy efficiency, could also address a much broader issue of patient wellbeing involving abuse that occasionally occurs in long term care facilities. They both recognized that LESA’s approach to sensing human motion, pose and location can automatically classify human activity in healthcare and eldercare environments while still preserving privacy. A video [camera-based] system can provide a lot more information than non-camera systems, but at the cost of invading patient privacy and requiring full-time monitoring by another person. The ‘indoor LIDAR’ approach now being developed at LESA can be designed to automate patient monitoring and report results directly to supervisors, record staff monitoring of patients while updating electronic medical record systems without cameras or additional staff input.
“Using IR light to measure distance and infer patient movement and wellbeing is not science fiction. We’re doing this now at the LESA Center where we’ve set up a small Patient Safety Testbed,” Karlicek explains. “It’s still in development but the initial findings are extremely promising. We hope to secure one or more development partners who can work with the Center’s researchers to commercialize indoor LIDAR sensing in healthcare applications. The healthcare specialists who have seen what we are up to feel it has a lot of potential, and we are eager to see this technology widely deployed in hospitals, assisted living centers and long term care facilities.”
In addition to testing the early prototype of the indoor LIDAR module, the research team is working to develop a low cost, integrated circuit version that will be compact enough (about the size of a U.S. quarter) to fit into existing light fixtures or other ceiling mounted systems. The target pricing for the final system is projected to be about $0.08/sqft of monitored space. LESA has filed initial IP on the ‘Indoor LIDAR’ system’s healthcare application, and eager to accelerate it through the commercialization pipeline with an industry partner. In the interim, LESA will engage with potential beta-test users to gather feedback on system design and performance, software development, and commercialization strategies.
For more information on the Patient Safety & Protection System prototype, or additional market segment applications for LESA’s Indoor LIDAR senor platform technology, contact Leah Scott at scottl2@rpi.edu or 518-276-4010.
By Leah Scott