Indoor Localization and Tracking Fazail Aslam, Jules Fakhoury, and Tho Le-Ngoc McGill University, Department of Electrical and Computer Engineering, Broadband Communications Research Lab Introduction The industry has been recently very interested in applying wireless real-time localization in mobile indoor environments for commercial applications such as: personnel and equipment tracking. In this research, we investigate the environmental effects on the performance of indoor localization and tracking systems and then find methods suitable to improve their reliability and accuracy. We also aim to deploy, improve, and use an ultra- wideband (UWB) based location system to carry out mainly workflow related studies in a hospital environment. Experiments were performed on the 3 operational UWB indoor localization systems, one installed in a hospital and the other two in our research labs. 0 1 2 3 4 5 6 7 8 Feb, 22 Feb, 23 Feb, 24 Feb, 25 Feb, 28 Marc… Marc… Marc… Marc… Marc… Marc… Marc… Marc… Marc… Marc… Marc… Time (in Hours) Zone A Zone B Zone C Zone D 48% 27% 25% 0% Zone A Zone B Zone C Zone D Zone A Zone B Zone C Zone D Zone A 0 6 0 0 Zone B 6 0 4 0 Zone C 0 4 0 0 Zone D 0 0 0 0 Time spent by the doctor in each zone during the monitored days Average time spent by the doctor Number of transits between zones for the doctor Tags in the Emergency Department of the Royal Victoria Hospital Zone A Zone B Zone C Zone D Conclusions •Our UWB-based location system performed with a reported accuracy in 3D of 15 cm. •The fingerprinting method yielded results with accuracy in the range of 1-3 m depending on the case •The trilateration is not particularly accurate, only being good for a rough estimation UWB Localization •UWB systems are characterized by a very large bandwidth and short-time pulses. •A large bandwidth improve the reliability as the signal contain different frequency components, which increases the probability that at least some of them can go through obstacles. •Spreading the signal energy over a large bandwidth decreases the power spectral density, thus reducing interference to other systems in the hospital. Combined AoA, TDoA technology: •The system performs measurement of both pseudorange (based on TDoA) and AoA •More measurements per sensor make the system more robust Green lines represent the AoA seen by each sensor, while blue curves represents the possible tag positions that would have generated the TDoA RTLS Architecture Tags •Transmits UWB radio pulses •Flexible update rate •Comes into 2 forms: slim and compact Filter-based location algorithm: •Sensors include information filters algorithms that use models of object dynamics to enhance location accuracy of dynamic and static objects in a variety of environments, eliminating reflections and ambiguous data •Algorithm works in an iterative manner by combining the previous estimate of the position with the current measurements Sensors •Detects UWB pulses from tags •Contain an array of antennas which can measure the Angle-of-Arrival (AoA) and Time-Difference-of-Arrival (TDoA) of tag signals. UWB RTLS Medical staff Device Equipment Tag Ubisense Sensor GUIs API Database (I,x,y,z) (I,x,y,z) RTLS Data Proc. Statistical data • Movement pattern • Site visit statistics • Total walking distance Movement heat map online offline Map Server Visual statistical results Workflow Analysis and Organizational Setup Improvement RTLS Data Capture 2.4GHz 6.0GHz Internet + VPN Emergency Area Personal Tag Results Applications Management of the Medical Staff Healthcare Asset Tracking and Management Patient Monitoring •The UWB system estimates the position of many doctors during their shift in real time. Measurements with 2 sensors and a static tag Wi-Fi Localization •This method can be implemented using the existing Wi-Fi setup or usb Wi-Fi detectors •First Wi-Fi access points/ detectors to be used are selected •A region is then selected and calibrated using these access point •Once a region has been mapped, a tag can be detected using the calibration maps and the access points The location coverage shown above measure he signal strength provided by various AP’s The survey calibration accuracy is shown above •This method can be implemented using usb Wi-Fi detectors •Position of at least 3 detectors known •Using the received signal strength and the path loss model, we can calculate the distance between the fixed detectors and the mobile node •Using simple geometry, we can then calculate the exact position. Fig (a) is the ideal case and (b) is with range estimation error Fingerprinting Method: Trilateration Method: •The information about the position of the doctors can be used to extract useful information such as the time spent in each zone and the number of transits between zones. •Using fingerprinting a location is mapped first and then the nearest neighbor method is used to estimate the location. 0 1 2 3 4 5 6 2 3 4 5 6 7 8 9 x (meters) y (meters) Position of Reference Points (blue) and Access Points (red) -2 0 2 4 6 8 10 0 2 4 6 8 10 12 x (meters) y (meters) Distance Estimate from AP1 Distance Estimate from AP2 Distance Estimate from AP3 Distance Estimate from AP4 True position Estimated Position 1 Estimated Position 2 Estimated Position 3 Estimated Position 4 Estimated Position 5 •Trilateration proves to be the simplest, yet the most inaccurate of the three approaches to localization. -5 0 5 10 -2 0 2 4 6 8 10 12 14 x (meters) y (meters) Distance Estimate from AP1 Distance Estimate from AP2 Distance Estimate from AP3 True position Estimated Posiion A calibrated map of the MC- 846 lab Original and estimated position of tags using fingerprinting. Original and estimated position of tags using trilateration The location system can be used for offline studies of clinical workflow patterns, helping improve efficiency. Tracking of specialized medical equipment would reduce the time wasted by clinicians looking for them throughout the facility. A precise location system in the ED would enable staff to track patients in a timely manner, ensuring accurate and consistent delivery of care.