LoRa is a low-power long-range radio technology for wide-area IoT connectivity with exceptional receiver sensitivity thanks to its chirp spread spectrum (CSS) modulation. However, insufficient data rate has always been an Achilles’ heel of LoRa. This paper proposes a novel PHY-layer design, BIC-LoRa, that leverages non-linear chirp shapes as pictograph to enhance LoRa’s data rate. It employs multiple non-linear chirps for modulation whose shapes encode additional bits to boost data rate while maintaining resilience to low SINR scenarios. To the best of our knowledge, this is the first attempt to enhance the data rate of LoRa by embedding bits in the chirp shapes. Furthermore, BIC-LoRa fully leverages the characteristics of non-linear chirps, allowing it to deal with packet collisions and increase network throughput. We implement BIC-LoRa in GNURadio and MATLAB, and compare its performance against CurveALOHA and standard LoRaWAN through real experiments on USRP B210 software-defined radios. Evaluation results demonstrate that BICLoRa achieves up to 29.4% improvement in data rate for a single link and 32% improvement in overall network throughput while retaining the low SNR robustness of LoRa in real-world scenarios.

This paper proposes transfer learning with spatiotemporal feature analysis using public facial images to build an automatic detection of facial paralysis caused by acute stroke. The overall process includes 1) facial detection and alignment network to extract major regions from the face, 2) transfer learning with feature extraction networks and gated recurrent unit, and 3) a classifier evaluating facial paralysis. We leveraged the Korean facial image data (K-FACE) from the public AI Hub to compensate for the insufficient data representing acute stroke symptoms. The experiment analyzed the effect of transfer learning and time series analysis using a gated recurrent unit with the deep learning models based on MobileNetV2, VGG16, and DenseNet121. Utilizing a facial big data system, transfer learning with spatiotemporal features showed a prominent performance with an accuracy of 0.925 and AUC of 0.924, which indicates the feasibility of real-time detection of stroke in daily living.

Medical big data has become important as many hospitals have been collecting massive amounts of medical information in daily treatment. We investigated the architecture of federated learning to construct the detection model of disease with blood test data formatted in flow cytometry standards to facilitate multi-site medical research. The big data characteristics of raw information in blood tests and privacy problems in sharing patient data make it hard to collect and share data into the central site to construct the generalized detection model. In this paper, we introduce the work-in-progress study, FedM-FCM, the federated learning of flow cytometry analysis with the pipeline from the data sources to the domain-shifted distribution of the federated learning model. We compose the major components of FedM-FCM with data representation of multi-dimensional flow cytometry, adoption of neural network models based on the data representation, and aggregation and distribution of learning parameters across the participating hospitals without data sharing.

This work investigates a run-time scheduling system for Bluetooth Low Energy (BLE). The Bluetooth specifications propose control of the anchor point (i.e., start time of each connection event) by changing the WinOffset parameter of the control protocol data unit (PDU), which can be very useful when the master schedules multiple connections simultaneously. Since we had no access to the controller (i.e., which wraps the physical layer and link layer of Bluetooth), we could not figure out how WinOffset works or use it the way we wanted. However, with Zephyr's advent, there is room to modify the controller of BLE, and we can adjust anchor points as we want using WinOffset. To the best of our knowledge, this work is the first systematic study of this regime in the research community. We experimentally study how commercial devices actually schedule multiple connections and how this inefficient scheduling degrades BLE's performance. Based on the preliminary study, we propose BLEX that dynamically adjusts anchor points of multiple slaves to satisfy quality of service (QoS) requirements without violating the Bluetooth specifications or intervening the BLE host (i.e., software-defined higher layer). Through extensive performance evaluation on off-the-shelf BLE chips, we show that BLEX provides stable performance regardless of the traffic requirements of applications and shows significantly reduced QoS failures compared to the state-of-the-art scheduling schemes for BLE.

The use of Bluetooth Low Energy (BLE) has been steadily increased since Bluetooth Special Interest Group (SIG) standardized BLE in 2010, and BLE performance has been improved owing to various studies. Bluetooth specification 4.2 defined only one data rate, i.e., 1 Mb/s. However, the specification 5 released in 2016 specifies 2 Mb/s and coded PHYs for higher throughput and more stable transmission, respectively. This version also allows the transmission power to be varied from -20 dBm to 20 dBm. Therefore we are enabled to use various data rates and transmission powers in BLE. However, no research has studied how to adapt these yet. In this paper, we propose an algorithm, termed $AdaptaBLE$, that selects data rate and transmission power of BLE by considering link stability and energy consumption. We implement $AdaptaBLE$ on real devices and measure its performance. As a result, we confirm that $AdaptaBLE$ successfully selects an appropriate TX power and data rate while keeping the BLE link stable.

The use of Bluetooth Low Energy (BLE) has steadily increased since 2010. Bluetooth specification 4.2 defined only one data rate, 1 Mb/s. However, Bluetooth specification 5 released in 2016 specifies 2 Mb/s and coded PHY for higher throughput and for more stable transmission, respectively. In addition, this version of specification defines transmission power from -20 dBm to 20 dBm. In this paper, we propose AdaptaBLE, an algorithm that selects the data rate and transmission power of BLE by considering link stability and energy consumption. We implement AdaptaBLE on real devices and measure its performance. As a result, we confirm that AdaptaBLE successfully lowers energy consumption while keeping the link stable.

The advance of deep learning has made huge changes in computer vision and produced various off-the-shelf trained models. Particularly, Convolutional Neural Network (CNN) has been widely used to build image classification model which allow researchers transfer the pre-trained learning model for other classifications. We propose a transfer learning method to detect breast cancer using histopathology images based on Google's Inception v3 model which were initially trained for the classification of non-medical images. The pilot study shows the feasibility of transfer learning in the detection of breast cancer with AUC of 0.93.

Predicting the functional outcomes of spinal cord disorder patients after medical treatments, such as a surgical operation, has always been of great interest. Accurate posttreatment prediction is especially beneficial for clinicians, patients, care givers, and therapists. This paper introduces a prediction method for postoperative functional outcomes by a novel use of Gaussian process regression. The proposed method specifically considers the restricted value range of the target variables by modeling the Gaussian process based on a truncated Normal distribution, which significantly improves the prediction results. The prediction has been made in assistance with target tracking examinations using a highly portable and inexpensive handgrip device, which greatly contributes to the prediction performance. The proposed method has been validated through a dataset collected from a clinical cohort pilot involving 15 patients with cervical spinal cord disorder. The results show that the proposed method can accurately predict postoperative functional outcomes, Oswestry disability index and target tracking scores, based on the patient's preoperative information with a mean absolute error of 0.079 and 0.014 (out of 1.0), respectively.

The need for fast treatment of patients in critical conditions motivates the use of mobile devices to provide prompt and consistent communication between hospitals and physicians. We propose a framework that supports ubiquitous access to medical systems using personalized mobile services and integrated medical systems. The proposed service-oriented medical framework provides dynamically composed services that are adapted to contextual variables such as the user's role, the network bandwidth, and resources available at mobile devices while supporting task allocation in distributed servers for massive resource-consuming services. It also manages accurate patient data by integrating local medical systems using medical information standards such as Digital Imaging and Communications in Medicine and Health Level 7. We have demonstrated the effectiveness of our framework by building a prototype of context-based adaptation of computerized tomography image retrieval for acute stroke treatments, which allows images to be viewed on mobile devices with WiMax wireless network. The proposed medical framework reduces hospital delays of patients and facilitates treatments in the absence of medical specialists.

The requirement of agile adaptation to varying resource constraints in mobile systems motivates the use of a service-oriented architecture (SOA), which can support the composition of two or more services to form a complex service. In this paper, we propose SOA-based middleware to support QoS control of mobile applications and to configure an energy-efficient service composition graph. We categorize services into two layers: functionality-centric services, which are connected to create a complex service to meet the user's intentions, and resource-centric services, which undertake distributed functionality-centric services in a way that increases the success rate of service composition while reducing contention at specific service nodes. We also present a service routing algorithm to balance the resource consumption of service providers on a service-overlay network. Through simulation of power-aware service composition using a realistic model based on ns-2 and traced data, we demonstrate that our approach can help both the mobile devices and the servers in a service-overlay network to reduce energy consumption without an increase in response time.