International audience

We present HandRate, the first smartphone-based system using a standard sensor (accelerometer) for opportunistically computing heart rate while a user holds their phone. Fundamentally, HandRate revisits ballistocardiography (BCG), a century-old technique for monitoring heart activity by measuring the body movement caused by the cardiac cycle. Traditionally performed using custom hardware, attached to a subject's body, revisiting BCG for the smartphone, held in hand, faces several challenges. The hand is an external organ furthest from the aorta and subject to motion artifacts, leading to a weak and noisy signal, while the position the phone is held in can impact which accelerometer axis best captures BCG. HandRate addresses these challenges by introducing a design involving two modules operating in tandem: the first aimed at transforming the accelerometer readings into a single-dimensional signal oblivious to how the phone is held, while the second module making heartbeat predictions based on this signal. Results from testing HandRate using data collected from 18 subjects show that it can estimate heart rate with accuracy similar to or better than systems requiring special sensors and/or active user participation.

The recent standardization by IEEE of Fine Timing Measurement (FTM), a time-of-flight based approach for ranging has the potential to be a turning point in bridging the gap between the rich literature on indoor localization and the so-far tepid market adoption. However, experiments with the first WiFi cards supporting FTM show that while it offers meter-level ranging in clear line-of-sight settings (LOS), its accuracy can collapse in non-line-of-sight (NLOS) scenarios. We present FUSIC, the first approach that extends FTM's LOS accuracy to NLOS settings, without requiring any changes to the standard. To accomplish this, FUSIC leverages the results from FTM and MUSIC - both erroneous in NLOS - into solving the double challenge of 1) detecting when FTM returns an inaccurate value and 2) correcting the errors as necessary. Experiments in 4 different physical locations reveal that a) FUSIC extends FTM's LOS ranging accuracy to NLOS settings - hence, achieving its stated goal; b) it significantly improves FTM's capability to offer room-level indoor positioning.

This paper addresses the problem of efficiently virtualizing NUMA architectures. The major challenge comes from the fact that the hypervisor regularly reconfigures the placement of a virtual machine (VM) over the NUMA topology. However, neither guest operating systems (OSes) nor system runtime libraries (e.g., Hotspot) are designed to consider NUMA topology changes at runtime, leading end user applications to unpredictable performance. This paper presents eXtended Para-Virtualization (XPV), a new principle to efficiently virtualize a NUMA architecture. XPV consists in revisiting the interface between the hypervisor and the guest OS, and between the guest OS and system runtime libraries (SRL) so that they can dynamically take into account NUMA topology changes. The paper presents a methodology for systematically adapting legacy hypervisors, OSes, and SRLs. We have applied our approach with less than 2k line of codes in two legacy hypervisors (Xen and KVM), two legacy guest OSes (Linux and FreeBSD), and three legacy SRLs (Hotspot, TCMalloc, and jemalloc). The evaluation results showed that XPV outperforms all existing solutions by up to 304%.