Conference
Workshop on Physical Analytics
About: Workshop on Physical Analytics is an academic conference. The conference publishes majorly in the area(s): Analytics & Computer science. Over the lifetime, 33 publications have been published by the conference receiving 584 citations.
Topics: Analytics, Computer science, Gesture recognition, Software analytics, Context (language use)
Papers
26 Jun 2016
TL;DR: The attempt to emerge with an approach which does not require any dedicated training inside the specific environment where the system is deployed, by suitably identifying a set of differential CSI feature candidates and selecting the (two) most effective ones via minimization of the summation of the Davies-Bouldin indexes.
Abstract: This paper focuses on the problem of providing a rough count of the number of people in a room using passive WiFi Channel State Information (CSI) measurements taken by a single commodity receiver. The feature which mainly distinguishes our work from others is the attempt to emerge with an approach which does not require any dedicated training inside the specific environment where the system is deployed. Our proposal stems from the intuitive observation that features which account for em variations of CSI are expected to be less sensitive to the surrounding environment as opposed to features which account for absolute CSI measurements. We turn such intuition into a concrete proposal, by suitably identifying a set of differential CSI feature candidates, and by selecting the (two) most effective ones via minimization of the summation of the Davies-Bouldin indexes. We preliminary assess the effectiveness of the proposed approach by training once for all the system in a room, and testing the system in two em different rooms having different size and furniture, and involving people freely moving in the rooms with no a-priori movement constraints.
75 citations
22 May 2015
TL;DR: A unified framework highlighting the challenges of smartphone-based driver behavior analysis is presented, and the most commonly employed driving events are reviewed, and some of the difficulties inherent in detecting these events are discussed.
Abstract: Insurance telematics programs are continuously gaining market shares in the automotive insurance industry. By recording data on drivers' behavior, the information asymmetry between the policyholder and the insurer is reduced, enabling a granular risk differentiation based on the true risk levels of the drivers. However, the growth of the insurance telematics industry is being held up by large logistic costs associated with the process of acquiring data. As a result, several market participants have started looking towards smartphone-based solutions, which have the potential of easing and improving the data collection process for both policyholders and insurers. In this paper, we present a unified framework highlighting the challenges of smartphone-based driver behavior analysis. Since all driver behavior analysis relies on access to accurate navigation data, we first address the intermediate step of smartphone-based automotive navigation. The considered topics include estimation of the smartphone's orientation with respect to the vehicle, classification of the smartphone owner as a passenger or driver, and navigation in GNSS-challenged areas. Once a driver-specific high-performance navigation solution has been obtained, it can be used to extract information on the driver's behavior. We review the most commonly employed driving events, and discuss some of the difficulties inherent in detecting these events.
50 citations
22 May 2015
TL;DR: WIBECAM passively leverages Beacon frames periodically emitted by a single off-the-shelf Wi-Fi access point, and its classification accuracy has been preliminarily assessed for four different activities in two different environments; the resulting confusion matrices show very promising performance.
Abstract: This paper presents WIBECAM, a Human Activity Recognition system which does not require neither user instrumentation, nor specialized infrastructure, nor active operation - it passively leverages Beacon frames periodically emitted by a single off-the-shelf Wi-Fi access point. As many other recent proposals, WIBECAM also exploits the different multipath conditions (and their temporal variations) induced by human activity. In most of the previously proposed systems, the classification is based on the characterization of the signal strength variations, caused by the human activity. WIBECAM's main distinguishing aspect is that it 'watches' the channel in the frequency domain where spectral metrics, calculated on the raw signal samples of the received Beacon frames, are like 'snapshots' of the channel taken in a regular and periodical way. The classification process uses properly selected features that measure the changes of consecutive 'snapshots'. WIBECAM adapts to any Wi-Fi access point (and may comply even with legacy 802.11b-only ones), as it does not exploit neither OFDM and CSI extracted from the receiver, nor MIMO/multiple antennas. WIBECAM has been built into USRP software radios. Its classification accuracy has been preliminarily assessed for four different activities in two different environments; the resulting confusion matrices show very promising performance.
39 citations
19 Jun 2017
TL;DR: PeriFi exploits multipath reflections as individual spatial sensors to increase the sensitivity of the conventional approaches, and analyzes each multipath component independently, increasing sensitivity so it can directly sense both moving and non-moving occupants.
Abstract: A large amount of energy could be saved by detecting home occupancy and automatically controlling the lights, and HVAC. Existing occupancy sensors can detect the motion of people but cannot detect people when they are stationary. In this paper, we present a system called Peripheral WiFi Vision (PeriFi), which exploits multipath reflections as individual spatial sensors to increase the sensitivity of the conventional approaches. PeriFi analyzes each multipath component independently, increasing sensitivity so it can directly sense both moving and non-moving occupants. Our evaluations for 6 physical configurations with 11 different occupancy states show that PeriFi can achieve 96.7% accuracy, which translates to nearly 30% improvement over the conventional approaches.
33 citations
11 Jun 2014
TL;DR: The GlimpseData framework is presented, which can be used as a filter to avoid sensing and processing video for face detection, and the mechanism avoids processing roughly 60% of video frames while missing only 10% of frames with faces in them.
Abstract: Emerging wearable devices provide a new opportunity for mobile context-aware applications to use continuous audio/video sensing data as primitive inputs. Due to the high-datarate and compute-intensive nature of the inputs, it is important to design frameworks and applications to be efficient. We present the GlimpseData framework to collect and analyze data for studying continuous high-datarate mobile perception. As a case study, we show that we can use low-powered sensors as a filter to avoid sensing and processing video for face detection. Our relatively simple mechanism avoids processing roughly 60% of video frames while missing only 10% of frames with faces in them.
25 citations
Performance Metrics
| Year | Papers |
|---|---|
| 2017 | 7 |
| 2016 | 9 |
| 2015 | 8 |
| 2014 | 9 |